EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: technix on December 26, 2015, 06:05:53 pm
-
Big idea: Use something to generate two 120kHz sawtooth waves, use a LM339 comparator to convert the input audio into two-level PWM using the two sawtooth waves, level shift with IR2110s to 12V gate voltage and switch 4 IRFP250 FETs to push out some crazy power.
Problems:
1) How to generate this two sawtooth waves used for two-level PWM?
2) What kind of inductor I will need to match IRFP250's current rating?
3) What scale of switch-mode power supply I am subscribing into?
4) How to tame the EMI?
5) How big of a speaker this thing can drive? aka how loud can it get?
-
http://seventransistorlabs.com/ClassD1/ (http://seventransistorlabs.com/ClassD1/)
(Minus the good linearity, and high power output. Relatively minor design issues, though.)
Also, there are transistors literally half an order of magnitude better than IRFPxxx, please try to forget the old ones exist! ;D
#4 is partly considered above. I haven't done a proper system level design and analysis, and I suspect it would fail when placed into such a test. But wouldn't need much help to pass, either. The two filter inductors provided do a pretty good job already, just "more of it" should finish the job. Some supply-input filtering would probably be desirable too, as well as input filtering and protection (which will be necessary to meet susceptibility demands of a commercial product). Since it's a 10W class device, a generic DC power adapter could be used, which is very handy (since an adapter doesn't have to be tested for surges or isolation, in an EMC test of the amplifier system).
As for #5, very large class D/PWM drives supply megawatts to motors in traction control systems (desiel-generator trains, cranes, ships, etc.). These are usually more specialized (using a lower switching frequency to save on switching losses, which are relatively large due to the losses in the high voltage transistors required), but the principle is largely the same, and most such systems could, in principle, be modified at some stage into basically being extraordinarily large subwoofer amplifiers! There is no limit on the scale, only the economics of need versus capital costs. (There is some limit to switching frequency, and consequent signal bandwidth, versus size, but that's not a hard limit. If nothing else, switching and RF systems can be scaled up using power combiner networks, it's just that the economics of all the extra components, plus having to ensure all the stages remain in perfect sync, aren't worth it for most applications.)
If you meant, just with a pair of IRFP250, the reasonable limit would be a few kW, up to perhaps 50kW for pulsed operation, give or take how much operating life you require (including as little as one switching cycle's worth of life...). :)
Tim
-
Big idea: Use something to generate two 120kHz sawtooth waves, use a LM339 comparator to convert the input audio into two-level PWM using the two sawtooth waves, level shift with IR2110s to 12V gate voltage and switch 4 IRFP250 FETs to push out some crazy power.
Problems:
1) How to generate this two sawtooth waves used for two-level PWM?
2) What kind of inductor I will need to match IRFP250's current rating?
3) What scale of switch-mode power supply I am subscribing into?
4) How to tame the EMI?
5) How big of a speaker this thing can drive? aka how loud can it get?
based on your evident lack of experience, I recommend just buy one and get crazy power out the box., but if you are hell bent doing a diy version, prepare for mega-watts of failures before your even get half a decent class-d amplifier with its critical sections to work reliably