Sansui AU-9900 / AU-11000 Repairs

dr*audio

Fish fingers and custard!
When repairing AU-11000 and AU-9900 amps the following work should be done:
I have found the pre - driver transistors in these and other Sansui amps to be failure prone, at least when they get this old. They are the ones that have a tab on the top, but they are not TO-220 case. Motorola used to call this case "Uni-Watt" because they had a 1 watt power rating. I have found that they go intermittent and sometimes fail outright, causing possible damage to the rest of the power amp. I recommend replacing them with KSC3503ESTU and KSA1381ESTU, available from Mouser Electronics.
There is a screw behind the front panel that provides a ground connection for the entire unit to the case. It gets loose and the ground is lost. I solder a braid from the input board (right hand side on the bottom) to the chassis.
Watch for loose or worn out connectors between boards. These cause intermittent problems.
When adjusting the bias, don't use the method in the manual. Instead, do the following:
The bottom panel and feet must be in place for this to work properly.
1. From the top, with a hemostat, you attach alligator clip jumper wires to the emitters of 2 output transistors; one NPN, one PNP on each channel. You can access the emitters on the transistor sockets on the heatsink.
2. Drape the wires over the back of the amp, out of the way, and make sure they can't short to each other or the chassis. Connect your meter to one set of wires for one channel. I use 2 meters so I can do both channels at once.
3. Place the top cover onto the unit with no screws. If you are careful you will not crimp or damage the testing wires.
4. remove the cover over the driver board.
5. Turn the amp on and allow it to warm up for a few minutes, then adjust the idle current for 20mA across the 2 emitter resistors. I don't remember the values of the resistors, but you multiply 20mA X the total resistance to get the voltage reading you want.
6. Take a piece of electrical tape and stick it to the top of the driver cover, to use as a handle. Place the cover over the drive board and attach with one screw.
7. Monitor the voltage and adjust after 10 minutes. Always replace the cover over the driver board.
8. After it has been stable for 30 minutes you are done.
 
Last edited:
Brilliant Warren ! - especially the carefully thought out bias setting procedure, I have used it and it works :thmbsp:

EDIT: By the above calculations, I make that 13.2mV across TWO emitter resistors of 0.33Ω - to give 20mA PER output pair PER channel.

I x R = V

0.020A x 0.66Ω = 0.0132v = 13.2mV
 
Last edited:
Warren, as far as the
KSA1381 transitors go will the KSA1381ESTU work ? I believe its just the packaging correct ?
And if so how many total will i need .
As far as the KSD3503 its coming up no results . What to get at this point ? Again how many ?
I havent took anything out of the amp so i havent got to deep into it at this point.
I want to place an order with mouser , just want to get everything i need at one time hopefully .

Thanks

Chris
 
I'll go out on a limb and guess the ksd3503 is a typo and should be ksc3503 (compliment to ksa1381)
 
Warren, as far as the
KSA1381 transitors go will the KSA1381ESTU work ? I believe its just the packaging correct ?
And if so how many total will i need .
As far as the KSD3503 its coming up no results . What to get at this point ? Again how many ?
I havent took anything out of the amp so i havent got to deep into it at this point.
I want to place an order with mouser , just want to get everything i need at one time hopefully .

Thanks

Chris

Sorry, guys, KSC3503ESTU and KSA1381ESTU. 2 of each. I will fix it above.
 
Warren thanks aton for your help here. Is there any other problomatic transistors in the 9900 that should be replaced ? If not ill get a order in real soon .

Thanks again

Chris
 
Last edited:
Just checked mouser
The KSC3503ESTU is unavailable .
They do have the KSC3503DS and KSC3503DSTU . These are the same correct ?

KSA1381ESTU they have plenty of .......................

Thanks
 
Just checked mouser
The KSC3503ESTU is unavailable .
They do have the KSC3503DS and KSC3503DSTU . These are the same correct ?

KSA1381ESTU they have plenty of .......................

Thanks

You will have to go with the 512-KSC3503DSTU. The D means the Hfe or gain is lower than that of the E part, but the E part is discontinued now, so there is no alternative. If you have a DMM with an Hfe tester in it, you could buy a large quantity and test the transistors, and select those with the highest Hfe.
 
Note that the basing is different on these transistors from the originals. For 2 or three of the places on the board, Sansui provided an extra pad for using transistors with a different basing. Just examine the board markings and you will see what I mean. I think also that the ones with heat sinks have to have the transistor installed on the opposite side of the heatsink. It's not difficult to figure out once you have it in front of you.
 
If the gain on the KSC3503D becomes an issue, you can still obtain the Sanyo 2SA1480/2SC3790 pair which appears to be the exact same transistor in a slightly different packaging. B&D has the E grade. 2SA1478/2SC3788 are the 200v versions comparable to the obsolete 2SA1380/2SC3502
 
Brilliant Warren ! - especially the carefully thought out bias setting procedure, I have used it and it works :thmbsp:

EDIT: By the above calculations, I make that 13.2mV across TWO emitter resistors of 0.33Ω - to give 20mA PER output pair PER channel.

I x R = V

0.020A x 0.66Ω = 0.0132v = 13.2mV

May I react to this calculation? Is it correct that it is for AU 9900? I see in the SM that idle current should be set to 50 mA (for channel). So this should be 50 mA per channel or 25mA per pair od transistors in one channel respectively?
Or it could be measured in mV as say the formula U=R*I? Then it should be 50*0,66=33 or 25*0,66= 16,5?

I presume the second choice as I am measuring only on one pair from two which make one channel. Is it true?

All reaction are much appreciated!
 
May I react to this calculation? Is it correct that it is for AU 9900? I see in the SM that idle current should be set to 50 mA (for channel). So this should be 50 mA per channel or 25mA per pair od transistors in one channel respectively?
Or it could be measured in mV as say the formula U=R*I? Then it should be 50*0,66=33 or 25*0,66= 16,5?

I presume the second choice as I am measuring only on one pair from two which make one channel. Is it true?

All reaction are much appreciated!

You are correct that the service manual calls for 25mA per channel. However, because the bias tracking diodes are located on the driver board and not on the heat sink, you don't get good bias tracking and the bias is not stable at 25mA, it will climb. I found that 20mA works well and is stable.
 
You are correct that the service manual calls for 25mA per channel. However, because the bias tracking diodes are located on the driver board and not on the heat sink, you don't get good bias tracking and the bias is not stable at 25mA, it will climb. I found that 20mA works well and is stable.

Thanks! I set it as you described and indeed the bias is quite stable. I set both channels after 20 minutes after turning the unit on to get reading od 13,2 mV. I think this time I have this unit completely restored and set to specs and its delight to listen to music coming from this jewel.
 
More:
On the AU-9900 and AU-11000 there is a screw behind the front panel (you have to remove the front panel to see it) that provides the ground path for part of the circuits in the amp. If it gets loose it causes all kinds of weird problems. Rather than tightening the screw, I solder a ground braid to the chassis from the input selector board. If you turn the unit upside down, with the front panel facing you, it's the board on the left side front. As I recall, the ground is connected to the housing of the switch on the right side of the board as you are looking at it in the position above. Scrape some plating off the chassis by the front panel and solder a braid from the board to the chassis.
 
DeOxit the push on board connectors and the pins on the boards themselves.

Ha! I've got a bad connection on the power supply sub board connector on my AU-9900. I was going to replace the connector but you just may have saved me some money, John! Thanks!
 
Did you ever see this Warren?

http://audiokarma.org/forums/showthread.php?t=398717

One of the previous owners of my AU-9900 must have had an intermittent connection (like you) but dealt with it in another way - have a look at the link - read it and weep :yes:

Happy to help you for a change :thmbsp:

I missed that one, but I worked on a CA-3000 where some a-hole did the same thing. I found some connectors. I think what I did was use 2.5mm spaced connectors and just used every other pin. Yeah, that's what I did, I just looked at them. Molex 0.1" spacing. That's what I was going to use on mine. I already tried cleaning it with a wound guitar string and it didn't work, but maybe the De-Oxit will work.
 
I know this is an older post but I would like to know if the onsemi mje350 and mje340 transistors would be suitable substitutes for failure prone motorola uni-watt transistors
 
Back
Top Bottom