Sound Practices Mailing List Files - Volume 2
=========================================================================
From: "Snoopy" <tcma@netvigator.com>
Subject: Re: [JN] Re: 400Hz or so Sinewave based PSU for DHT's
Date: Sun, 27 Feb 2000 18:23:43 +0800
Source: Sound Digest Archive v02.n466
Hi all,
I would like to share my own experience on a hi-fi system of my friend,
who has employed 400 Hz sinewave power supply. In my opinion, the sound
from his system is one of the best that I have heard. I am a layman on the
above subject. I would try my best to describe the power supply as below.
2. In principle, the supply consists of 2 units. One as an rectifier to
rectify the AC voltage from mains (220 volts 50 or 60 Hz in HONG KONG) into
DC voltage. Another unit is a signal generator to produce a sinewave AC
voltage at 400 Hz from the DC produced by the first unit.
3. My friend said that the 2 units which supply electricity for his whole
hi-fi system is costing around US$10,000. Some others may be cheaper.
4. He also said that with the power supply, the sound from his system is
more clear, stable and quiet.
5. For your information, he is using EMT 930 LP player, Studer 730A CD
player, EMT console as pre-amplifier, ?? power amplifier and 4-way homebread
horn speaker (Electro-voice EV-350 and GAUSS 1502 as high, Philips drivers
and horns as mid, 2 x 15" GAUSS 4583 woofers as low).
Regards,
Tung-cheung MA
HONG KONG
=========================================================================
From: "T. Loesch" <ezee_e@hotmail.com>
Subject: [JN] 400Hz or so Sinewave based PSU for DHT's - over-rating transformers????
Date: Tue, 22 Feb 2000 18:49:08 GMT
Source: Sound Digest Archive v02.n459
Hi all,
I just got this crazy Idea. I have recently been looking into PSU's that
re-synthsize the Mains Voltage, primarily because of me getting ready to get
a Garrard 401 going.
So, using four cheap and cheerfull Chip Amplifiers (say Nat Semi LM3875) we
can quite easily make a 50Hz, 200VA PSU using Stepup Transformers (and how
about an Auto Transformer as Stepdown for our Chip's?).
So, let's take this a step further. Why not use normal Toroid Mains
X-Foermers which have a quite wide bandwidth, use the same chips and a
suitable set of Transformers to get use our HT(s) and LT(s).
Filtering at these frequencies is easy. Using two 1H Chokes and two 22uF
Capacitors in a Choke Input Supply suggests less 0.5mV remaining noise on
the HT line, after the Output Transformer we get about 0.03mV noise, that is
about 100db below 1 Watt....
But best of all, I think a transformer normally rated for only a few dozend
watt should work at much higher power for this application. And how about
using cheap toroidal mains X-Formers as "common mode" chokes with
cancellation of the whole DC in the Windings?
So Joe's, tell me, am I on to something here?
No hum, perfect regulation and cheap? That surely cannot be, or can it?
Later Thorsten
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
=========================================================================
From: "Danielak, Robert M" <robert.m.danielak@lmco.com>
Subject: RE: [JN] 400Hz or so Sinewave based PSU for DHT's - over-rating t ransformers????
Date: Tue, 22 Feb 2000 18:55:57 -0500
Source: Sound Digest Archive v02.n459
i have been playing with this idea on and off over the last year
or so.
however, i did not stop at 400hz.
i have been thinking more like 100kHz and up... ;)
now you're talking *small* transformers and filters...
i've been building some prototype switching power supplies
to run tube amps from my car.
but i've also been doing a couple offline (mains) supplies.
the idea was to reduce the size, weight, and eventually cost
of conventional 50/60hz transformers and related items...
so i have had some fun with conventional square wave
switchers (flyback/push-pull/etc.)
mixed results...
i've also heard from a number of folks that switching
power supplies *kill* the sound of audio equipment.
i can believe it.
i work with a number of so-called power supply "experts".
their goals have practically *nothing* to do with "good sound".
high efficiency, low ripple, tight regulation, low mass, small size, and
wide load tolerance
(stability) are all required for the space industry.
add radiation tolerance, temperature extremes, etc.
and you can see that they are going in the wrong direction...
in an audio amp, especially a tube amp, i'm sure we can sacrifice
a few of these parameters for:
not necessarily low ripple, but kind harmonic structures
not great regulation, but decent bandwidth
and efficiency... shit, i'm burning more power in
my filiments...
80% is fine. but >90% is actually easy with >250V output...
oops, off on a tangent...
in any case, i wonder how much power you can really get out of
a LM3875, LM675, LM3886
i found that the LM675 gets *very* hot trying to do a 30W
power supply.
what kind of efficiency could we expect from
such a circuit.
like 75% in practice...?
so the output transistors (chip) will dissipate 25W for
a 100W supply.
not saying that efficiency is an issue... ;)
just that the devices are only meant to handle so much
dissipation.
at this point my thoughts were diverted to class C amp.
but haven't continued on...
my free-running flyback seems to work well in the car, so
i've been concentrating on other square-wave and
quasi-resonant approaches.
my latest design for the car consists of a *linear*
pre-regulator with a free-running flyback or push-pull
output stage.
the flyback transformer has the advantage of doing
double-duty as step-up and output inductor.
the supply seems to have all the benefits of a standard
choke input supply.
but a 100W supply is about the size of a car's ashtray...
hmmm...
since i don't smoke, that's where i plan to install the supply
in my car.
bob.d.
> ----------
> From: T. Loesch[SMTP:ezee_e@hotmail.com]
> Reply To: Thorsten@tnt-audio.com
> Sent: Tuesday, February 22, 2000 1:49 PM
> To: sound@lists.io.com
> Subject: [JN] 400Hz or so Sinewave based PSU for DHT's - over-rating
> transformers????
>
> Hi all,
>
> I just got this crazy Idea. I have recently been looking into PSU's that
> re-synthsize the Mains Voltage, primarily because of me getting ready to
> get
> a Garrard 401 going.
>
> So, using four cheap and cheerfull Chip Amplifiers (say Nat Semi LM3875)
> we
> can quite easily make a 50Hz, 200VA PSU using Stepup Transformers (and how
>
> about an Auto Transformer as Stepdown for our Chip's?).
>
> So, let's take this a step further. Why not use normal Toroid Mains
> X-Foermers which have a quite wide bandwidth, use the same chips and a
> suitable set of Transformers to get use our HT(s) and LT(s).
>
> Filtering at these frequencies is easy. Using two 1H Chokes and two 22uF
> Capacitors in a Choke Input Supply suggests less 0.5mV remaining noise on
> the HT line, after the Output Transformer we get about 0.03mV noise, that
> is
> about 100db below 1 Watt....
>
> But best of all, I think a transformer normally rated for only a few
> dozend
> watt should work at much higher power for this application. And how about
> using cheap toroidal mains X-Formers as "common mode" chokes with
> cancellation of the whole DC in the Windings?
>
> So Joe's, tell me, am I on to something here?
>
> No hum, perfect regulation and cheap? That surely cannot be, or can it?
>
> Later Thorsten
> ______________________________________________________
> Get Your Private, Free Email at http://www.hotmail.com
>
=========================================================================
From: connlyra@gol.com
Subject: [JN] Re: 400Hz or so Sinewave based PSU for DHT's - over-rating transformers?
Date: Wed, 23 Feb 2000 05:12:38 +0900
Source: Sound Digest Archive v02.n459
Thorsten,
I don't see any mention of 400Hz in your post. What were you thinking
of using as the signal generator?
The idea itself is sound - not crazy at all. For quite a few years
now, there have been two competing product lines of synthesized
100V/120V AC power sources here in Japan with adjustable line
frequency control. One is by a company called CSE (using analog
signal generation and discrete SS power amplifiers), the other is by
Shinano Denki (which uses digital signal generation, not sure about
the output amplifiers).
I think that PS Audio in the US is now doing something similar, but
while the thinking looks to be the same, I am not sure what sort of
specific implementation they use.
I have used the CSE line extensively. From experience, not all audio
products take kindly to 400Hz. The power transformers start to
vibrate and hum if the line frequency gets too high. Likely for that
reason, CSE's original product lineup included a 200 and 400Hz
setting, but most of their more recent versions don't go up so far.
Assuming that your power transformers are up to the task, the sound
will improve at the higher line frequencies. Definitely. It is a big
plus to have frequency control, because in practice you will want to
raise the line frequency as much as possible without forcing the
power transformers into their shake, rattle and roll region.
Like you, I have thought about using power ICs in a BTL configuration
rather than a discrete circuit for the output amplifiers. For lots of
power, the 3875 should be fine. For less power, considering that you
will be using a steup transformer to get the requisite voltage, even
something like the 1875 or 675 may be sufficient.
However, I have been more interested in doing a 3-phase version.
3-phase 400Hz _should_ sound damn good. But while single-phase signal
generators and sine/cosine SGs do exist, I don't know of anyone who
makes a ready-packaged three-phase signal generator chip. And if that
is the case, you need to roll a low-distortion, low-drift signal
generator circuit that offers three outputs that are 120 degrees
apart at all frequencies, with a single master frequency control.
Thoughts?
jc
=========================================================================
From: "T. Loesch" <ezee_e@hotmail.com>
Subject: [JN] Re: 400Hz or so Sinewave based PSU for DHT's - over-rating transformers?
Date: Wed, 23 Feb 2000 10:50:48 GMT
Source: Sound Digest Archive v02.n460
Jonathan,
>I don't see any mention of 400Hz in your post. What were you thinking of
>using as the signal generator?
Any old Wien bridge followed by a batch of LM3875 or LM3886....
The Choice of LM38XX is primarily down to the fact that these things are
near indestructable, cheap and need a minimum of external parts, also the
excellent >120db Hum rejection allowing a fairly "dirty" PSU to be used, so
small value reservoir Capacitors and hence less noise pumped back into the
mains.
>However, I have been more interested in doing a 3-phase version.
>3-phase 400Hz _should_ sound damn good. But while single-phase signal
>generators and sine/cosine SGs do exist, I don't know of anyone who makes a
>ready-packaged three-phase signal generator chip.
Nope, but why not make a crystal locked system having the classic
"staircase" Generator for a near sinewave. Then you can link these counters
up in such a way that all your three phases are prefectly locked. A modest
(4th order) Gaussian Lowpass Filter should round up the Job and leave us
with very nice sinewaves....
Later Thorsten
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
=========================================================================
From: "T. Loesch" <ezee_e@hotmail.com>
Subject: RE: [JN] 400Hz or so Sinewave based PSU for DHT's - over-rating t ransformers????
Date: Wed, 23 Feb 2000 11:57:47 GMT
Source: Sound Digest Archive v02.n460
Hi there,
>i have been thinking more like 100kHz and up... ;)
The Idea with 400Hz is to use readily available ready made transformers.
But yesterday evening I also stareted to think:
"Why not higher...?" Looking at the Datasheet, the LM3886 and similar should
allow > 50W RMS at low distortion for something like 50khz (100kHz will be
pushing it.
>so i have had some fun with conventional square wave
>switchers (flyback/push-pull/etc.) mixed results...
Not surprised.
>i've also heard from a number of folks that switching
>power supplies *kill* the sound of audio equipment.
Exactly, too much harmonics and noise.
>i work with a number of so-called power supply "experts".
Perhaps you could get them to help with the transformers? Let's say we have
a Source offering around 50kHz, 35V RMS and about 6A (RMS) max giving around
200W, enough to light up a pair of 300B's, a 5U4 or so and a pair of 437A
Input valves, including LT and HT.
I need to know the best readily available Core for low distortion and what
primary winding is needed. I can figure the rest myslef....
>in any case, i wonder how much power you can really get out of
>a LM3875, LM675, LM3886
The 3886 will easily do 50W RMS prolonged, assuming a decent heatsink.
>what kind of efficiency could we expect from
>such a circuit.
50%
Later Thorsten
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
=========================================================================
From: "Danielak, Robert M" <robert.m.danielak@lmco.com>
Subject: RE: [JN] 400Hz or so Sinewave based PSU for DHT's - over-rating t ransformers????
Date: Wed, 23 Feb 2000 13:54:05 -0500
Source: Sound Digest Archive v02.n460
> >i have been thinking more like 100kHz and up... ;)
>
> The Idea with 400Hz is to use readily available ready made transformers.
>
understood...
> But yesterday evening I also stareted to think:
>
dangerous when that happens... ;)
> "Why not higher...?"
>
two word: core loss.
we can look to more exotic materials, though.
the nickle crowd would be happy to hear that high nickles
like supermalloy would be up to the task (barely) at 50khz.
and amorphous, like magnesils, could do 50khz with room to spare.
i'd have to dissapoint those folks a bit and jump to ferrite cores, myself.
then again, maybe i should take another
look at those tape cores again...
> Looking at the Datasheet, the LM3886 and similar should
> allow > 50W RMS at low distortion for something like 50khz (100kHz will be
>
> pushing it.
>
yeah, 50k is what i ended up with, for a number of reasons.
including 50k being the maximum bandwidth of any amplifier
i'd design...
> >so i have had some fun with conventional square wave
> >switchers (flyback/push-pull/etc.) mixed results...
>
> Not surprised.
>
yes, i fell into many of the conventional power supply designers
traps.
but i'm learning... ;)
> >i've also heard from a number of folks that switching
> >power supplies *kill* the sound of audio equipment.
>
> Exactly, too much harmonics and noise.
>
actually, it's not all that black and white.
in fact, as much as guido will object (grin)
this is probably one of the least of my worries...
although, i'm sure many of the worst offenders in the commercial market
sacrificed proper
snubbing/damping for the last little bit of efficiency...
> >i work with a number of so-called power supply "experts".
>
> Perhaps you could get them to help with the transformers? Let's say we
> have
> a Source offering around 50kHz, 35V RMS and about 6A (RMS) max giving
> around
> 200W, enough to light up a pair of 300B's, a 5U4 or so and a pair of 437A
> Input valves, including LT and HT.
>
i think i can handle that....
i'll stick with what i have data for, but there are certainly other options.
i like pot cores.
they're easy to wind (bobbins), assemble, and mount.
they also provide very good self shielding.
for a 200W sinewave, bipolar application a size 3622 should
do nicely.
this core is 36mm (1.4in) in diameter and 22mm (0.86in) tall.
(like about 6 american 1/2 dollars stacked up)
magnetics inc. type P material (ui=2500) of philips 3c8 (?)
or equivalent.
using Bmax of about 2kG, and Ac=2cm^2, f=50khz
we need a minimum of 4 turns on this core.
this low number is nice, but it is a double-edged sword.
unless you want to play with fractional turns,
it does not give you a lot of granularity in your secondary
voltages.
we're talking about 35V/4T = 8.75 volt per turn.
if you want to do low voltages like 5VAC, 6.3VAC, etc
then you have to increase this number a bit...
i'd look at the maximum differential i'd allow between
the closest voltages.
i.e. 4.8VAC for the 5V and 6.4V for the 6.3V.
that's 1.6V /T
then 3T for 4.8VAC and 4T for 6.4VAC.
and about 22T primary.
(ok, it's getting ugly. probably best to do separate
trannies for 5V and 6.3V)
i'd shoot for 1T or 2T for the 5VAC output.
back into 7T or 14T primary.
keep in mind that you'd adjust the input for the right
trim...
so lets stick with 7T on the primary. ditch the 6.3VAC
winding for now.
(maybe rectify/regulate
a 5VAC (1T) or 10VAC (2T) winding the 6.3V needs)
i'd be a little nervous about the symmetyry of our
signal generator/amp (offsets will cause unbalanced DC).
so, to avoid flux-walking, i'd use a modest gap in the core.
the 3622 comes with various standard gaps
and are identified by their specific inductance value
in mH/1000T.
suppose we use a -1000 (1000mH/1000T).
with 7 turns, the primary inductance will be about 0.05mH.
this is about 15k at 50khz. so the magnitizing inductance
will be low...
so we have a core
size: 3622
material: P of 3C8, 3C81
Al/Gap: 1000
what about the wire?
6Arms suggest about 15AWG wire. but you can skimp without
major penalty down to 16AWG.
it just so happens that 7T of #16AWG fits in one layer
on the 3622 bobbin.
in practice, though, we'd probably use something like
3 #20AWG in parallel to make the wind easier.
you can also interleave these windings between the
secondaries for better coupling.
i've done #16 on an even smaller core, so let's stick
with that.
the rest of the wind, is simply a matter of applying the
5V/T ratio. there will be a physical limitation to the
number of secondary turns.
this is where pot cores suffer. not as much window area as,
say, an EC41.
but the 7T of #16 and a layer of tape will only take about
20 to 25% of the total area.
you should easily fit 100T of #28AWG for B+. and several
filament windings...
> I need to know the best readily available Core for low distortion and what
>
> primary winding is needed. I can figure the rest myslef....
>
yeah, there's the rub.
it is pretty difficult to buy onesies and twosies of these cores.
mostlty available thru wholesale distributers in qty 50, 100, 1000, etc...
> >in any case, i wonder how much power you can really get out of
> >a LM3875, LM675, LM3886
>
> The 3886 will easily do 50W RMS prolonged, assuming a decent heatsink.
>
yeah. but i thought you were talking 200W...
so you'd need more than one...
no biggie.
i'm thinking more in terms of using separate transformers
and even separate amps for filaments, driver B+, output B+, etc.
this way you can optimise the amp/tranny for the application.
youcould even build in a delay on the B+ to prevent
cathode stripping.... ;)
oh well
back to work
(modelling the power system for our next probe to europa)
bob.d.
=========================================================================
From: "Mike Wurtz" <mikewurtz@mindspring.com>
Subject: Re: [JN] 400Hz or so Sinewave based PSU for DHT's - over-rating transformers????
Date: Wed, 23 Feb 2000 14:10:16 -0600
Source: Sound Digest Archive v02.n460
If you guys really want the ultimate switching regulator, then look no
further than the Linear Technology LT1533 and LT1534.
they are "square wave" switchers of sort, but are speacialized low noise
types. they have both voltage and current slew limiting.
you can do supplies with less noise than a linear with these.
(unless you filter the linear with an LC :^)
the site is linear.com
just pump in 1533 or 1534 and you'll get the part.
mike wurtz
- -----Original Message-----
From: Danielak, Robert M <robert.m.danielak@lmco.com>
To: sound@lists.io.com <sound@lists.io.com>; 'Thorsten@tnt-audio.com'
<Thorsten@tnt-audio.com>
Date: Wednesday, February 23, 2000 12:53 PM
Subject: RE: [JN] 400Hz or so Sinewave based PSU for DHT's - over-rating t
ransformers????
> >i have been thinking more like 100kHz and up... ;)
>
> The Idea with 400Hz is to use readily available ready made transformers.
>
understood...
> But yesterday evening I also stareted to think:
>
dangerous when that happens... ;)
> "Why not higher...?"
>
two word: core loss.
we can look to more exotic materials, though.
the nickle crowd would be happy to hear that high nickles
like supermalloy would be up to the task (barely) at 50khz.
and amorphous, like magnesils, could do 50khz with room to spare.
i'd have to dissapoint those folks a bit and jump to ferrite cores, myself.
then again, maybe i should take another
look at those tape cores again...
> Looking at the Datasheet, the LM3886 and similar should
> allow > 50W RMS at low distortion for something like 50khz (100kHz will be
>
> pushing it.
>
yeah, 50k is what i ended up with, for a number of reasons.
including 50k being the maximum bandwidth of any amplifier
i'd design...
> >so i have had some fun with conventional square wave
> >switchers (flyback/push-pull/etc.) mixed results...
>
> Not surprised.
>
yes, i fell into many of the conventional power supply designers
traps.
but i'm learning... ;)
> >i've also heard from a number of folks that switching
> >power supplies *kill* the sound of audio equipment.
>
> Exactly, too much harmonics and noise.
>
actually, it's not all that black and white.
in fact, as much as guido will object (grin)
this is probably one of the least of my worries...
although, i'm sure many of the worst offenders in the commercial market
sacrificed proper
snubbing/damping for the last little bit of efficiency...
> >i work with a number of so-called power supply "experts".
>
> Perhaps you could get them to help with the transformers? Let's say we
> have
> a Source offering around 50kHz, 35V RMS and about 6A (RMS) max giving
> around
> 200W, enough to light up a pair of 300B's, a 5U4 or so and a pair of 437A
> Input valves, including LT and HT.
>
i think i can handle that....
i'll stick with what i have data for, but there are certainly other options.
i like pot cores.
they're easy to wind (bobbins), assemble, and mount.
they also provide very good self shielding.
for a 200W sinewave, bipolar application a size 3622 should
do nicely.
this core is 36mm (1.4in) in diameter and 22mm (0.86in) tall.
(like about 6 american 1/2 dollars stacked up)
magnetics inc. type P material (ui=2500) of philips 3c8 (?)
or equivalent.
using Bmax of about 2kG, and Ac=2cm^2, f=50khz
we need a minimum of 4 turns on this core.
this low number is nice, but it is a double-edged sword.
unless you want to play with fractional turns,
it does not give you a lot of granularity in your secondary
voltages.
we're talking about 35V/4T = 8.75 volt per turn.
if you want to do low voltages like 5VAC, 6.3VAC, etc
then you have to increase this number a bit...
i'd look at the maximum differential i'd allow between
the closest voltages.
i.e. 4.8VAC for the 5V and 6.4V for the 6.3V.
that's 1.6V /T
then 3T for 4.8VAC and 4T for 6.4VAC.
and about 22T primary.
(ok, it's getting ugly. probably best to do separate
trannies for 5V and 6.3V)
i'd shoot for 1T or 2T for the 5VAC output.
back into 7T or 14T primary.
keep in mind that you'd adjust the input for the right
trim...
so lets stick with 7T on the primary. ditch the 6.3VAC
winding for now.
(maybe rectify/regulate
a 5VAC (1T) or 10VAC (2T) winding the 6.3V needs)
i'd be a little nervous about the symmetyry of our
signal generator/amp (offsets will cause unbalanced DC).
so, to avoid flux-walking, i'd use a modest gap in the core.
the 3622 comes with various standard gaps
and are identified by their specific inductance value
in mH/1000T.
suppose we use a -1000 (1000mH/1000T).
with 7 turns, the primary inductance will be about 0.05mH.
this is about 15k at 50khz. so the magnitizing inductance
will be low...
so we have a core
size: 3622
material: P of 3C8, 3C81
Al/Gap: 1000
what about the wire?
6Arms suggest about 15AWG wire. but you can skimp without
major penalty down to 16AWG.
it just so happens that 7T of #16AWG fits in one layer
on the 3622 bobbin.
in practice, though, we'd probably use something like
3 #20AWG in parallel to make the wind easier.
you can also interleave these windings between the
secondaries for better coupling.
i've done #16 on an even smaller core, so let's stick
with that.
the rest of the wind, is simply a matter of applying the
5V/T ratio. there will be a physical limitation to the
number of secondary turns.
this is where pot cores suffer. not as much window area as,
say, an EC41.
but the 7T of #16 and a layer of tape will only take about
20 to 25% of the total area.
you should easily fit 100T of #28AWG for B+. and several
filament windings...
> I need to know the best readily available Core for low distortion and what
>
> primary winding is needed. I can figure the rest myslef....
>
yeah, there's the rub.
it is pretty difficult to buy onesies and twosies of these cores.
mostlty available thru wholesale distributers in qty 50, 100, 1000, etc...
> >in any case, i wonder how much power you can really get out of
> >a LM3875, LM675, LM3886
>
> The 3886 will easily do 50W RMS prolonged, assuming a decent heatsink.
>
yeah. but i thought you were talking 200W...
so you'd need more than one...
no biggie.
i'm thinking more in terms of using separate transformers
and even separate amps for filaments, driver B+, output B+, etc.
this way you can optimise the amp/tranny for the application.
youcould even build in a delay on the B+ to prevent
cathode stripping.... ;)
oh well
back to work
(modelling the power system for our next probe to europa)
bob.d.
=========================================================================
From: "T. Loesch" <ezee_e@hotmail.com>
Subject: RE: [JN] 400Hz or so Sinewave based PSU for DHT's - over-rating t ransformers????
Date: Wed, 23 Feb 2000 19:27:39 GMT
Source: Sound Digest Archive v02.n460
Yo Dude,
> > "Why not higher...?"
>
>two word: core loss.
How about these Ferrite EE42 Cores in my Catalogue where the text proclaims
"up to 300VA and 300kHz" and notes a Al of 3500?
>using Bmax of about 2kG, and Ac=2cm^2, f=50khz
>we need a minimum of 4 turns on this core.
>
>this low number is nice, but it is a double-edged sword.
Wind it so that our three 5V secondaries can be one turn each,
centertapped.... (300B X 2, 5U4) and hence with a few turns more, like seven
(35V : 5V) or even 14 Turns Primary (split into two X 7 Turns) allowing us
two turns for the 5V Secondaries.
This way all those who wire everything in solid silver can afford to make
the 300B Heater Winding from Solid Silver in Teflon without breaking the
Bank.... ;-)
Now, with about 350 to 370V RMS X 2 being the output for the HT, we need 2 X
140 or so Turns, still very managable.
>i'd be a little nervous about the symmetyry of our
>signal generator/amp (offsets will cause unbalanced DC).
I'm not.
>it is pretty difficult to buy onesies and twosies of these cores.
I can get quite a nice selection from www.rswww.com.....
> > The 3886 will easily do 50W RMS prolonged, assuming a decent >heatsink.
> >
>yeah. but i thought you were talking 200W...
Parallel four per side, two batches of four in Bridge - gives 35 - 40V and
current till the get go....
Later Thorsten
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
=========================================================================
From: "Mike Wurtz" <mikewurtz@mindspring.com>
Subject: Re: [JN] 400Hz or so Sinewave based PSU for DHT's - over-rating transformers????
Date: Wed, 23 Feb 2000 15:16:21 -0600
Source: Sound Digest Archive v02.n461
for the filament? oh,
i thought you were looking for super low noise regulated DC power supplies.
never mind.
mike wurtz
- -----Original Message-----
From: T. Loesch <ezee_e@hotmail.com>
To: mikewurtz@mindspring.com <mikewurtz@mindspring.com>;
robert.m.danielak@lmco.com <robert.m.danielak@lmco.com>; sound@lists.io.com
<sound@lists.io.com>
Date: Wednesday, February 23, 2000 2:23 PM
Subject: Re: [JN] 400Hz or so Sinewave based PSU for DHT's - over-rating
transformers????
>Hi there,
>
>>you can do supplies with less noise than a linear with these.
>>(unless you filter the linear with an LC :^)
>
>Actually, with the 50kHz linear supply i hope to do full Choke Input
>Supplies.... No Polarised Capacitors, AC on the Filaments and the first
>"hum" component at 100khz....
>
>And all air-core Inductors in the Filter....
>
>See where I'm comming from?
>
>Later Thorsten
>______________________________________________________
>Get Your Private, Free Email at http://www.hotmail.com
>
=========================================================================
From: "Danielak, Robert M" <robert.m.danielak@lmco.com>
Subject: RE: [JN] 400Hz or so Sinewave based PSU for DHT's - over-rating t ransformers????
Date: Wed, 23 Feb 2000 15:31:50 -0500
Source: Sound Digest Archive v02.n461
> Yo Dude,
>
> > > "Why not higher...?"
> >
> >two word: core loss.
>
> How about these Ferrite EE42 Cores in my Catalogue where the text
> proclaims
> "up to 300VA and 300kHz" and notes a Al of 3500?
>
those should be good for conservatively 350W at 50khz.
way more at 300khz.
these appear to be un-gapped. Al suggest a u2500 mat'l.
should be good to go...
but i'd still recommend at least a small gap...
> >using Bmax of about 2kG, and Ac=2cm^2, f=50khz
> >we need a minimum of 4 turns on this core.
> >
> >this low number is nice, but it is a double-edged sword.
>
> Wind it so that our three 5V secondaries can be one turn each,
> centertapped.... (300B X 2, 5U4) and hence with a few turns more, like
> seven
> (35V : 5V) or even 14 Turns Primary (split into two X 7 Turns) allowing us
>
> two turns for the 5V Secondaries.
>
ok.
you don't really want to center tap a single turn, do you?
fractional turns can be done, but it's a pain.
so we'd do 2T center-taped. which means 14T.
on the taller EE42 core you could probably fit it in a single layer.
> This way all those who wire everything in solid silver can afford to make
> the 300B Heater Winding from Solid Silver in Teflon without breaking the
> Bank.... ;-)
>
?!?
> Now, with about 350 to 370V RMS X 2 being the output for the HT, we need 2
> X
> 140 or so Turns, still very managable.
