I am using two matched quads of Soviet EL84 equivalents that others use with success. No problems with them in the cheapo Chinese amplifier kit either. Only problems here.
Last edited:
The snag you have is you cannot be sure the mods are not the source of the issue. Can you bypass the bias board, for example? I am not sure I would prefer having 2 resistors and a trimmer on the cathode of each tube instead of a single resistor. That is 2 mods on its own (shared cathode bias replaced, bias current fixed).
They are said to be, but are they ? Can't tell until tested. Besides ST-35 is made for EL84 not some clone.
Throwing this out there as another thing to check. C6, C7 and R10 are part of the feedback network. It’s in the heart of this ST35 design. C6, and C7 are picofarad values and it’s important to get them close to right. Your DMM probably doesn’t have the capability of verifying this unless it’s a capacitance meter, a bench meter or expensive Fluke. I noticed you typed ‘unreliabe’ in your didactic Excel datasheets.
I have seen in other designs, builders getting these values wrong or wrongfully putting one value for the other, and getting a pitchy sound. It can cause the amplifier to oscillate if wrongfully installed.
That being said, carry on with the biasing adventure.
Best,
Anand.
I have seen in other designs, builders getting these values wrong or wrongfully putting one value for the other, and getting a pitchy sound. It can cause the amplifier to oscillate if wrongfully installed.
That being said, carry on with the biasing adventure.
Best,
Anand.
The positive voltage on the grids of the output tubes needs further consideration, IMHO. I don’t know if it matters to have a small voltage, indicating the grid is drawing some grid current in this situation. I am on vacation and can’t measure it. It would be good if someone with access to a Dyna SCA 35 or ST 35 could measure what your grid (pin #2) voltages are.
If the output tubes are not well balanced, and many “matched quads” are not, this might be another avenue to persue: Perhaps the Dyna Z-565 output transformers tolerate imbalance less than most outputs. Dave Gillespie has a piece where he argues against individual bias for the ST35. I don’t remember this OT tolerance, or lack of it, was mentioned, but he had reasons why he did not recommend individual bias. OP will recall that I recommended Daves EFB bias and PS earlier in the thread. My ST35 with his board sounds wonderful to me and biased to only ~25 mA.
If the output tubes are not well balanced, and many “matched quads” are not, this might be another avenue to persue: Perhaps the Dyna Z-565 output transformers tolerate imbalance less than most outputs. Dave Gillespie has a piece where he argues against individual bias for the ST35. I don’t remember this OT tolerance, or lack of it, was mentioned, but he had reasons why he did not recommend individual bias. OP will recall that I recommended Daves EFB bias and PS earlier in the thread. My ST35 with his board sounds wonderful to me and biased to only ~25 mA.
Sold by sellers I trust, with the test results, and cause no issues with them in my cheapo Chinese amp. We can rule out the tubes.
I don't want that, thanks. I have gone in this direction instead.
I used the ones that came with the kit. I have checked the labelling several times. They are in the right spots. Whether the smallest value pF caps are truly accurate I have no idea because I couldn't get a reliable measurement of them. The others were at least close, even with my inexpensive meter. I'll look at the labels on them one more time.
These tubes work fine for other owners in the ST-35 and are durable. "Some clone" from the Soviet-era military made in Russia to military specs is better than the majority of what's produced today. We can rule out the tubes.
I might buy another meter I suppose. What would cause those millivolts there?
Why would turning the bias all the way up solve that problem if it exists? Why would this amplifer have an oscillation problem when there are no reports of problems with the kit, and no known issues with any of the components?
The mystery is this: why turning the bias pots all the way up does give good sound. It sounds great when running like that, with either quad of tubes installed.
Last edited:
This small DC voltage isn't significant. Some comes from leakage through the coupling caps and some comes from grid current (which is never zero, and increases with current).
You're resisting the obvious possibility of parasitic oscillations, but that's your stone to roll up the hill.
All good fortune,
Chris
No, I'm not. I just have no way to test for or correct that. Also, how would turning the bias pots all the way up still yield good sound if that's the problem? Please explain because I don't understand the issue or the possible causes or remedies. I'm obviously no expert here.
