OK, thanks, will do tomorrow:
- 470pF to 1nF across Emitter / Collector of T101. Does it have to be directly on the pins, or can I find a spot on the correct sides of R116 and R114 / R118 resp., if I can't get to the pins ?
- I'll probably leave the 1nF at B/E of Q103 in
- Can I leave R125 (gate stopper of IRFP9140) at 220R ? It didn't restrict bandwidth according to before / after on scope. Or should I revert to 100R ?
Regards, Claas
- 470pF to 1nF across Emitter / Collector of T101. Does it have to be directly on the pins, or can I find a spot on the correct sides of R116 and R114 / R118 resp., if I can't get to the pins ?
- I'll probably leave the 1nF at B/E of Q103 in
- Can I leave R125 (gate stopper of IRFP9140) at 220R ? It didn't restrict bandwidth according to before / after on scope. Or should I revert to 100R ?
Regards, Claas
Update on my R.3
After implementing the mods, the left side of my R.3 behaved nice and stable on the bench. I didn’t try to touch the washer of the 9140 anymore, either.
I then implemented the mods on the right side as well. No offset instability observed after multiple extended runs of the amp.
To recall:
- Raised gate resistor for 9140 (R125 / R225) to 221R
- Set level shifter resistors to 72R (R115) resp. 62R (R215)
- 1 nF across base / emitter of Q103 / Q203
- 1 nF across collector / emitter of T101
The last mod (1 nF C-E of T101) I did only on the left side, because I couldn’t get to a suitable spot for connection on the right side without pulling the board from the heatsink.
The SissySIT R.3 is in my system now and making sweet music.
The right side is dead quiet. One issue remains with the left side, though:
On the left side, however, the previously offending one, I have a low-level wideband noise, sounds the same to me as tube rush does. So no hum, more like white noise. I can hear it with my ear near the loudspeaker. It comes on when the bias comes on.
I didn’t see anything on the distortion plots, but these were from before the mods. Both channels behave identical in terms of bandwidth, square waves, etc. Distortion and harmonic profile is also the same between channels.
From the circuitry, the only difference between the two sides is the additional 1 nF on C-E of T101 on the left side. Can this have something to do with it, or would it be a suspect component ?
Thanks and regards, Claas
After implementing the mods, the left side of my R.3 behaved nice and stable on the bench. I didn’t try to touch the washer of the 9140 anymore, either.
I then implemented the mods on the right side as well. No offset instability observed after multiple extended runs of the amp.
To recall:
- Raised gate resistor for 9140 (R125 / R225) to 221R
- Set level shifter resistors to 72R (R115) resp. 62R (R215)
- 1 nF across base / emitter of Q103 / Q203
- 1 nF across collector / emitter of T101
The last mod (1 nF C-E of T101) I did only on the left side, because I couldn’t get to a suitable spot for connection on the right side without pulling the board from the heatsink.
The SissySIT R.3 is in my system now and making sweet music.
The right side is dead quiet. One issue remains with the left side, though:
On the left side, however, the previously offending one, I have a low-level wideband noise, sounds the same to me as tube rush does. So no hum, more like white noise. I can hear it with my ear near the loudspeaker. It comes on when the bias comes on.
I didn’t see anything on the distortion plots, but these were from before the mods. Both channels behave identical in terms of bandwidth, square waves, etc. Distortion and harmonic profile is also the same between channels.
From the circuitry, the only difference between the two sides is the additional 1 nF on C-E of T101 on the left side. Can this have something to do with it, or would it be a suspect component ?
Thanks and regards, Claas
forgot to say - I believe now you can just chill
some faint noise here and there, you can expect with those big buggers ( or Germanium, as in Old Soul)** but nothing to write home about, if that is only misbehave
just for giggles - try to compare DC voltages across sit gate resistors (R vs. L)
anyway, I hope that Foot is still tapping ......