>
yes, but wouldn't mess with anything smaller than #30AWG.
> >i'd be a little nervous about the symmetyry of our
> >signal generator/amp (offsets will cause unbalanced DC).
>
> I'm not.
>
ok. suit yourself....
(in flame retardant gear and goggles)
well, of course, i'm exaggerating. but without a gap,
and especially with more and more turns we're adding,
it's not going to take much to saturate the core.
even with a relatively large gap, we'd still have plenty
of primary inductance for our application.
sure, the gain doesn't have to be too large, so input offset
drift might not be too bad, but you never know...
> >it is pretty difficult to buy onesies and twosies of these cores.
>
> I can get quite a nice selection from www.rswww.com.....
>
i'll have to check them out...
> > > The 3886 will easily do 50W RMS prolonged, assuming a decent
> >heatsink.
> > >
> >yeah. but i thought you were talking 200W...
>
> Parallel four per side, two batches of four in Bridge - gives 35 - 40V and
>
> current till the get go....
>
hmm....
can you really parallel these things?
i guess i've seen it done in some SS car amps.
unless that was four channels with bridge outputs...
good luck,
bob.d.
=========================================================================
From: "Danielak, Robert M" <robert.m.danielak@lmco.com>
Subject: RE: [JN] 400Hz or so Sinewave based PSU for DHT's - over-rating t ransformers????
Date: Wed, 23 Feb 2000 17:01:39 -0500
Source: Sound Digest Archive v02.n461
> Bob,
>
> >but i'd still recommend at least a small gap...
>
> Okay, so EE42 Core, small Gap. How many turns minimum on the primary for
> this shooting match? Are 14 turns enough?
>
well.
there are two listings for the E42.
there's a 42/15 which has an Ac of 1.83cm^2
and a 42/20 with Ac = 2.37cm^2.
the 42/15 will need about 5 turns, min.
the 42/20 will need <4.
so 14 should be fine.
unfortunately, these cores don't seem to come pre-gapped,
except by special order.
> >?!?
>
> Well, if we keep the heaters as AC (should be no big deal with the first
> product from the heater thermal modulation being 100kHz AND way lower than
>
> anything you get with 50Hz or 60Hz AC heating), then our cathode return
> goes
> via our Transformer Winding (hence Centertapped). Make this a suitable
> piece
> of Silverwire, voiala.
>
yes, i'm with you.
keep filaments, AC
(although i seem to recall someone experimenting with HF AC who found
serious radiated emission problems).
my only ?!? was "why silver wire? why teflon?"
> Now try this with a 60Hz based transformer....
>
i guess you figure it would be more turns.
not many, really. but each turn *would* be longer... ;)
> >yes, but wouldn't mess with anything smaller than #30AWG.
>
> I think we will need something ever so slightly larger for >200mA draw
> from
> the HT anyway.
>
not really. a wise man once told me that "500cma is for wimps".
but it's a moot point now since you have a lot more window area
in that E42 core than in my pot core example.
the problem with smaller than 30AWG is simply that it's
difficult to work with.
i was winding a small IT using 39AWG a few years back.
i must have gone 100 turns before i realized the wire broke...
> >can you really parallel these things?
>
> Yes. You use a 0.1 Ohm non Inductive Resistor in each output, independednt
>
> Feedback Loops and then a overall feedback loop.
>
ugh.
i'd just assume doing a discrete current booster at the output
of a single (much better) op-amp.
> I'd say wire everything 47
> Lab's style with minimum leadlength and no bandwidthlimiting and we are in
>
> business.
>
sure. why not.
i started doing my switchers that way, too.
no PCB. just a metal heatsink/substrate and solder lead-to-lead.
3 dimensional is the way to go...
> With 8 Chips we either cut the thermal load (probably a good idea) or we
> can
> get around 400VA form the whole joint, aka 11A.... I personally would go
> for
> the lower thermal load.
>
probably not a bad idea...
let me back up a minute, though.
what are we using for a power source up front.
we still need a nice stiff +/-40V or so to feed this monster.
so what are we accomplishing here?
we have double conversion going on here.
we're sweeping the water uphill so it can come down again ;)
actually, i get the point that we want to try HF AC
on the filaments to avoid LF hum problem without resorting to
DC.
and i suppose we want to get rid of the 120Hz ripple on our
B+ without resorting to regulation.
then we're counting on the PSRR of the amplifier to reject
crap from the up-front supply.
all good stuff, i suppose.
but why not go all out and do an "offline" version of this.
bypass the intermediate +/-40V supply.
your final amp stage must then work from 170VDC (US) and 320VDC
(EU).
gotta run,
bob.d.
=========================================================================
From: "PEARL Cust Serv" <custserv@pearl-hifi.com>
Subject: Re: [JN] Re: 400Hz or so Sinewave based PSU for DHT's - over-rating transformers?
Date: Wed, 23 Feb 2000 18:18:37 -0700
Source: Sound Digest Archive v02.n461
connlyra@gol.com wrote:
> Thorsten,
>
> Man, you sure change direction a lot. Big difference between 400Hz and 50kHz.
Hi Thorsten, All:
Watching this thread, I have to agree with that which appears above.
To me, the notion of a 3 phase, 400hz-to-a-few-kHz solution seems to be the
one . . .
At 400 Hz a full-wave rectified, 3-phase output appears as a DC Level
with a 4% ripple at 2400Hz. This is less than 8% of the ripple of a similar,
single phase implementation and this ripple appears at TWENTY times the
frequency of a 60Hz, single phase implementation
Combining these two factors, it can be seen that the ripple component in
the "DC" achieved by rectifying a 400Hz, 3 phase source will be something on
the order of 200 times easier to deal with than that of the standard, 60 Hz,
single phase case.
That's 46dB improvement right from the git go . . . . and never mind the
fact that the 400Hz lines have been derived by clean, power oscillators.
Wrt to the requirement for 3 voltage sources 120 deg apart, this is a
nearly trivial task as there is no apparent need at all for the trio to be
in precise phase relationship nor track to in precise amplitude.
Two oscillators respectively driven from a master thru 12 and 24dB/oct
filters selected on their frequency response curves for the 120 and 240 deg
points of phase shift would easily satisfy the phase requirements and
simple, gain massage or resistive voltage division of the outputs of
the buffers following the three oscillators would put the trio in
sufficient "amplitude lock" for a listener with even a highly over-active
imagination . . . :>)
So, my vote is that we stop tilting at windmills, call perfectly
adequate just that . . . . and get on with it.
Btw, due to a crack addicted landlord we were forced to move my
extensive wood/metal shop away from its 30 yr. home into a space that, for
the moment, does not have three phase power.
In response we just whipped up a 5hp, rotary, single- to 3-phase,
convertor one afternoon. It uses a 5Hp "idler" motor to "generate the third
phase and with a few compensation caps to provide voltage and phase
balancing once the idler is up to speed, it works just FINE! ! !
So if you REALLY want three phase power, you can have in . . . easily
and cheaply.
Bill - PEARL, Inc.
=========================================================================
From: "T. Loesch" <ezee_e@hotmail.com>
Subject: Re: [JN] 400Hz or so Sinewave based PSU for DHT's - over-rating transformers????
Date: Wed, 23 Feb 2000 20:22:51 GMT
Source: Sound Digest Archive v02.n461
Hi there,
>you can do supplies with less noise than a linear with these.
>(unless you filter the linear with an LC :^)
Actually, with the 50kHz linear supply i hope to do full Choke Input
Supplies.... No Polarised Capacitors, AC on the Filaments and the first
"hum" component at 100khz....
And all air-core Inductors in the Filter....
See where I'm comming from?
Later Thorsten
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
=========================================================================
From: "T. Loesch" <ezee_e@hotmail.com>
Subject: RE: [JN] 400Hz or so Sinewave based PSU for DHT's - over-rating t ransformers????
Date: Wed, 23 Feb 2000 20:49:08 GMT
Source: Sound Digest Archive v02.n461
Bob,
>but i'd still recommend at least a small gap...
Okay, so EE42 Core, small Gap. How many turns minimum on the primary for
this shooting match? Are 14 turns enough?
>you don't really want to center tap a single turn, do you?
Hey, IT CAN BE DONE - but I agree It ain't fun.
> > This way all those who wire everything in solid silver can afford > to
>make the 300B Heater Winding from Solid Silver in Teflon > without breaking
>the Bank.... ;-)
> >
>?!?
Well, if we keep the heaters as AC (should be no big deal with the first
product from the heater thermal modulation being 100kHz AND way lower than
anything you get with 50Hz or 60Hz AC heating), then our cathode return goes
via our Transformer Winding (hence Centertapped). Make this a suitable piece
of Silverwire, voiala.
Now try this with a 60Hz based transformer....
>yes, but wouldn't mess with anything smaller than #30AWG.
I think we will need something ever so slightly larger for >200mA draw from
the HT anyway.
>can you really parallel these things?
Yes. You use a 0.1 Ohm non Inductive Resistor in each output, independednt
Feedback Loops and then a overall feedback loop. I'd say wire everything 47
Lab's style with minimum leadlength and no bandwidthlimiting and we are in
business.
With 8 Chips we either cut the thermal load (probably a good idea) or we can
get around 400VA form the whole joint, aka 11A.... I personally would go for
the lower thermal load.
Later Thorsten
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
=========================================================================
From: "T. Loesch" <ezee_e@hotmail.com>
Subject: RE: [JN] 400Hz or so Sinewave based PSU for DHT's - over-rating t ransformers????
Date: Wed, 23 Feb 2000 22:29:39 GMT
Source: Sound Digest Archive v02.n461
Bob,
>my only ?!? was "why silver wire? why teflon?"
Well, some people like Silver/Teflon, I like Copper Polyethylen.... We have
all got our favourite wires for good sound....
> > Now try this with a 60Hz based transformer....
> >
>i guess you figure it would be more turns.
Yes, and where four (overall) small turns in silver don't really count
(moneywise), a few dozend large (at the very least) count quite heavily....
> > >can you really parallel these things?
<snip>
>ugh.
>
>i'd just assume doing a discrete current booster at the output
>of a single (much better) op-amp.
Well, there you have no overvoltage, overtemperature, SOAR and god knows
what protection. Only the superfast fuse, thet to save your transistors will
lay heroically down their live quite easily, just give them halve a
chance.... ;-)
Especially on stuff for more or less experimentation I like to have any
conceivable protection circuits (and then some).
>what are we using for a power source up front.
>
>we still need a nice stiff +/-40V or so to feed this monster.
>
>so what are we accomplishing here?
Well, we keep the 50Hz mostly out of our Amp Chassis, use only sonically
good non polarised Capacitors (I go for my specially made Ansar Supersound
Polypropylen - the oilers can use their PCB filled relics), make for a
fundamentally unstressed supply and get exceptionally good regulation
without regulators. Sounds like a "Win/Win" situation to me.
>we have double conversion going on here.
>we're sweeping the water uphill so it can come down again ;)
Of course. But if we otherwise make a 50 Hz (60Hz for the US) syntesizer to
get clean power? And remeber, I wanted to do an Autotransformer to give use
a single +70V or so under full load (remeber - we are bridged, so one
Voltage will do fine), with very small value capacitors in the PSU to keep
the noise send back into the mains low.... With > 120db PSRR on the Positive
Pin I don't worry much about the Ripple on the PSU line.
>actually, i get the point that we want to try HF AC
>on the filaments to avoid LF hum problem without resorting to
>DC.
Yup.
>and i suppose we want to get rid of the 120Hz ripple on our
>B+ without resorting to regulation.
Yup. Even better, check out the regulation with a 2.2mH Air Core Choke as
input Choke and a 22uF Film Capacitor in the HT....
>then we're counting on the PSRR of the amplifier to reject
>crap from the up-front supply.
And our HF transformer which will also make sure very little low level AF
range crap comes through.
>your final amp stage must then work from 170VDC (US) and 320VDC
>(EU).
Yes. And there we have no longer got a Voltage that is quite safe to touch
(assuming - is earthed), no convenient fully protected Chip's to make our
Amp and so on.
Now for you and me, that is an Option.... Good old Joe "I wonder what
happens if I plug this Valve in here" wouldnt get much milage out of
this.... The Chip-Version with fully protected Chips will just shut down if
it gets too hot, will be mostly idiot proof and pretty easy to assemble....
Sound cool to me.
Later Thorsten
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
=========================================================================
From: connlyra@gol.com
Subject: [JN] Re: 400Hz or so Sinewave based PSU for DHT's - over-rating transformers?
Date: Thu, 24 Feb 2000 08:32:24 +0900
Source: Sound Digest Archive v02.n461
Thorsten,
Man, you sure change direction a lot. Big difference between 400Hz and 50kHz.
At any rate, you may be interested in checking out some of Akihiko
Kaneda's MJ (Musen to Jikken) designs from the early 80s. He was
using designing switching preregulators that operated at around 400 -
550kHz, using Soshin pot cores, and listening tests played a big part
in his designs. I recall something like 200 or 210uH, but it has been
a long time since I looked at any of those schematics.
In more recent designs, have you looked at the Linn Klimax? It too
uses switching technology (although I am pretty sure that it does not
use an off-the-shelf solution), and sounds pretty decent for a
high-powered SS amp. The newer Jeff Roland power amps combine
switching regulators with PFC, but I have heard that they are
utilizing commercially available solutions. Not absolutely sure,
though...
=========================================================================
From: "Mike Wurtz" <mikewurtz@mindspring.com>
Subject: Re: [JN] Re: 400Hz or so Sinewave based PSU for DHT's - over-ratingtransformers?
Date: Thu, 24 Feb 2000 08:05:03 -0600
Source: Sound Digest Archive v02.n462
truly, if you guys don't know about the LT1533 and lt1534, and you want
clean DC supplies you should look at them.
they won't work at 400kHz, but man, that's a whole can of worms that i don't
think is worth getting into. maybe 200khz for off line.
these controllers are not huge power becuase they have integrated switches.
but for preamps and such are perfect. (off line has been done with these,
but its not for novices.)
they switch at like 30-50kHz if you truly want low noise. the harmonics are
very low, and the effeciency is 15% lower than what you can do if you switch
like a raped ape.
