Another Marantz 18 issue.

[tonebells]

Sorry, I'm not really sure what you are saying.

+ and - 1.4V where? We need to know where you are probing. Guessing won't be helpful.
When you say "goes low" what do you mean? Do both go towards zero or negative?
"The turn on delay" what is that? As far as I know there is no such thing.
Are you saying that if you turn off and then immediately back on you don't have the problem?

+1.4 at pin 10G and -1.4 at 10K. Low would be 0 volts. When the hum happens both go as high as +6 volts. See picture below from the owners manual about the turn on delay.

Yes after the unit plays for ~ five minutes you can turn off then back on with no hum. If you leave it off for ten minutes the hum comes back.

Tony

[nickb]

Thanks for the clarification.

The delay comes form the 350uF cap C21 and associated resistors. I'd missed that. The funny thing is that it feeds both channels so you'd expect them to behave the same. The idea is that initially C21 is discharged and so no current flows in the driver and so turns of the power amp. The time constant is not long enough when turned off to account for ten minutes so it has to be something after that. That sort of a time constant strongly suggests a thermal effect yet no transistor showing any sensitivity. Since you have freezer you can try it on other components especially the ceramic caps.

If that doesn't show anything I think I try defeating the delay by removing C21 and no speaker. You can then power up and down and compare the two channels to see what shows up.

[tonebells]

The 28 volts at 10A and the 26 volts at 10B both turn on when power is applied. So I would guess there is something wrong in the delay. I will order some 350UF caps to replace C20/21/22 and see what happens. Thanks nickb.

Tony

[nickb]

The 28 volts at 10A and the 26 volts at 10B both turn on when power is applied. So I would guess there is something wrong in the delay. I will order some 350UF caps to replace C20/21/22 and see what happens. Thanks nickb.

Tony

Since that's common to both channels I don't think it could cause a problem in just one. If you remove it that will defeat the delay and you can then probe the voltages during power on to see more clearly what it going on. Save your money for now. Also check the behavior of the voltage across R1003 and R1004.

[tonebells]

Since that's common to both channels I don't think it could cause a problem in just one. If you remove it that will defeat the delay and you can then probe the voltages during power on to see more clearly what it going on. Save your money for now. Also check the behavior of the voltage across R1003 and R1004.

R1003 and R1004 stay at 0 volts until the delay ends then there is .6 volts across them. I guess I don't understand how the delay works. Anyhow, I unhooked the negative end of C21 and the delay is disabled. Audio starts as soon as power is applied and no hum. Voltage on outputs jump to just over 1 volt then down near 0 volts.

Tony

[nickb]

Are you saying that both channels now behave the exactly same way?

[tonebells]

Are you saying that both channels now behave the exactly same way?

With C21 disconnected they both act the same. No hum from the right channel. No dc jumping around.

Tony

[tonebells]

I may be wrong but here goes. The delay is caused by the base of Q1001 being more positive then the emitter so no 1 volt at collector. Looking around Q1001 to see what's up.

Tony

[tonebells]

Going to try replacing Q1001.

Tony

[nickb]

I may be wrong but here goes. The delay is caused by the base of Q1001 being more positive then the emitter so no 1 volt at collector. Looking around Q1001 to see what's up.

Tony

Q1001, which is the current source for the entire driver, is on when the base is negative wrt the the emitter. C21 and associated resistors cause the voltage on the base on power on to be close to the +ve rail and then slowly charges toward the -ve rails.

I would not suspect Q1001 since it works OK when you don't have C21 fitted. I would consider possible that one of the differential pair transistors has gone low gain, i.e. QQ1006/Q1004 or Q1003/Q1005 or you that have instability. The heating/cooling was a rather inexact test for low gain transistors.

But stepping back a bit it's the hum that worries me. This suggests that far too much current flows through the output transistors during the bounce and it could be very close to destroying them. This is one of the reasons I said not to connect a speaker i.e. to avoid damaging that too. What I'd really like to know is what is happening to the current in the output transistors during the bounce. To do that you'd need to use a digital storage 'scope to measure the volts across R36. Or it could be a sign of instability.

I raised the possibility of instability before and that remains to be checked as far as I know. Again you need a scope to see that. There is a real likelihood that the presence of the speaker could make it switch from stable to unstable so making the whole thing rather tricky to work on.

[tonebells]

I do have an analog scope and a PC based scope. Do you have an idea what good subs for Q1006/Q1004 ans Q1003/Q1005 would be? I would guess they need to be matched?

Tony

[nickb]

They might work.

With C21 installed, use the analog scope to check for instability on power on. not having storage might make it hard to see. The PC scope might also work if the bandwidth is high enough 1MHz+ ought to do. The key thing is to set the trigger to normal and pick the level to grab what you want.

Use the PC scope in differential mode to measure the voltage drop across R36. Diff mode is where you use two probes, one at either end of R36 and both ground leads to OV. You use the math function on the PC scope to subtract the two. You'll be measuring a small voltage in the mV range.

[tonebells]

Well the PC scope is junk. The channels are not even close to reading the same voltage. I hooked up my analog scope and R36 is as it should be (same as left). I looked at a few things on the output and it all looks OK. Is there some thing specific I should be looking for?

Thanks Again

Tony

[nickb]

What we are looking for on the output is a burst of oscillation and note the max (+ or -) DC volts during power up.

[tonebells]

So far all I have seen is the voltage. A flat line up to 5 volts then low. I will look some more over the weekend.

Tony