Cross-conduction - how to avoid?

Status
Not open for further replies.
Lots of useful thoughts there cheers.

Lumanauw: I haven't quite managed to figure out how that circuit works but I like the theory. I'm planning something that monitors the two sides of the input LTP like the ESP distortion monitor and drives some kind of gain control on the input of the amplifier.

sajti: Thanks for the reassurement. I'm not planning to use more than 4 pairs of output devices so hopefully the MJE15032/33's should handle it. If I want any more power I'll be going for multiple rails to keep the efficiency sensible. But that's all for a few projects down the line!

megajocke: That is a very good point, I am using almost entirely OnSemi models. I'll see if I can find some other suitable devices in the library from another manufacturer to try instead.

Edit: Just tried swapping the MJ15022/3 models for 2SC3281 and 2SA1302 models in PSpice and miraculously the cross-conduction has disappeared! Looks like it could be due to dodgy OnSemi models after all! Unfortunately I can't work out in the bloatware that is Orcad and PSpice how to add/change models easily. Might have to go to LTSpice!
 
As I put in the edit of my last post, it seems the problem might be the OnSemi models I've been using. With a change of output devices I've now got no cross-conduction, even abusing the output stage.

So it may be we've been chasing a non-existent problem. It has however been an interesting and useful discussion.
 
Hi, Ceharden,

I found this, by Edmond. We can see the sensing in VAS is feeded back to -input near clipping.
 

Attachments

  • edmondamp-clamp.png
    edmondamp-clamp.png
    16.6 KB · Views: 254
Status
Not open for further replies.