Linear technology is working on a controller using the same current and slew
switching.
they are in effect "linear" switching regulators. they don't just try to
slam the switch to the next state, they do in under closed loop control to
minimize all the offenses.
don't know if many of you have tried, but if you switch at 400kHz, you
cannot get rid of this.
you can filter with god's own capacitors, and radiation from the swich node
will appear.
the LT1534 is truly low noise. like 50uV out to 100Mhz. i think that's on
the data sheet example that williams did. i didn't see, nor can measure
that, but it looked pretty darn close to a grounded probe to me!
mike wurtz
- -----Original Message-----
From: T. Loesch <ezee_e@hotmail.com>
To: connlyra@gol.com <connlyra@gol.com>; sound@io.com <sound@io.com>
Date: Thursday, February 24, 2000 4:46 AM
Subject: Re: [JN] Re: 400Hz or so Sinewave based PSU for DHT's -
over-ratingtransformers?
>Jonathan,
>
>>Man, you sure change direction a lot.
>
>Not really.
>
>>Big difference between 400Hz and 50kHz.
>
>Yes and no.
>
>>At any rate, you may be interested in checking out some of Akihiko
>>Kaneda's MJ (Musen to Jikken) designs from the early 80s. He was
>>using designing switching preregulators that operated at around 400
550kHz,
>>using Soshin pot cores, and listening tests played a big part in his
>>designs.
>
>Did he ever try non switching types (as proposed here?).
>
>>In more recent designs, have you looked at the Linn Klimax?
>
>No, but I listened to one and was deepely unimpressed.... Perhaps just a
>problem with Linn's audio circuitry, but I thought this amp made Krell
sound
>good....
>
>Then again, ever since Linn's "Brilliant" Powersupplies the sound of their
>gear went from just not very good to dowright offensive.... Hence my view
of
>not having a SWITCHING supply, but a linear one, albeit at relatively high
>frequencies.
>
>Later Thorsten
>______________________________________________________
>Get Your Private, Free Email at http://www.hotmail.com
>
=========================================================================
From: "spence barton" <ence-ack@rio.com>
Subject: Re: [JN] Re: 400Hz or so Sinewave based PSU for DHT's - over-ratingtransformers?
Date: Thu, 24 Feb 2000 09:02:33 -0800
Source: Sound Digest Archive v02.n462
Frank said:
>Ever sat in a jetliner
>and noticed the background noise with the engines off? That out-of-tune Ab
>chord is pretty darned annoying to me! (Aircraft use 400Hz powersystems to
>save weight in powersupply components.)
As one who has a earpiece with 400hz by-products in his ear all of his
working day, I can truly say it is not something anyone would want in an
audio system. 60/120 is way less intrusive. Having said that I have never
heard power supply by-products that I like.
>My advice: git thee out of the audio band!
That would be really good.
spence
=========================================================================
From: Frank_Deutschmann@trepp.com
Subject: Re: [JN] Re: 400Hz or so Sinewave based PSU for DHT's - over-rating transformers?
Date: Thu, 24 Feb 2000 10:44:17 -0500
Source: Sound Digest Archive v02.n462
> At 400 Hz a full-wave rectified, 3-phase output appears as a DC Level
>with a 4% ripple at 2400Hz. This is less than 8% of the ripple of a
similar,
>single phase implementation and this ripple appears at TWENTY times the
>frequency of a 60Hz, single phase implementation
> Combining these two factors, it can be seen that the ripple component
in
>the "DC" achieved by rectifying a 400Hz, 3 phase source will be something
on
>the order of 200 times easier to deal with than that of the standard, 60
Hz,
>single phase case.
> That's 46dB improvement right from the git go . . . . and never mind
the
>fact that the 400Hz lines have been derived by clean, power oscillators.
That's all fine and good, but do keep in mind that according to Mssrs.
Fletcher and Munson and a host of other people who have followed them, a
signal in the realm of 2kHz would have to be roughly 40-50dB down in
amplitude from a signal at 60Hz in order to be experienced as having a
similar subjective volume (i.e.: equivallent amplitude in phons). This, of
course, assumes we are concerned with small amplitude 60/120Hz sound.
And that's not to mention that 2400Hz is an "annoyance" that we are not
accustomed to: many/most of us have heard so much 60/120Hz hum that we
literally "tune it out"; I doubt spurious 2400Hz noise would be experienced
similarly, even once we accomodate Fletcher-Munson. Ever sat in a jetliner
and noticed the background noise with the engines off? That out-of-tune Ab
chord is pretty darned annoying to me! (Aircraft use 400Hz powersystems to
save weight in powersupply components.)
My advice: git thee out of the audio band!
- -frank
=========================================================================
From: "T. Loesch" <ezee_e@hotmail.com>
Subject: Re: [JN] Re: 400Hz or so Sinewave based PSU for DHT's - over-rating transformers?
Date: Thu, 24 Feb 2000 10:46:13 GMT
Source: Sound Digest Archive v02.n462
Jonathan,
>Man, you sure change direction a lot.
Not really.
>Big difference between 400Hz and 50kHz.
Yes and no.
>At any rate, you may be interested in checking out some of Akihiko
>Kaneda's MJ (Musen to Jikken) designs from the early 80s. He was
>using designing switching preregulators that operated at around 400 550kHz,
>using Soshin pot cores, and listening tests played a big part in his
>designs.
Did he ever try non switching types (as proposed here?).
>In more recent designs, have you looked at the Linn Klimax?
No, but I listened to one and was deepely unimpressed.... Perhaps just a
problem with Linn's audio circuitry, but I thought this amp made Krell sound
good....
Then again, ever since Linn's "Brilliant" Powersupplies the sound of their
gear went from just not very good to dowright offensive.... Hence my view of
not having a SWITCHING supply, but a linear one, albeit at relatively high
frequencies.
Later Thorsten
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
=========================================================================
From: "PEARL Cust Serv" <custserv@pearl-hifi.com>
Subject: Re: [JN] Re: 400Hz or so Sinewave based PSU for DHT's - over-rating transformers?
Date: Thu, 24 Feb 2000 11:15:16 -0700
Source: Sound Digest Archive v02.n462
Frank Deutschmann@trepp.com wrote:
- -----------------
Snipped bit added back in:
Hi Thorsten, All:
Watching this thread, I have to agree with that which appears above.
To me, the notion of a 3 phase, 400hz-to-a-few-kHz solution seems to be the
one . . .
End of addition.
- -----------------
At 400 Hz a full-wave rectified, 3-phase output appears as a DC Level
with a 4% ripple at 2400Hz. This is less than 8% of the ripple of a similar,
single phase implementation and this ripple appears at TWENTY times the
frequency of a 60Hz, single phase implementation
Combining these two factors, it can be seen that the ripple component in
the "DC" achieved by rectifying a 400Hz, 3 phase source will be something on
the order of 200 times easier to deal with than that of the standard, 60 Hz,
single phase case.
That's 46dB improvement right from the git go . . . . and never mind the
fact that the 400Hz lines have been derived by clean, power oscillators.
>
> That's all fine and good, but do keep in mind that according to Mssrs.
> Fletcher and Munson and a host of other people who have followed them, a
> signal in the realm of 2kHz would have to be roughly 40-50dB down in
> amplitude from a signal at 60Hz in order to be experienced as having a
> similar subjective volume (i.e.: equivallent amplitude in phons). This, of
> course, assumes we are concerned with small amplitude 60/120Hz sound.
I'm more concerned with the elimination of the huge "gaps" in fullwave
rectified, single phase, pulsating "DC". These are VERY difficult to "fill
in" in sonically transparent manner. . .
> And that's not to mention that 2400Hz is an "annoyance" that we are not
> accustomed to: many/most of us have heard so much 60/120Hz hum that we
> literally "tune it out"; I doubt spurious 2400Hz noise would be experienced
> similarly, even once we accomodate Fletcher-Munson. Ever sat in a jetliner
> and noticed the background noise with the engines off? That out-of-tune Ab
> chord is pretty darned annoying to me! (Aircraft use 400Hz powersystems to
> save weight in powersupply components.)
I'm going to suggest that those systems were not designed with the
requirements of high-end audio in mind . .
> My advice: git thee out of the audio band!
>
> -Frank
- --------------------
Hi Frank:
I must direct your attention to the opening paragraph of my original
post, inserted here as an addition to your post.
Therein, I stated, "To me, the notion of a 3 phase, 400hz-to-a-few-kHz
solution seems to be the one . . ."
While your comments vis the audibility of a 2.4kHz artifact are well
taken, I chose the 400Hz case to illustrate some potential benefits in the
worst case.
The master frequency could easily be raised up to several kHz before
transformer-core losses became an issue but such a problem is easily
scotched by the use of .004" tape in the c-cores for both the PT and
following chokes.
I have an issue with going much higher than that which is, simply
stated, why?
As frequency is increased spurious EMI tends to "spray" to a much
greater extent and this can be VERY difficult to control. Attempts at such
control are seen as the tremendously tight shielding around "switchers."
This polyphase, linear PS idea is VERY intriguing and one that I will
build at some point.
We're all involved in a bit of a theoretical speculation here, never the
less my gut feeling is that this will work very well.
. . . . and, changing the frequency of the master oscillator in such a
supply could be made as simple as adjustment by a pot so one could play at
some length.
Bill - PEARL, Inc.
=========================================================================
From: Frank_Deutschmann@trepp.com
Subject: Re: [JN] Re: 400Hz or so Sinewave based PSU for DHT's - over-rating transformers?
Date: Fri, 25 Feb 2000 12:51:56 -0500
Source: Sound Digest Archive v02.n464
Not to get us too far off the topic of roaches, I'd like to just point out
a few things....
"PEARL Cust Serv" <custserv@pearl-hifi.com>@lists.io.com on 02/24/2000
01:15:16 PM
Sent by: owner-sound@lists.io.com
To: sound@lists.io.com
cc:
Subject: Re: [JN] Re: 400Hz or so Sinewave based PSU for DHT's -
over-rating transformers?
Frank Deutschmann@trepp.com wrote:
- -----------------
Snipped bit added back in:
Hi Thorsten, All:
Watching this thread, I have to agree with that which appears above.
To me, the notion of a 3 phase, 400hz-to-a-few-kHz solution seems to be the
one . . .
End of addition.
- -----------------
At 400 Hz a full-wave rectified, 3-phase output appears as a DC Level
with a 4% ripple at 2400Hz. This is less than 8% of the ripple of a
similar,
single phase implementation and this ripple appears at TWENTY times the
frequency of a 60Hz, single phase implementation
Combining these two factors, it can be seen that the ripple component
in
the "DC" achieved by rectifying a 400Hz, 3 phase source will be something
on
the order of 200 times easier to deal with than that of the standard, 60
Hz,
single phase case.
That's 46dB improvement right from the git go . . . . and never mind
the
fact that the 400Hz lines have been derived by clean, power oscillators.
>
> That's all fine and good, but do keep in mind that according to Mssrs.
> Fletcher and Munson and a host of other people who have followed them, a
> signal in the realm of 2kHz would have to be roughly 40-50dB down in
> amplitude from a signal at 60Hz in order to be experienced as having a
> similar subjective volume (i.e.: equivallent amplitude in phons). This,
of
> course, assumes we are concerned with small amplitude 60/120Hz sound.
I'm more concerned with the elimination of the huge "gaps" in fullwave
rectified, single phase, pulsating "DC". These are VERY difficult to "fill
in" in sonically transparent manner. . .
>>>>> Bill, you missed my point: your 46dB of improvement is already eaten
up by the fact that our ears/brain are roughly 40-50dB more sensitive to
2400Hz compared to 60/120Hz; i.e.: its a wash moving to the higher
frequency, in terms of audible ripple from the power supply, measured in
phons.
> And that's not to mention that 2400Hz is an "annoyance" that we are not
> accustomed to: many/most of us have heard so much 60/120Hz hum that we
> literally "tune it out"; I doubt spurious 2400Hz noise would be
experienced
> similarly, even once we accomodate Fletcher-Munson. Ever sat in a
jetliner
> and noticed the background noise with the engines off? That out-of-tune
Ab
> chord is pretty darned annoying to me! (Aircraft use 400Hz powersystems
to
> save weight in powersupply components.)
I'm going to suggest that those systems were not designed with the
requirements of high-end audio in mind . .
>>> Again, you miss the point: all the high-end tweakery in the world will
not make 400Hz and its harmonics sound good! This is an extremely
discordant frequency: it is roughly half way between the Ab and G above
middle-C, and naturally none of the higher harmonics are in tune, either.
Perhaps 400H is a natural note in some (eastern?) music scale, but for
those of us vastly more familiar with western music, 400Hz, and any chord
based on it, sounds extremely off key. Consequently, this frequency and
its harmonics will stick out even more than Fletcher-Munson indicate; the
smallest amount of these frequencies leaking into the signal chain will
wreak havoc with the music: from this perspective, 60/120Hz is a far better
choice, as a) we're less sensitive to those frequencies, and b) we're so
used to it that we tune them out anyway.
[snip]
I have an issue with going much higher than that which is, simply
stated, why?
As frequency is increased spurious EMI tends to "spray" to a much
greater extent and this can be VERY difficult to control. Attempts at such
control are seen as the tremendously tight shielding around "switchers."
>>> Pro tape decks almost invariably (IME) use an erase oscilator of
200kHz; they don't use an erase osc anywhere near the audio band, though
they could, for *very* good reasons that are readily aparent if you think
about them for a while. Namely, anything within the audio band runs great
risk of being easily perceived, and that anything below 200kHz poses a
problem of audio harmonics mixing with the osc and ending up back in the
audio baseband: 200kHz gives protection from even substantial levels of the
5th harmonic of an audio signal.
I'm confident that if you experiment with this at all, you would rapidly
find that "EMI spray" is a far lesser evil than tossing inharmonic trash
into the audio spectrum is! Further, as concerns switchers, you're not
building a switching supply, now are you?! Very different problem: your
linear supply has an inherent advantage over a switcher in this regard, and
the three-phase approach improves further on the linear single phase
approach; diode switching noise should be vastly reduced, even vis a vis a
60Hz linear. More specifically, I think you will find that 200kHz is an
excellent frequency for this project.
And now, back to the roaches....
- -frank
=========================================================================
From: "PEARL Cust Serv" <custserv@pearl-hifi.com>
Subject: Re: [JN] Re: 400Hz or so Sinewave based PSU for DHT's - over-rating transformers?
Date: Fri, 25 Feb 2000 14:29:30 -0700
Source: Sound Digest Archive v02.n465
Frank Deutschmann@trepp.com wrote:
SNIP
> Bill, you missed my point: your 46dB of improvement is already eaten
> up by the fact that our ears/brain are roughly 40-50dB more sensitive to
> 2400Hz compared to 60/120Hz; i.e.: its a wash moving to the higher
> frequency, in terms of audible ripple from the power supply, measured in
> phons.
