Amp with high pitched static noise

[G6ONEDave]

What is printed on D1? I wonder if D1 has got reverse leakage. If it was me I'd change it for a known good replacement. We need to know what type of diode it is.

Dave

[G6ONEDave]

Here's a sketch of the what I can glean from the component side of the pcb. Can't do any more without seeing the requested track side.

Dave

[Station X]

Much appreciated Dave.

Wasn't the track side shown in post #82?

[spacebiscuit]

The track-side of the pcb as requested, plus the bulb-holder i made.

Also note regarding the diode test, the image showing the capacitance is from the good working channel. the one without capacitance is the channel wth the fault.

[G6ONEDave]

Thanks for the track side image, that should help. Graham the image in post 82 whilst being of the track side had coloured arrows added that were obscurring several tracks, especially those concerning D1, R7 and C5 and what they connect to.

I hope to amend my sketch circuit (post 102) in due course using the track image in post 104 and the earlier component side layout pic.

Dave

[G6ONEDave]

This is my attempt at doing a rough sketch of the circuit. I profess to have got it perfect but it might help others to see what is happening.

Dave

[Diabolical Artificer]

Bless you Dave for taking the time to draw that, I know it can take hours to do over several iterations. I'm no expert on transistor amp design so will leave analysis to others, but from what I can see, D1 biases the fets so they're turned on, so, pretty typical AB1 design.

So as you can Mr SB the only things that can really cause a big short is the mosfets. To get the amp working as mentioned previously is to power up the board, current limited power supply without the fets in. If that's ok, solder in a N & P type fet, power up, ok/not ok? Next solder in the remaining fets, same again. If at any point your LL glows bright or a fuse blows, remove one fet & repeat.

I can't think of another way the OP can fix this amp with the limited test gear he has, if anyone else has a better idea, please chip in. Thanks again Dave for all your work.

Andy.

[G6ONEDave]

Thanks Andy.
I just thought that in agreeance with you that the only way forward was to draw the circuit. Hopefully this won't become a habit. That was the third redraw but I saved a bit of time by not drawing the symbols for the mosfets.

This might be a silly comment but I thought that all diodes exibited some capacitance, as shown in the test for D1 from the working amp. However the test on D1 from the faulty amp has no capacitance value, that's why I wondered if it had reverse leakage resistance making the capacitance too large or non existant. It's just my thoughts and I could be totally wrong with my suspicion.

Dave

[spacebiscuit]

Also wanted to say thank you Andy for taking the time to draw the schematic, obviously it is something beyond my skill set (at least a meaningful effort not full of errors).

My plan was to not power the amp up again until I have the new Mosfet which are in the post.

In the meantime would it be worth my while ordering a replacement Diode (D1). Presumably these are cheap and can only help?

[Diabolical Artificer]

I didn't draw it, Dave did, I've too much on at present, as Dave says it takes ages and probably even longer if you don't have the physical board to check your work.

You can power the amp up with two FET's, IE one N & one P type.

Depends what D1 is, is theire any writing on it? From your test results they had a large ( to my eyes) forward bias voltage, that could be your tester though.

Andy.

[spacebiscuit]

The Mosfets arrived this morning, sooner than I hoped.

I have put them into my tester - attached photos show the results:

Image 1 - 2SJ352
Image 2 - K2221
Image 3 - K2221 (from another replacement batch I purchased but did not fit - included for comparison)

[Diabolical Artificer]

They look OK, nothing I can see shows a fault, but how they fare in your amp is another thing entirely. Without a professional curve tracer it's a case of suck it and see.

As mentioned previously try one pair and see what happens. To reiterate, power up without any FET's, if OK, no bangs, smoke or blown fuses...pop a pair of FET's in, OK? Pop another pair in. If everything's OK then get your good board, take some DC voltage measurements at the gates, drain & source of each FET. After that connect aa speaker or better still a dummy load.

Andy.

[spacebiscuit]

Presumably that would that be a matched pair, ie the 2SJ352 and one K2221 - not the pair of N-Mosets or the P-Mosfet pair?

[Diabolical Artificer]

The former, ie the 2SJ352 and one K2221 yes.

Andy.

[spacebiscuit]

I've not tried to fit the new parts yet - trying to be measured and not rush things.

I've been back over the board again and came across that scratched out track again. The 2nd photo shows the same point on the component side.

I've checked using my DMM between the negative connector and that small resistor on the good board, since I only have access to the component side of the board I can conclude that it doesn't have that track scratched out since the dmm shows continuity where as on the bad channel board there is no continuity.

So I'm wondering if I should make that track good - could that tin itself be causing the fuses to trip.

I also checked that diode (D1) - with a dmm I got the same consist readings albeit with it stil on the board.

[Herald1360]

A picture of the same area of the good board would be interesting. Does it have the same capacitor mod as the scratched out track board?

I wonder what that mod was for? Something taking off at RF, perhaps?

[spacebiscuit]

Originally Posted by Herald1360

Does it have the same capacitor mod as the scratched out track board?

That is the question. If that scratched out track is intentional it wasn't done by me and therefore it's been like that for more than 25 years and the amp has functioned faultlessly.

[Diabolical Artificer]

AFAIK that trace connects the 10k resistor from -Vcc to the green LED. It doesn't do anything else but feed that 10k. With the trace cut the green LED will still light because the blue cap's leg is shorting out the fuse. Hope I have that right, that blue cap is still weird.

Andy.

[spacebiscuit]

I've temporarily fitted a pair, see the attached photo, I think I have used the correct points for a pair - J2221 on the left and the 2SJ352 on the right. I've checked that the solder points are good, no continuity where it should not be.

I also popped the old J2221 into my tester and at one point I was getting results similar to the new part N-MOSFET. However since I got the good reading I've not been able to repeat it. Now I get diode instead of N-E-MOSFET. I've tried opening and closing the gate with the DMM without success.

Anything else I can check before I power this up with my bulb tester?

[Radio Wrangler]

The soldering looks suspect on those two transistors. I would definitely redo them with some fresh lead-based solder and a hot iron to get fully-wetted, shiny joints before trying any power.

David