9090DB, Went from Playing Normally to Quiet

[BearsFan]

I have a couple comments.

1. If the fuses blew, there's a reason - find out this reason before changing them.
2. If this were my receiver to fix, and it came to be like that, the first thing I'd do is test the output transistors in the channel which blew fuses. It's really simple on a 9090 to remove the outer heatskink, pull out the transistors one by one and check them with the diode test function of a DMM.
3. If transistors are shorted, you have to go to the driver board, and check all the transistors for that respective channel, change any bad/shorted ones.
4. Measure resistors and replace those which are out of tolerance. Replace all fusistors as a matter of course, and any bad capacitors.
5. Clean the card connector with deoxit.
6. Install fuses, power up unit, set bias and offset following factory service manual procedures
7. Connect to speakers and enjoy some music.

The only way you will have blown fuses if the transistors are still fine, is if you have something way off causing the bias to skyrocket, OR, sustained overload, or maybe a shorted cap or wire. Since you weren't blasting it when it failed, I suspect there's been a semi catastrophic failure. Don't worry though, this isn't that hard to fix. A tech could get it back out there in no time. If any of the above seems too daunting, do consider taking it to a tech, since this is quite a valuable receiver and not the best one for learning.

Any guidance as to the pinout of the connector in relation to the schematic would help. I can eyeball the board and see what components go to what pin but using the schematic to connector correlation would make it much easier to read the transistors in circuit and see differences between right and left sides.

 

[maxhifi]

Any guidance as to the pinout of the connector in relation to the schematic would help. I can eyeball the board and see what components go to what pin but using the schematic to connector correlation would make it much easier to read the transistors in circuit and see differences between right and left sides.

Download the repair manual and look at the pcb layout, I believe they are numbered.

 

[BearsFan]

Thank you, should have known this ;-( ... before calling it a day, I did some left right comparisons last night and there were some differences in readings. Will make a spreadsheet and send in awhile, need coffee! Thanks

 

[Skywatcher]

Hi "Max" , as Skywatcher mentioned, welcome to the party! I've been very conservative chasing this issue since it began. Just to quickly recap:
> Receiver was playing loud, not cranked, about 85-90 db on my decibel meter from 10 feet from the speakers, about "10 o'clock" on the volume knob.
> I went over to investigate and could hear a very faint tinny sound out of the right channel. Left channel was dead
> Put the receiver in mono to see if i could get anything out of the left side, no go.
> Popped the cover and found both F1 and F2 open, F3,4 and 5 were OK. There is no smell of burining or overheating inside.
> Removed the driver board and looked for obvious physical signs of component failure, none found.
> Put a meter across both big filter caps, both act the same, voltage climbs to about a volt and then drops to 0 volts. Ohmed out , the resistance bounces around then stabilizes at 2K ohms on both big caps.
> Did the same test ground to positive side of big caps, similar result.
> I have not ohmed out thedriver transistors; I could do a quick check in the morning and compare readings from all 4 transistors in circuit to see if there is an indication of a short. At a minimum I can compare the reading from the left and right pairs.

The plan is to try three steps suggested by Skywatcher to isolate power supply issues from driver / output stage: install fuses, power up with F4 and F5 out, F4 and F5 in, and Driver board in, nothing attached, volume minimum, switch in aux.
If, between the two of you the consensus is I should pull the output transistors and do a diode bias test on them first is the correct sequence, I can do that. I am assuming that the case is the collector and the two pins are emitter and base.

One last thought I have is based on what I discovered when I opened the unit up. As noted, the workmanship on the driver board was pretty piss poor and the dirt and surface corrosion should have been taken care of when the unit was refurbed IMO. The four pots at the top of the driver board were not replaced and they look 40 years old. Is there any logic to reaching out to the community and seeing who might have the time and expertise to rebuild these two boards? I have the technical ability to remove and replace them, set up biases, etc, and test and check the output transistors, install replacements, silicon the metal contact, etc.

It's just a thought. I don't have a thousand dollars to redo a recap, rebuild of the whole unit, but these two boards in particular look ratty and may be the rot cause for this issue. I'm getting ahead of myself but just thought i would put it out there. Thanks a bunch for the support! Skywalker, I will hold off on power up until I hear back from you tomorrow. In the meantime I am going to poke around the unit and take some further resistance readings from the driver board connector. Hopefully I can figure out the pinout between the board and the schematic, I don't see any pin number references on the board itself. Here are some updated, and hopefully better pictures of the driver board.

Don't worry, you won't spend any where near a thousand smackers to get this going. Probably the most expensive *single* component will be each output transistor. Depending on which may be shorted, I have a couple of the original Toshiba 2SB554 and 2SD424 here. However, the modern ON Semi replacements MJ21193G, MJ21194G we use are (some others work too) available, though the price has shot up a couple of bucks each to an average of about $8 each when you figure in shipping. Do *not* attempt to buy transistors on eBay. 99.9% will be chinese fakes. I usually buy from Mouser, or Digikey, but on occasion B&D Enterprises. Mouser is my first choice unless they don't have what I need.

