Reinventing the N-channel wheel?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
The images were not deleted, they were just moved to a case sensitive host and the original files happen to have the name in upper case while I used lower case in my posts. I think that I have fixed all the file names. The same problem happens with the images that I posted in other old threads.

The breadboards of this project are still resting in one corner covered with a plastic bag to avoid dust. I may continue some day, but now I'm deeply involved into class D development, as lumanuaw cleverly predicted almost two years ago. Some of my current class D work may be seen in the class D forum.
 
Sorry it is long ago, just to say thanks.

Actually I'm searching for cheap amp project, but I've found this kind of input stage. It is new for me and I think I've got the point, a strong differential.
May I ask you your experience with this stage? Is there ever a problem caused by this input? Because it is very rare, might there some errr.
I have planning to use this kind in my follower amp, it need stronger differential than currently used.
Sorry this thread was long ago, but thanks for this post.


I'm stopped designing class-D but I'll just looking at yours.
Sadly this project ended here. I've enough experiencing myself with DMOS in power conversion and Amps, Floating Driver is the best way DMOS Amps for me.
 

Attachments

  • evanch_rr0_bw.PNG
    evanch_rr0_bw.PNG
    12.5 KB · Views: 373
I'm realize some disadvantage in this differential, now.
Yes it has stronger gain, but it is not much better. Some transistor already with almost constant Vbe at certain current region.
Both input will interfere, and some more problem, like stability and performance in stressed condition.
 
MOSFET's are limited by the package they are in for power dissipation.
Some specs seem a bit optimistic to me especially the IRFP260.

I usually use IRFP240's (£1-50) in my designs as they are cheap and not optimistic with power rating.

I believe an amplifier needs cross conduction to get rid of crossover distortion.
Simulations can be used as general pointers but are rarely exact.
 
I'm realize some disadvantage in this differential, now.
Yes it has stronger gain, but it is not much better. Some transistor already with almost constant Vbe at certain current region.
Both input will interfere, and some more problem, like stability and performance in stressed condition.

I think you dismiss this circuit a bit too quickly: it is very smart and clever indeed (what else would you expect from Eva?..).
It looks like a variation on the Xquad, but without some of its problems.
The accuracy and linearity, compared to a plain differential, are outstanding: practically one order of magnitude improvement.
 

Attachments

  • EvaQuad.gif
    EvaQuad.gif
    47.8 KB · Views: 341
That's pseudo-class-A stuff. It keeps the ouput devices always biased to an adjustable current level (R79 and R80) and there is another "AC cross-conduction" adjustment (R78). That reduces gate-charging spikes (the ones in Ic(Q17) and Ic(Q18)) to an absolute minimum. These waveforms also show the extreme non-linearity that MOSFET devices show in class-B at high frequencies, even in the harmless PSpice models with linear capacitances.

In PSpice it looks beautiful, as always...

I have a much simpler diode crossed output stage that does exactly this.
But an entirely different method, and drastically lower parts count. Search
my recent posts... Granted, I havn't yet tried grafting your input stage
with my output. That combo might be a winner... My methods are kinda
screwy: I do let Schottky devices sometimes turn off, but only Schottky.

You demand the right set of always-on AB curves for sure. I agree with
need for active pull up and down when it comes to driving MOSFET gate
capacitance with any symmetry, you got that right too. Just throwing a
few too many parts at an analog computing and driving problem that can
be solved an easier way.
 
Last edited:
I think you dismiss this circuit a bit too quickly: it is very smart and clever indeed (what else would you expect from Eva?..).
It looks like a variation on the Xquad, but without some of its problems.
The accuracy and linearity, compared to a plain differential, are outstanding: practically one order of magnitude improvement.

Sorry, I don't know about Eva, there are many amp marked EVA and don't know them. I've seen Eva's class-D thread, but not all of it.
I'm really can't placing this good stuff in my follower amp. This Amp is class-H multi step with MOSFET as final, driven by fast 5A peak driver to eliminate crossover and switch distortion. Unlike BJT, big MOSFET has big drain-source capacitance that makes difficult for class-H. 10-15MHz small oscillation always happened in crossover and rail switching recovery, and this Xquad can't handle this condition. Currently I'm using 4x2sc3355 for my input stage.
 

Attachments

  • H6691L iut.PNG
    H6691L iut.PNG
    29.3 KB · Views: 251
hi umut, you ask the same.

Hi elvee, this is H6660L(older) driver stage faillure, 1A peak driver. PNP input stage. Btw, do you know when this project continued? might I'll help.

I've moved from class-D because the component limit make me stack in no improvement. Then I'm trying to create digital power source, and this become component limited too.
There also EEEngine from yamaha starting beating class-D amps, and other like tripath no news (may be stacked too). I've heard that crown class-D driving motors at industries, this is bad.
 

Attachments

  • resp@BF970 faili.PNG
    resp@BF970 faili.PNG
    4.6 KB · Views: 200
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.