Sansui 9090db cycling in/out of protection

[bobok7787]

Hello,
I've got a lovely 9090db I'm working on that is showing some odd signs I'm hoping someone has dealt with before.

To start, the unit had no output in the left channel, I found 2 defective output transistors and 1 bad transistor on F-2624 (TR05). It's worth noting I tested all transistors on the left channel (and compared them to right channel readings), most out of circuit (only removed one of the transistors attached to heatsink).

Powering up on a DBT (100w bulb), the light goes bright but quickly dims and comes out of protection, but after 3-4 seconds it kicks back in to protection and the light glows bright again, then it cycles continuously like this in and out of protection. It does this consistently if left on (although I've never left it on for more than 30 seconds like this).

I monitored voltage at points 40 and 41 on F-2656 prior to the protection relay and sure enough, both channels show huge swings of voltage right as it kicks into protection (up to 1.4-5 volts DC, monitoring both at the same time with two meters).

Interestingly enough, if I disconnect the left channel from the protection board, the unit functions perfectly. It kicks (and remains) out of protection and DC offset and bias all measure within spec on both channels.

So would the problem be somewhere on the protection board? Or F-2624?

And because I know you'll ask F-2624 was fully rebuilt before I had it, all fusible resistors were replaced and measure within spec, new capacitors, DC trim pots (the bias pots work perfectly). Here is a picture of both sides of the board if you care to see it.

Thanks so much in advance for any help, what a great community Audiokarma has become!

 

[Tom B]

Check the left channel output (DC balance) to see if you get a DC voltage swing as it goes out of normal into protect. If you do, the problem is most likely the driver board. I would suspect the 2 differential amp transistors, or the larger grey transistor (TO126 case), but you already replaced that one (TR5). So try the TR1 and TR3, replacing them with some KSC1845 or similar. If you had bad outputs or drivers or pre drivers, or even some of those small signal diodes, you would have a more solid problem. Drifting in and out is usually noisy 2SC1400's differential amp transistors, or a flakey fusable resistor. Seeing the fusables were replaced, I would suspect the TR1 and TR3

 

[bobok7787]

Thanks for your help Tom! I did monitor DC offset while it was cycling in and out of protection and it does spike up to >1.3V right as it kicks into protection.

I pulled TR1 and TR3 from the circuit, both passed the 6 point check. While poking around I decided to re-check the diodes and found D13 to be defective, measuring 0.386V forward, and 0.572V backward. Aha! Could this tiny diode be whats been causing these issues?

 

[Tom B]

I am not sure what the 6 point check is, as I use a transistor tester or analog meter to check them. But noisy transistors many times test fine, but are usually sensitive to cold and heat. I test diodes similar so I am not sure what voltage measurements you are taking. I usually measue on ohms with a triplett meter and look for about 15 ohms forward and no more than megohms reverse. I suppose that D13 could cause protect issue, although I would have expected more of a not coming out of protect rather than cycling. But change it and see.

 

[avionic]

I am not sure what the 6 point check is

Really..Sure you do..
http://audiokarma.org/forums/index.php?threads/testing-transistors.201303/#post-2362176

 

[Tom B]

Really..Sure you do..
http://audiokarma.org/forums/index.php?threads/testing-transistors.201303/#post-2362176

Correct, I never saw this a a six point check, but that is exactly how I check transistors usually in circuit, although I use a Triplett 630 analog MM. I should have realized that seeing 6 individual tests are done on the transistor, it can be referred to as the 6 step.

Thanks
Tom

 

[bobok7787]

Alright I changed the diode. It still cycled as it did before, bias was still steady at 30 mA on both channels, but there were still huge voltage swings on both channels.

As I was poking around more, I noticed C09 and C10 appeared to have their polarity reversed (referencing the schematic, the board doesn't have polarity markings). Or should this be a bipolar cap? I can't seem to find a matching schematic for the diode version of F-2624, does anyone know of one?

 

[DougBrewster]

courtesy Tolitoms and the F2624 Renaissance thread. Resolution is best I can do at this time of night...

