Class d sub

Status
Not open for further replies.
Jump to Latest
#2
It is clear that you haven't searched the forums...It is a regularly discussed topic...

However, the likes of Tripath, Zetex, TI all do chips. The Zetex one is the ZXCD1000 (http://www.zetex.co.uk)

Strictly speaking, the Tripath in not Class D at all, it's called a class T. I believe the differences are that it has variable sample rate or something. It also uses predictive DSP algorithms etc etc. I challenge you to find a review of an amp using the Tripath that doesn't get a good write up. I know you're not interested in that...but there you go - a bonus. Find those at http://www.tripath.com.

Obviously TI is at http://www.ti.com...Do a search on their website.

Also Maxim may do a Class D driver...I know they do class D amplifiers for headphones but you may be able to get the PWM output...http://www.maxim-ic.com.


Like I said before, search the forums, there is loads. But for now, I hope that helps,
Gaz
 
#8
"is it not possible to build a class d sub with normal components?

how come all come in chip form?"

https://www.onsemi.com/pub/Collateral/AN1042-D.PDF

Linear Technology also has a design for a full bridge amplifier but it looks even more complex.

The voltage balance circuit may be eliminated from the Motorola (ON semiconductor) design if using a non-regulated supply. Use a 40-0-40V 300VA transformer for 200W/4R, 55-0-55V 300VA for 200W/8R, use IRF640/9640 for outputs.

If an input diode clipper (NAD 'soft clip') is used the duty cycle limit switches could be eliminated.
 
#9
djk wrote:

If an input diode clipper (NAD 'soft clip') is used the duty cycle limit switches could be eliminated.
Click to expand...

I like it if an amp can be driven as close to the rails as possible, so I used a different approach on a class-d amp 10 years ago, to overcome latchup during overdrive situations (although you might still use some circuitry to achieve a soft clipping characteristic):

The MOSFET drive-circuit is basically the same as with Motorola AN1042 (though the rest of the circuit is different), but I used a parallel DC-path using a transistor in common base circuit. This has to supply only miliamperes to work properly.

Regards

Charles
 
#10
#11
It is not easy to build a class d from discrete components. Firstly, you need an extremely fast sawtooth or triangle wave. That usually requires an opamp at least. You need a comparator to generate the PWM there's another high speed opamp. After that you need under/over sampling protection etc etc and then you have the discrete output that all the circuits have. In other words you still have to have chips unless you make up discrete opamps...but these are not really in the signal path (apart from the comparator). You shouldn't really take noise into consideration (by comparing discrete/chip) since you are totally dismantling the signal anyway! You will get far better results from an H-bridge driver or dedicated Class D driver...and it'll be a whole lot safer!

Gaz
 
Status
Not open for further replies.
Top Bottom