My very first Class D pwm (switching) amplifier.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Thanks, IVX, for your comments.

A) Rfilter and Cfilter value is are no successfully chosen.

About Rfilter and Cfilter, some simulations have been made and the desired cutoff (33KHz aprox) was met.

B) R14, C18, R6, Q1, Q2 is can be eliminated.

The Slow-start components (C18, R6) perhaps can be eliminated, but proved very useful in my previous version, so I prefer to keep them.

About Q1 and Q2, I guess you are talking about Q1 and Q5 and the circuit you showed in your post #353. Well, perhaps, but I also proved this set-up in my previous version and worked very well, so IMHO... if it works, don't touch it! :) But it is only an opinion!

Does someone want to build and test it????

Thanks!
 
ok, but 100ohm input impedance is too low!

Sorry, but IMHO that's not correct: the input impedance is not determined by Rfilter, but by R14. Thus it is 47K (That's what R14 is for, to fix Zin). Just think in what would happen if I don't put the input filter... the impedance won't be 0 Ohms, will it?

To test it, I have simulated what would happen if the signal generator had an output resistance of 100 Ohms: if Zin=100 Ohms as you said, the output of the filter at low frequencies should have a drop of 6dB (half amplitude), but it doesnt. It only has that drop when the output impedance of the generator is 47K, as expected -> so the input impedance is 47K aprox.


Best regards.
 
Hi subwo1,
aha.. not only my English bad so much, but your Russian is even worse. :) Actually in this article.htm i wrote about some my experiments at 1995 - 99. (the earliest extremely stupid and 28 irf mosfets have gone to smoke!) Regarding the fig. ABD - for that moment I didn't see _crest_lt_scematic, but made it for safety minimization Dead Time as well. However had more EMI problems with FM radio, and I have thrown this approach.
Best regards IVX
 
Just tested my new Class D Amp

Today was the day of the big testrun... after some adjustments, it's really working nicely.

This is actually my first post to this forum, but I've been following it quite a bit.

I'd really like to thank Circlotron for his description of his BCA-Amp earlier. It was a great help for my design! I hope you read this...

But now to my amp, if anyone is interested. My plan was to build a powerful Class D amp for my subwoofer. I already built Elliot's Sub-Preamp, that compensates the subresonant response of closed subwoofers. Due to the significant boost of subsonic frequencies down to 10Hz, serious power is needed to maintain a linear response.
After experimenting with different modulation techniques, I read Circlotron's posting about his BCA-Amp, and I realised that this (patented) principle would solve my problems at the zero-crossing.

To make it short, I just finished building the amp on Veroboard, using a LF347N Quad-Opamp to generate 250kHz triangle waves, a KA319 Comparator for PWM generation, and in the power stage, 2 IR2112 Hi and Lo MOSFET drivers, and 4 BUZ11. The Power stage is a full bridge (4 MOSFETs, 4 torodial chokes, 220mH, and 100nF, as filters). The whole thing (2 small pieces of veroboard) is not even a handfull.

I can't really measure THD or S/N, but checked with an Audio Generator and an Oscilloscope. The output really looks nice, a cute sinewave, without visible deformations, even at full power (shortly before clipping). Lacking a good powersupply, I only could drive it with 30V (2Amp DC), which gave me an almost-30V-peak sine into 8Ohm load resistors, hence ~50 Watts RMS. The resistors got fairly hot, so I cooled them in a glass of water (which also got fairly hot after a while...). The MOSFETs were merely impressed. They don't even have a heatsink, and just got a bit warm. The freewheeling diodes also didn't have a problem. I think there is much more potential...
I should mention, though, that it only has a linear frequency response from 4Hz to 5kHz. This should be sufficient for a subwoofer, though. I'm going to experiment a bit, to make it full range (for my next project). Anyway, except for the missing high frequencies, what I could hear up to now through my speakers didn't sound to bad. No noticable distortion, and a powerful bass response.
If anyone is interested, I can try to update my schematic (most of the design process happened with a soldering iron... ;) and post it. And maybe also some pics...

Cheers

Felix (sfx)
 
Some pictures

Hi IVX,

Haven't had time yet to finish the schematic. But good news are, I have some pictures for you :)

http://syseng.anu.edu.au/~felix/iar/Projects/Thunderball/

The schematic will follow soon...

Which software do you use to measure THD, S/N, etc.? I used my soundcard already to measure the frequency response (spectral analysis, i.e. FFT of white noise) of my speakers, but haven't found a software yet that could do all the other measurements.

Cheers

Felix
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.