Strange Distortion Problem - Please Help

[LinuxGuru]

Hi, tube folks !

Having run into strange problems I'm now have to call for community help.

I have built 40W PP AB1-class 6L6GC amp 485V B+ 5K, original Sansui AU-111 schematic and mod (made by someone in Japan) both attached. The guy from Japan built a power amp using original AU-111 transformers. I stayed upon original schematic (I kept RC 150pF + 22K in cathode of 6AQ8 which seem to help improve stability).

Since I did not have original Sansui AU-111 transformers, I used my own 60W unit, which also have 5K primary, separate CFB (same as original), 4/8 Ohm output (no 16 Ohm tap), and linear frequency response from 20Hz - 100 KHz. Amp is being powered by stabilized regulated power supply capable of 360 - 500V 0.5A. For a while I used 6P3S-E (40 mA idle current) and Shuguang 6L6GC (50 mA idle current).

First check I did with oscilloscope and generator. Amp is capable of 16 - 17V RMS output voltage (which corresponds to 36 - 40W of output power) on 8 Ohm load (50W resistor), very nice smooth looking sine-wave. Square-wave also seem to be good even on 10 KHz frequency. Unfortunately, I have single-beam oscilloscope and cannot compare two sine-waves side by side.

As it often happens during tests, once I have overheated the tube, and once even occasionally shorted two wires while measuring voltage (fuse was blown and PCB developed some burn marks).
I have cleaned PCB and replaced one of the 10 ohm cathode protection resistors which was seem to be measuring fine but looked strange due to the physically close location to recent short.

Now the problem. At relatively small signal level (several V) THD are 0.8% which is what it supposed to be (measured with RMAA and ESI Juli sound card), and IMD 1.8% (should be 1% or less). Unfortunately, once output level increases distortions also go unacceptably high. They are clearly visible on oscilloscope. What is really strange, 200 HZ/14V sine-wave looks perfectly fine, while 20 Hz is visible distorted crossing zero region (switching tubes ?).
Frequency response is +0.1/-0.12dB in 20Hz - 20 KHz range (look at the image attached, its perfect), noise level -94dB.
+1dB/-0.5dB peaks/dips above 30 KHz do not matter, not audible anyway.
Small overshoot/undershoot at square wave test is absolutely normal for big tranformer.

The first thing come into mind is to replace tubes what I of course did (I have tube tester so I can check them, all were OK). Second, it could be wrong polarity of CFB winding, but in this case it is not (amp oscillates if CFB connection polarity is wrong). Disconnecting CFB also yields no changes.

Suspecting power supply problem, I did the following. I disconnected amp from 6080-based regulated voltage stabilizer, and connected straight to the CLC filter (200 uF + 5H + 50uF). Then, I even removed 5H choke. Nothing changes.

Output transformer was another possible culprit - or to be precise, saturation of core at low frequencies. Reducing idle current of one of output tubes down from 50mA to 40mA, I got the following results at 20Hz:

Idle current VL1/VL2 --- V(out) --- W(out) --- Drop (dB)
I1=I2=50 mA --- 15V --- 28.13W --- (0dB)
I1=50 mA, I2 = 40mA --- 14V --- 24.5W --- (-0.6dB)
I1=50 mA, I2 = 33mA --- 13V --- 21.13W --- (-1.24dB)

Although distortions went visibly higher, output level dropped insignificantly, only -1.24dB. And this with imbalance which is simply huge by any means ! Thus, transformer core is very far from saturation.

Anyone have an idea what could be wrong?
Thanks in advance for any suggestion or help.

 

[battradio]

Hi ,
Noticing the the original trnsformer has a cathode winding for the output tubes , the feedback and drive would have to be adjusted to to alow for this . What transformer did you use for the output .

 

[LinuxGuru]

Hi ,
Noticing the the original trnsformer has a cathode winding for the output tubes , the feedback and drive would have to be adjusted to to alow for this . What transformer did you use for the output .

I use custom-built transformer with CFB, 5K primary, and 4/8 Ohm secondary. 60W rated instead of original 45W. BTW, even with CFB disconnected, the same distortions are still here.

 

[battradio]

Try swaping in a conventional transformer to see if that makes a differance .

 

[BinaryMike]

I also say it's likely to be an OPT problem. Obvious distortion occurs at LF only, and when the signal slews in one direction only. That sounds like a core magnetization issue to me. Even so, you should add screen stopper resistors of at least 100R to this design. RF parasitic oscillations can be VERY sneaky.

 

[seta1]