Don't worry about the pinout, at least not for the moment. No sense making things more complicated than they need to be. You would need to pull the transistors to test them properly anyways. I have a complete F-2624, rebuilt, and tested in my own 9090db, available if need be. And, if you like, I could repair yours inexpensively. And yes, those 4 pots have a habit of going bad, especially the BIAS pots. I always change them with Bourns

Check your output transistors one at a time. Pull one, check it, and put it back (even if bad) so you don't lose track of their position. Just mark the bad one(s) if any. Do this before powering it up just in case any are shorted. If any are shorted, we'll replace them first before moving on to the next step. In the mean time, we can work on the driver board.

Here you will find the pinout of the TO-3 transistor. https://www.el-component.com/bipolar-transistors/2sb554

 

[Skywatcher]

Thank you, should have known this ;-( ... before calling it a day, I did some left right comparisons last night and there were some differences in readings. Will make a spreadsheet and send in awhile, need coffee! Thanks

MMMmmmmm coffeeeeeee, the elixir of life!

 

[Skywatcher]

Hi "Max" , as Skywatcher mentioned, welcome to the party! I've been very conservative chasing this issue since it began. Just to quickly recap:

One last thought I have is based on what I discovered when I opened the unit up. As noted, the workmanship on the driver board was pretty piss poor and the dirt and surface corrosion should have been taken care of when the unit was refurbed IMO. The four pots at the top of the driver board were not replaced and they look 40 years old. Is there any logic to reaching out to the community and seeing who might have the time and expertise to rebuild these two boards? I have the technical ability to remove and replace them, set up biases, etc, and test and check the output transistors, install replacements, silicon the metal contact, etc.

The soldering job on the driver board looks alright. They didn't clean the glue off (that becomes corrosive and conductive), didn't replace the VD1212 diodes, and used "generic" capacitors of unknown quality, and smeared the thermal compound all over. Of course, I'm a bit "picky" with "neatness". lol!

Be *very* careful of those STV-3H diodes! They are the two brown "things" mounted piggyback on two of the driver transistors. The leads are fragile and are easily broken. They are used for thermal BIAS control. If they are "suspect", "carefully" unsolder one lead from the board and check it with a meter. They should read about a 1.4 forward voltage in diode test function.

Checking things from the connector on the driver board isn't really all that helpful since everything is in circuit. A bad resistor, or diode, etc can lead you on a wild goose chase. If you are comfortable with it, test resistors and diodes by pulling one lead, and transistors by pulling out altogether and putting them back one at a time. Refer to your photo(s) for correct orientation. This is a very handy device to have, and it is cheap too. You can do a quick check of almost anything. But be aware when you test a capacitor, *discharge* it first or you'll kill it. There is no "protection". I forgot once and POOF! lol! I also have a more expensive Peak DCA55 that only tests semiconductors, but this little Mega328 tests a lot more than semis. And it comes with a case.
https://www.amazon.com/KOOKYE-Mega3...N2M2?source=ps-sl-shoppingads-lpcontext&psc=1[/QUOTE]

 

[Skywatcher]

Those "cheapie" testers don't come with leads at that price. I made a short set out of 20awg silicone wire that are nice and flexible and used good minigrabbers. The ones that do come with test (hook) leads, well those leads are not so nice and don't last very long, but will do the job. You just have to keep reattaching the grabbers to the wires now and then.

If you don't mind not having a fancy-dancy case, this tester works great too. These aren't "critically accurate", but they will let you know if a component is bad or not, and identify the leads as well. https://www.amazon.com/diymore-Mult...011&rnid=2470954011&rps=1&s=industrial&sr=1-3

 

[BearsFan]

OK, output transistors have been removed and checked and it appears that one is bad. I did a simple tone test on the meter then a forward / reverse bias. It just crossed my mind that i did not do a collector to base / emitter test, just emitter to base...duhh.

In any event, here are the results - if I need to go back and recheck the collector bias let me know. This guy loves his silicon grease....

 

[BearsFan]

Skywatcher, please PM me on the board rework. Thanks.

 

[Skywatcher]

OK, output transistors have been removed and checked and it appears that one is bad. I did a simple tone test on the meter then a forward / reverse bias. It just crossed my mind that i did not do a collector to base / emitter test, just emitter to base...duhh.

In any event, here are the results - if I need to go back and recheck the collector bias let me know. This guy loves his silicon grease....

Here is a thread by Echowars on using the 6-way test on bipolar transistors. http://audiokarma.org/forums/index.php?threads/bipolar-junction-transistor-testing-basics.43186/ Any time I get within a foot of the white stuff, it seems to get on my hands somehow, then spreads like wildfire to other things.

 

[BearsFan]

Here is a thread by Echowars on using the 6-way test on bipolar transistors. http://audiokarma.org/forums/index.php?threads/bipolar-junction-transistor-testing-basics.43186/ Any time I get within a foot of the white stuff, it seems to get on my hands somehow, then spreads like wildfire to other things.

I will redo this check, I want to do it correctly. Thanks

 

[Skywatcher]

Skywatcher, please PM me on the board rework. Thanks.