**if some Gremlin made bad solder point is not well hidden
some faint noise here and there, you can expect with those big buggers ( or Germanium, as in Old Soul)** but nothing to write home about, if that is only misbehave
just for giggles - try to compare DC voltages across sit gate resistors (R vs. L)
anyway, I hope that Foot is still tapping ......
**if some Gremlin made bad solder point is not well hidden
Thanks - I'm actually a bit hesitant to schlepp the amp down to my basement workshop and take it apart again ... so, if the noise doesn't get more over time ...
With a bad solder joint I would expect some temperature-dependency, or it getting worse with multiple heat cycles - let's see.
Amp is now so nicely stable and sounds so good at the moment.
With regard to THF-51S gate current - I didn't measure that on the right side, but on the left side (the one with the slight noise) it had been well below 1 uA all the time. I believe even below 500 nA when the amp was warm, but not hot, but my DMM doesn't resolve enough at those tiny voltages (50 uV is guesswork). Gate current I actually can measure again quite easily up here in the music system, so will do that again.
Regards, Claas
With a bad solder joint I would expect some temperature-dependency, or it getting worse with multiple heat cycles - let's see.
Amp is now so nicely stable and sounds so good at the moment.
With regard to THF-51S gate current - I didn't measure that on the right side, but on the left side (the one with the slight noise) it had been well below 1 uA all the time. I believe even below 500 nA when the amp was warm, but not hot, but my DMM doesn't resolve enough at those tiny voltages (50 uV is guesswork). Gate current I actually can measure again quite easily up here in the music system, so will do that again.
Regards, Claas
This is a super nice amp. I did experience a period during break in where things weren't sounding too good, hard to explain, but I wasn't enjoying it as much. It was just a bit fatiguing or something. Maybe 25 or 30 hours in, I would guess, didn't really keep track. Cleared up after a few long evenings and it's good now for sure.
Offset issue still not solved
My offset instability is still not solved. After playing music for several hours yesterday, I found that the offset on the left side had again risen to 100-140 mV and was fluctuating.
Also, I tried Botte's suggestion this morning and added some solder to the gate connector of the SIT on the noisy side. Alas, no dice unfortunately, noise (like tube rush) is still there.
I guess I should try a different SIT next.
After that, if that doesn't help, I would take the amp board off the heat sink, and reflow all joints related to the gate drive of the IRF510, 9140 and SIT. I would also change the offset pot (P103) and the two gate Zeners (ZD104 and ZD105) when I have the amp board off, just as a precaution.
Sounds reasonable ?
Regards, Claas
My offset instability is still not solved. After playing music for several hours yesterday, I found that the offset on the left side had again risen to 100-140 mV and was fluctuating.
Also, I tried Botte's suggestion this morning and added some solder to the gate connector of the SIT on the noisy side. Alas, no dice unfortunately, noise (like tube rush) is still there.
I guess I should try a different SIT next.
After that, if that doesn't help, I would take the amp board off the heat sink, and reflow all joints related to the gate drive of the IRF510, 9140 and SIT. I would also change the offset pot (P103) and the two gate Zeners (ZD104 and ZD105) when I have the amp board off, just as a precaution.
Sounds reasonable ?
Regards, Claas
well, fact that you have some shot noise and also mild DC offset issue - most logical explanation is misbehaving SIT
remember - DC offset is set in purely passive manner - there is no any trace of active correction of same and - taking that in account , DC Offset behavior of Sissy and similar amps is (now) rock solid
I'm writing that having on mind your other, well behaving, channel
and, yup, pretty much simplest and fastest way of further troubleshooting is changing SIT on that channel or - even better - exchanging them between channels
that way situation is clear from start - you're not introducing 3rd SIT with unknown nature
remark - if it stays that way , 150mV of DC offset is nothing to be worried of
remember - DC offset is set in purely passive manner - there is no any trace of active correction of same and - taking that in account , DC Offset behavior of Sissy and similar amps is (now) rock solid
I'm writing that having on mind your other, well behaving, channel
and, yup, pretty much simplest and fastest way of further troubleshooting is changing SIT on that channel or - even better - exchanging them between channels
that way situation is clear from start - you're not introducing 3rd SIT with unknown nature
remark - if it stays that way , 150mV of DC offset is nothing to be worried of
Thanks, Zen Mod.
