• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

DynaKitParts ST-35 power supply mods?

Hi all, I'm considering building a new DynaKitParts.com ST-35, and I am wondering about the power supply. I am not interested in the Enhanced Fixed Bias board for the ST-35 and have ruled that out for reasons that I need not go into here.

1) DynakitParts.com does offer an optional choke as shown in the bottom half of the ST-35 schematic below. How and why is the choke an improvement?

2) Other capacitors are available online to increase the power supply capacitance by as much as double. Is it worth the trouble, and if so, which values (A, B, or C) are most important because triple and quad caps are available in all sorts of values and combinations, some of which can be combined to increase only A, or A and B, or all three. For example, I have found 80/70/20 and 120/40/20 and 100/47/22. It also is possible to use two caps to get 120/80/40 - there is room on the tiny chassis for two caps believe it or not - however, a second cap would prohibit the use of the choke since there isn't room for both.


1681270403798.png
 
Last edited:
I'd find a place for individual cathode resistors and bypass caps for the output stage.

That will be done. They have an add-on board for that, and the bias board and the choke both fit on the chassis. PhotoShopped together because I can't find a pic with them all in place at once, it all fits perfectly like this:

1681272671536.png



I'd make C8A much smaller, to reduce B+, and C8B as large as possible.

I don't understand this part.
 
Hi all, I'm considering building a new DynaKitParts.com ST-35, and I am wondering about the power supply. I am not interested in the Enhanced Fixed Bias board for the ST-35 and have ruled that out for reasons that I need not go into here.

1) DynakitParts.com does offer an optional choke as shown in the bottom half of the ST-35 schematic below. How and why is the choke an improvement?

2) Other capacitors are available online to increase the power supply capacitance by as much as double. Is it worth the trouble, and if so, which values (A, B, or C) are most important because triple and quad caps are available in all sorts of values and combinations, some of which can be combined to increase only A, or A and B, or all three. For example, I have found 80/70/20 and 120/40/20 and 100/47/22. It also is possible to use two caps to get 120/80/40 - there is room on the tiny chassis for two caps believe it or not - however, a second cap would prohibit the use of the choke since there isn't room for both.


View attachment 1163908
The most important "mod" is a quad of well matched EL84, next is an EFB board that you of some reason
refuse to install.
 
quad of well matched EL84

The DynakitParts adjustable bias add-on board allows individual tube bias tweaking, so the tube quad need not be perfectly matched.

EFB board that you of some reason
refuse to install.

I don't like the EFB board, so I won't be using it. Yes, I get the idea behind it, and I have read all about it. I won't be using it. I prefer the board from DynakitParts instead.
 
You cannot get tubes matched by individual bias, Gm should also be equal if dist is as low as possible. Besides
a matched quad of el84 is peanuts compared to other suggested "mods"

you still need replacement tubes in the future, beeing able to simply drop them in is more appealing then
needing adjustment each time tubes are replaced.

But that is my opinion , your seem to differ. Good luck with the build you will have first class components and
the result will be good.
 
re: power supply caps: The Dynas were designed in the days of horsepower races and cheap good quality valves. Like others of its day, abusing the valves' rating was common and expected. ST35s run at 350 - 400 VDC B+ which is way over best practice of maybe 300 VDC tops. My personal choice is to forego a dB or 2 of peak output capability for better reliabilty, but that's just me.

FWIW, I'd also prefer adjustable individual output valve biasing ability, if easily available. Even better in some respects than individual cathode bias. I'd keep the grid leak Rs down to 200K or 300K just to not tempt fate.

All good fortune,
Chris
 
  • Like
Reactions: 1 users
You cannot get tubes matched by individual bias

No, but it helps. A common feature on power amps for a reason. Matched pairs still required b/c push-pull.

ST35s run at 350 - 400 VDC B+ which is way over best practice of maybe 300 VDC tops.

My present amp is running 6P14P at 340V (minus 10V across the tube), which is considered high, but it does sound good. What's the simplest way to reduce the voltage in the ST-35 circuit without using the EFB power supply board?

... individual cathode resistors and bypass caps for the output stage.

Easily accomplished with their add-on board during the build. This schematic was sent to me by someone:

individual bias circuit.jpg
 
That looks like it 'll work fine. So, ABCD connect to the cathodes? R1 - R4 need to be a couple watts each, maybe even a small brick resistor, tolerance isn't real critical. Don't count on a lot of adjustment range with given values.

If you have a choke (to control ripple) B+ can be adjusted wherever you want it by changing the value of C8A - smaller cap, less B+, cooler running power transformer, but more ripple.

All good fortune,
Chris
 
Last edited:
  • Like
Reactions: 1 user
Don't count on a lot of adjustment range with given values.

Yes, I have been made aware of that. It should help tweak the values, but not make major adjustments.

