Help with a Class D amplifier / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Help with a Class D amplifier

1 2 next

Move Thread

LAN_403

Inducktion

Wed Jul 20 2011, 04:28AM

Registered Member #3637 Joined: Fri Jan 21 2011, 11:07PM
Location: Buffalo, NY
Posts: 1068

So, as I've learned a while ago that Class D amplifiers aren't as simple as I thought they were, I decided to set out on building a REAL one.

But, I've ran into a few problems along the way....

For the audio signal, I've decided to amplify it with a KA358 Op amp from Fairchild. I made it variable amplification using a 50k potentiometer for feedback. The output signal (when watched on the scope using a tone from my computer) it amplifies it from a low voltage, (roughly 200 mV or so?) To around 10 volts. I have it also DC offsetted, so the sine wave is 10 v peak to peak.

My problem arose, when I tried to use the comparator to generate the PWM signal. Feeding the amplified audio into a LM393 comparator, and using a 555 timer square wave signal, the comparator output absolutely nothing. Huh?

I looked back, and doubled checked everything to make sure it's all connected right. I scoped the inputs again, to make sure they were still kicking, and yeah, they were....

I've also tried using a low pass filter on the square wave generated by the 555, which made a nice 10 P-P sine wave. However, this was not DC offset, but I didn't think that made a difference, although, prove me wrong. When I tried feeding that into the comparator, still nothing was output by it.

But, when I switch out the comparator with a KA358 op amp, it outputs a square wave, but the square wave is the tone's frequency.... However, I can change the duty cycle with the 555 timer.

For powering all of the logic, I'm just using a 15 volt wall wort, with a LOT of extra filter capacitance (like in the order of 40000 uF)
Would not having a negative rail create some problems?

This is making no sense whatsoever to me, and has been extremely frustrating to me. Could someone please point out the errors of my ways? Also, as far as the signal strength for the comparator, do the two input signals need to be roughly the same voltage? If so, how would I get that?

I appreciate the help!

Myke

Wed Jul 20 2011, 05:53AM

Registered Member #540 Joined: Mon Feb 19 2007, 07:49PM
Location: MIT
Posts: 969

Would you mind drawing up a schematic with the pin numbers included? (I've made a mistake of wiring things to the wrong pins before) The 555 would generate an exponential growth and decay wave with the peak at 2/3 Vcc and 1/3 Vcc. You could use two op-amps that are designed for rail to rail voltage outputs and make a simple triangle wave generator with that. It's made by a comparator with hysteresis and a voltage integrator.

Wolfram

Wed Jul 20 2011, 11:42AM

Registered Member #33 Joined: Sat Feb 04 2006, 01:31PM
Location: Norway
Posts: 971

You're not supposed to feed the comparator with a square wave, but with a triangle or sawtooth wave. You can get a triangle wave from the classical two op-amp oscillator (integrator and scmitt trigger)

and you can get a sawtooth wave from a 555 with some constant current source trickery

The linearity of the ramp determines the distortion of the amplifier, so an exponential sawtooth from a standard RC oscillator won't work very well.

Inducktion

Wed Jul 20 2011, 07:36PM

Registered Member #3637 Joined: Fri Jan 21 2011, 11:07PM
Location: Buffalo, NY
Posts: 1068

Here it is!

Arcstarter

Wed Jul 20 2011, 08:58PM

Registered Member #1225 Joined: Sat Jan 12 2008, 01:24AM
Location: Beaumont, Texas, USA
Posts: 2253

I would also do what Anders suggested, like i suggested before. If you can't get the triangle wave generator to work with the op amps, you can make a simple inverter oscillator with a hex inverter, and the voltage across the timing capacitor is a fairly clean triangle wave. however, it is low amplitude and it has alot of DC offset, but you could just use a DC block cap and an op amp.

Edit:

Wolfram

Wed Jul 20 2011, 09:51PM

Registered Member #33 Joined: Sat Feb 04 2006, 01:31PM
Location: Norway
Posts: 971

An inverter oscillator will make an exponential triangle wave which will lead to a lot of distortion.

Also, GDTs do not work to well for duty cycles deviating too much from 50%, as this will add a "DC offset" to the output.

Inducktion

Wed Jul 20 2011, 10:10PM

Registered Member #3637 Joined: Fri Jan 21 2011, 11:07PM
Location: Buffalo, NY
Posts: 1068

Anders M. wrote ...

An inverter oscillator will make an exponential triangle wave which will lead to a lot of distortion.

Also, GDTs do not work to well for duty cycles deviating too much from 50%, as this will add a "DC offset" to the output.

What am I supposed to use instead?

Adam Munich

Wed Jul 20 2011, 11:24PM

Registered Member #2893 Joined: Tue Jun 01 2010, 09:25PM
Location: Cali-forn. i. a.
Posts: 2242

Gate drive IC?

Myke

Thu Jul 21 2011, 09:19AM

Registered Member #540 Joined: Mon Feb 19 2007, 07:49PM
Location: MIT
Posts: 969

Some things I noticed from the schematic itself is that the comparator is driving the bases. Comparators aren't designed to source all that much current so you'd need to use an inverter to square things up (fed by the comparator and a pull up resistor). This would then go to the bases of the transistors.

Anders M., the non-50% duty cycle shouldn't be a problem because he has a 1uF DC blocking cap which would prevent core saturation. Using a GDT is nice in this case because then you don't need to deal with level shifting the gate drivers so that the voltages on the MOSFET gates looks right.

On the datasheet for the BDW93C, it says it's designed for low speed switching applications. At a 40kHz switching speed, the switching time is about 16% of the cycle

Inducktion

Thu Jul 21 2011, 06:35PM

Registered Member #3637 Joined: Fri Jan 21 2011, 11:07PM
Location: Buffalo, NY
Posts: 1068

Where does it say that it's designed for low speed switching?... I'm looking at fairchild's datasheet and it says nothing. :P

Also, I might end up using UC3710N low side gate drivers to solve the driving problem. They're two gate drivers in one package, one inverting, and the other non inverting ^^

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob