T-Amp 5Hz oscillation

[rudiloos]

Hi to all

I have recently built an amp with a Tripath 2350 chipset. The problem I'm having is that it oscillates at 5Hz (rail to rail square wave). The board was drawn up by myself but is almost identical to the Tripath reference board with the only exception being the connectors.
This happens with or without a load, with or without signals and have adjusted the DC offset to various positions. Here is my setup:
33-0-33V toroidal transformer, rectified through a 35A bridge and 14100uF (3x4700u) caps on each rail - Produces ~49.5V per rail.
5V supplied by PC PSU (Used for many other projects without problems)
250W 12" speakers (These are fully functional on my other (linear) amps)
What I have done:
Replaced the TC2001 (Input and control chip)
replaced the TK2350 (Driver chip)
replaced all STW34NB20 Mosfets
Tried with a 28-0-28V toroidal.
Triple checked all component values and orientation.
Searched the internet with only this reference that might have some similarities but no solution -
http://www.diyaudio.com/forums/showthread.php?s=&threadid=73120&highlight=
Observations: (With scope)
The 5Hz switch outputs is coming out of the TC2001
The TC2001 is stable without the main power (with only the 5V)
The TC2001 does mute and unmute.
Power supplies are stable (5Hz not present)

I wanted to put a post on 41Hz.com but seeing that this is my own venture, I don't think that Jan would appreciate it much ????
I do have other boards and components but would like someones input before having to build up another one. (Might be my only choice to confirm board integrity)

Any type of input would be appreciated.
Thanks, Rudi
PS: An excellent job by all you DIY gurus

 

[BWRX]

Are you sure the scope probe is properly grounded?

You said you drew up the board, so you used the same basic schematic as the Tripath reference board but is your layout similar as well?

I'll try to look at the datasheets for the chipset later and see if I can come up with some more specific questions!

 

[rudiloos]

Hi brian

I hooked up the amp to a 12" speaker that I can sacrifice. Sounds like a flasher unit on steroids. This confirms that the mosfets and driver circuitry is working well.
I built up a 5V linear supply today to check if it was not an issue with the PC supply. Identical results.
I also used a heavy copper board and the same layout as the reference board.
I suspect something in the feedback circuit Did find a resitor that was open circuit in the feedback. Replaced it which also prompted me to check all the other associated components of both channels. I did a check between the two channels and all values are now identical.
That fact being that I am getting the 5Hz on both channels, doesn't help me to pinpoint the fault.
What I did find in the datasheet was that it takes 200ms to de-assert the mute(This is the only thing I can see that could relate to 5Hz), but my mute circuit is stable.
I have decided to start building a second board now to see if I get the same results.
If somebody out there has a reference board or an AMP2, I would really appreciate a couple of voltage readings on the control chip (TC2001)

Thanks for the input.
Rudi

 

[BWRX]

Hmmm.... can you measure the switching frequencies of each channel coming out of the TC2001? Are the frequencies very close or close enough to interact and cause a beat frequency of 5hz? What values are you using for the CFB caps for each channel?

 

[rudiloos]

Using 150pF and 300pF.

While biulding the second board, I found a fault on the board itself. This was the only one with this specific fault. Now I suspect the board.
I will have to do an electrical test on the next board before I stuff it.

Trying to get a thing done here in Africa without a grinder and hammer is tough . There has to be at least one board out the 10 that should pass.

Anyway, I'm seeing double at the moment - need some rest.
I will need a day or two to stuff the other board. I might post you a scope shot of the output. You might be able to see something that I'm missing or might not even know that it's there in the first place.

Thanks so far
Rudi

 

[rudiloos]

Brian:
Thanks for the help. I found the fault with the help of Tripath support.
I got the feedback1 and feedback2 swapped around.
It's up and running. In a way I'm quite glad all this happened, because now I know this amp quite well.
Cannot comment on the quality, as I have a couple of backyard components in at the moment, but so far it sounds quite good.

Thanks again
Rudi

 

[rudiloos]

Did mention that Tripath helped me out with this problem, but what I did not mention was that they were able to help me with just 3 emails and a total of 2 days.
This is excellent considering that time difference.

Hats off to Tripath support on this one.
Rudi