If you have a choke (to control ripple) B+ can be adjusted wherever you want it by changing the value of C8A - smaller cap, less B+

Interesting. I built a model of the power supply in PSUdesigner, and I can see the effect, but it only shaves off a few volts before the ripple goes way up. So, not sure that would be a viable option.

However, rugged output tubes are available these days should tube life become a concern. I haven't seen anyone complain about tube life in the DynaKitParts version of the ST-35. Does anyone here have any experience with it?
 
As Chris Hornbeck mentioned, with the choke installed, you should add Lots more capacitance After the choke. That helps in the ripple department.
(you already have the 60uF that you are removing from before the choke, put it in parallel with the 40uF; and add another 100uF for a 200uF total).

Removing or reducing the value of the first cap will also reduce the very high frequency hash and very large transient current in that potential ground loop. Hummm, Hsssss.

I most often use the Hammond 5H 200mA choke. Works great, and gives a better amount of ripple reduction than a choke with less Henries.

And, Chris also mentioned a few other good points:
Lower B+ Voltage, more reliability, tubes operated within their maximum ratings, a cooler running power transformer.

Want more power output, then use another amplifier design.
The old "rule of thumb" was to Increase the power by at least 3 dB (2X the power, requires that you get another output tube type, more powerful power transformer, new output transformers, etc.).
That is a completely new project.

And, by the old "rule of thumb", you will not even notice if the modified EL84 amplifier has 1 dB less output power.

Bias Adjustments? Why? Just use individual fixed self bias resistor values according the the B+ voltage and OPT DCR voltage drop.

Who says you can not get very well matched EL84 tubes?
Go to eurotubes.com (they re-test extensively, and for multiple characteristics; and do excellent matching too).
They ship to many countries.

I am using matched JJ output tubes, with a single bias resistor for the 2 cathodes of the input stage, and with a single bias for the output tubes of a Balanced Amplifier (and No bypass caps across the self bias resistors). How about that?

Some like the EFB board, But I will never use one.
 
  • Like
Reactions: 1 user
Wow, with no modifications, dissipation is 12.6 to 13 watts. o_O

Yet, I don't find anyone online complaining about tube life with actual examples, just people stating theoreticals about tube life. Does anyone have recent, hands-on experience with an ST-35 running as originally designed?

Sovtek EL84M tubes are supposed to be able to handle a lot of abuse and also Russian 6P14P-EV tubes.

If I were smart enough to figure out how to fit EFB inside the chassis like Art Grannell did, I might do it, but I will never use the giant external EFB power supply board that is now sold, which spans the whole width of the amplifier.

Art is no longer with us unfortunately. Art made this little EFB board to fit inside an original ST-35. He mounted the regulator to the chassis instead of a heatsink.

I am not sure if this mod would apply exactly to the current DynaKitParts.com ST-35 kit as I don't know if any changes have been made. Also, there is no schematic from Art in the article. http://www.tronola.com/html/efb_in_a_dynaco_st-35.html

st35mod3A_low.jpg
 
Last edited:
I've run an SCA-35 (with ST-35 innards) with factory cap values, individual 375R cathode bias, triode connected "Made in USSR 6BQ5" output valves since the late 1980s. The output valves look like they've been smoking 2 packs a day but still perform. Doesn't get used much, runs hot, but still runs. Recommendations to strangers on diyAudio must, by their nature, be conservative and best practice. Also, don't stick your finger in the light socket!

All good fortune,
Chris
 
Also, don't stick your finger in the light socket!

Why not? How about a fork instead? :LOL:

The output valves look like they've been smoking 2 packs a day but still perform.

Thank you. (y) I'm now interested in actual experiences vs. theoretical discussions of ST-35 tube life after I was unable to find widespread complaints of output tube failure ("premature" or not) in the ST-35. It all seems to be hypothetical discussions of driving tubes "hard" and what "might" happen with no proof or actual values like consistently getting x hours of tube life instead of y hours. I don't see facts or actual experiences backing up any of the theoretical discussions of tube life.

If we are only talking about a theoretical reduction in tube life of say 10% or even 20%, then it's not worth discussing as far as I am concerned, especially if a rugged tube like the Sovtek EL84M or 6P14P-EV eliminates any "problem" anyway. I did find one and only one complaint about the ST-35 and current production output tubes. One person put in an inexpensive set of JJ EL84 tubes, and in the dark, the plates did have a faint red glow. Ok, fine. Use something else. I don't consider that one example a problem that requires remediation in the amplifier.

As I said, I would install the EFB if I were smart enough to figure out how to do it inside the DynaKitParts ST-35 like Art Grannell did in his original ST-35, but since I'm not, I think I'll just try the DynaKitParts add-on bias board that allows monitoring and tweaking the 4 tubes individually.

