AMZ tweed stack sputtering after some time

[KarenColumbo]

Dear fellows, I breadboarded me an AMZ tweed stack from my CD. Sounds great and all but it happens to start sputtering after some 10 minutes or so - signal gets chopped like very badly biased fets. Which is strange because - well - this Mu Amp or SRPP or totem pole whatchacallit is said to be self-biasing.

- I tried 3 different breadboards taking extra care to not overlook anything.
- swapped out the fets after the breakdown started, tried j201 and 5457 .. to no avail
- wallwart delivers steady 9.44 Volts, voltage at the upper jfets' gate is more or less exactly half of it
- the signal loss happens at the second stage when I probe the lower jfet's drain. First stage is crystal clear and loud a.f.

Mysterious.

[thermionix]

If you power off, power back on, does it work fine for the next 10 minutes?  Or do you have components failing and being replaced?

[KarenColumbo]

If you power off, power back on, does it work fine for the next 10 minutes?  Or do you have components failing and being replaced?

I plugged off and on, no change. It stays dead. Will try yet another breadboard, they come cheap, and that may be the problem. Switching fets from one stage to the other and inter-stage switching had the same result. first stage cool, second not cool. So it may not be those three legged basterds

[Kipper4]

Scheme pls so we can see what you have got.

[KarenColumbo]

Well that's the problem - it's on the AMD-CD which is subject to charge. Or is it okay if I post this here?

[amz-fx]

Since the two mu-amp sections are very similar, if the first one works then the second should work as well. The layout of both should be very close in design on your breadboard.

Measure the dc voltages on the first mu-amp and compare to the second one. They should be quite close.

The upper fet on the second amp should be around 9.44v (d), 4.7+v (s) and 4.7v (gate). Source is often very slightly higher voltage than the gate.

I suggest a close inspection of the layout of the second mu-amp.

Best regards, Jack

[KarenColumbo]

Since the two mu-amp sections are very similar, if the first one works then the second should work as well. The layout of both should be very close in design on your breadboard.

Measure the dc voltages on the first mu-amp and compare to the second one. They should be quite close.

The upper fet on the second amp should be around 9.44v (d), 4.7+v (s) and 4.7v (gate). Source is often very slightly higher voltage than the gate.

I suggest a close inspection of the layout of the second mu-amp.

Best regards, Jack

thanks jack, will do

[amz-fx]

For those interested in this project, the basic info is here:

http://www.muzique.com/tech/tweed.htm

No pc boards are currently available for the project. I'm kicking around ideas for a v2.0 at this time.

Best regards, Jack

[KarenColumbo]

Well I teared it down and rebuilt it several times and arrived at something slightly different:
Medved's Fet Twin which has exactly the amount of clean-ness I expected (18 volts provided):

I guess my stress-addled brain just sent wrong information to my hands while sticking parts to the breadboard.

However.

I didn't want to let this go and made a layout for the Tweed Stack - checked it about 20 times - and sent it off AllPCB. Sometimes attack is the best form of defense. Boards will arrive in 3, 4 days' time so time will tell ...