Yes, and I have a capacitor on order to do that if absolutely necessary. It involves desoldering 4 wires from the tube sockets and installing one cap and a resistor to configure the amp with one bias resistor and cap for all four tubes as designed. If I get desperate, I'll have to try that.
See post #180.
The cathodes are already well bypassed by C1, C2, C3, C4 on the bias board.
All good fortune,
Chris
Ok. Before I resort to that mod, I'll remove the entire bias board if I have to go that far and try the original design with one capacitor and one resistor for all four tubes and see what happens. The cap is already on order and the resistor came with the kit.
"All tubes are created equal, but some tubes are more equal than others" - George Orwell's "Animal Farm", with my changed words.
The fact that a different tube type, other than a real EL84, sounds real good in someone else's amplifier does not prove anything.
Tubes are variable, even from the same manufacturer; lot codes, date codes, different foundries, etc.
But they can be widely different if they are of a different tube type.
I can remember when someone posted that an SV83 was equivalent to an EL84.
Lots of persons following that advice did not get to hear the difference, because the screens of the SV83 burned out before the music came on.
Do you want to tube roll a Dyna ST35 push pull EL84 amplifier?
Then get a matched quad of 6BQ5, and have a go at it; or get a completely different matched quad of EL84.
. . . Give the amplifier a chance.
Everything I said above is a generalization.
"All Generalizations Have Exceptions".
Is your output transformer an Exact copy of the original Dyna outputs (or maybe an original amplifier, or NOS)?
If is is different, ALL bets are off.
Have fun making it perform just like the original.
The fact that a different tube type, other than a real EL84, sounds real good in someone else's amplifier does not prove anything.
Tubes are variable, even from the same manufacturer; lot codes, date codes, different foundries, etc.
But they can be widely different if they are of a different tube type.
I can remember when someone posted that an SV83 was equivalent to an EL84.
Lots of persons following that advice did not get to hear the difference, because the screens of the SV83 burned out before the music came on.
Do you want to tube roll a Dyna ST35 push pull EL84 amplifier?
Then get a matched quad of 6BQ5, and have a go at it; or get a completely different matched quad of EL84.
. . . Give the amplifier a chance.
Everything I said above is a generalization.
"All Generalizations Have Exceptions".
Is your output transformer an Exact copy of the original Dyna outputs (or maybe an original amplifier, or NOS)?
If is is different, ALL bets are off.
Have fun making it perform just like the original.
We have others here who have built this kit, and nobody but me has a problem with the bias or the sound apparently. Lots of others have used these output tubes in this kit, and if there is any problem using them (2 different versions) in this amplifier, I don't see anyone online complaining about it. I would hate to invest in a third set of output tubes only to find that I have the exact same problem, which is what I would expect.
As for the transformers, they are supposed to be wound using the original process with the change being that they have 4 ohm outputs instead of 16 like the originals. Yes, that makes them different. However, the kit with these transformers has been for sale for many many years with no transformer problems.
This is a reputable kit from a reputable seller: https://www.dynakitparts.com/shop/st-35-kit-120-vac/
Is there an inexpensive meter for sale online that's reliable and trustworthy?
About the 6P14P:
https://effdubaudio.com/6p14p-versus-el84/
https://www.tdpri.com/threads/anyone-tried-the-russian-nos-6p14p-ev-el84-tubes.653181/
https://www.thetubestore.com/reflector-el84-6p14p
6P14P-ER is the most rugged-ized version.
6P14P-EV is another rugged-ized version:
https://www.vivatubes.com/nos-soviet-era-military-spec-6p14p-ev-7189-6bq5-el84-e84l-vacuum-tube/
As for the transformers, they are supposed to be wound using the original process with the change being that they have 4 ohm outputs instead of 16 like the originals. Yes, that makes them different. However, the kit with these transformers has been for sale for many many years with no transformer problems.
This is a reputable kit from a reputable seller: https://www.dynakitparts.com/shop/st-35-kit-120-vac/
Is there an inexpensive meter for sale online that's reliable and trustworthy?