Sansui never updated their schemas to show the introduction of the diode biassing scheme.

 

[petehall347]

did you try just the right channel ? if it then behaves the same as just left side connected it could be power supply flaky .

 

[DougBrewster]

I've seen this cycling before and it came down to excessive bias and the current limiting (TR15/17) catching it,reducing the output and the whole thing repeating itself.

 

[bobok7787]

Thank you DougBrewster! According to this I was wrong about the polarity change, back to the drawing board

 

[Tom B]

Be careful with the replacements you are doing. It seems your first problem was issues on the left channel, and if you disconnected left channel from the protect board the unit worked fine on the right. You now seem to have big voltage swings on both channels. Have you changed anything on the right side? If you are having similar swings on both channels, check the regulated power supplies + and - feeding the drive board (as others have suggested), to see if one is swinging? This is the voltage sitting on the collectors of the driver transistors. ot sure but it is probably + and - 60volts. Those voltages should be rock solid.

 

[bobok7787]

Thanks again Tom! With the amp board unplugged and pulled from the unit, I get a rock solid +/- ~60V (and the unit comes out of protection and remains stable).

When I plug the amp board in, I do indeed get huge voltage swings at the +/- 60V power rails (also measured at the collectors on the output transistors, same swings).

 

[DougBrewster]

When I plug the amp board in, I do indeed get huge voltage swings at the +/- 60V power rails (also measured at the collectors on the output transistors, same swings).

I take it that these swings are on the left channel only. If so, and you're satisfied that TR01/03 are good, take a look at your bias diode in this channel; check it for open circuit connections (or diode) and reverse polarity. The cathode is marked with a yellow dot but I can't remember which way they face when mounted..

 

[Tom B]

The big swings at the collectors of the outputs is unusual, as that + and - 57vdc DC voltage comes directly from Caps sitting on a seperate set rectifiers fed from a seperate winding of the transformer. The + and - 60v comes from its own winding, its own set of rectifiers, and regulator transistors.
It could be Something on the driver board creating an overcurrent condition, sucking down the power supplies, or a combination of issues.
Things I would suggest
Set the bias for minimum on both channels to see if you go into protect
Check the main filter caps to see if one is open, or very low, or leaky (12000uf @80v)
Check the filter caps for the regulator supplies (100uf @80v)
Make sure the 330uf aps that were replaced are good, and in the right way.

 

[bobok7787]

Aha! It took looking at with at with a magnifying glass, but I found C05 and C06 both had broken traces at their connections to R19/R20.

I fashioned some direct connections with some clipped leads, soldered them in, powered on with the DBT, relay click, DIM BULB, and we have a stable amp! I biased both channels to 20 mA (while on the DBT), and I can adjust the DC offset to a stable +/- 5 mV.

But I'm still not out of the water. It will play music on very low volume in both channels without distortion, but if I turn it up to a medium volume the DC offset in the left channel spikes and it goes into protect, then clicks back on.

DougBrewster, I checked the bias diodes, both measure ~1.62V/OL, I think those are okay. I also checked TR01/TR03 (in circuit, compared to right channel TR02/04) and they seem fine as well.

Could TR01/TR03 be the culprits even though they test fine?

Thanks so much for everyone's help, I feel much closer after today!

See attached for some pics of the broken traces if anyone cares to see them, after looking at the pictures I'm thinking I need stronger glasses :/

 

[petehall347]

is it still on DBT ?

 

[DougBrewster]

Wow, same defect both channels. Who'd have guessed that? Far as next move is concerned Tom B's suspicion about TR01/03 has a new significance and as PeteH says is it still on DBT and what DBT wattage are you using?

 

[bobok7787]

It's still on DBT with a 100w light bulb. I'm also noticing the caps on the left channel are slightly warm to the touch, compared to the right channel which feel cold.

 

[petehall347]

try balance to left then to right .. it will behave odd on DBT . but caps shouldn't warm up that i know about . maybe something close by is warming them up from radiated heat ?