I beg to differ, I didn't miss your point at all, rather you seem to
have missed mine:
I simply picked the 400 Hz example to give a worst case appraisal of the
potential for improvement. As I said in my earlier reply, this polyphase
supply could easily be made to run at 10Khz in which case the ripple
frequency would be 60kHz and that is, to me, the frequency of interest.
>> And that's not to mention that 2400Hz is an "annoyance" that we are not
>> accustomed to: many/most of us have heard so much 60/120Hz hum that we
>> literally "tune it out"; I doubt spurious 2400Hz noise would be experienced
>> similarly, even once we accomodate Fletcher-Munson. Ever sat in a
> jetliner
>> and noticed the background noise with the engines off? That out-of-tune
> Ab
>> chord is pretty darned annoying to me! (Aircraft use 400Hz powersystems
> to
>> save weight in powersupply components.)
> I'm going to suggest that those systems were not designed with the
> requirements of high-end audio in mind . .
>
> Again, you miss the point: all the high-end tweakery in the world will
> not make 400Hz and its harmonics sound good! This is an extremely
> discordant frequency: it is roughly half way between the Ab and G above
> middle-C, and naturally none of the higher harmonics are in tune, either.
> Perhaps 400H is a natural note in some (eastern?) music scale, but for
> those of us vastly more familiar with western music, 400Hz, and any chord
> based on it, sounds extremely off key. Consequently, this frequency and
> its harmonics will stick out even more than Fletcher-Munson indicate; the
> smallest amount of these frequencies leaking into the signal chain will
> wreak havoc with the music: from this perspective, 60/120Hz is a far better
> choice, as a) we're less sensitive to those frequencies, and b) we're so
> used to it that we tune them out anyway.
Well . . . that's fine as far as it goes but I was not talking about
dealing with 400Hz AC issues, I was discussing 2.4kHz issues and those only
as a worst case example AND - although I didn't say so - prior to extensive
filtering thru an "H" section r+C / L / C / L /C.
I'm not arguing your points, they are valid . . . . they are just not
relevant to the notions I'm trying to convey
>
> [snip]
>
> I have an issue with going much higher than that which is, simply
> stated, why?
> As frequency is increased spurious EMI tends to "spray" to a much
> greater extent and this can be VERY difficult to control. Attempts at such
> control are seen as the tremendously tight shielding around "switchers."
>
>
> Pro tape decks almost invariably (IME) use an erase oscilator of
> 200kHz; they don't use an erase osc anywhere near the audio band, though
> they could, for *very* good reasons that are readily aparent if you think
> about them for a while. Namely, anything within the audio band runs great
> risk of being easily perceived, and that anything below 200kHz poses a
> problem of audio harmonics mixing with the osc and ending up back in the
> audio baseband: 200kHz gives protection from even substantial levels of the
> 5th harmonic of an audio signal.
Certainly . . . there is no mystery here.
> I'm confident that if you experiment with this at all, you would rapidly
> find that "EMI spray" is a far lesser evil than tossing inharmonic trash
> into the audio spectrum is! Further, as concerns switchers, you're not
> building a switching supply, now are you?! Very different problem: your
> linear supply has an inherent advantage over a switcher in this regard, and
> the three-phase approach improves further on the linear single phase
> approach; diode switching noise should be vastly reduced, even vis a vis a
> 60Hz linear. More specifically, I think you will find that 200kHz is an
> excellent frequency for this project.
This needs to be built and listened to . . . .
>
> And now, back to the roaches....
Good idea . . . . now I recall that I have one around here somewhere or
other
> -frank
Bill - PEARL, Inc.
=========================================================================
From: connlyra@gol.com
Subject: [JN] Re: 400Hz or so Sinewave based PSU for DHT's - over-rating transformers?
Date: Sat, 26 Feb 2000 10:11:39 +0900
Source: Sound Digest Archive v02.n465
>All the high-end tweakery in the world will not make 400Hz and its
>harmonics sound good! This is an extremely discordant frequency: it
>is roughly half way between the Ab and G above middle-C, and
>naturally none of the higher harmonics are in tune, either. Perhaps
>400H is a natural note in some (eastern?) music scale, but for those
>of us vastly more familiar with western music, 400Hz, and any chord
>based on it, sounds extremely off key. Consequently, this frequency
>and its harmonics will stick out even more than Fletcher-Munson
>indicate; the smallest amount of these frequencies leaking into the
>signal chain will wreak havoc with the music: from this perspective,
>60/120Hz is a far better choice, as a) we're less sensitive to those
>frequencies, and b) we're so used to it that we tune them out
>anyway.<
Frank,
I assume that you have good reasons for the statements above, and
they may be valid ones.
However, I have actual experience in comparing a 400Hz AC power
system against a 60Hz one (and 50, 80, 120, and a few other
frequencies), thanks to the kind cooperation of the CSE company in
Japan and one of their chief stockholders.
In a nutshell, my findings are that 400Hz will sound better, _if_ the
mains transformer of the powered-up audio product is up to the task -
ie, will not hum, buzz or otherwise misbehave at the higher
frequency, and this latter point is by no means guaranteed.
I concede that different people and different systems have different
requirements and sensitivities. Also, not all implementations of
400Hz AC power sources have been designed with listening purposes in
mind. But when the 400Hz AC power source has been suitably
engineered, and the mains transformer can take it, I can verify that
400Hz AC power can work quite well.
Your own mileage, of course, may vary.
jonathan carr
=========================================================================
From: "Daniel J. Marshall" <danmarshall@worldnet.att.net>
Subject: Re: 414 in Petit Onken Was:Re: [JN] Push-pull 515Bs for Onken Cabinet
Date: Thu, 10 Jun 1999 20:13:46 -0700
Source: Sound Digest Archive v02.n175
Johari Yip wrote:
>
> ----- Original Message -----
> From: Daniel J. Marshall <danmarshall@worldnet.att.net>
> To: <sound@lists.io.com>
> Sent: Friday, June 11, 1999 2:40 AM
> Subject: Re: [JN] Push-pull 515Bs for Onken Cabinet
>
> Dan, fellow joenetters
>
> I'm reading your post with great interest because I have just acquired a
> pair of 414-8C in PA enclosure. I must agree with what you have said
> regarding your preference of the 414 over the 515. The 414 also has a
> "faster" response with lesser overhang because of a light cone with the same
> size of magnet. I have intention to partner they with the Oris150 crossing
> at about 200-220 Hz.
> Dan, one question that I would like to have an answer to is whether the
> Petite Onken, as appeared in SP issue #1Vol4, really is the optimum
> enclosure for the 414?
>
> Johari
Hi Johari,
I don't recall the volume of the Petit Onken. You should be able to
easily calculate it. And as far as optimum volume goes, I guess it
would depend on your preferences. If going for a "fast" sounding bass,
one might prefer a somewhat smaller enclosure volume tuned to a hoigher
frequency than if going for maximum bass extension. I built mine some
time ago, maybe fifteen or twenty years ago and did not do the
calculations myself. However, the chap who did went for maximum bass
extension which required about 7 cu ft. and I would say that he achieved
his goal, as they had excellent sounding bass response, being quite
extended, smooth and clean, though maller enclosures, tuned to a higher
frequency might suit some listeners preferences where the bottom octave
is not so important.
Dan Marshall
=========================================================================
From: Douglas Purl <dcp@selway.umt.edu>
Subject: Re: 414 in Petit Onken Was:Re: [JN] Push-pull 515Bs for Onken Cabinet
Date: Thu, 10 Jun 1999 20:41:48 -0600 (MDT)
Source: Sound Digest Archive v02.n175
On Fri, 11 Jun 1999, Johari Yip wrote:
> I'm reading your post with great interest because I have just acquired a
> pair of 414-8C in PA enclosure. I must agree with what you have said
> regarding your preference of the 414 over the 515. The 414 also has a
> "faster" response with lesser overhang because of a light cone with the same
> size of magnet.
Intuitively correct, factually incorrect.
Doug Purl
=========================================================================
From: "Johari Yip" <hfyip@pacific.net.sg>
Subject: 414 in Petit Onken Was:Re: [JN] Push-pull 515Bs for Onken Cabinet
Date: Fri, 11 Jun 1999 09:30:52 +0800
Source: Sound Digest Archive v02.n175
- ----- Original Message -----
From: Daniel J. Marshall <danmarshall@worldnet.att.net>
To: <sound@lists.io.com>
Sent: Friday, June 11, 1999 2:40 AM
Subject: Re: [JN] Push-pull 515Bs for Onken Cabinet
Dan, fellow joenetters
I'm reading your post with great interest because I have just acquired a
pair of 414-8C in PA enclosure. I must agree with what you have said
regarding your preference of the 414 over the 515. The 414 also has a
"faster" response with lesser overhang because of a light cone with the same
size of magnet. I have intention to partner they with the Oris150 crossing
at about 200-220 Hz.
Dan, one question that I would like to have an answer to is whether the
Petite Onken, as appeared in SP issue #1Vol4, really is the optimum
enclosure for the 414?
Johari
If you want to opt for a smaller enclosure, I would either go
> for a pair of 414B/Cs, or perhaps a different speaker altogether. The
> 414s sound a lot better to my ear than do the 515s anyway, especially in
> the mifrange, and, properly baffled, have quite excellent bass
> performance flat to 30 Hz with usable output below that. They do need a
> fairly good sized enclosure for this performance. I was using 7 cu ft
> with mine, when I had them, and I wish I still did. This enclosure
> volume was a value determined by Jonas Renkus, designer of the
> Renkus-Heinz and Emilar line and was done back before I regained a keen
> interest in audio. He measured the TS parameters of the actual speakers
> and then did an optimum enclosure design. I tried them in four cu ft
> ewnclosures to see how they sounded, but four cu ft doesn't hack it for
> 414-8Cs.
>
I would consider something
> like this, especially if you use something like 803B (or other) horns
> crossed over at a low frequency, or perhaps use a smaller, highly
> efficient 8 or 10 inch midrange, perhaps horn loaded, if that is your
> preference, crossed over at something like 200 or 300 Hz (depending on
> the horn characteristics, then pick up the top end with a pair of good
> one-inch horns crossed over at, say 800 to 1,600 Hz.
=========================================================================
From: "Paul De Raymond Leclercq" <Pderl@btinternet.com>
Subject: [JN] 416a
Date: Wed, 13 Jun 2001 00:27:25 -0700
Source: Sound Digest Archive v02.n908
This is a multi-part message in MIME format.
- ------=_NextPart_000_0062_01C0F39F.9E952CE0
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Thanks to all who have offered information on this one; I'm still =
waiting for the odd bits and pieces - and for the valves too.
I could be interested in buying a couple more as well.
Please e-mail me if you wish - paul@13e1.com
Paul
- ------=_NextPart_000_0062_01C0F39F.9E952CE0
Content-Type: text/html;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META content=3D"text/html; charset=3Diso-8859-1" =
http-equiv=3DContent-Type>
<META content=3D"MSHTML 5.00.3211.1700" name=3DGENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=3D#ffffff>
<DIV><FONT face=3DArial size=3D2>Thanks to all who have offered =
information on this=20
one; I'm still waiting for the odd bits and pieces - and for the valves=20
too.</FONT></DIV>
<DIV> </DIV>
<DIV><FONT face=3DArial size=3D2>I could be interested in buying a =
couple more as=20
well.</FONT></DIV>
<DIV> </DIV>
<DIV><FONT face=3DArial size=3D2>Please e-mail me if you wish - <A=20
href=3D"mailto:paul@13e1.com">paul@13e1.com</A></FONT></DIV>
<DIV> </DIV>
<DIV> </DIV>
<DIV><FONT face=3DArial size=3D2>Paul</FONT></DIV></BODY></HTML>
- ------=_NextPart_000_0062_01C0F39F.9E952CE0--
=========================================================================
From: "Paul De Raymond Leclercq" <Pderl@btinternet.com>
Subject: [JN] 416a/b/c/d
Date: Thu, 14 Jun 2001 02:27:31 -0700
Source: Sound Digest Archive v02.n909
This is a multi-part message in MIME format.
- ------=_NextPart_000_0209_01C0F479.9056C640
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Hello again everybody.
I would be interested in purchasing a couple more of these.
Please let me know if you have any to sell.
Thanks to everyone who has responded
Best wishes
Paul
- ------=_NextPart_000_0209_01C0F479.9056C640
Content-Type: text/html;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META content=3D"text/html; charset=3Diso-8859-1" =
http-equiv=3DContent-Type>
<META content=3D"MSHTML 5.00.3211.1700" name=3DGENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=3D#ffffff>
<DIV><FONT face=3DArial size=3D2>Hello again everybody.</FONT></DIV>
<DIV> </DIV>
<DIV><FONT face=3DArial size=3D2>I would be interested in purchasing a =
couple more=20
of these.</FONT></DIV>
<DIV> </DIV>
<DIV><FONT face=3DArial size=3D2>Please let me know if you have any to=20
sell.</FONT></DIV>
<DIV> </DIV>
<DIV><FONT face=3DArial size=3D2>Thanks to everyone who has =
responded</FONT></DIV>
<DIV> </DIV>
<DIV><FONT face=3DArial size=3D2>Best wishes</FONT></DIV>
<DIV> </DIV>
<DIV><FONT face=3DArial size=3D2>Paul</FONT></DIV></BODY></HTML>
- ------=_NextPart_000_0209_01C0F479.9056C640--
=========================================================================
From: "Chris Vryonides" <chrisvryonides@hotmail.com>
Subject: Re: [JN] 416/E810F Many Thanks
Date: Wed, 13 Jun 2001 09:23:55 -0000
Source: Sound Digest Archive v02.n908
Thanks very much to all who contributed here. Will keep you posted of any
progress with these puppies.
Best,
Chris
_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.
=========================================================================
From: "T. Loesch" <ezee_e@hotmail.com>
Subject: Re: [JN] 417,437 ,7788 as front end MC phono.
Date: Thu, 31 Aug 2000 10:14:53 GMT
Source: Sound Digest Archive v02.n648
Hi all,
>At this point I decided I was just kidding myself trying to get stellar
>performance out of low output moving coils by just using a sole tube on the
>front end-
After several years of wrestling with this I also came to this conclusion.