I will change SIT next.
First, some more data points from today, just to document:
- Offset rises first very gradually from 2-4 mV (where the other channel stays the whole time) to may be 20-25 mV over the course of two hours, then takes off to 100-150 mV and fluctuates by maybe 50 mV. Fluctuations go to over 200 mV.
- Gate curent of SIT stays very low. Below 0.5 uA at first (resolution threshold of my most sensitive DMM) to 0.5 uA and now 0.6 uA when offset is high. Gate current rock-steady.
- I didn't hear anything different in the music.
- Noise stays the same
- This SIT didn't exhibit any noise in original SissySIT. Offset was so much temperature- and rail-voltage dependent then that I can't say if I had an offset issue in original SissySIT.
Regards,
Claas
I will change SIT next.
First, some more data points from today, just to document:
- Offset rises first very gradually from 2-4 mV (where the other channel stays the whole time) to may be 20-25 mV over the course of two hours, then takes off to 100-150 mV and fluctuates by maybe 50 mV. Fluctuations go to over 200 mV.
- Gate curent of SIT stays very low. Below 0.5 uA at first (resolution threshold of my most sensitive DMM) to 0.5 uA and now 0.6 uA when offset is high. Gate current rock-steady.
- I didn't hear anything different in the music.
- Noise stays the same
- This SIT didn't exhibit any noise in original SissySIT. Offset was so much temperature- and rail-voltage dependent then that I can't say if I had an offset issue in original SissySIT.
Regards,
Claas
Offset
Here’s an update on the first of two issues with the left side (R channel board) of my SissySIT R.3 – offset instability. Noise will follow in a separate post.
I changed in another SIT at the left side and adjusted R115 to 87R to accommodate its 2.62V measured Vgs. I know that I was supposed to exchange SITs between channels, but hoped to get away with only one SIT change, in case this worked already.
At first, no change – offset instability still crept in after a few hours. I then snipped out the 1 nF across C-E of T101 and re-adjusted offset so that it stays slightly negative or gets single-digit positive at max. Offset seems to be stable now even after many hours of operation, several cycles, variations of rail voltages, and also after disturbances by, for example, switching off the pre-amp.
So I will continue to watch the offset of the left side, but don’t feel the need to act on the offset at the moment.
Regards, Claas
Here’s an update on the first of two issues with the left side (R channel board) of my SissySIT R.3 – offset instability. Noise will follow in a separate post.
I changed in another SIT at the left side and adjusted R115 to 87R to accommodate its 2.62V measured Vgs. I know that I was supposed to exchange SITs between channels, but hoped to get away with only one SIT change, in case this worked already.
At first, no change – offset instability still crept in after a few hours. I then snipped out the 1 nF across C-E of T101 and re-adjusted offset so that it stays slightly negative or gets single-digit positive at max. Offset seems to be stable now even after many hours of operation, several cycles, variations of rail voltages, and also after disturbances by, for example, switching off the pre-amp.
So I will continue to watch the offset of the left side, but don’t feel the need to act on the offset at the moment.
Regards, Claas
White noise
Now to the second issue – the low-level noise on the left side.
It was absolutely unchanged after changing the SIT on that side.
I then did some more diagnostics to try to locate the problem. To cut to the chase, with Jumper JP101 open (before the transformer) and the bridge R111 connected to ground:
No noise whatsoever, total silence like on the right side ! This would probably mean that the whole output stage is out of the picture as a cause for the noise, or not ?