About the 6P14P:
https://effdubaudio.com/6p14p-versus-el84/
https://www.tdpri.com/threads/anyone-tried-the-russian-nos-6p14p-ev-el84-tubes.653181/
https://www.thetubestore.com/reflector-el84-6p14p
6P14P-ER is the most rugged-ized version.
6P14P-EV is another rugged-ized version:
https://www.vivatubes.com/nos-soviet-era-military-spec-6p14p-ev-7189-6bq5-el84-e84l-vacuum-tube/
Last edited:
Perhaps there is something about your loudspeakers that works well with Dyna ST-70, and many other amplifiers;
But perhaps that same loudspeaker load makes your ST-35 act abnormal.
Without a test bench full of equipment, there is no way to measure:
Your amplifier characteristics
Your loudspeaker characteristics, like the low impedance frequency areas, high impedance frequency areas, and impedance phase angle versus frequency, etc.
I think you said that you do not trust your DMM.
Otherwise, I would suggest accurately measuring the DCR of your loudspeaker.
Minimum [AC, Audio] impedance can be as low as the DCR, often at 2 or more frequency areas.
Good luck finding the cause of your less than perfect sound with that amplifier and loudspeaker combination.
But perhaps that same loudspeaker load makes your ST-35 act abnormal.
Without a test bench full of equipment, there is no way to measure:
Your amplifier characteristics
Your loudspeaker characteristics, like the low impedance frequency areas, high impedance frequency areas, and impedance phase angle versus frequency, etc.
I think you said that you do not trust your DMM.
Otherwise, I would suggest accurately measuring the DCR of your loudspeaker.
Minimum [AC, Audio] impedance can be as low as the DCR, often at 2 or more frequency areas.
Good luck finding the cause of your less than perfect sound with that amplifier and loudspeaker combination.
I have had one of these for 3 years and it has performed satisfactorily. I just purchased two more and they arrived within a week or two. They have some useful functions and are good value, I think, for $15 shipped.
Yes, very high frequency oscillations caught me out, even after I was strongly warned about them by Rikaro. I'd not experienced them previously, they're elusive (mine came and went, depending on which components were being touched), caused unusual voltage readings and other bad behaviour - and made troubleshooting difficult. I tried, but couldn't identify them with a multimeter, it took an oscilloscope. This was in a Tubelab SPP - well tried over many years. But I believe it was the combination of my choice of Hammond 1650FA OTs and feedback cap that allowed the gremlins in.
The key lesson I learned is: when troubleshooting, keep an open mind, only change one thing at a time, test methodically, and keep good records. But I think that's you all over, anyway.
Good luck NGO - ST35s are often touted as people's favourite amp amongst many.
Simon
You do not have a lot of test equipment, you do not even trust your DMM, right?
1. A loudspeaker that works well with a Dyna ST-70, and many other amplifiers, might possibly . . .
Have load impedance versus frequency variations, and load phase angle versus frequency variations . . . that could cause "Heartburn" for an ST-35.
1A. If you had a good DMM, you could accurately measure the DCR of your Loudspeaker.
At 1, 2, or 3 frequency areas, the minimum impedance of your loudspeaker may be as low as the DCR of the loudspeaker.
Manufacturer's loudspeaker impedance ratings are Nominal.
2. You seem willing to try reasonable things to get the sound you want. Triode wire the output tubes.
Please get four 100 Ohm or 200 Ohm resistors.
Remove the connections from the Ultra Linear Taps, mark them so you know where they came from (you might want to go back to original UL).
Connect the screen of an output tube to its respective plate, using the 100 or 200 Ohm resistor from screen to plate.
Repeat for the other 3 output tubes.
This time, you do need the individual bias board; because the triode wired plate impedance, rp, is much lower than UL wired.
So, when you adjust one cathode current, then another cathode current, etc., you will have to go back and re-adjust each one till they are the same (trust me on this one, the four adjustments do interact if the amp is triode wired).
2A. Now, with Triode wired, the power output will be less.
(And perhaps the gain will be lower if the new open loop gain is near to the negative feedback targeted gain.
But, you might like the sound better.
Listen at low and middle power levels.
See what you think.
If that fixes the problem (admittedly with less power capability), then . . .