Hi,

Ok, having looked at your square waves, you have some serious issues at the lead edge. This lead edge bump translates into one nasty sounding amp. This can be caused by a lot of things. First do a test with your scope, while injecting the amp with a square wave of around 10Khz at it's input, look for the place where you first see that lead edge bump in the wave form. That is your problem area. You will need to test basically at the plate of each tube with a probe capable of blocking at least 1000 VDC. The common way of lowering this or eliminating this distortion of the square wave is to use an anti ringing capacitor in the negative feedback network. The original schematic shows a 250P capacitor across a 4.7K resistor which should do the trick. The amps I build do not have this even at frequencies as high as 30K and as low as 8hz. Of course, most tube amps at under 150hz make a square wave look more like a saw tooth wave. Your 200hz wave looks great with minimal tilting, but it still has that nasty lead edge spike. Now, when testing, if you don't see this spike, even at the plates of the 6L6's, then you need a better quality output transformer. But if you see it anywhere before the output transformer, then your circuit needs some reworking. I see this a lot in cheap Chinese output transformers. It is probably why a lot of their amps (many of which I hear people talking about buying right here on Audio Karma) exhibit this when I run tests on them when servicing them. You can only tweak the circuit so much before you run out of options when trying to compensate for a poorly wound and designed output transformer. Your amp also looks as if it's phase splitter or something is slightly out of balance. The top half of the square wave is slightly different as compared to the bottom half. This can be the phase splitter, mismatched output tubes and or the output transformer. Also you have nasty crossover distortion in that one sine wave. You have a couple issues here. I'm betting your phase splitter is not balanced, your output tube bias or the tubes themselves are not balanced and your output transformer is less then perfect for this application. The amp will most probably sound reasonably good as is but not world class. But, tinkering with a design is what it is all about in this hobby. It is humbling to find how much thought needs to go into perfecting such a seemingly simple amplifier design. I ran into this problem once in a push pull ultralinear class A amp I designed. It is a 22 watt amp using a pair of EL34 output tubes biased at full plate dissipation at idle. I traced a similar problem to an imbalance in my phase splitter and a poor quality output transformer. I replaced the transformer with a Hammond 60 watt 6,600 ohm primary job then added an AC balance pot to the 6SN7 phase splitter. I adjusted everything and voila! problems solved. The amp is cathode bias and I did use separate bias resistors funneled into a 25 ohm 5 watt pot which allows some slight DC balance of the output tubes. All this 'fixed' issues that were of a similar nature to yours (although of a much smaller magnitude).

 

[FalconEddy]

Some good thoughts here from everyone.

seta1 had mentioned phase issues, and based on that sine wave at 20Hz I tend to agree.

If you look really closely, you'll see that the anomaly starts around 135 degrees and continues another 135 degrees all the way to the 270 degree point.

At first glance I thought it also had a phase jump at 180 degrees, but a closer look will show the issue starts at 135. This point in the sine wave cannot be matched to the 270 degree point. So even though the wave appears to push ahead at 180 degrees, the waveform is still completely shifted (phase lag) from it's normal path at 135 until it hits 360. At that point the waveform starts off normal again until it's interrupted by that phase lag again at 135 degrees. The crossover distortion must sound really nasty at those frequencies.

Since this appears to be frequency specific, the issue can most like be resolved using a tweak to balance the phase and/or the OPT's aren't optimal for this amp design.

Could I get a 24 bit (or 16 bit) .wav file (a small 1 second snip) to analyze it?

Regards,
FalconEddy

 

[battradio]

Could the cathode windings be out of phase

 

[FalconEddy]

Could the cathode windings be out of phase

Based on his description, probably not.

He has no oscillation issues, and disconnecting the CFB section from his custom-wound OPTs has no effect on the anomaly.

The cause of this issue has camoflaged itself pretty well, but since he has a fairly good handle on it, it will eventually be uncovered.

The nice thing about using different avenues of troubleshooting for a specific issue that's a real dog to pinpoint is that eventually you'll eliminate all the possibilities. Once all have been eliminated you start getting creative by branching out into the 'not so specific' issues, which allow you to focus in multiple directions. That's usually the point in time where you find out that 2+2=3, and knowing that's incorrect, the cause is somewhere very nearby.

However, with that said, tough issues that require a process of elimination method are sometimes excrutiatingly slow and time consuming to identify; but always give you a much higher level of satisfaction once they're resolved.

. . Falcon

 

[LinuxGuru]

Thanks a LOT for everyone offered help and advices!

Call me whatever, but after long search I found source for this not so obvious problem with distortions on low frequency/high signal level. The most difficult part was the fact that these distortions appear only below 50Hz starting from 10V of output RMS voltage.
When assembling amplifier, I occasionally swapped 33K and 47K resistors in phase splitter. Unbelievable, how so simple mistake lead to so long solving & difficult problem.
Additionally, amplifier seem to be very sensitive to twin triode mismatch in phase splitter, as well as mismatch of output tubes. As suggested Japanese tweaker, I have decreased NFB (down from 22dB), so amplification factor went from 10 to 15, THD increased a little bit but are not audible anyway.
After all, I've got very nice unit. Frequency response 20Hz-50KHz, +0.1/-0.36dB in range 20Hz-20KHz, noise level -89dB. Depending upon output power (and 6L6GC idle current 40mA), THD is in range of 0.1% - 0.96% (with pair of Shuguang 6L6GC which were overheated countless number of times, and draft assembly with very long interconnect wires all over the table).

Additionally, this amp proves push-pull output transformer design concept (method of equivalent generators with check at several critical points with and without DC BIAS) which allows to construct device which is very immune to idle current imbalance (without micro-air gap). Most units tolerate max 5mA, main has only -1.24dB drop at 20 Hz/40W and enormous imbalance, I1 = 50mA, I2 = 33mA (of course distortions are visible on oscilloscope, but sine-waves have symmetrical peaks and output voltage drop is very small). Transformer rated at 60W, I'm currently do not have amplifier to test at this power, but even at 40W this numbers are impressive.

A little more about square-wave test. I don't think it have much of usefulness, especially taking into account that interpretation of the result is quite difficult. Square wave composed of leading/falling edge (rise time of several MHz), and DC between them. This is not something large audio transformer supposed to work with. Take for example overshoot. Formula for calculating overshoot is quite complex, but combination of lower leakage inductance and lower distributed capacitance lead to higher overshoot. So, almost perfectly looking 2KHz square-wave on output may be a sign of not so good transformer after all.