Self Oscillating Amps
I use three of the ICE Power self oscillating amplifiers in my product. See the PM-3 amplifier at www.rh.com. I have never had the slightest problem from intermodulation or anything else fromy having the amplifiers unsynced. And in heavy duty professional use, with over 1000 modules deployed we have had only one field failure.
I use three of the ICE Power self oscillating amplifiers in my product. See the PM-3 amplifier at www.rh.com. I have never had the slightest problem from intermodulation or anything else fromy having the amplifiers unsynced. And in heavy duty professional use, with over 1000 modules deployed we have had only one field failure.
Right. Shoot-through.ZeroTHD said:Remember, I'm new to this, so don't be too hard on me.
Regarding choice of components. I picked components I have on the shelf for easy and cheap prototyping.
Thanks,
Anders
The diodes on q5(?) and q11 won't be very effective... It's a baker clamp. You need a lower forward voltage than that of the cb diode of the transistor ie one schottky. Not a standard pn diode and certainly not two. The simulation won't catch this because recovery isn't very well modelled in spice.
Simulate one single driver with a mosfet with an ohmic load to get a feel for the kind of resistor values you need. Hint: only the 10 ohm resistors are correct.
Look at the gate wave form. If you want to prevent shoot-through the gate waveform should fall quickly.
Self-oscillating
Self-oscillating amps are indeed liable to produce intermod tones. So far we've gone up to 5 channels x 100W crammed on a single PCB. The layout is a ***** but you can get the tones below a few tens of microvolts, which is acceptable. Besides, if you're a hobbyist - who cares if you try the water with mono blocks first?
A propos of this thread, Jan-Peter was wondering if the SODA modulator couldn't be released into the wild by now. I guess it can. It isn't nearly as good as the current state of the art UcD but I can't give you that yet of course. On the other hand it's already got one important thing: post-filter feedback. A disadvantage is that it doesn't like to be clipped without load. I'm leaving it up to your own imagination how to detect if the amp's gone oscillating at fres, but simply disabling and re-enabling the power stage is enough to get it back onto its feet.
Self-oscillating amps are indeed liable to produce intermod tones. So far we've gone up to 5 channels x 100W crammed on a single PCB. The layout is a ***** but you can get the tones below a few tens of microvolts, which is acceptable. Besides, if you're a hobbyist - who cares if you try the water with mono blocks first?
A propos of this thread, Jan-Peter was wondering if the SODA modulator couldn't be released into the wild by now. I guess it can. It isn't nearly as good as the current state of the art UcD but I can't give you that yet of course. On the other hand it's already got one important thing: post-filter feedback. A disadvantage is that it doesn't like to be clipped without load. I'm leaving it up to your own imagination how to detect if the amp's gone oscillating at fres, but simply disabling and re-enabling the power stage is enough to get it back onto its feet.
Attachments
Thank you all for your input.
I was a little bit in a hurry when I posted the schematics- some of the resistor values are *way* off my best guesses (which probably are way off the ideal anyway ).
Jan-Peter,
Do you think It's time already to build a prototype? I'll think i have to do some more simulations first...
Bruno Putzeys,
Thank you for the hints. Although my lack of experience in the field I can already say that your solution is elegant.
IVX,
No problems with shoot-through? With the current resistor values?
And again: thanks!
.Anders
I was a little bit in a hurry when I posted the schematics- some of the resistor values are *way* off my best guesses (which probably are way off the ideal anyway
). Jan-Peter,
Do you think It's time already to build a prototype? I'll think i have to do some more simulations first...
Bruno Putzeys,
Thank you for the hints. Although my lack of experience in the field I can already say that your solution is elegant.
IVX,
No problems with shoot-through? With the current resistor values?
And again: thanks!
.Anders
Pabo,
I would do this first, (quote from Bruno):
Afters this build a proto with a double layer, one layer with very BIG GROUNDPLANE....
Succes!
Regards,
Jan-Peter
www.hypex.nl
I would do this first, (quote from Bruno):
Afters this build a proto with a double layer, one layer with very BIG GROUNDPLANE....
Succes!
Regards,
Jan-Peter
www.hypex.nl
Yep, will do. We have pretty good equipment for building prototypes at work and I feel more confident laying out a PCB than I do with analog design.
I've tried Bruno's advice before, and it looks pretty good with a single MOSFET driving a single 6 Ohm resistor. The Should I alter the symmetry between the high- and low side driver to?
Regards,
Anders
I can only get one leg of the driver to switch at all by driving it with a 5+ amp current source.....10 Amps gives an idea switch..?
Anything else (voltage source or less current than 5amps) gives me drive voltages which float all over and only a veeeeery low valued active pullup resistor brings gate drive to appropriate levels, but I think it's far too low as I then lose amplitude of the signal.
Can see I have some work ahead of me. Should probably start by finding a better mosfet model ...the one I'm using has almost 1n CISS, could explain why it likes ten amps drive at 500khz.
Cheers.
Chris
Anything else (voltage source or less current than 5amps) gives me drive voltages which float all over and only a veeeeery low valued active pullup resistor brings gate drive to appropriate levels, but I think it's far too low as I then lose amplitude of the signal.
Can see I have some work ahead of me. Should probably start by finding a better mosfet model ...the one I'm using has almost 1n CISS, could explain why it likes ten amps drive at 500khz.
Cheers.
Chris
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.