Since I aquired a pair of very nice antique Grampian Ribbon Mic Stepups (25
Ohm to about 50kOhm) I have taken to using these with a Shindo Claret Copy.
So (IMHO) there is something better than FET's out there. A good
Transformer.
And the world has (again) come full circle, as Allen Wright worked on that
Fet-Mod to get rid of exactly that transformer.... ;-)
>Allen Wright's been telling us this for some years now (and on the evil's
>of cap bypassing) so respect to you Allen.
Even if it is a lower number, he doesn't live in Einstein Strasse for
nothing, as Simon remarked earlier.
Much more interresting is however that my Housenumber for a long time was
101 (take him to room 101 - get it?)....
>The circuit is below- the guru's will say it isnt correct
Funny. I like it. But then, I hardly qualify as Guru....
Later T
_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.
Share information about yourself, create your own public profile at
http://profiles.msn.com.
=========================================================================
From: "Michael (Mick)Maloney" <supra@cantech.net.au>
Subject: [JN] 417,437 ,7788 as front end MC phono.
Date: Thu, 31 Aug 2000 14:56:46 +0800
Source: Sound Digest Archive v02.n648
I put this on my web page -some joes may be interested in text.
Moving Coil Phono Preamp-417, 437 or 7788 tubes as front end.
I wanted to determine for once and all just which tube was the top
performer as front end for moving coil phono.
I did a lot of circuit testing and listening using my dedicated phono
ptototype which has sockets to accept all of the candidates. The phono
circuit is basically a Loesch although there are significant differences.
All circuits were actually hardwired into place-no clips or potentially
sonic damaging methods were used.
Linestage incorporated on same chassis is a "monkey" type- 6F8-G direct
coupled via plate choke from cathode of 6J5 which is transformer coupled
,600 ohm output with a bit of feedback back to the unbypassed cathode of
6F8-G.
Power supply is dual 5881/TL431 shunt regulator in separate chassis.
Cartridge is Denon DL-103 -0.39mV output.
Reference is my 417/2SK369 fet phono amp- I feel I've got the best possible
sound I can get from this combination so I determined to see if I could
improve on it by either using the 437 (actually I used 3A/167 -the British
Equivalent) or a triode wired 7788.
From a purist point of view I wanted to solely use a tube without a fet on
the front end and I started with the 7788 as it's extremely high
transconductance looked like it could deliver the gain and low noise.
Which it could -I was able to get more than enough gain and close enough
noise levels to the performance of the 417/2SK369 as to be of no difference
when more than a foot away from the speaker .
Unfortunately I had to bypass the cathode of the 7788 with a 1000uF
electrolytic cap to get this result and I could clearly hear the damage the
cap was wreaking - I've become very sensitive to the sound of caps on
cathodes in the last few years and whilst most punters in retail land would
still be astounded by the sound it was way behind the tube/ fet combo.
Unfortunately a small cap ,giving a choice of better quality wouldn't work
here.
With the cathode unbypassed ,noise and gain levels were a problem- the
sound was good ,but not as good as the reference.
On to the 437 -this can also be used without a fet -pretty much the same
performance as the 7788-a little noisier in terms of hiss at high levels,
but the same problem with that damn cap needed to give flat frequency
response. And as with the 7788 the same problems when unbypassed-noise and
gain. Although it sounds very nice unbypassed (ignoring the noise at high
levels) still not as good as the 417/fet combo.
At this point I decided I was just kidding myself trying to get stellar
performance out of low output moving coils by just using a sole tube on the
front end- the fact is that the best performance is out of the fet and by a
long way- it's black and white- ok if you want to live in tube land the
frame grids sound very good but the fets sound even better again. At this
level the tubes just cant deliver the resolution of the fets.
Allen Wright's been telling us this for some years now (and on the evil's
of cap bypassing) so respect to you Allen.
I then tried the 7788 and 437 with fet combos to find the top gun - I used
three matched and paralleled 2SK369s to give about 20ma through the tube.
There is no top gun -apart from the fet that is -the sound is dominated by
the fet and that is the sound you hear -there are subtle differences
between the 7788 ,437 and 417 but they are just differences , not
improvements.
I got the various front ends to sound as good as the reference 417/fet but
not better . 417's are still available at under $20 -the 437 is costing
upwards of $100 and the 7788 is also costing more than it's worth.
The circuit is below- the guru's will say it isnt correct but I've built so
many phonos now I'm inclined to quote FZ on gurus-"the price of meat has
just gone up and
"
7788 and line stage circuit is at
<http://www.cantech.net.au/~supra/417-437-7788.html>
http://www.cantech.net.au/~supra/
SUPRATEK TRIODE " valve amplifiers for those who demand absolute sound and
finish-worldwide "
Mick Maloney
3 Parkside Ave Mt Pleasant Western Australia 6153
Ph/Fax 61 8 93644385 supra@cantech.net.au
=========================================================================
From: Jeremy Epstein <jepstein@shwd.com>
Subject: [JN] 417A anthology - long (Courtesy Carter Hendricks)
Date: Fri, 14 Jan 2000 11:00:00 -0500
Source: Sound Digest Archive v02.n402
Carter hendricks sent me a batch of material he had archived on taming
the 417A for audio use. I am applying very minimal editing and hopefully
who said what will remain clear, if not, tough noogies.
Carter wanted me to point out that some of the material is specific to
the ongoing discussions from which they were pulled, ymmv. I hope this
is of general interest, I realize it is pretty long.
- -j
================================
Carter wrote :
Hi Jeremy--
I'm writing this from the shop / no
link to the joelist. I'll be a bit
more informal...
The 417A just has two heater pins
no center tap. I used like a .01
mono ceramic from each heater lug
directly to the body of the tube
socket. Reid used some Spanglish
expression to say that the heater
required some reference to ground.
Nail one side to power supply ground
near the trannie or build a center
tap with two resistors. Or use your
windings center tap if you have one.
There are multiple grids plates etc.
Tie both grids together with neat
short small wires etc.
I think the consensus was 20mA for
sure. 100R, 2200uf, Muses ok BG
better but oh $$.
Remember what Lev wrote about a series R on a cathode bypass a week ago?
[Some guys were using battery
bias schemes in order to avoid the
HUGE bypass caps Necessary for bass.]
I built a potato with Black Art
transformers but went off in this
other direction after hearing Tom
Ronan's with 46's. The simplicity
of the 417A was wonderful though.
Taught me to hate holcos and like
2watt AB's.
I don't know how much of this was
before the archives got going... It
is usually funny how the list
revolves.
--Carter
=================================
(Takuji Yamamoto) (by way of Reid Welch) wrote :
>Subject: Re: Takuji on cathode battery bias
>Reid asked Takuji:
>
>"Have you heard of biasing a tube by means of a battery in the cathode
>circuit?
>Do Japanese experimenters use "ni-cad" cathode bias batteries? This is the
>type of battery Spence prefers. It gives a good tone to the 417A, and with
>nearly the ideal bias voltage. We have not tried a lead acid battery
>because the voltage rises too high for a 417A."
>
>
>Japanese speak of "Kotei bias", which has the same meaning. It
>is a form of fixed bias.
>
>I haven't heard lead battery sound but many Japanese do not like the tone
>of a lead battery.
>
>Many amplifier builder use the Kotei bias method with DH tubes.
>
>Sakuma uses the method only for the screen grid current supply of the 4P55
and
>other tubes.
>
>He hasn't used it for DHT cathodes, because a little trouble could destroy
>DH tubes.
>
>Although Kotei bias tone is very nice, we supply the SG power by means of a
>circuit using the power transformer. It is only practiced by Sakuma.
>In almost amplifiers current for the SG is dropped from the B+ rail by
>resistors.
>
>I haven't heard lead battery sound but many Japanese do not like the tone
>of a lead battery.
>
>"How about ni cad batteries?"
>
>When I built a simple FET equalizer, I tested current supplied by many
>kinds of battery.
>
>Through my test, I learn that the battery has great influence on the
>amplifier's tone; like capacitors.
>
>The battery I used for my amplifier came by the recommendation of builders
>of DC transistor amplifiers.
>
>If you want to find the best battery for your amplifier you had better test
>many kinds.
>
>Soredewa mata
>
> Explore the Wonders | Takuji Yamamoto
> of | http :// www.big.or.jp/~dh/
>
> Direct Heating | takuji@cc.osaka-kyoiku.ac.jp
> Reid Welch (rewelch@earthlink.net)
>3901 Hardie Road Miami, Florida 33133
> (305) 665-4972
===================================
>From: Roscoe Primrose <roscoe@aiko.com>
>To: Dave Stagner <dstagner@icarus.net>
>Subject: Re: 417A operating points, et al
>Dave Stagner wrote:
>>
>> Whaddaya say, kids? Where do YOU like to run the 417A, and why? Better
>> yet, what's a good, clean method for tweaking the bias?
>>
>> Here's my situation... i'm using a single 417A at 160v, with 2.4v battery
>> bias via stacked AA NiCads on the cathode. The load is a 10k:10kct
>> interstage transformer (a fairly hefty little thing rated for 30ma dc...
>> and cheap, too!). This drives push-pull 6B4Gs. Yeah, i know this input
>> can't swing enough voltage to really max out the 6B4Gs, but it was simple
>> to build and i don't need all that much power.
>>
>> Anyway, what sort of operating points have others found useful? And is
>> there a good way (say, with a pot) to adjust battery bias on the
cathodes?
>> Or perhaps some recommended cathode resistor values?\
>
>I'm running the 417A/5842 at approx. 160V with a 75R Caddock as a
>cathode resistor, bypassed with a 3500uf Elna, for about 20mA, direct
>coupled to a 45. Operating point seems nearly ideal, have played with
>more/less current amd been less happy, especially with less, more seemed
>pretty OK up to about 25mA, but I went with 20mA mostly beacause I had
>the 75R Caddocks... I've used the same operating point with the same
>Allied ITs you're using, driving SE 2A3s and 45s, and that seemed the
>best operating point with the ITs as well.
>
>>
>> Finally, any suggestions for an alternative driver tube that would work
>> well with cathode bias and the 10k interstage, to drive push-pull 6V6
>> (triode wired)? I plan on making a variant of this amp for a friend.
>> He's a non-techie, poor, and will be moving away soon. I'd like to build
>> a similar amp for him out of junk parts, and 6V6 is what i have. :}
>> However, i'd like to use regular cathode bias rather than battery bias
for
>> his amp, just so there is one less thing that can go wrong.
>>
>
>You might try the 5687...
>
>Peace
>--
>Roscoe Primrose -- mailto:roscoe@aiko.com --
===============================================
>From: Reid Welch <rewelch@earthlink.net>
>Subject: Unruly 5842
>
>Biasing 417A/5842 tubes with a cathode battery results in large variations
>of standing current between seemingly identical tubes. Cathode battery
>enthusiasts use identical 1.36V nicad potentials, yet the reported
>5842/417A cathode currents vary greatly from each other and from the 1948
>WE curves. Until the present time I have operated a single pair of Amperex
>5842's with cathode battery bias (CBB) - actually only for the last month
>or so. Amperex currents were about in line with the published data. -But
>what about joelisters' results ranging from 7mA @ 170V from a WE 417A to
>about 25mA @ 150V from a Raytheon? Swapping those plate voltages, 150V on
>the WE would only coax a few mA while the Raytheon would fry.
>
>A check of my preamp's second stage (the only nicad biased stage in the
>preamp) revealed large deviations in standing current between different
>brands of 5842 tubes and similarly large differences _even between tubes
>from the same batch_. Natural variations in intangible factors such as
>contact potential and polarity/magnitude of grid currents may account for
>the unruly behavior. Actual causes remain a bit mysterious. Bias troubles
>may be a little less problematic if a grid battery or a bypassed cathode
>resistor is employed instead.
>
>CBB perhaps causes greater bias discrepancies than do either grid biasing
>or cathode resistor biasing. CBB obviates a coupling cap. However, a toll
>is charged by the inflexibility of the 1.36 bias voltage.
>
>Nonetheless, critical listening reports indicate that a nicad cathode
>battery sounds much better than the Old Reliable cathode resistor. I don't
>think a cathode battery makes a circuit sound much different than fixed
>bias. It will be interesting to someday measure the relative effects CBB
>upon 5842 distortion. I expect THD to be greater if the tube is CBB, or if
>it is lightly biased. Especially if the grid return contains appreciable
>resistance.
>
>Tabulated data collected from 19 tube samples follows this text. The data
>demonstrates the importance of care in choosing tubes... especially for CBB
>when the Ep is not adjustable. An individual 5842 may not be very suitable
>for the particular application. If plate voltage is reduced to avoid too
>much CBB current, then insufficient signal headroom might be experienced in
>large-signal drivers. OTOH, underbiasing in order to obtain desired
>current -as is possible with fixed grid bias or with cathode resistor bias,
>will abruptly limit the useable voltage output.
>
>At any rate, 5842's run considerable grid current because of their light
>biasing (thanks, Lynn). Below 1V of bias, the normal negative grid current
>may change directions erratically. A large grid resistance is therefore
>avoided because it may impose additional, unpredictable shifts upon the
>operating point. It's a good idea (-not a mandate-) to avoid grid resistors
>larger than 100k. Try to design with no more than half that resistance.
>Ideally there will be no grid resistance -only impedance. This tube was
>designed for grounded grid operation. It owes us no apologies when we
>ground the cathode instead.
>
>5842's which greatly over run or under run the desired standing current at
>the target plate voltage should be swapped out. An exception to the grid
>current guideline is a tube run at very low levels. In my preamp, stage two
>has a 680k grid resistor, _in circuit only when the prior, phono stage is
>driving it_. The large resistor was chosen to avoid loading down the RIAA,
>the interstage coupling cap, and the stage one tube. This resistor's
>loading effect was dovetailed into the RIAA calculations.
>
>A benchtop testing setup would be better for assessing 5842 bias points
>than my preamp. For the survey I selected a number of new old stock 5842
>and 417A tubes at random and plugged them into the same stage two socket of
>the preamp. The first column shows the plate current resulting from 1.36V
>of nicad battery to CBB the tube. The battery is a paralleled pair of
>heavy-duty nicad "D" cells). The last column indicates the negative grid
>bias needed to set the current to 25mA at 130V. For that test the tube was
>plugged into the (fixed bias) socket of the third stage.)
>
>
>-Plate to cathode voltage of 130V was maintained for each tube except for
>the test in column III. It measured the plate voltage needed to draw 25mA
>from a CBB tube.
