TriodeUSA ST-70 amp clipping below rated output

Two issues needing attention, perhaps pointing to same problem.

My (TriodeUSA) ST70 is clipping on both channels at around 22W as carefully measured by oscilloscope. Only the phase splitter has been changed (to a nicely balanced 12AU7A) since buying the amp used. Came with new looking JJ EF806Ss and EL34s. I changed out the phase splitter because the negative half of the signal was distorting long before the positive. With the RCA clear top 12AU7A the waveform is symmetrical but still clipping around 22W.

Lately, the left channel bias is moving day to day. I aim for .700 on each side. I check after about two hours or so of listening. One day the left will be at .675 (right moved up slightly as expected but not by as much) and the next day it will be at .715, right lower by a bit again as expected. Is this excessive movement day to day?

Is it time for new power tubes, new EF806Ss, or both?

Or should I be investigating other areas?

Thanks,
Jeff

jeffjmr wrote:Two issues needing attention, perhaps pointing to same problem.

My (TriodeUSA) ST70 is clipping on both channels at around 22W as carefully measured by oscilloscope. Only the phase splitter has been changed (to a nicely balanced 12AU7A) since buying the amp used. Came with new looking JJ EF806Ss and EL34s. I changed out the phase splitter because the negative half of the signal was distorting long before the positive. With the RCA clear top 12AU7A the waveform is symmetrical but still clipping around 22W.

Lately, the left channel bias is moving day to day. I aim for .700 on each side. I check after about two hours or so of listening. One day the left will be at .675 (right moved up slightly as expected but not by as much) and the next day it will be at .715, right lower by a bit again as expected. Is this excessive movement day to day?

Is it time for new power tubes, new EF806Ss, or both?

Or should I be investigating other areas?

Thanks,
Jeff

Try new matched power tubes.

bias variation will depend on voltage variation on your AC wall outlet, your variation seems within normal range (maybe +/-2%)
As for the clipping, from your description it would seem to be the driver board isn't really optimized for the tubes, and some resistors would probably have to be changed
so that the phase splitter creates equal outputs, you'd probably need an oscope to see those waveforms and then adjust the circuit accordingly.
Well, you do say the waveform is now symmetrical, so maybe it just doesn't have enough output swing. I'm not familar with that board, but might possibly need a higher voltage
to the phase splitter, or the phase splitter circuit parameters need to be adjusted (plate & cathode resistors).

I should clarify.

The signal at the splitter output was asymmetrical. I replaced the tube with a nice balanced one and the output from the splitter is now symmetrical.

The amp’s speaker output clipped at 22 watts before and after the splitter tube replacement.

The bias on the left channel is swinging day to day quite a bit more than the right side. I know some variation can be accepted but the left moves by a factor of two or three times the swing of the right. And as expected, the right is moving opposite the moving of the left, but not an equal amount.

I have tested the output tubes with an uncalibrated tester and they are not well matched from a transconductance viewpoint. But both channels clip at the same level, so your point about the splitter voltage is well taken and I will check it. It’s not that I need the extra watts; the Altecs are 103db+ efficient. Just suspect something is wrong that may be affecting fidelity.

Thanks,
Jeff

jeffjmr wrote:I should clarify.

The signal at the splitter output was asymmetrical. I replaced the tube with a nice balanced one and the output from the splitter is now symmetrical.

The amp’s speaker output clipped at 22 watts before and after the splitter tube replacement.

The bias on the left channel is swinging day to day quite a bit more than the right side. I know some variation can be accepted but the left moves by a factor of two or three times the swing of the right. And as expected, the right is moving opposite the moving of the left, but not an equal amount.

I have tested the output tubes with an uncalibrated tester and they are not well matched from a transconductance viewpoint. But both channels clip at the same level, so your point about the splitter voltage is well taken and I will check it. It’s not that I need the extra watts; the Altecs are 103db+ efficient. Just suspect something is wrong that may be affecting fidelity.

Thanks,
Jeff

Loss of power is one symptom of worn-out power tubes. Get a new matched set of tubes!

You don't say if you had previously seen 35 watts per channel on your scope.  It sounds to me like the El34s are triode connected.

sayntjack wrote:You don't say if you had previously seen 35 watts per channel on your scope.  It sounds to me like the El34s are triode connected.

The amp has clipped at around 22w since I bought it, and it is as I bought it as far as power tubes is concerned. The only changes I have made to the amp are the phase splitter tube to the RCA 12AU7A, an inrush current limiter, and the “yellow sheet” diode mod. B+ did not change with these mods. It has always been 445-450.

