Sansui B-77 Stability Issues

jahmadeus

jahmadeus

New Member
I am currently working on a Sansui B-77 that has been giving me a lot of grief.

When I received the amp, the outputs on the left channel were shorted and a few nearby resistors (kR22, kR24, kR25, kR29) had smoked. One of the drivers (kQ9) had been replaced with a substitute. Everything else seemed to ohm ok.

I replaced the outputs, the blown resistors and swapped the drivers with the parts listed in the schematic (2sb528 and 2sd528). I also checked all nearby components in and out of circuit against the schematic and other (seemingly good) channel.

I then powered the amp on. Relay clicked on and the amp was running with minimal current draw.

I checked the bias next and discovered the bias on the blown channel was low (1mV instead of the 3mV it should be set to). I then tested the amp with some audio. Everything appeared to be running nicely. Low current draw, cool heatsinks, etc. The amp ran for about 20 minutes before the current draw began to slowly rise until it trips my variac (set to trip at 2A). The reading across the emittors also grew to a few hundred mV when this happened. It appears the amp goes into thermal runaway.

I have gone through the circuit with a fine tooth comb and cannot find any bad components. All voltages check good against the schematic and the other channel with the exception of one thing. The range of the bias adjustment for the previously blown channel is significantly lower than the other channel. Its bias pot adjusts the bias voltage from .7mV (ccw) to 3.3mV(cw). The other channel has a range of 1.1mV to 5mV. The adjustment pot for the bad channel needs to be set almost fully clockwise to achieve a bias of 3mV. All of the devices in the bias network have either been replaced or checked out of circuit as good. There must be something I am overlooking.

I have read that these particular amps can be hard to stabilize once they have a major failure. If there is anyone who has any insight to this particular model, I would love to hear it.

The service manual for reference.
https://www.vintageshifi.com/repertoire-pdf/pdf/telecharge.php?pdf=Sansui-B-77-Service-Manual.pdf
 

Attachments

  • sansuib77-unchecked.jpg
    sansuib77-unchecked.jpg
    127.4 KB · Views: 11
Register to hide this ad
You may still have a burnt resistor or two somewhere, Q5, Q6 & Q7 need checking, and don't forget to check the Protector Circuit as well for this, including its semiconductors.

Obviously the need to turn the bias up so much on the previously failed channel is causing this 'instability' - keep checking. ;)
 
Last edited:
Hyperion said:
You may still have a burnt resistor or two somewhere, Q5, Q6 & Q7 need checking, and don't forget to check the Protector Circuit as well for this, including its semiconductors.
Click to expand...

I suspect this is the case but so far everything checked tests fine against the schematic/other channel. I swapped out Q5, Q6 and Q7 to see if that would make a difference and it did not.

I have not tested (out of circuit) many of the passive devices in the protector circuit, I think I might try that next.
 
Hyperion said:
Don't limit your search/checks - you may miss something... ;)
Click to expand...
Yes indeed. The problem is im running out of things to check. I went ahead and shotgunned all the semiconductors in the protector circuit and the passives all seem to read normal. Maybe its a resistor that is opening up under load? The weird thing is that no matter where i probe for voltage, everything seems to read exactly the same as the other channel (except the voltage between the output emitters).

The search continues...
 
The plot thickens...

So I returned my focus to the three transistors in the bias network (Q5,Q6,Q7). I decided to swap Q7 with one from the working channel. The problematic channel now has a normal bias range and the previous working channel has the low bias. However this is weird as I already had replaced this transistor (2sa992). Ive also tried a couple different 2sa992's with different packages and suffixes (most of what i have are the 'F' hfe class, which is specified in the schematic). Verified with my Peak DCA Pro that the hfE/Vce curves looked very close to the working transistor. The working transistor has faded lettering and is hard to read but i can make out 'A9', cannot read the full part number or the suffix code.

Seems to clearly be related to this transistor but I am lost as to why the replacements I have tried are not working. Is it possible that 2sa992 F is listed incorrectly and the working part is not actually a 992 F? Does anyone have any guess to what could be throwing me for a loop?

Any input is appreciated.
 
This effect is often seen when the hfE of the replacement is low, but you say you have checked this, so I don't know why swapping the transistor made that difference. Anything bought recently marked '2SA992' should be regarded with great suspicion, as they are likely to be counterfeit. And as regards KSA992 (the modern current production direct replacement), I haven't been able to buy these with hfE consistently above about 420, most being around 330-370, whereas originals can show hfE of 500-650.
 
Hyperion said:
This effect is often seen when the hfE of the replacement is low, but you say you have checked this, so I don't know why swapping the transistor made that difference. Anything bought recently marked '2SA992' should be regarded with great suspicion, as they are likely to be counterfeit. And as regards KSA992 (the modern current production direct replacement), I haven't been able to buy these with hfE consistently above about 420, most being around 330-370, whereas originals can show hfE of 500-650.
Click to expand...

I had a few different types laying around as I work for a shop that has a pretty huge stock of NOS parts. Some that were purchased recently (slightly smaller package, marked as 'F' class) that tested lower for hfe (350 - 450) and some older ones (marked as 'E' class) that tested closer to 600. The ones i purchased recently had a slightly different hfe/Vce curve when tested, the older ones were almost identical. Neither worked. Perhaps all my replacements are no good, but I feel like I must be overlooking something critical here.

Thanks for your response!
 
You must log in or register to reply here.
Back
Top Bottom