My findings, noise level not measured, but relative indication how loud I hear it:
Right side at normal state – noise level 0
Right side jumper open – noise level 2
Left side normal state – noise level 5
Left side jumper open – level 15, so, much louder
Left side jumper grounded, before transformer – level 1 – only very slight noise
Left side, bridge R111 grounded – silence
I guess I should focus my inquiries on the area around the transformer, from the jumper through the transformer pins to R111 ? Loud noise with the jumper open, does this point to something ?
Gain, waveforms, frequency response is identical between channels. I have Cinemag HPC transformers, by the way.
Regards, Claas
Now to the second issue – the low-level noise on the left side.
It was absolutely unchanged after changing the SIT on that side.
I then did some more diagnostics to try to locate the problem. To cut to the chase, with Jumper JP101 open (before the transformer) and the bridge R111 connected to ground:
No noise whatsoever, total silence like on the right side ! This would probably mean that the whole output stage is out of the picture as a cause for the noise, or not ?
My findings, noise level not measured, but relative indication how loud I hear it:
Right side at normal state – noise level 0
Right side jumper open – noise level 2
Left side normal state – noise level 5
Left side jumper open – level 15, so, much louder
Left side jumper grounded, before transformer – level 1 – only very slight noise
Left side, bridge R111 grounded – silence
I guess I should focus my inquiries on the area around the transformer, from the jumper through the transformer pins to R111 ? Loud noise with the jumper open, does this point to something ?
Gain, waveforms, frequency response is identical between channels. I have Cinemag HPC transformers, by the way.
Regards, Claas
By the way, DC resistance (jumper open)
- Jumper "X" side to ground: 30R
- R111 to ground: 117R
Also, upon further thought, the fact that the noise goes away if R111 is grounded probably does not automatically exculpate the output stage. C103/C104 do block DC, but they might let noise from the output stage through nevertheless ...
- Jumper "X" side to ground: 30R
- R111 to ground: 117R
Also, upon further thought, the fact that the noise goes away if R111 is grounded probably does not automatically exculpate the output stage. C103/C104 do block DC, but they might let noise from the output stage through nevertheless ...
Last edited:
From the evidence, noise is generated in some semiconductor at the output stage. Connecting transformer input to the ground makes AC grounding of IRF510 and IRFP9140 gates through three transformer windings with 90 Ω remaining DC resistance. Connecting to the ground at R111, what is actually transformer output, AC connects those gates directly to the ground. That’s why noise is lower in that case.
It looks that noise is present at IRF510 and IRFP9140 gates. IMO, possible sources are those MOSFETS or PC817 optocoupler. Damn, any semiconductor in the OS except SIT and replaced Zener diodes is suspect. Not funny.
It looks that noise is present at IRF510 and IRFP9140 gates. IMO, possible sources are those MOSFETS or PC817 optocoupler. Damn, any semiconductor in the OS except SIT and replaced Zener diodes is suspect. Not funny.
Thanks, tombo56 !
- 9140 has been replaced already because of suspected role in offset instability
- SIT has been replaced
That would leave the IRF510. When I take the amp apart and the board off the heatsink, I will probably also replace the three Zeners to the Gates (ZD101, ZD104, ZD105) and the offset pot P103, while I'm at it.
Actually, first I will change the two optocouplers. Because they are socketed, that I can do without taking the amp apart. I did change those already eraly in the process, but with two from the same batch. Will try a different make now.
Regards, Claas
- 9140 has been replaced already because of suspected role in offset instability
- SIT has been replaced
That would leave the IRF510. When I take the amp apart and the board off the heatsink, I will probably also replace the three Zeners to the Gates (ZD101, ZD104, ZD105) and the offset pot P103, while I'm at it.
Actually, first I will change the two optocouplers. Because they are socketed, that I can do without taking the amp apart. I did change those already eraly in the process, but with two from the same batch. Will try a different make now.
Regards, Claas