SA-8800 Trim Pots

look at the DC offset voltages at both channel outputs when protection is in either state (good or disconnected)
 
look at the DC offset voltages at both channel outputs when protection is in either state (good or disconnected)

5.1mV on left (hot) channel. 12.9mV on right channel. I measured from the board, as I cannot activate the speaker terminals because of the broken switch. All the transistors appear to be original including the 2SC1913s and 2SA913s
 
Do you more experienced AK members think I should start a new thread, as the issue is no longer about the bias pots, but instead about high bias? I don't know the protocol on this kind of thing here. Thank you.
 
Sorry for the delay. Here are the measurements on the other transistors:

Q27: E = 1.61V / C = 38.55V / B = 2.02V
Q28: E = 1.58V / C = 38.05V / B = 2.18V

For Q27 and Q28, all voltages are negative.

Q29: E = 1.61V / C = 38.44 / B = 2.2V
Q30: E = 1.59V C = 38.2V / B = 2.2V

For Q29 and Q30, all voltages are positive
am not understanding why all base voltages are high . i would have expected differences in bad channel . but looking at base voltages its not making sense . are you sure you looked at correct transistors and your meter is working properly ?
 
polarity is switched . must be user error . unless i am seeing it wrong . have only got laptop at moment and it not easy
 
Sounds like I must be a total idiot, and with a bad meter to boot. But thanks for the helpful input. Appears like it may be time to take it to a repair shop. Hopefully Mark will have time to weigh in before I am forced to concede.
 
Last edited:
I've now repeated the transistor tests, using Mark's very clear instructions...even an idiot would have a hard time screwing up this straightforward process, given the specificity of the instructions. I believe my meter is working correctly, having just tested a few other voltages on another unit that I had opened up for cleaning.

The results of the second (actually third) round of testing are so close to the ones I already posted that there is no reason to repost them. Your conclusion that the results can only be due to user error - while perhaps correct - is not very helpful. What error is it you think I am making? Throw me a bone here:).
 
Pete - did you notice that the voltage measurements I called for are NOT ground referenced??

Q27: E = 1.61V / C = 38.55V / B = 2.02V E=R99 1K 1.6mA
Q28: E = 1.58V / C = 38.05V / B = 2.18V E=R100 1K 1.6mA

For Q27 and Q28, all voltages are negative.
Q29: E = 1.61V / C = 38.44 / B = 2.2V E=R105 1K 1.6mA
Q30: E = 1.59V C = 38.2V / B = 2.2V E=R106 1K 1.6mA

I have got so much going on it's HARD to give appropriate and extended attention to all these threads.

Excessive bias is a result of the base voltages of Q39 (Q40, Q41 & Q42) NOT matching up with the voltage turn on requirements of the Drivers (Q39-Q42) and output transistors, as expressed in PAIRS - Each driver for a specific output.

Example - I used some ON-Semi transistors that had a SLIGHTLY LOWER Vbe turnon and the idle current range was excessively high and would NOT adjust low enough.
Normally it is assumed 1.2v at the drivers and 0.6v at the outputs, but it might only need 1.15 volts and the extra 50 mV
It is 7mA through the bias diodes for the MINIMUM voltage, and they stick in a 100 ohm resistor that will at 7 mA give us an extra 0.7 volts that is divided between the two channels for 0.350v.


and ya know what:
Look at all the transistors on the "hot" channel especially the heat sunk transistors and the output transistor.
See if any are not factory. Transistor markings, solder condition etc...
WHAT ARE THE TRANSISTOR NUMBERS of the driver and it's specific output transistor. THIS IS CRITICAL and I would not have asked it if it wasn't.
 
Last edited:
Well, I spent some more time examining the main amp section, and identified two missing capacitors Specifically, C33 and C34...they show up on the schematic and board diagram, but are missing. They are 47uf/25V NPs. Interestingly, the holes in the board for them are soldered over, altho not as neatly as factory joints. If I had to give an answer based on my poor abilities with schematics, these would not play a role in bias, but I could certainly be wrong. So yet another can of worms is opened (potentially).

EDIT: This was posted before I saw Mark's post above...that's why it seems out of context.
 
Last edited:
and ya know what:
WHAT ARE THE TRANSISTOR NUMBERS of the driver and it's specific output transistor. THIS IS CRITICAL and I would not have asked it if it wasn't.

I apologize...I read your question about the drivers and outputs (in Post #119) as simply asking me to verify that they were "factory". I must have overlooked your request for the part numbers. I will get them in just a few minutes.
 
OK, here is the info on outputs and drivers:

Outputs: 4 Fairchild 2SC2525 KY(G) and 4 Fairchild 2SA1075 KY(G). All solder joints look very good...shiny, clean, solid and original
Drivers: 2 2SC1913 and 22SA913. Solder joint look very good....shiny, clean, solid and original.

Is this what you wanted?

I recognize that you (Mark) are very busy, and appreciate the amazing level of help you are providing. There is no urgency at my end. I just respond quickly because I have a PC in my workshop/music room, and I'm down here most evenings playing guitar or listening to music and messing with electronics.

Again, sorry I didn't give you the part numbers earlier...I just didn't understand that you wanted them. They are precisley as shown on the board diagrams.
 
Last edited:
ok, they look like factory parts which is even more puzzling given the conditions and readings.

and there may have been assumptions on my part about wanting numbers and nox explicitly saying it. I'd have to backtrack and analyze...
 
...OK, here is the info on outputs and drivers:

Outputs: 4 Fairchild 2SC2525 KY(G) and 4 Fairchild 2SA1075 KY(G)...

The 'F' is Fujitsu. They are original Ring emitters.

I have some failed ones here (used in the SA-9800 as well) that will test fine on basic tests, but will fail at voltages well under their rating.
 
whoops .sorry i missed that part .
Ok, I know how it is, not having time to dissect the instructions posted.
But when readings don't make sense, I try to figure WHAT they did that I didn't expect.
We gave him a scare about his meter and taking the readings.

BUT I STILL don't like those base voltages
Q27: E = 1.61V / C = 38.55V / B = 2.02V E=R99 1K 1.6mA
because 2.02 - 1.61v = 0.41v which had me saying "he know something I dont?" and I checked and did the math.
still has me scratching my head, because it is APPARENTLY showing 1.6mA through a biased OFF transistor!!
 
Are you sure that you are on the correct test points? 278mV across the emitter resistors should have your 60W bulb lighting up the room.
Have you measured the voltage drop on the emitter Rs themselves? Are they all more or less the same?
Test all of the resistors. I've seen plenty that looked perfectly fine, only to fall into two pieces when touched with probe leads.
Do you have an analogue meter? I have found them to be superior to digital meters for finding reverse leakage in semiconductors (out of circuit) in the absence of a fancy transistor tester.
I'll try to give Mark a little rest from this project. You will just have to methodically start at the outputs and work your way backwards.
 
What do you mean by "correct test points"? I followed Mark's directions, and did the testing 3 different times with the same results. Or are you referring to something else? To measure the transistor voltages, I connected the black lead of my Fluke multimeter to the specified test point, and then used the red lead to probe each leg of the specified transistors.

As regards 278mV...I assume you are referring to the 280mV bias reading I reported in Post #104? That was on full power, not on the DBT. It is in the low 50mV range when on the 60 watt DBT. The bulb does not go completely dim...the filament still glows, but certainly nowhere close to "lighting up the room". On the DBT, the right channel (the "good" channel) measures around 14mV.

I have not measured the voltage drop across the emitter resistors themselves, as that has not been requested. I did check their resistance earlier, at someone else's suggestion; they were on spec.

By "all" the resistors, do you mean every resistor on the board? That will take some time, as mapping the resistors is difficult with this amp because the location numbers (such as R10, Q18, etc) are not printed on the board. I will be glad to do it if necessary for solving the problem, but wanted to be sure that's what you are asking for first.

I do not have an analog meter, nor have I done any testing for reverse leakage. In fact, I don't know what that means nor how to do it.

I'm sorry, but your recommendation to methodically "start at the outputs and work your way backwards" isn't helpful to me. Specifically, what is it that you are suggesting I do? Pull each output? Do what with them? Pull each transistor on the board? And do what with them? Lift every resistor and test resistance? Or measure voltage drops only? Test every diode? I'm not even sure I know how to do that, other than simple diode testing. What is it that I would be looking for?
 
Back
Top Bottom