>
>-The stage is plate loaded by a Tango 60H plate load choke. It works into
>an external load of 50k ohms (volume pot).
>
>A 5mV sine wave was input to the grid. The output was measured at the
>plate. Input and output were simultaneously monitored by a dual channel AC
>millivoltmeter. Mu calculations were made at 2kHz and 20Hz. Mu was derived
>by dividing the AC output by the grid's AC input. Bass loss is expressed as
>percentage gain loss at 20Hz compared against the mu at 2kHz.
>
>Gain figures are valid only as far as they relate to each other. The gain
>in most instances was higher than the rated mu of 43 despite the 50k
>external load.
>
>Reduced gain at 20Hz results from the combined effects of falling inductive
>reactance, choke DCR and 5842 rp. is expressed as a percentage of the gain
>at 2kHz. It's a lumped term because I can't accurately separate the terms.
>
>In general...
>
>-Excessively low current increases tube rp. THD rises as the current is
>lowered. Also, low current causes proportionately more H3 distortion.
>
>-Tube brands are ID'd by initials as listed. These samples are new,old
>stock 5842's or 417A's of various vintages:
>
>_W_ estern 417A 1952-1976 manufacture
>_A_ mperex 5842 1983 manufacture
>_T_ ungsol 5842 unknown date, but old
>_E_ ricsson 5842 unknown date, but lovely to look at.
>_R_ aytheon 5842 mixed "Q" and "WA" suffix
>
>
>
>
>
> I III V
>nicad Ep which neg.grid bias to
> CBB draws 25mA@1.36V obtain 25mA @ 130V
>130VEp II IV
> mu at 2kHz, mu @ 25mA,
> %mu loss @ 20Hz %mu loss @ 20Hz
>
> I II III IV V
>
>E1 12mA 45.0 -11% 157V 45.6 -7 % .746
>E2 18mA 41.6 -8% 143V 42.4 -6.6% 1.081
>E3 18mA 46.6 -5.5% 146V 47.0 -6.8% .893
>E4 20mA 46.0 -8.3% 140V 46.0 -7.8% .860
>E5 22mA 46.0 -7% 134V 46.0 -7% 1.174
>
>W1 04mA 38.4 -16.6% 190V 44.4 -5.4% .500 (est.)
>W2 08mA 47.0 -12.7% 180V 49.6 -8% .600 (est.)
>W3 12mA 43.0 -7.9% 155V 45.6 -7.5% .728
>W4 12mA 46.0 -7.8% 152V 48.0 -7.1% .839
>W5 17mA 46.6 -9% 149V 47.0 -7.2% .923
>W6 18mA 44.4 -6.3% 143V 45.0 -6.6% .958
>
>T1 17mA 44.6 -6.7% 147V 45.0 -6.6% .920
>
>R1 08mA 39.6 -11.6% 170V 43.0 -7% .400 (est.)
>R2 17mA 39.6 -7% 145V 40.4 -5.9% .915
>R3 25mA 49.6 -8% 130V 49.6 -8% 1.310
>
>A1 19mA 47.0 -8.55 143V 48.0 -7.5% .857
>A2 25mA 46.6 -7.7% 130V 46.6 -7.7% 1.400
>A3 28mA 47.0 -6.4% 124V 47.8 -7.1% 1.325
>A4 28mA 48.0 -7.5% 122V 47.8 -7.1% 1.420
>--------------------------------------------------------
>
>Best,
>
>Reid
>
>
> rewelch@earthlink.net
>========================================
==================================
>From: Reid Welch <rewelch@earthlink.net>
>Subject: Re: potato (417A)
>Dave, BTW- was an ancestor of yours in the radio business 'way back in the
>'thirties? I have a good reason for asking. Nothing to do with Bridey
>Murphy, either.
>
>At 05:23 PM 2/12/97 -0500, Dave wrote:
>
>>I was itching to try out my new 030's, you know get some current running
>>through them, and wired up a single tube amp. I am using a pair of
Amperex
>>5842's with 16V on the plate and a bias of -1V into the 030's with a 15
ohm
>>load on the 9 ohm taps.
>
>160V with only -1V bias?
>
><snip>
>
>
>
>
>>How can current rating be determined???
>
>You mean the rating of power transformer? If you are only drawing a
>twentieth of an amp, i doubt it matters.
>
>If you parallel you'll get a better sound, says Tom and Doc . Lower rp
>giving better bass and all that. What about the dreaded parasitic
>oscillation of paralleled high gm tubes?
>
>Tom, Dan; Don't you need a bit of damping resistance in series with the
>plates and grids?
>
>
>
>>Reid, if your watching, I seem to remember you mentioning your thoughts on
>>the 5842's and 417A's and the plates starting to cherry, care to re-post
>>some of them??? My plates are a cherrying had to increase the bias a bit
>>to get a reasonable glow, cuz my initial settings from the curves were
>>gettin pretty orange :-0
>
>
>Yeah, one volt is not enough at 160V. suggest you place a sensing resistor
>of one ohm or ten ohms in the plate supply lead. Maasure the drop across it
>and multiply by 100 or by 10, respectively, to quantify your plate current.
>
>About "cherry red" reposting, I'll send what I have to anyone who asks
>rather than burden the list. The core of those posts was not much more than
>this:
>
>At 4W dissipation an Amperex plate glows dull red.
>At 4.5W Raytheon and WE417a plates glow dull red.
>Glow, within reason, does not seem to doom the tube. I've run them
>Amperexes at 4.5W for thousands of hours.
>
>-Maybe plate color adds to the tone color. <wink>
>
>Above 150V on the plate, residual gas will ionize giving a blue tinge
>visible inside the plate structure. This is no problem for a high-level
>tube; noise for a phono input tube.
>
>I might add that If one runs a 5842 with fixed bias, don't stand the grid a
>-megohm- above ground, because it might lose its bias through the effects
>of gas current. There's consuiderable gas current in these little hotheads.
>
>With either cathode or fixed bias any value of grid resistance causes
>_some_ change in the quiescent current. The effect of even a 1k ohm grid
>resistance is perceptible.
>
>Cathode bias has the great advantage of self-compensation.
>
>Just wished I liked cathode bias sound as much as I like the stages with
>fixed bias in my system. I have tried Black Gates to bypass the cathode.
>Good results; just not the best for this contrarian.
>
>Sorry, Doc. You and the rest of the work must be mistaken... IMBO
>
>(i have a lot of that)
>
>reid
>
>
> rewelch@earthlink.net
>========================================
==================================================
>From: Reid Welch <rewelch@earthlink.net>
>Subject: Re: To SRPP or not to SRPP
>
>
>I said:
>>> Paul, I do wish my 5842 would swing more volts. It would help.
>>> Nevertheless, in my ITC amp I get the VV 30B to deliver 15W at 5% THD.
>>> That's with 185V/24mA on the 5842 and 420V/80mA for the VV30B.
>>
>>You're getting 15w at 5% and still want more?!?! The 5842 spec says
>>180v max, but you might try 1/2 5687 as Spence Barton has suggested -
>>it will take 300v per spec. Plate dissipation is 4.2w so 14mA at 300v
>>is the limit unless you parallel the sections but I'll bet it would
>>work great.
>>
>>-Paul Joppa
>
>
>Cool! But actually I want less output from the next amp... Improve that
>first watt, and all. Some 5842 specs go as high as 190/200V Ab.max. Plate
>dissipation is probably the limiter.
>
>Casual observations: The 5842s I've tried (Amperex, Raytheon, WE) don't
>blow up or arc even when zapped cold with 500 volts. At plate voltages
>above about 150V they all show a bit of violet glow between grid and plate.
>(ionizing gas). In the Acme preamp, this ionization point, which makes
>objectionable noise in the preamp gain circuit, means I must not run the
>preamp's 5842s above about 135 volts. There is no point of such a high
>voltage in a preamp anyway. A microammeter in the grid return shows
>considerable gas current in all these tubes. Lightly biased 5842s, such as
>in the Acme preamp vary quite a bit in the effectiveness of a given
>biasing voltage. Reason:the serendipity of contact potentials, plus gas
>currents, add or subtract slightly from the bias. As I said...the tube is
>very lightly biased here (about a quarter volt in stage 1 and stage 2). At
>one time I had the first stage (MC phono)operating on contact potential
>bias only. The cathode was grounded, and the grid resistor 47K. Limited
>adjustment was available by manipulating the plate supply. Well, it was
>disconcerting to learn that an Amperex might run 22 mils,and 55V while
>another Amperex - 22mils and 65 V. A WE 417A could hit 95V for that
>current. The phenomenon wasn't all that brand-specific, but WEs tended to
>have more "contact-potential" than the other tubes.
>
>Plate dissipation: At 4.5W, an Amperex "cake pan" plate shows color. It's a
>small-ish, sandblasted nickel plate. The other brands have larger,
>blackened plates which only begin to glow at about 5W.
>
>Pretty tasty,this tiny tube with glowing entrails: yellow heater, orange
>cathode, cherry plate and a splash of rich blue.
>
>Short term bench experiments don't indicate this abuse destructive. Were
>5842/417A really rare and expensive, I might be more conservative. New
>Sensor sells Raytheons today at $3 a head. So I have no qualms about baking
>the beastie for the good of "science". Maybe try 5 or 5.5W. I do believe,
>with its gold plated grid smack-up (about 6 mils)to an incandescent
>cathode; that however lightly this tube is run its grid gets very, very
>hot! The tiny plate and the glass seem pretty well de-gassed in
>manufacture. Furthermore, I expect to be able to _back down_ on the 2 W
>heater dissipation. At 4.5W plate dissipation, the radiant plate feeds
>considerable heat back to the grid and cathode. By dropping a half W from
>the heater dissipation, it's plausible that permissible plate dissipation
>is increased by at least part of that heater reduction.
>
>A 5842 has such good reserve emission that at 3000+ hours, reducing heater
>voltage to 5.4V does not alter the (30 mA) space current drastically. So, I
>ask myself, why would I want to heat at greater than 6V at any time?
>
>Enough rambling... Thanks!
>
>
>
>
>
> Reid Welch (rewelch@earthlink.net)
=======================================
>From: "Reid Welch" <rewe@msn.com>
>Subject: Deeper into 5842s
>
>---Miscellaneous Thoughts---
>
>Amperex 5842 plates are physically smaller than WE, Raytheon, and L M
>Ericsson. They have an unusual, stamped shape; like a cake pan, with the
>bottom of the pan facing the cathode. The plate has a rough gray finish.
>Other 5842s, and the WE417A have folded, formed, blackened anodes.
Hmmm...
>Amperex: Less surface area + gray color = a hotter plate. That's why its
>reddening at 4 watts.
>
>I guess plate temperature didn't concern Amperex engineers. Amperex
>manufactured this 5842 unchanged for many years. But why does the plate
step
>in towards the grid? The bottom of the plate "pan" facing the frame grid
has
>about the same surface area as the cathode. Whatever the reason, Amperex
>certainly looks different than WE and the others.
>
>A couple of subscribers have suggested that an over-heated plate may cause
>noise if the plate releases occluded gas. No disagreement there. But I'll
>wager the plate was well purged in manufacture. At any rate, a bit of noise
>has no consequence in a high level driver.
>
>More to the point, the 5842 in the first stage of my phono preamp is a very
>different tube. About two watts dissipation. Tiny signals. Textbooks led
me
>to expect that high current/high gm = ultra low noise. Equivalent noise
>resistance of this tube is vanishingly low. Realities proved silence is
not
>guaranteed.
>
>One MC input tube may hiss like a snake. Its brother from the same carton
may
>be 20dB quieter. What causes such variabilty? Some plausible reasons can
be
>offered, but whatever the answers, they'll apply equally well to other tube
>types. At this junction I ask readers to send corrections and add insight
to
>the ideas presented here
>
>
>---True Facts?---
>
>5842 is not an audio tube. It was intended for grounded grid, high
frequency,
>multiplex repeater service. The input capacitance is high, and as with all
>triodes, effective capacitance is multiplied by the amount of mu achieved
in a
>grounded-cathode audio application. That is to say- the more gain, the
less
>HF extension.
>
>For best frequency response, source impedance must be low and the load
>resistor kept very small. Unfortunately neither best gain nor good
linearity
>is achieved that way. Choosing instead a larger load resistor of about
seven
>times Rp (about the largest practical multiple), gives adequate linearity
for
>line stage use and around 3/4 of the tube's rated mu. Unfortunately, FR
>suffers in proportion because insufficient power is available through the
7Rp
>load resistor to drive the tube's capacitance.
>
>Add a 100k ohm volume pot to the grid circuit and what do you have? A 3dB
or
>greater rolloff before you reach 20kHz. In one stage.
>
>The grounded cathode 417A / 5842 it is not an RF tube.
>
>5842 is a lackluster high frequency tube.
>
>Drive it from a 50k or 100k pot? Expect severe HF losses.
>
>
>---Donkey Serenade---
>
>What about those infamous 5842 parasitic oscillations? Here's a shameful
>confession of my salad days, 10 months ago--- When I first got the Acme up
>and running, it played a little music. More often it hissed, popped,
snapped
>and sizzled. Probing the guts, I learned to tune in several different radio
>stations. One plate lead tuned in Cuban Salsa. (Well, I *am* in Miami)
>Touching a finger to a phenolic panel (on which were mounted two
>well-insulated coupling caps), elicited eerie moans and squeals. It was
>fairly alive with eddy currents.
>
>The plate volts meter on the third stage langorously bobbed to and fro. I
>added RF chokes to all the leads and attempted all sorts of RF decoupling.
>Days and weeks passed. Nothing I tried worked. No cure. All forlorn and
lost.
> Then one happy day, the umpteenth expert suggested checking the heater
>supply.
>Zooks! He was right! That supply was floating! Out came my hammer and I
>nailed the negative heater leg to the ground. Perfecto! ...... It was no
>fault of the other experts. They couldn't think as ignorantly as RW.
>
>Take a few simple steps to prevent parasitics: 1) Avoid repeating my
history.
>2) Choose a grid pin far from the plate pin for your input. 3) A small
series
>resistor soldered right onto the grid lug is wise, but not absolutely
>necessary. 4) Don't let the grid and plate leads get close or parallel to
>each other. 5) Remove the unused clips from the socket. If a tie point is
>desired, the central post of a socket can safely be used for a ground.