It is your loudspeaker characteristics that are giving your ST-35 so much trouble.
Or, perhaps your original circuit of the ST-35 with its damping factor that is giving your loudspeaker so much trouble.
Please try this, and let us know what you hear with triode wired.
1. A loudspeaker that works well with a Dyna ST-70, and many other amplifiers, might possibly . . .
Have load impedance versus frequency variations, and load phase angle versus frequency variations . . . that could cause "Heartburn" for an ST-35.
1A. If you had a good DMM, you could accurately measure the DCR of your Loudspeaker.
At 1, 2, or 3 frequency areas, the minimum impedance of your loudspeaker may be as low as the DCR of the loudspeaker.
Manufacturer's loudspeaker impedance ratings are Nominal.
2. You seem willing to try reasonable things to get the sound you want. Triode wire the output tubes.
Please get four 100 Ohm or 200 Ohm resistors.
Remove the connections from the Ultra Linear Taps, mark them so you know where they came from (you might want to go back to original UL).
Connect the screen of an output tube to its respective plate, using the 100 or 200 Ohm resistor from screen to plate.
Repeat for the other 3 output tubes.
This time, you do need the individual bias board; because the triode wired plate impedance, rp, is much lower than UL wired.
So, when you adjust one cathode current, then another cathode current, etc., you will have to go back and re-adjust each one till they are the same (trust me on this one, the four adjustments do interact if the amp is triode wired).
2A. Now, with Triode wired, the power output will be less.
(And perhaps the gain will be lower if the new open loop gain is near to the negative feedback targeted gain.
But, you might like the sound better.
Listen at low and middle power levels.
See what you think.
If that fixes the problem (admittedly with less power capability), then . . .
It is your loudspeaker characteristics that are giving your ST-35 so much trouble.
Or, perhaps your original circuit of the ST-35 with its damping factor that is giving your loudspeaker so much trouble.
Please try this, and let us know what you hear with triode wired.
I'm not going down the road of a third quad of output tubes. Why? Because others verify that these tubes (two types) do work in this amplifier kit.
My first step, one that doesn't cost anything and a good place to start, is to retest all of the voltages with the 6P14P-ER and the bias set to what may be (if measured correctly by my meter) 40mA where it sounds great instead of the specified 35mA where it has a harsh edge. I find it interesting that just a 13% increase makes that big a difference in the sound, but it does. Someone pointed out Steph's similar experience when increasing by 5-10%, even though her settings were significantly lower.
To date, nobody here has addressed (as far as I remember) exactly what "problem" or "flaw" (be it design, component value, or whatever other problem type), I am overcoming by setting the bias to where it sounds best. It sounds fantastic with that 13% increase in the current, even though it is driving the tubes even harder than the 100-105% of the original design.
I could do that in the future. Before I do a design change like that, my second step going forward, which I think is very reasonable, is to return the amplifier to its original design from Dynaco, which is one capacitor and one resistor for all four output tubes. That will remove from the equation a whole lot of possible issues that might come from the add-on bias kit sold by the manufacturer, which although it is supposed to work, you never know what might go wrong. When troubleshooting, I think simplifying is a good start. So, I have the proper capacitor on order and the correct resistor came with the kit. I will have to desolder four wires and add a few more, but that's only a few hours of work and easily reversible. I can do that. If it suddenly sounds fine, then we can delve into the bias circuitry.
It's not that I don't trust my meter or that it's known to be bad. It measured all of the resistors and capacitors exactly to their label specs, except the smallest of the feedback caps in the low pF range, where I wouldn't expect it to work well. Someone here also questioned its ability to measure millivolts accurately. I can't say that I have any way to verify its accuracy in the millivolt range. So, the option of getting another meter was presented, and that's manageable. I do trust the one I have to accurately measure the DCR of the speakers. I have no reason to suspect it measuring that.
As I noted previously, I did try the 4 ohm output from the ST-35. All traces of harshness disappeared when biased per spec at 35mA, and we know that my 8-ohm rated Vifa drivers wired in parallel should yield about 4 ohms on paper. So, why not just use the amplifier's 4 ohm output? Sub-bass response wasn't good off the 4 ohm output. It simply wasn't there, making the overall sound unacceptable. I have had the same experience with the 4 ohm outputs on my other two tube amps. The 8 ohm output always sounds better overall, an experience duplicated by many other people with many tube amps over time.