But I just realized my test setup may have had an influence. More experienced folks may be able to tell me if it matters: I used 16ohm 50w non-inductive resistors for my load. My Altec speaker/crossover network are a nominal 10-12 ohm impedance. I use the 8 ohm speaker outputs. And frequency doesn’t appear to have much of an effect across the audio spectrum.

If it matters, I can test again and record the input power that resulted in the 22w clipping. I did not record that value the two times I tested. I can only say it was below 5vpp since that is the limit of the signal generator that I used.

As I said, with the original JJ EF83S I had a significant distortion on the negative half of a sine wave signal which prompted the change to the RCA 12AU7A at which point that distortion disappeared. But clipping remained at 22W at the speaker outputs. Drivers are JJ EF806Ss, and the power tubes are JJ EL34s which are not well matched.

I am currently trying to decide between EH 6550s, EH 6CA7s, Genelex KT66s, KT77s and KT88s. So many choices!

Thanks for any advice,
Jeff

jeffjmr wrote:

sayntjack wrote:You don't say if you had previously seen 35 watts per channel on your scope.  It sounds to me like the El34s are triode connected.

The amp has clipped at around 22w since I bought it, and it is as I bought it as far as power tubes is concerned. The only changes I have made to the amp are the phase splitter tube to the RCA 12AU7A, an inrush current limiter, and the “yellow sheet” diode mod. B+ did not change with these mods. It has always been 445-450.

But I just realized my test setup may have had an influence. More experienced folks may be able to tell me if it matters: I used 16ohm 50w non-inductive resistors for my load. My Altec speaker/crossover network are a nominal 10-12 ohm impedance. I use the 8 ohm speaker outputs. And frequency doesn’t appear to have much of an effect across the audio spectrum.

If it matters, I can test again and record the input power that resulted in the 22w clipping. I did not record that value the two times I tested. I can only say it was below 5vpp since that is the limit of the signal generator that I used.

As I said, with the original JJ EF83S I had a significant distortion on the negative half of a sine wave signal which prompted the change to the RCA 12AU7A at which point that distortion disappeared. But clipping remained at 22W at the speaker outputs. Drivers are JJ EF806Ss, and the power tubes are JJ EL34s which are not well matched.

I am currently trying to decide between EH 6550s, EH 6CA7s, Genelex KT66s, KT77s and KT88s. So many choices!

Thanks for any advice,
Jeff

There is no such animal as JJ EF83S. Possibly ECC83S. But replacing this with a
12AU7 will upset all tube alignment, If there should be an ECC83 / 12AX7 the
trouble shoothing should start with using this.

Usually lack of power comes from worn out power tubes. Have you tried replacing them ?

Senior moment, Peter. You are correct. ECC83S.

Tube alignment? Please explain. The 12AU7A comes highly recommended from other ST70 users as a suitable replacement for a 12AX7, albeit “slightly” less gain. Is 35w to 22W considered slight? All voltages have been carefully checked against the manufacturer’s recommendations and all are in spec.

I am taking your advice to replace the power tubes once I settle on a type.

Thanks,
Jeff

jeffjmr wrote:Senior moment, Peter. You are correct. ECC83S.

Tube alignment? Please explain. The 12AU7A comes highly recommended from other ST70 users as a suitable replacement for a 12AX7, albeit “slightly” less gain. Is 35w to 22W considered slight? All voltages have been carefully checked against the manufacturer’s recommendations and all are in spec.

I am taking your advice to replace the power tubes once I settle on a type.

Thanks,
Jeff

No circuit designed for a 12AX7 will function well ( in a hifi amp) with a 12AU7
Guitar amps where distortion is wanted in a controlled manner could accept this.

Now for lack of power. As this does not seem to be a Dynaco st70 could you
supply schematics for the amp ? No Dynaco uses 12AX7 as phase splitter.

Jeff ---- for what it is worth, (and from what I remember of the Triode St70) in the late 1990s or early 2000s there were several versions of that PCB along with Triode's Kit Version.

I purchased one of the PCBs (6AU6 version) but had trouble getting it to work properly.  Seems like there was a series of component updates and after awhile I gave up and bought another board from someone else.

Here are some links that might be helpfull.

Stew

The SDS Sheldon Stokes ST70 PCB was originally sold thru Triode in the Chicago area.....

https://www.quadesl.com/refurbishment/refurb_dynaST70.html

and there were 2 versions, with EF86 or 6AU6 tubes.

https://www.quadesl.com/refurbishment/pdf/st70board_ef86.pdf
https://www.quadesl.com/refurbishment/pdf/st70board_6au6.pdf


There was another "release or version" of the Stokes/Triode PCB at this link.

http://site.triodestore.com/ST70Instructions.pdf


See the following link for TRIODE's ST70 Full Kit..........

http://www.triodeelectronics.com/st70buildkit.html

I am beginning to understand that I’ll need to be more careful when scanning the web for info on “ST70s”.