PM sent. I hope you got it.

 

[Skywatcher]

OK, output transistors have been removed and checked and it appears that one is bad. I did a simple tone test on the meter then a forward / reverse bias. It just crossed my mind that i did not do a collector to base / emitter test, just emitter to base...duhh.

In any event, here are the results - if I need to go back and recheck the collector bias let me know. This guy loves his silicon grease....

If the 2SD424 is the only one bad, you're in luck. I have one good 424 R though yours are O (the gain range designator) but will still work as long as yours measures close. hFE (gain) measured on my Atlas DCA55 is 34

 

[BearsFan]

Thanks for sending that EchoWars writeup! There is a second bad transistor, collector / base short. Testing all 8, the forward bias was about .56v and reverse was OL as expected. The two bad ones read .2 and 3 ohms respectively across the bad junctions.

 

[BearsFan]

Thanks for sending that EchoWars writeup! There is a second bad transistor, collector / base short. Testing all 8, the forward bias was about .56v and reverse was OL as expected. The two bad ones read .2 and 3 ohms respectively across the bad junctions.

MaxHifi, thanks for pushing the outputs, good stuff.

 

[Skywatcher]

Thanks for sending that EchoWars writeup! There is a second bad transistor, collector / base short. Testing all 8, the forward bias was about .56v and reverse was OL as expected. The two bad ones read .2 and 3 ohms respectively across the bad junctions.

Yep, two shorted for sure! I looked and unfortunately I do not have any more 2SB554. I must have used them up. Just for giggles, I peeked on FleaBay and sure enough, every one I saw are fakes! Put a wanted post in Barter Town. Perhaps someone has some. Try to get the O version to match what you have, but either O or R will work according to the service manual. With them shorted, we definitely have to work on the driver board. If that fails, we'll have to replace all 4 in that channel with ON Semi. All 8 would be better for "balancing" the system, but that will cost more.

 

[maxhifi]

OK, output transistors have been removed and checked and it appears that one is bad. I did a simple tone test on the meter then a forward / reverse bias. It just crossed my mind that i did not do a collector to base / emitter test, just emitter to base...duhh.

In any event, here are the results - if I need to go back and recheck the collector bias let me know. This guy loves his silicon grease....

Good job! You can track down originals, or, just change the 4 in that channel with the excellent ON Semi substitutes. Then go to the board. Progress always feels great!

 

[Skywatcher]

Good job! You can track down originals, or, just change the 4 in that channel with the excellent ON Semi substitutes. Then go to the board. Progress always feels great!

Good call! Right after you said that in your first post, I smacked myself in the forehead. Must not have had enough coffee yet.

 

[maxhifi]

Good call! Right after you said that in your first post, I smacked myself in the forehead. Must not have had enough coffee yet.

I have to say that 9 times out of ten, when I get an amp brought to me for repair which "keeps blowing fuses", or just shows up with blown fuses, it's got at least one bad output transistor. It's just a statistic based thing - looking at the schematic says that there could be a variety of problems, but that's the one I keep seeing over and over again across all brands of amp. At least the 9090 is a discrete amp, and doesn't need some obscure STK module changed to make it work!

Similarly, when I get an amp which has had work done to it in the past - that work is more than likely replacement of one or more output transistors. Sometimes it is caused by something in the driver, and other times it's caused by abuse. The good thing about the 9090 is it's SO well documented. People love them and there's a ton of info online.

 

[maxhifi]

It's just a thought. I don't have a thousand dollars to redo a recap, rebuild of the whole unit, but these two boards in particular look ratty and may be the rot cause for this issue. I'm getting ahead of myself but just thought i would put it out there. Thanks a bunch for the support! Skywalker, I will hold off on power up until I hear back from you tomorrow. In the meantime I am going to poke around the unit and take some further resistance readings from the driver board connector. Hopefully I can figure out the pinout between the board and the schematic, I don't see any pin number references on the board itself. Here are some updated, and hopefully better pictures of the driver board.

I can think about this in a couple ways. I don't see anything terrible about this board, it just looks like an old Sansui board which has seen a bit of service work, to me. I think the previous dude did change most of the bad parts. If you're not comfortable repairing it yourself, by all means send it to someone else who can fix it.

I don't usually change parts which aren't bad unless I'm doing a full restoration, if it were me I'd leave the caps and trim pots alone. The capacitors could possibly have seen too much voltage during the failure, so one or more of them may need replacement. Failure analysis of the board will tell you this. Changing all of them is cheap and easy, but may not really be needed.

You have to decide - do you want to pull the transistors and check them one by one, and change the bad ones, or do you want to do a total rebuild? Last failed 9090 I had, had two bad transistors, and a couple shorted outputs. I think I billed the client two hours for all of the work, and that included ordering parts for it, changing all the fusistors, and changing the bad caps. It hasn't come back to me, so I'm assuming it still works. That's a "repair" as opposed to a "restoration".

I'd say send me the board, but I don't have a 9090 to test it with right now - so if you do send it out make sure it goes to someone who has an amp which can be used to test it before they send it back to you.