>
>Suggested pinout: 1=plate, 3=heater+, 6=K, 8=G, 9=heater-
>
>
>---Choking... is fun!---
>
>Chokes are wonderful. Chokes and ITs are the best loads for your 5842.
>Chokes give power, utmost linearity, maximum mu and first-rate frequency
>response. A choke is like a transformer with a fraction of the compromises
>and a portion of the cost. ITs for preamps are excellent, but more costly
and
>can't beat a c-core plate choke in measured performance. Oh, but it is a
>shame about that coupling cap. I compromise fairly happily with capacitors
>linking small-signal tubes.
>
>Given finite inductance, the 1800 ohm Rp of the 5842 produces better LF
>response than higher Rp tubes. I've gotten good results from two special
>plate load chokes from the Tango line. I use a compact, handsome-looking
60
>Henry Tango TC-60-35W for the preamp output stage. I'd swear it makes the
>preamp sound much, much better than the non-inductive ww Mills it replaced.
>The -3dB point falls at 5 Hz and 400kHz. Actual in-circuit HF response is
>down somewhat because that tube is driven through a 10k pot. A 1dB
resonance
>dip and peak occurs between 120kHz and 150kHz.
>
>Should you have enough tubes to select from, a quiet 5842 is a wonderfully
>rich sounding MC amplifier; and choke-loaded, the 5842 is over qualified to
>drive any passive RIAA in a single lump. My choice, for cost and size
>considerations was the compact TC-30-50, a 30H, 50mA choke. ( -3dB @ 10 Hz
and
>? kHz).
>
>I bought all my Japanese iron from April Audio, NYC. Other quality
sources
>include Sound Shop Big and SJS Acoustics.
>
>Reid Welch
>rewe @ msn.com
================================
>From: Joe Roberts <jroberts@tpoint.net>
>Subject: Re: VHF / Oscillations
>
>The 417 is a little firecracker. My 417 preamp went nuts at VHF.
>Even when used in a preamp, the tube wants to perform its intended job of
>high gain VHF amplifier.
>
>I suggest the following measures:
>
>Decouple the heater lines with microhenry value RF chokes and put a .005
>cap across the heater pins.
>
>Use a 470 ohm grid stopper (1/2 W carbon) with the body of the resistor
>right next to the grid pin.
>
>Solder a shield across the 417 socket if the above doesn't work.
>
>You must keep all leads ultra short.The 417 is still hot at frequencies
>where your preamp wiring is a decent antenna.
>
>If you check out ham projects using the 417A as a preamp for VHF
>receivers, you'll see that they used to treat this tube as the wild
>oscillator that it could easily become.
>
>Just because it sounds interesting doesn't mean that it is not cooking up
>on Channel 6 at the same time.
>
>Joe
======================================
>From: jc@interport.net (JC MORRISON)
>Subject: Re: CV 5112
>
>hey tom,
>lucky you. the "advanced" series vackooooom tubes, which were designed over
>a period spanning the close of the korean war and the early 1960's, are the
>most sophisticated tubes ever made. from this work, much of it done by
>western electric engineers (although not entirely!), all of the really
trick
>innovations in tube design evolved. the actual devices have numbers which
>range from "400" and "5000" (the 5000# series are the industrial
equivalents
>of the mil#). the CV5112, is the euro version of the WE437A which was
>intended for use in repeater huts for line carrier amps and fm vhf/uhf
>repeaters. the problems associated with triodes at high frequencies were
>addressed particularly with three designs: the 416A B C, 417A and 437A. a
>frame grid with incredibly tight spacing (less than one mm) between plate
>and cathode coupled with special materials and coatings pushed the
bandwidth
>and sensitivity to previously unthinkable performance. one thing common to
>all three is a sickly high transconductance and low plate resistance. you
>would have to parallel 22 12AX7's to achieve a similar grid sensitivity as
a
>437A (and the combined capacitance would render the tubes useless without
>feedback or some kind of neutralization). this fact gives these tubes a
>really wide potential bandwidth without feedback and exposes small signals
>to the most impressive s/n ratio, as well as leaving no stone unturned in
>its amplification of subtle artifacts. and therein lies the rub. used at
>audio frequencies, these tubes can be considered the electrical equivalent
>of nitro injected, supercharged drag racing engines, absolutely content to
>release all of its energy at a twist of the fuel line. many people (perhaps
>even joe roberts) have come to despise these tubes for their difficult
>nature. they are not easy devices to implement and require care and
>attention un-nescessary in commercial receiving tubes. for all that, they
>offer the ultimate in resolution and subtle detail, tasks that do not
appeal
>to everyone equally. i have spent a number of years learning to use these
>tubes and wish that there were more available than there are. i keep
>thinking that there also must be some east block military equivalents which
>are known only to russian or chinese engineers (from their space and
>ballistic missile programs, the same place these tubes were applied
>here...). it would be fascinating to me to know this for a fact (and get
>some cheap!). again, i think most people will prefer to stick with the
>accepted audio device dogma, and make mu followers out of whatever, to gain
>some of the advantages of these "advanced" series tubes without the
trouble.
>send your tubes to ME! send your tubes to ME! send...
>jc
=========================================
>From: joe.roberts@tpoint.com
>Subject: RE: AL'S PHONO PREAMP
>
>RE: The Loesch 417A preamp, JC sez:
>
>
> > you can replace the 6GK5 with a 6C4 for use as a MM. have you built
> > many
> > preamps? you should know that this is one of the most tricky things
> > you will
> > ever attempt... layout, power supply, and parts quality are all
> > crucial. a
> > "budget" version will take out it's cost on your nerves and
> > patience. if you
>
>Yep, even though the preamp is intended for use at AF, if the devices you
>use are VHF/UHF capable, you gotta treat the thing like a UHF circuit.
>There can be all sorts of funky goings on way above the upperlimit of
>your 25 mHz scope.
>
>I was talking to Bill Kleronomos, Hollow State radio guru, about 417s and
>he confessed that he found them to be a real bear to work with. Take off
>like crazy. If you want to mess with 417As read up on how the old geezer
hams
>of yore built their 144 mHz rigs. One thing you will definitely want to
look
>into is carefully bypassing the filanment lines with ceramic .01 and .001
caps
>right at the tube socket to the closest ground point you can get and you
will
>want to use a few RF chokes rated for the filament current at hand to break
>up the filament circuit at RF.
>
>Joe
=================================
>From: jc morrison <jc@interport.net>
>Subject: Re: 5842 Basing
>At 04:55 PM 2/26/96 -0500, you wrote:
>>>
>>> Nope. The 417A/5842 is a tube for use in UHF grounded grid
>>> amplifiers. It has four grid pins so that the lead inductance to
>>> ground can be kept under control. At the frequencies it was designed
>>> for, one lead would be a large inductance, leaving the grid less than
>>> properly grounded. (Although I will admit that the plate dissipation
>>> is a rather surprisingly high 4 watts.)
>>>
>> So what's the best way of using these 4 grid leads? Parallel all of
them?
>>some of them? Or just use one g input?
>>
>>Kevin
>>
>>
>hey kevin,
>connect them all together! and also use two polystyrene .001 50v caps from
>the filament pins directly to a lug on the chassis. a 100 ohm grid stopper
>won't hurt either...
>jc
===========================
Manfred Huber wrote :
> Jeremy asked,
>
> >What is up with the 417A? I think the spec sheet said 9 Vbias with a 360
> >ohm cathode R will get 25mA, and 130V at the plate. I'm not getting
> >anything like this current and I have only 1.5Vbias - why is it running
> >so cold? It seems way off. Am I missing something here?
>
> The operating point you mention has +9V at the grid while you have 0V
> (note the sign!). The cathode potential is 27mA*360Ohms=9.72V. The
> resulting voltage from cathode to grid is 9V-9.72V=-0.72V
> This is a technique often used with high mu/high gm tubes. The grid is set to a
> positive voltage so one can use a larger value for the cathode resistor.
> This yields more cathode degeneration and thus a more stable operating point.
> I have never seen this in a HiFi amp but the 417A was never meant to be an audio triode.
============================
++++++++++++++++++++++++++
=========================================================================
From: blackie <tubesville@yahoo.com>
Subject: [JN] 417a & "blackie box"
Date: Mon, 1 May 2000 03:20:21 -0700 (PDT)
Source: Sound Digest Archive v02.n536
the first stage on the "blackie box" has 100 ohms grid
stopper
AFAI can see, no wiggly! snooped all over w/ my scope
(up to 400mHz) also, no sonic evidence of RF...sound
is smoooooth
100k input impedance is a carbon pot, i feed direct
from cd
careful layout- 417 is pretty far from
everything...input cables are braided silver plated
teflon wire, lo capacitance...
=====
blackie
blackie@tubesville.com
www.tubesville.com
__________________________________________________
Do You Yahoo!?
Talk to your friends online and get email alerts with Yahoo! Messenger.
http://im.yahoo.com/
=========================================================================
From: blackie <tubesville@yahoo.com>
Subject: Re: [JN] 417a & "blackie box"
Date: Mon, 1 May 2000 07:49:52 -0700 (PDT)
Source: Sound Digest Archive v02.n536
- --- Allen Wright <AllenVSE@compuserve.com> wrote:
> Blackie wrote:
>
> >>>100k input impedance is a carbon pot, i feed
> direct
> from cd<<<
>
> You don't need this high a resistance with a CD
> source - I'd suggest you
> try something W-A-Y lower, like 2K. Because if you
> have _any_ miller (or
> stray) capacitance after the pot, 100K will kill the
> treble at any volume
> position except full on, or almost all the way off.
>
> Allen (VSE)
hmmmm...yes, it's what i had on hand, so i used it
gave no problems, so i made no changes...HF rolloff is
negligible at settings i normally use until 80k...
probably would sound better w/ a lower value though...
>
=====
blackie
blackie@tubesville.com
www.tubesville.com
__________________________________________________
Do You Yahoo!?
Talk to your friends online and get email alerts with Yahoo! Messenger.
http://im.yahoo.com/
=========================================================================
From: blackie <tubesville@yahoo.com>
Subject: Re: [JN] 417a & "blackie box"
Date: Mon, 1 May 2000 07:55:25 -0700 (PDT)
Source: Sound Digest Archive v02.n536
- --- "Martin E. von Lindenberg" <mvon@pop.mail.rcn.net>
wrote:
> At 03:20 AM 5/1/00 -0700, blackie wrote:
> >the first stage on the "blackie box" has 100 ohms
> grid
> >stopper
> >
> JC, Blackie,
>
> Thanks for your continued help. The grid stopper
> does not appear on the
> schematic. I notice how far the 417A was from
> everything else. I had
> assumed *very* short leads to the input jack was the
> reason.
ya know, maybe i didn't use a stopper...though
normally i would...haven't been under the hood of that
baby in 5 yrs or so...
i'll check if you want...
i just bang 'em together and only make changes if i
have a problem...i never seem to build the same amp
twice
usually my troubleshooting curve is short, cause
basically that's what i do all day every day...yawn
input lead is very very short, goes right to the
grid...ha ha, nice hi-tech cloth insulated wire
throughout too hehehe
HTH
=====
blackie
blackie@tubesville.com
www.tubesville.com
__________________________________________________
Do You Yahoo!?
Talk to your friends online and get email alerts with Yahoo! Messenger.
http://im.yahoo.com/
=========================================================================
From: "Martin E. von Lindenberg" <mvon@pop.mail.rcn.net>
Subject: Re: [JN] 417a & "blackie box"
Date: Mon, 01 May 2000 08:57:21 -0400
Source: Sound Digest Archive v02.n536
At 03:20 AM 5/1/00 -0700, blackie wrote:
>the first stage on the "blackie box" has 100 ohms grid
>stopper
>
JC, Blackie,
Thanks for your continued help. The grid stopper does not appear on the
schematic. I notice how far the 417A was from everything else. I had
assumed *very* short leads to the input jack was the reason. In my
practical experience as a kid (no, not now, the first time around!) all
runs to a volume control on a TV were either extremely short by laying out
the chassis so that the audio section was right by the volume control, or
always the cable runs were shielded. Apparently this was so second nature
that it wasn't discussed.
I looked through a lot of prints I have, particularly of the "classic"
amplifiers, and did not see grid stoppers on any of them except for a
Williamson that had a 10K (!!??) in front of a 6SN7. Wouldn't this large
of a value cause a large phase shift? Was it used because of overall
instability of a multistage w/ global feedback?
Lastly -- why not use a mu<1 buffer in front of the normal input stage?
This shouldn't have to be bandwidth limited because of the fractional gain,
and then it can be optimized for low output z of the broad bandwidth triodes.
Martin E. von Lindenberg
=========================================================================
From: jc@izone.com
Subject: Re: [JN] 417a & "blackie box"
Date: Mon, 01 May 2000 10:21:47 -0400
Source: Sound Digest Archive v02.n536
martin wrote:
>
>I looked through a lot of prints I have, particularly of the "classic"
>amplifiers, and did not see grid stoppers on any of them except for a
>Williamson that had a 10K (!!??) in front of a 6SN7. Wouldn't this large
>of a value cause a large phase shift? Was it used because of overall
>instability of a multistage w/ global feedback?
>
>Lastly -- why not use a mu<1 buffer in front of the normal input stage?
>This shouldn't have to be bandwidth limited because of the fractional gain,
>and then it can be optimized for low output z of the broad bandwidth triodes.
>
>Martin E. von Lindenberg
>
yes,
the idea was to bandwidth limit the frequency content ahead of the amp and
the feedback loop. most of the mods you see on later versions of the
williamson, such as the heath-kit stuff, involve even more radical steps to
insure stabiliity. and as for the buffering question, yes, of course...
this is something you see in test equipment and rf stuff all the time. i
think that buffering the volume control is a first class solution to the
inconsistencies of pots. i'm not going to get into a thing about cathode
followers vs. transformers though... its true that less is generally more,
and transformers are WAY more complex than first order RC networks, but
i've heard good results with both. i don't hate cathode followers, they can
be done right. most of the people i know, have had to go the route of a
stepped attenuator to get away from the dulling and roll off you get in the
middle of a pot (where most people insist on leaving it! thats a funny
psycho babble thing to think about... why couldn't the thing be turned up
all the way all the time where it sounds best because its out of the
circuit?! i know one guy who flips out if the pot is anywhere other than 12
noon for his normal listening! no amount of explanation will convince him
otherwise... this is the way he wants it! i think its a sure sign of OCD,
oops, have i offended someone else?). the larger the value of the pot, the
more you lose. ev