If the current version of the ST-35 can't properly drive a very simple three-way bass reflex speaker from either output, then it's just a poor design, possibly the output transformers which have been changed from 8/16 ohm to 8/4 ohm with the NFB therefore moved from the 16 ohm tap to the 8 ohm tap. I hope that's not the case, but I'm open to the possibility, in which case I'll simply sell it. No other amp has had a problem with my speakers, and that's a half dozen solid state amps and three tube amps. "Which one of these is not like the others, which one of these doesn't belong?" as the Sesame Street song goes.
I think this idea also deserves additional consideration. The specs for C6, C7, and R10 are the exact correct value, verified by both Dynakit and D. Gillespie. I verified the value of R10 before installing. My meter may not be the most accurate for measuring pF caps, but C7 came in at 34pF, at least close to 39pF and likely within my meter's margin of error. I could not get a good read on the 18pF caps. My meter just isn't that good. It jumped all over from about 6 to about 12 and never settled down. It was the same for both of the 18pF caps. No reason to suspect the ones that came with the kit are bad. They are very easily replaced, but I hate to replace components without having a serious reason to suspect them. Thoughts anyone?
The included feedback capacitors are mica. I am a firm believer that different cap types perform differently, and I'm not going to have that argument here. Is it possible that mica is not the best choice here? I'm not a fan of PS Audio, but I did find a video from Paul where he says that mica caps are "terrible" in negative feedback circuits. Interesting. My note: ceramic caps are much maligned in hi-fi. I was amazed when I worked on an old Fisher tube receiver about 30 years ago for a friend. It had some bad caps that I removed and replaced for him. That receiver was absolutely packed with ceramic disc capacitors, and it sounded wonderful. I'm not convinced that they are the root of all evil, and I am wondering if the ceramic disc caps used in the original design might actually be a better choice than the mica caps. I suppose we also have changed the sound using modern film resistors throughout instead of the old carbon types. Maybe that wasn't such a good idea. Who knows.
So, I have a plan, be it ideal or not. 1) Retest voltages with the 6P14P-ER set at 40mA where I like them, 2) If needed, then use the cap that I have on order to go back to the original design, a change that would remove the bias board add-on/mod, 3) Probably get a better meter just because I would like to have the piece of mind and it would not go to waste, 4) Reconsider the negative feedback caps, 5) Test the DCR of the speakers since there seems to be some interest in that and it's easy. 6) Consider this option:
and ... 7) Keep hoping that I can find someone with a scope without shipping the amplifier somewhere, which I'm not going to do.
My first step, one that doesn't cost anything and a good place to start, is to retest all of the voltages with the 6P14P-ER and the bias set to what may be (if measured correctly by my meter) 40mA where it sounds great instead of the specified 35mA where it has a harsh edge. I find it interesting that just a 13% increase makes that big a difference in the sound, but it does. Someone pointed out Steph's similar experience when increasing by 5-10%, even though her settings were significantly lower.
To date, nobody here has addressed (as far as I remember) exactly what "problem" or "flaw" (be it design, component value, or whatever other problem type), I am overcoming by setting the bias to where it sounds best. It sounds fantastic with that 13% increase in the current, even though it is driving the tubes even harder than the 100-105% of the original design.
I could do that in the future. Before I do a design change like that, my second step going forward, which I think is very reasonable, is to return the amplifier to its original design from Dynaco, which is one capacitor and one resistor for all four output tubes. That will remove from the equation a whole lot of possible issues that might come from the add-on bias kit sold by the manufacturer, which although it is supposed to work, you never know what might go wrong. When troubleshooting, I think simplifying is a good start. So, I have the proper capacitor on order and the correct resistor came with the kit. I will have to desolder four wires and add a few more, but that's only a few hours of work and easily reversible. I can do that. If it suddenly sounds fine, then we can delve into the bias circuitry.
