^Suspect that the 'base stopper' resistors you added might be the ticket here since my drivers are modern but my outputs are old..
LLM suggests checking zobel network..
New to me AU-666
- Thread starter bhairavi
- Start date
LLM suggests checking zobel network..
Well I tried adding resistors across the BE of the outputs, and it helped significantly, but didn't get rid of the ringing completely.
Alternative suggestion was snubber caps to the drivers. I don't have any small enough. At this point I'm just working off what AI is telling me, doesn't feel very good, very likely to send me down the wrong rabbithole.
I've spent way too much time on this and half feel like throwing this thing in the the trash.. Probably should walk away for now.
Alternative suggestion was snubber caps to the drivers. I don't have any small enough. At this point I'm just working off what AI is telling me, doesn't feel very good, very likely to send me down the wrong rabbithole.
I've spent way too much time on this and half feel like throwing this thing in the the trash.. Probably should walk away for now.
I can’t remember on the 666 but I recall on the 999 it had quiet a few carbon resistors that had drifted a lot on the amp modules, might be worth checking some values on any carbon resistors here
Scott Muni
Well-Known Member
My au 999 had a open carbon resistor on F1230, which supplies negative voltage. No visual damage, found when ohm checking all resistors during the re-cap.
I humbly suggest it might be a good idea to try a test run with the feedback path open. Oscillating starts on the negative half cycle and quits when the upper driver transistor has taken over on the other half cycle. I'd pull R819 and C823, run the amp as before and look at the outputs to see if you've killed the oscillation. TR803 and TR805 have a fun little positive local feedback thing going which tends to excite oscillation. I think there's a value off in one of the feedback network components or possibly too much gain in one of those transistors.
Unfortunately (for this issue, not for me
) I'm off to go fishing so my response time will lag....a lot
Checked all the resistors on both boards and also the negative/positive filter boards. On the filter boards the large resistors have drifted a bit, around 20%.
I humbly suggest it might be a good idea to try a test run with the feedback path open. Oscillating starts on the negative half cycle and quits when the upper driver transistor has taken over on the other half cycle. I'd pull R819 and C823, run the amp as before and look at the outputs to see if you've killed the oscillation. TR803 and TR805 have a fun little positive local feedback thing going which tends to excite oscillation. I think there's a value off in one of the feedback network components or possibly too much gain in one of those transistors.
Unfortunately (for this issue, not for me
Well when I took out R819 and C823 (15pf, on the underside of the board) the oscillation went away but now there is a major hum.
R819 tests good. I can't tell if the 15pf tests bad or if my transistor tester is unable to read such small values. The fluke DIMM definitely can't resolve to this level.
I'll be able to get some replacements tomorrow and see that does.
How would I tell if the gain was 'too' high for the circuit on any of these transistors?
How different is the gain from the originals? I remember the flying saucer to be low gain. I’ve never tried ksc2383 in that circuit can you try out ksc2690 or bd139?
gslikker
Super Member
I am wondering about C809. As there is a 33 Ohms resistor to the base of the NPN driver, C809 may not help that much turning it off, who knows maybe a capI humbly suggest it might be a good idea to try a test run with the feedback path open. Oscillating starts on the negative half cycle and quits when the upper driver transistor has taken over on the other half cycle. I'd pull R819 and C823, run the amp as before and look at the outputs to see if you've killed the oscillation. TR803 and TR805 have a fun little positive local feedback thing going which tends to excite oscillation. I think there's a value off in one of the feedback network components or possibly too much gain in one of those transistors.
Unfortunately (for this issue, not for me
of 1 microfarad or something between collector and emitter of the TR809 is better. In the end the new driver transistors are rather fast. Of course, my comment is just an uneducated guess.
BTW my way to protect these Sansui amps is using power resistors during tweaking bias and offset, no need for the dbt then.
I do not know whether it would affect oscillation. If fighting oscillation, you will have to try different loads (or at least load and no load) anyway.
I used blown fuses having 390 ohms resistors soldered sideways.
Last edited:
Don't rush to judgement about C823 &/or R819 being 'the' problem, it's more that you now know more about where to look. Breaking the feedback loop is a way to remove some very hard to intuitively understand circuit behavior (frequency dependent) by simplifying things. One effect of removing the feedback is that the now 'open loop' gain is higher, so no surprise that the hum is louder. But where is the hum coming from and what's different about the components in that feedback signal path between channels are a couple of good things to dig in to.
I see I missed a connection for so my comment about positive feedback between TR803 and TR805 is of the mark.
Closing the loop again is the way to test your possible fixes
Did you do hfe matching for the 2 transistors at the front end of the amp? TR801/803?
Last edited:
JyriK
Active Member
That it will certainly do. It will create solutions for you based on nothing that will lead to nowhere. Whenever it tells you to check the Zobel, you know it doesn't have a ****ing clue..
Tr801/803 are 992 and hfe is 330/338
Tr805/807 are 2383 and hfe is 202/202, and I tried replacing them with BD139 with hfe of 130/130.
Replaced 15pf caps across R819 with new ones.
Also tried adding 47pf caps across BC of the drivers (TTA/TTC004B) (this was an LLM suggestion).
No change in oscillation on the rise.
The amount of oscillation almost completely disappears when I increase the volume of the input to max (using main in function so nothing in the preamp muddies the picture).
Last edited:
Wait so is that an alternate schematic?
And I would be trying different values across the BC of the VAS transistors Tr805/807? Meaning try changing values at C805 or C811?
Thank you! Will try adding the miller cap if not already there. And increasing c811. And then additional cap across BC of tr805 and 807.
It worked! So on my amp boards there is a 100pf at C811, but an identical cap in parallel on the underside = 200pf. There was no 47pf, and when I added this (BC of Tr807) the oscillation disappeared, and sine wave is clean at low-medium volumes.
BUT, at high volume, there is still distortion, and the output looks like this:
This is BOTH channels, and the preamp is clean. So what the heck could it be now that is affecting both poweramp channels??
As in probe using the scope? Will do that next. I wonder if it's the PNP outputs? I replaced the 2SD188 with 2SD188A that I had from another amp but left the PNP as their hfe was fine in the 70s...
If it's the outputs then I should see a clean wave everywhere else in the amp board, yeah?