But as my schematic below shows, the 12AU7 is actually recommended though the assembly manual references several other options for the splitter as well as power tubes and triode mode with the attendant wiring and resistor changes.

That being said the amp as I bought it came with the JJ ECC83S. Peter has me concerned that though my amp matches the schematic, was the ECC83S a bad match or does it still look like a 12AU7 is a mistake?

But given the 12AU7 lower gain, perhaps that is the reason for the decreased output? As I mentioned before, I don’t need the extra power. Just want to be sure the rest of the amp is performing as expected before I invest in new output tubes.

BTW, I’m leaning towards Tung Sol 6L6GCs. Any opinions?

Jack here again.  From the schematic that you posted of the Triode Elec. Dynaco ST70, your amp requires an EF86 in the left and right positions on the PC board and a 12AU7 in the center.  A good option to the 12AU7 is a 12BH7 which just gives you a little more of everything.  Stick with the EL34s for output tubes as that is what the circuit and output transformers work best with.  My suggestion for EL34s is the current production TungSol EL34Bs.  They are a slight variation of an EL34 with better dynamics and highs and lows.  The EL34s sound best operating at 45-50ma per tube.

Now, as to your measured max 22W output. The likely possibilities are: It is wrongly measured, there is an issue with circuitry / components or, as I suggested before, the EL34s are wired in triode mode which is often done.  I suggest that you check the wiring at the EL34 sockets to verify if this is the case.  That is, establish if the green and green/white wires on the output transformers are connected to pin 4 of the output tubes.  If they are, then you have the ultralinear connection with 35 Watts available.  If not, then you will probably find that pin 4 is connected to pin 3 via a 100 ohm resistor.  This would be a triode connection which would produce near the 22W that you measured.

jeffjmr wrote:I am beginning to understand that I’ll need to be more careful when scanning the web for info on “ST70s”.

But as my schematic below shows, the 12AU7 is actually recommended though the assembly manual references several other options for the splitter as well as power tubes and triode mode with the attendant wiring and resistor changes.

That being said the amp as I bought it came with the JJ ECC83S. Peter has me concerned that though my amp matches the schematic, was the ECC83S a bad match or does it still look like a 12AU7 is a mistake?

But given the 12AU7 lower gain, perhaps that is the reason for the decreased output? As I mentioned before, I don’t need the extra power. Just want to be sure the rest of the amp is performing as expected before I invest in new output tubes.

BTW, I’m leaning towards Tung Sol 6L6GCs. Any opinions?

The triode ( ECC83/ECC82) is connected as a concertina. Simply stated is that
the cathode will follow it's input as a cathode follower i.e. with amplification slightly
below 1. As the cathode and plate resistors are equal the plate voltage will
vary exactly like the cathode. As long as the tube can cope with the current req
it does not matter what tube to use. But ECC83 / 12AX7 will have problem
with the app 3mA that flows, causing unlinearity or possibly early clipping.
Stay with ECC82/12AU7

The low power could be caused by worn EL34 or triode connected EL34 as suggested by jack.

As you can see the amp matches the schematic and does not appear to me to be in triode mode. The missing plastic end of the cap is due to sliding the (heavy to me) amp on and off the base and catching an edge. The cap tests good. The three mods I made, yellow sheet diode, inrush current limiter and mains rf filter are evident. Oh, and new larger speaker output terminals on the 8 ohm taps.

I threw the original ECC83S back in temporarily and measured and it performed similarly, 22watts at clipping.

BUT, I also retested using a load closer to my actual speaker/dividing network impedance, 10-12 ohms. . The only non-inductive resistors I have are 16 ohm 50w. Testing with the 16 ohm I read 18.7 vrms, as clipping just begins to be noticeable on the scope. I paralleled a 24 ohm 10w resistor with my 16, for a combined 9.6 ohms and ran the text again and got 18.2 vrms before clipping! That’s pretty close to 35w! (I don’t own a programmable load but hope to get one soon). Perhaps this indicates all is well?

Also, let me take this opportunity to thank Bob for this terrific, helpful forum. Feel a bit guilty taking up so much resources for a non-VTA amp but it is highly appreciated. And wish you the best in your retirement. I’m not far behind you!

Thanks all for your input.
Jeff