It's not that I don't trust my meter or that it's known to be bad. It measured all of the resistors and capacitors exactly to their label specs, except the smallest of the feedback caps in the low pF range, where I wouldn't expect it to work well. Someone here also questioned its ability to measure millivolts accurately. I can't say that I have any way to verify its accuracy in the millivolt range. So, the option of getting another meter was presented, and that's manageable. I do trust the one I have to accurately measure the DCR of the speakers. I have no reason to suspect it measuring that.
As I noted previously, I did try the 4 ohm output from the ST-35. All traces of harshness disappeared when biased per spec at 35mA, and we know that my 8-ohm rated Vifa drivers wired in parallel should yield about 4 ohms on paper. So, why not just use the amplifier's 4 ohm output? Sub-bass response wasn't good off the 4 ohm output. It simply wasn't there, making the overall sound unacceptable. I have had the same experience with the 4 ohm outputs on my other two tube amps. The 8 ohm output always sounds better overall, an experience duplicated by many other people with many tube amps over time.
If the current version of the ST-35 can't properly drive a very simple three-way bass reflex speaker from either output, then it's just a poor design, possibly the output transformers which have been changed from 8/16 ohm to 8/4 ohm with the NFB therefore moved from the 16 ohm tap to the 8 ohm tap. I hope that's not the case, but I'm open to the possibility, in which case I'll simply sell it. No other amp has had a problem with my speakers, and that's a half dozen solid state amps and three tube amps. "Which one of these is not like the others, which one of these doesn't belong?" as the Sesame Street song goes.
I think this idea also deserves additional consideration. The specs for C6, C7, and R10 are the exact correct value, verified by both Dynakit and D. Gillespie. I verified the value of R10 before installing. My meter may not be the most accurate for measuring pF caps, but C7 came in at 34pF, at least close to 39pF and likely within my meter's margin of error. I could not get a good read on the 18pF caps. My meter just isn't that good. It jumped all over from about 6 to about 12 and never settled down. It was the same for both of the 18pF caps. No reason to suspect the ones that came with the kit are bad. They are very easily replaced, but I hate to replace components without having a serious reason to suspect them. Thoughts anyone?
The included feedback capacitors are mica. I am a firm believer that different cap types perform differently, and I'm not going to have that argument here. Is it possible that mica is not the best choice here? I'm not a fan of PS Audio, but I did find a video from Paul where he says that mica caps are "terrible" in negative feedback circuits. Interesting. My note: ceramic caps are much maligned in hi-fi. I was amazed when I worked on an old Fisher tube receiver about 30 years ago for a friend. It had some bad caps that I removed and replaced for him. That receiver was absolutely packed with ceramic disc capacitors, and it sounded wonderful. I'm not convinced that they are the root of all evil, and I am wondering if the ceramic disc caps used in the original design might actually be a better choice than the mica caps. I suppose we also have changed the sound using modern film resistors throughout instead of the old carbon types. Maybe that wasn't such a good idea. Who knows.
So, I have a plan, be it ideal or not. 1) Retest voltages with the 6P14P-ER set at 40mA where I like them, 2) If needed, then use the cap that I have on order to go back to the original design, a change that would remove the bias board add-on/mod, 3) Probably get a better meter just because I would like to have the piece of mind and it would not go to waste, 4) Reconsider the negative feedback caps, 5) Test the DCR of the speakers since there seems to be some interest in that and it's easy. 6) Consider this option:
and ... 7) Keep hoping that I can find someone with a scope without shipping the amplifier somewhere, which I'm not going to do.
Last edited:
There is a lot of useful information in the diytube forum posts on the st-35.
You may find these to be useful
http://www.diytube.com/phpBB2/viewtopic.php?f=1&t=390
http://www.diytube.com/phpBB2/viewtopic.php?f=1&t=6585
Wondering if your increased current is pushing the amp towards Class A operation...(refer the post by "paart" above)
You may find these to be useful
http://www.diytube.com/phpBB2/viewtopic.php?f=1&t=390
http://www.diytube.com/phpBB2/viewtopic.php?f=1&t=6585
Wondering if your increased current is pushing the amp towards Class A operation...(refer the post by "paart" above)