After fiddling with the power supply in PSUDesigner, I encountered some results that I didn't really expect (novice here). It seems like using a resistor - as designed - instead of the optional choke - yields a significantly lower voltage than I expected. Increasing the resistor from 50 to 100 or even 150 ohms might actually be an improvement due to the voltage drop, although wasting power and heat. Maybe it's best to just use a resistor as intended in the design instead of the optional choke. 100 or 150 ohms makes it cleaner than 50 ohms and also drops the voltage a little bit more. I did this in my Chinese 6P14 amp, changing the resistor from 100 ohms to 200 ohms. It may not be the most efficient way, but it works fine and it's dirt cheap and so simple that a novice (ME) can't get it wrong. ;)

... or ... maybe the tube life "problem" is all theoretical anyway. Hmmmm ... :unsure:
 
Last edited:
6A3sUMMER consistently encourages individual cathode bias resistors, and I've climbed on that bandwagon too. This is the single most important thing to do for classic cost-constrained amplifiers.* Your adjustable-individual-cathode-resistor board is an improved version of that, and the EFB board is just an automatic version of that. If set for the same resting bias and same bypass caps, the manual-adjust version will have identical static performance but a lower time constant with varying signal level.

It's possible to argue the "best" time constant in various ways. How long is the increased current demand? How much change? How much does the changed bias effect performance? How do we humans hear that changing bias? etc.

*David Hafler famously said "The best part is the cheapest part". Way ahead of his time. But they did make some fine transformers and some very elegant circuitry. Very elegant.

All good fortune,
Chris
 
@Chris Hornbeck that is some of the best information and advice I have received to date! 🏆

I will fiddle with the power supply in PSUDesigner some more before making a final decision on the power supply cap/choke/resistor values and selecting what I think might be the "best" combination, which likely will be the original design as intended, or some very small adjustment to it to possibly reduce the voltage a tiny bit. I can't justify changing a time-tested and beloved design (and sound) by too much when there are no widespread complaints about tube life, just hypotheticals. I don't ever run output tubes to end of life anyway. I use the sweet spot in the middle of tube life and then discard them. A 10% life reduction or something like that is irrelevant to me.

The individually adjustable bias board add-on from DynaKitParts.com with four trim pots isn't expensive, and it makes perfect sense to me. It has been tested extensively, so not a new product and no complaints about it. That add-on probably is as far as I will go with modifying the original design.
 
With those design goals, I'll give you my blessing (don't knock it - I can make raccoons talk) and leave you in peace.

But! If in future you for some reason want to reduce B+ some a good place to include resistive losses is in series with the rectifiers, more bang for the buck. Please keep the choke; they were a luxury in 1966 when grown men earned $2 an hour but easy enough now.

All good fortune with your project,
Chris
 
If in future you for some reason want to reduce B+ some a good place to include resistive losses is in series with the rectifiers

Why not just add a series resistor as soon as the power comes out of the diodes then? OR add a series resistor after the choke between the choke and the OPT CT so as not to affect the rest of the circuitry that follows the choke? Am I missing something basic here? Educate a novice.

In my Chinese 6P14 amp, which does not use a choke, increasing the R in the CRC filter from 100R to 200R solved several issues at once, including the fact that the power transformer was designed for 110 instead of 120 VAC so the entire power supply needed to be brought down all at once. It worked fine.
 
Last edited:

Yes, that's the idea. But series resistances reduce stress to the power transformer. Adding some resistance before the first cap reduces peak-to-average current ratio, which transformers like, and generally are less wasteful than resistance downstream of the first cap. Waste is heat, so it matters in a tiny chassis.

There is a downside to these attempts to make the power supply less "perfect" - voltage regulation, meaning the power supply's ability to keep a constant voltage with varying current loads, is significantly decreased. In the ST-35, biased almost into true class A (ideally zero current draw change with signal level), this doesn't matter enough to matter. In other amps, it could matter a lot, so yours is a special case. Modern SETs are also in that special case. But just FYI.

Since you're already familiar with PSUD and have a choke, you can tailor B+ about wherever you want it. In your projected use, B+ of 350 VDC at 30 or 35mA, doesn't seem excessive at all. So don't mind me and my conservative approach - every case is particular. As Jill Conner Brown's daddy said "Be particular."

All good fortune,
Chris
 
  • Like
Reactions: 1 user
"ST35s run at 350 - 400 VDC B+" . . .

Just build it per original specification, and see how it sounds.
Then, after an hour of operation in a warm room, see how hot the EL84 tubes are and how hot the power transformer is (Summer is coming).
Leave the amplifier un-modified; or modify the B+.

Lowering B+
A choke input has less heat (the heat is mostly from the choke DCR); But a series resistor with enough resistance to drop enough voltage (V = I x R) has more heat (Watts = I-squared x R).

I once experienced a Red Plate. It was from exceeding the tube's maximum specifications.

Just saying
 
Last edited: