G-7500 Issues

Sriskie

Well-Known Member
I've been struggling a bit with a G-7500. I think I have it mostly sorted out but am unsure of some things. I've been wading through numerous threads, particularly Anderslober's epic 27 pager and still have some concerns.

Without getting into every detail I'll get right to where I'm at now. The unit idles perfectly and is stable. This is the first time I've monitored the bias while playing music so I don't know what's "normal". As I increase volume, bias rises proportionately. I've seen conflicting posts on this and am searching for some clarification. At a fairly high volume (around 11:00), the bias is rising to about 40mv and drops right back down with the volume. It does this on both channels equally. I've read that the bias should rise with signal and also that it's a problem. I'm a little confused here. Perhaps it should rise slightly but not that much?:scratch2: The outputs and emitters are staying cold as well as 2 of the drivers. TR 17/18 (MJE15033G) run about 10 degrees warmer.

There was a mention about runaway bias being caused by the same drivers I've used, but they've been recommended several times. I'll list all of the subs I have in there now in case there's something that's amiss.

Drivers;
2SD381-MJE15032G
2SB536-MJE15033G

Outputs;
2SA1106-NJW0302
2SC2581-NJW0281

Others;
2SD313-MJE15032G
2SC2071-KSC3503D
2SA939-KSA1381
2SB507-MJE15033G
2SA992-KSA992
2SC1845-KSC1845

All of the E.C.s, 10D2 diodes, and N.I.R. resistors have been replaced and all of the other resistors checked so I'm thinking I may have made a poor choice somewhere on the TR subs if it's not behaving itself like it should.
 
Yes......I remember the G7500 ordeal! Anyway....to me your tr subs look good. If you have the volume at 11 o'clock without anything blowing or getting HOT.....then I think you are ok.
Now.......did you set the dc offset and the bias? Were they both easily adjustable and stable? If so, I think your are ok.
In some instances bias can/will rise with increasing volume.......it just must not runaway!
A guru will check in soon........
 
Yes, everything set normally with Bournes 25 turn trim pots. I'm just unsure what "runaway bias" constitutes.


Sent from my iPhone using Tapatalk
 
I think I'm going to try installing the set of drivers I removed from the channel that didn't blow and put them into the channel that did. I don't know how much bias it takes to blow the thing up again, and I'd rather not find out.

Anders... If I followed your thread correctly, I think you ended up installing original drivers in your G-7500 for similar reasons but there were so many other things going on, I don't recall if it did anything for you. I guess I should give it another read to try get get more of it to sink in.


Sent from my iPhone using Tapatalk
 
Make sure that you don't reverse the main outputs. The manual, IIRCC, is not correct in this regard. This fooled me quite a bit. I ended up manually following where the emitters, collectors and bases for each output were connected. This worked.
Be careful..........if you mess up the main output order......it blows stuff on the driver board again!
 
Thanks Anders. I'd love to hear from you what the bias was doing once you finally got yours fixed.

Sansui service manuals can be fun, eh? It doesn't help that someone had already tried to fix this thing either. When I pulled the first output, I drew a "1" on the board and made a note of the transistor number. I put the subs in the same way and the the DBT stayed bright. After some research, Overundr1's thread in particular, I realized I had them in wrong. There were also 8 N.I.R. resistors on F-2980 that didn't match the list in the manual. I replaced with the same value as I pulled. That could be wrong too...they looked original but how can you tell? Whoever was in here before seems to have pulled a lot of components for testing before giving up. The solder pads for the TR15 driver do not exist anymore, he built some wire bridges, and I redid and tested them. That's on the side that didn't blow.

I found a recommendation for using MJE15030/150301 as subs for the drivers, they'll be here for the weekend.
 
Sansui manuals have a notoriety for being inaccurate. I have personally run into numerous differences between as drawn and as built. Also there have been a number of times that a part that was listed, shown on the foil diagram and schematic and simply was not there on the board. This is why asking on AK is so great because you can compare with others who have working original units. I have a dead one on the shelf, which I could pull and take pics for you if needed.
 
Last edited:
Thanks KingBubba.

I think I have everything oriented correctly. As for the different resistors, it could just be production changes. eg. The service manual states there should be 3.3k and I have 1.8k or 470 and I have 150. Pretty big discrepencies but they're all wrong if one is. The TRs all match the manual.
 
R15 & R16 are marked as 3.3KΩ in schematics for G-5700/6700/7700 and your G-7500, and also the value I have in two 6700s. These are in the direct path of the DC Offset adjustment and feeding the two FETs. I'm not sure why anyone would replace them with 1.8KΩ, unless they had modified the P/S.

I'd suggest looking at the reference voltage at the junction of R15 & R16 and see if it is -18.3vdc as advertised.

R606 is the only one I see marked as 470Ω(5W) and it is paralleled with R605, 330Ω(5W) for 194Ω. If R606 has been changed to 150(?) it would make the total = 279Ω, or a nearly 50% increase. The schematic shows that the supply side of R605&R606 should be +57vdc, and the other side is not marked. Schematics for the G-5700/6700/7700 label that point and the collector of TR604 as +22vdc, and the emitter as +12.5vdc.

You might want to check these voltages and see if the 150(?) has changed any of them.
 
Thanks Steve!

R605/606 were not the ones I had doubts about. I'll list the problem ones below.

I've measured across at least one of each pair of the questionable resistors and the voltages match the schematic pretty closely. The biggest discrepancy I found was 4%, most were closer.

The R15/16 junction measures -19.05v instead of -18.3...acceptable?

R49/50 are listed as 470ohm...schematic shows 150ohm...actual 150ohm

R616/617 are listed as 3.3kohm...schematic shows 3.3kohm...actual 1.8k

R43/44 are listed as 220ohm...schematic shows 82ohm...actual 82ohm

R612 is not listed...schematic shows 150kohm...actual 1.8kohm

That's pretty messed up! I ordered them all based on the parts list but confirmed and tested them as they were pulled.

I think these resistors are fine but still have doubts about the rising bias readings while playing music.
 
First: If you have the 7500 on a DBT then the voltage measurements will be a little lower than values marked in the schematic.

Line Voltage: This varies depending on conditions, and drives the voltage readings you take. I try to start each session by measuring B+ & B- and record that in my notes.

If I have agreement between the schematic and what is installed I feel confident. Agreement between channels helps, but it can just mean that a previous tech was thorough, not necessarily right.

Agreement between F-2980 schematics of different receivers is a pretty good indicator, although not all will be the same.


Thanks Steve!

R605/606 were not the ones I had doubts about. I'll list the problem ones below.

I've measured across at least one of each pair of the questionable resistors and the voltages match the schematic pretty closely. The biggest discrepancy I found was 4%, most were closer.

The R15/16 junction measures -19.05v instead of -18.3...acceptable? Could be result of different values for R616/617 and R15/16, and or B+ & B-.

R49/50 are listed as 470ohm...schematic shows 150ohm...actual 150ohm These are 22Ω(½) in the 6700/7700, and 33Ω(½) in the 5700

R616/617 are listed as 3.3kohm...schematic shows 3.3kohm...actual 1.8k These are 3.3KΩ in 5700/6700/7700

R43/44 are listed as 220ohm...schematic shows 82ohm...actual 82ohmThese are 22Ω(½) in the 6700/7700, and 33Ω(½) in the 5700, so for that series R43/44=R49/50

R612 is not listed...schematic shows 150kohm...actual 1.8kohm R612 is 1.8KΩ(½) in 5700/6700/7700

That's pretty messed up! I ordered them all based on the parts list but confirmed and tested them as they were pulled.

I think these resistors are fine but still have doubts about the rising bias readings while playing music.
 
Well that was stressful. I was unsuccessful in my attempt to blow this thing up again. :banana:

I installed the original 2sb536 and 2sd381 drivers in the channel that previously blew and left the MJE15033 and MJE15032 in the other. No difference between the channels as far as bias after I adjusted the replacements the same as the other side. Both sides go up and down with the volume like crazy. So without anything else to try, I put the power to it. At about 10 watts on the meters, bias is jumping up, at times well over 300mv (my DMMs can't keep up). I managed to get the outputs up to about 130 deg. F after a while but the drivers and emitters didn't warm up too much.

It looks to me that the MJEs are indeed suitable substitutes. I guess I had it sorted out a week ago but lacked the guts to risk blowing it up again. Live and learn, I suppose. At least that gave me the chance to go through the F-2852 board with new caps, KSC1845s, KSA992s, and MV-12s (2 X 4148s in series). I sounds much better now.
 
Great! A significant accomplishment.

What did you do about the questionable resistor values?
 
The bias is not supposed to be adjusted with signal. Period. It goes up with signal or the amp wouldn't be doing what it's supposed to do. You put more voltage into a load, you get more current by Ohm's Law.
Runaway bias is when you have no signal and the bias keeps rising and you can't correct for it by adjusting the pot. You adjust the bias with no signal and monitor it for 20 minutes. It should eventually stabilize and stop drifting. If it continues to rise you have a problem.
Sometimes the amp oscillates with no signal and that causes large current in the output stage and eventually blows the outputs. You have to monitor the output with a scope to see that. Any time you replace the outputs or drivers with different transistors it's a good idea to check the amp for stability to make sure it won't oscillate. You can run a square wave at 20KHz into it and turn the output up to no more than 10V peak to peak, into an 8 Ohm load, and look at the output with a scope. There should be minimal or no ringing on the corners of the square wave. If you see large ringing, the amp could be unstable.
You can then run a sine wave into the amp set to 10V peak to peak out, and sweep it upwards in frequency from 1KHz on up until the output starts to drop off. When you get to 7.07V peak to peak, that's the -3dB point of the amp. Continue to raise the frequency and the output will continue to drop. If it's stable it will just continue to drop until there is a very low signal or nothing. Continue to raise the frequency to make sure it doesn't come back up. If it does, the amp is unstable.
 
Since the voltages around them matched the schematic, I used the same values it had when it came to me.

The service manual for this thing is scarey. On the F-2852 board it lists one pair at 33 ohm and it's actually 33 k ohm. I didn't bother checking the schematic on that one. I wouldn't change it either way.


Sent from my iPhone using Tapatalk
 
The bias is not supposed to be adjusted with signal. Period. It goes up with signal or the amp wouldn't be doing what it's supposed to do. You put more voltage into a load, you get more current by Ohm's Law.

Runaway bias is when you have no signal and the bias keeps rising and you can't correct for it by adjusting the pot. You adjust the bias with no signal and monitor it for 20 minutes. It should eventually stabilize and stop drifting. If it continues to rise you have a problem.

Sometimes the amp oscillates with no signal and that causes large current in the output stage and eventually blows the outputs. You have to monitor the output with a scope to see that. Any time you replace the outputs or drivers with different transistors it's a good idea to check the amp for stability to make sure it won't oscillate. You can run a square wave at 20KHz into it and turn the output up to no more than 10V peak to peak, into an 8 Ohm load, and look at the output with a scope. There should be minimal or no ringing on the corners of the square wave. If you see large ringing, the amp could be unstable.

You can then run a sine wave into the amp set to 10V peak to peak out, and sweep it upwards in frequency from 1KHz on up until the output starts to drop off. When you get to 7.07V peak to peak, that's the -3dB point of the amp. Continue to raise the frequency and the output will continue to drop. If it's stable it will just continue to drop until there is a very low signal or nothing. Continue to raise the frequency to make sure it doesn't come back up. If it does, the amp is unstable.




Thank you very much for clearing the part about the bias rising up for me Dr. I wasn't sure what was normal. When I blew the right channel the bias was doing what it is now and the "good" channel changed only slightly. That and some conflicting posts I've found really had me confused as to what I should be seeing. Like I said, this is the very first time I've ever monitored bias while playing music. It provided enough information to get my addled and inexperienced mind spinning.

AudioKarma continues to home school me. That's awesome!


Sent from my iPhone using Tapatalk
 
Another thing that can happen; in an amp that has response down to
DC, if you have a leaky cap on the input of the amp, it will amplify any DC leaking into the input from the previous stage, and then if you measure the bias with the signal on, that channel can have higher current as you raise the volume, (if the DC is coming from before the volume control) because it is passing DC. In this case, you could have no audio signal connected to the amp, but the bias would increase with the volume control, even though there is no signal.
 
Another thing that can happen; in an amp that has response down to
DC, if you have a leaky cap on the input of the amp, it will amplify any DC leaking into the input from the previous stage, and then if you measure the bias with the signal on, that channel can have higher current as you raise the volume, (if the DC is coming from before the volume control) because it is passing DC.

:yikes:

That sounds like a disaster waiting to happen :yes:
 
Back
Top Bottom