Hello, I am currently working on a Sansui B-77 that had the left channel fail. The outputs, drivers and various other transistors in the channel were either completely shorted or measured bad compared to good parts. Went through all the components in the channel and replaced any that seemed questionable.
After replacing the damaged components I was able to fire up the amp. But upon checking the bias I noticed that the previously blown channel had a much lower bias. The adjustment on the bad channel varied from .7mV (counter clockwise) to 3mV (clockwise). The good channel had a range of about 1.2mV to 6mV.
I set bias for the bad channel to 3mV, but this required me to turn the adjustment pot to its max (the good channel only needed to be turned halfway, for reference). The amp ran fine for about 25 minutes and drew basically nothing for current (~250mA). Then the bias would increase and the amp would start rapidly increasing its current draw (eventually tripping my variac's set current limit).
After a lot of searching around and re-checking components I found that Q7, a 2sa922 (F) in the bias network seems to be the cause of this. When swapping in Q7 from the working channel, the problematic channel now has the proper bias range and is stable. The previously working channel now has the same low bias range the other channel had.
However, swapping in a new 2sa992 does not effect the problem. I tried several 2sa992's that I attempted to hfe match to the working Q7.
This is where I am a little lost, is there an attribute of this transistor that is critical to consider for this circuit? Have I been fooled somehow?
Is anyone familiar with this particular unit that could shed some light on what I might be experiencing?
Any input is greatly appreciated.
Schematic for reference.
https://www.vintageshifi.com/repertoire-pdf/pdf/telecharge.php?pdf=Sansui-B-77-Service-Manual.pdf
After replacing the damaged components I was able to fire up the amp. But upon checking the bias I noticed that the previously blown channel had a much lower bias. The adjustment on the bad channel varied from .7mV (counter clockwise) to 3mV (clockwise). The good channel had a range of about 1.2mV to 6mV.
I set bias for the bad channel to 3mV, but this required me to turn the adjustment pot to its max (the good channel only needed to be turned halfway, for reference). The amp ran fine for about 25 minutes and drew basically nothing for current (~250mA). Then the bias would increase and the amp would start rapidly increasing its current draw (eventually tripping my variac's set current limit).
After a lot of searching around and re-checking components I found that Q7, a 2sa922 (F) in the bias network seems to be the cause of this. When swapping in Q7 from the working channel, the problematic channel now has the proper bias range and is stable. The previously working channel now has the same low bias range the other channel had.
However, swapping in a new 2sa992 does not effect the problem. I tried several 2sa992's that I attempted to hfe match to the working Q7.
This is where I am a little lost, is there an attribute of this transistor that is critical to consider for this circuit? Have I been fooled somehow?
Is anyone familiar with this particular unit that could shed some light on what I might be experiencing?
Any input is greatly appreciated.
Schematic for reference.
https://www.vintageshifi.com/repertoire-pdf/pdf/telecharge.php?pdf=Sansui-B-77-Service-Manual.pdf
Last edited:
Your link takes me to Google to sign in---I don't have Google on my PC,s -OTOH-- hifi engine has the schematic .
The gain range is quite large (hfe) for a 2SA922 and with the letters supplied on the schematic it looks like Sansui handpicked the gain to suite the bias , as you have said swapping it round cured it but not several you tried .
The conclusion is you measure the gain of the "good one " and fit one with the same gain.
Ah good ! I found Sanyo,s old specifications for a 2SA922 and the higher the letter the more gain .
The gain range is quite large (hfe) for a 2SA922 and with the letters supplied on the schematic it looks like Sansui handpicked the gain to suite the bias , as you have said swapping it round cured it but not several you tried .
The conclusion is you measure the gain of the "good one " and fit one with the same gain.
Ah good ! I found Sanyo,s old specifications for a 2SA922 and the higher the letter the more gain .
Yes I saw the gain rating in the service manual.
I have a bunch of 'F' class 992s and just a couple E's (higher gain).
The working 992 has a gain of about 525. I matched the 992 F's I had as close as I could and got about 475, although the hfe/Vce curve looked different on these than the working one. The package on these were also slightly smaller than the working device, so maybe they are knockoffs?
I then tried the two 992 E's that I had laying around, they had a much higher gain and close to the working one (575 and 615). The curve on these looked very close to the working one.
Maybe the match needs to be even closer? Or maybe my tester doesnt test under the right conditions to correctly match (I use on of those handheld Peak Atlas DCA Pro testers).
Perhaps I need to source some more 992s and keep trying for a closer match? I have never encountered something so sensitive to hfe.
Thanks for your input.
I have a bunch of 'F' class 992s and just a couple E's (higher gain).
The working 992 has a gain of about 525. I matched the 992 F's I had as close as I could and got about 475, although the hfe/Vce curve looked different on these than the working one. The package on these were also slightly smaller than the working device, so maybe they are knockoffs?
I then tried the two 992 E's that I had laying around, they had a much higher gain and close to the working one (575 and 615). The curve on these looked very close to the working one.
Maybe the match needs to be even closer? Or maybe my tester doesnt test under the right conditions to correctly match (I use on of those handheld Peak Atlas DCA Pro testers).
Perhaps I need to source some more 992s and keep trying for a closer match? I have never encountered something so sensitive to hfe.
Thanks for your input.
do I get a small chance to have a list with the things you replaced so far ?
close pictures could also be amusing
close pictures could also be amusing
Hello,
Here are the components I have replaced and/or tested out of circuit.
Green = Replaced
Blue = Tested good out of circuit
Every other component was checked in-circuit.
I shotgunned all the transistors with NOS parts according to the schematic (matched the hfe class for each). I also replaced a handful of components that seemed to be from a previous dodgy repair (several under-rated resistors). Tested all elec. caps and found a couple to be way off (2uf instead of 10uf). Definitely a shotgun approach compared to how I normally repair amps, but there was a good amount of carnage from the previous failure so I decided to go heavy handed.
Solder joints were re-flowed, cleaned and the board closely inspected for bridges, broken traces, etc.
As far as pictures, I have the amp closed up until my 2sa992(E)'s come in, but I can take pictures if that would be helpful. What would you want to see?
Any input is appreciated.
Thanks
Sorry for the small resolution on the previous post, this link should work better...
sansuib77-checked-v2 — ImgBB
sansuib77-checked-v2 — ImgBB