Designed an amp, it oscillates!!!!!

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Well, this is kind of an addendum to my older thread titled "much less crazy idea" in which I was coming up with a good design.

I have revised the design a great deal and built it into a four channel high power amplifier that will to 900W bridged or parallel, depending on your speaker impedance.

After building a single channel, just for prototype and to get some ideas, I was very very happy with my design, except that it oscillates too easily.

The great virtue of my design is very little noise, since the gain is very low. I'm thinking about 10V input for full output so it doesn't get much noise. Also, I find that the amp stabilizes very fast. Upon connection to a power outlet, there is a barely audible click, with a mellow and smooth sound. And the output voltage is instantly stable and what seems to be about >4mV.

The sound of the amplifier is simply musical and true, very very clean, enjoyable, and pleasant, with an exceptional noise floor. So I have reached my main design goals of good power-up stability and low noise.

The only problem is that the amplifier oscillates very easily on some loads and draws great current when doing so. It is oscillating above audio frequency. I need some help with how to make the amp less susceptible to oscillation.

Any ideas anyone?
 
Well, fortunately, this thread has become redundant. I just realized what my design was missing.

If you look at ESP 3A, there's a CFP output stage very similar to what I'm using. Look at the lower driver transistor and theres a cdom cap on it from base to collector.

I added this to my design and it stopped oscillating completely, even without a zobel network or anything, and I'm driving piezo tweeters directly too. That's nice!

The sound of the amp is now even nicer, with a more able approach because it's not wasting power to oscillate. It sounds like it has tons of headroom, fast response, and very musical.

I'm so darn happy with this amp!!! :D
 
I used 100pF caps in mine.

I'd imagine that 100nF would reduce your slew rate and frequency response a bit. Are you sure you aren't using 100pF?

Anyway, the cap made a world of difference.


To everyone else that reads this. There's still a problem. I measured the output offset incorrectly. There's about 300mV, not <4mV as I mentioned before:bigeyes: I need some pointers on eliminating this since my brain is fried after all that work on this thing.

Thanks.
 
Oops! You're right!!

Bad typo!!

It is 100pF.

As for the large output offset voltage, this can be addressed by (1) matching your DC input impedances as seen by the diff amp (i.e., input reistors and feedback resistors, and (2) gain-matching the differential pairs in the amp.

The scenario is worst if the amp is completely DC-coupled, because the closed-loop gain then also includes the DC component.]

Other than this, you may have to add an offset adjust trimmer.

Can we see your latest version schematic?
 
No diag yet. I've got to get time to draw one on the computer.

If you mix D.Self's blameless amp, and some of Elliot's designs of the p68 and the p3a you'll get something close to my design.

I used a single ltp input stage with CCS and current mirror, CCS on a class A VAS, and a CFP output stage. Very simple ideals actually, but a very linear and stable design.

Really my own version of the blameless so to speak.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.