Class D amp in N&V Q&A

[Bob Scott]

Thank you Ceasar. Yup, this won't work.

This schematic needs to get rid of IC6, the transformer and rectifiers. It's not a battery charger.
Replace IC6 with an IR IRS2011. Double check that the logic nets labeled A and B are exactly the waveforms that the IRS2011 needs for inputs. I have not checked yet. I have never ever used 555 timers in my life 'cause my chief engineer once told me not to use them 'cause the are "inconsistent". Leave L1, C4, C7 Move C5, C6 as discussed below. Delete R10 and R11.

http://www.irf.com/product-info/datashe ... 011pbf.pdf

This amplifier has DC on the output, half of the voltage of the 60-100V supply. This is common in older linear amps in the late '60s, like my old Lafayette LA-85T. It was normal design to ground the speaker negative terminal and install a large capacitor in series with the positive end of the speaker. That way, no DC is present on the speaker terminals. If the cap is a 2,000uF cap, its capacitive reactance is 4 ohms at 20Hz. With a 4 ohm speaker, response will be down 3Db at 20Hz. So, remove C5 and C6. Ground the neg side of the speaker. Install C5, C6 in parallel with each other and in series with the + end of the speaker.

It'll be one of those amps that go "BOOOF" when you turn it on as C5 and C6 charge up, like my old LA-85T. Hope it does not blow the speaker coil right out of the gap as the caps charge. Watch for this if your speaker enclosures have ports. The only way to cure this is to use a split power supply for the 60-100V supply and eliminate the caps C5 and C6, which would be possible, but this amp has NO NEGATIVE FEEDBACK!

Come to think of it, the way Russ had C5, C6, and the speaker wired eliminates the "BOOOF", but you have DC floating on the output terminals. If that's OK with you, then ignore everything except the first paragraph and leave C5, C6 here they are.

MAn, this editor is impossible to work with once you go past a couple of paragraphs. The whole reply shakes up and down.

[CeaSaR]

What's interesting is that class D is a PWM of the audio signal. PWM is usually
associated with motor control. Well, what is a standard moving coil speaker but
a linear electric motor. With the insanely quick reaction time of the speaker, it
makes sense to use class D to drive them with the utmost in efficiency. I'm
waiting to see what Russ can devise that "should" work.

In the meantime I'll keep reading up on amps of all kinds.

CeaSaR

[Robert Reed]

Hi Caesar
Correct if I am wrong, but isnt the PWM converted to a sine wave before the speaker by filtration. I can't imagine a PWM directly driving a speaker without causing huge distortion. I thought that I had read somewhere that even with filtering, the class D's had more distotion than an "audiophile" would accept

[CeaSaR]

According to the "literature", the final conversion to analog is performed by the
speaker itself due to the momentum of the cone and the lag caused by the mass
of the cone trying to change direction. That smooths off the sharp edges of the
PWM signal into a relatively close approximation of a sine wave.

As for distortion, I would hazard to say probably, but with better designs it can
be minimized. I have yet to actually hear one (that I know of), so first hand
experience is not available.

Check out my links earlier in the thread. Lots 'o info out there.

CeaSaR

[Bob Scott]

Correct if I am wrong, but isnt the PWM converted to a sine wave before the speaker by filtration.

The PWM is converted to audio. Audio isn't always a sine wave but we normally audio equipment is analysed with sine waves in the lab. eg: the distortion measurement analyser. Actually the average voltage of a PWM signal IS audio with high frequencies added.

Good PWM is done at a high frequency where intermodulation products do not appear in the audio band, but well above it so as to make filtering easy without filtering any audio frequencies. If the high frequency harmonics are filtered out, only the original audio is left. If the high frequencies "RF" are not filtered out, the crossover passes them on to the tweeter, which may burn out because the tweeter cone and its mass cannot respond fast enough to ultrasonic frequencies.

I can't imagine a PWM directly driving a speaker without causing huge distortion. I thought that I had read somewhere that even with filtering, the class D's had more distotion than an "audiophile" would accept

The "distortion", the carrier and its harmonics are not in the audio band. Anyone who says he can hear them is a charlatan, a liar.

Don't most computer sound cards run in class D? They don't have heat sinks on the audio output.

[Robert Reed]

Caesar
Yes, that info on the site you posted was quite interesting and very well presented, but unfortunately not for the DIYer as the author pointed out. At least not for the faint of heart DIYer!
Guess I will stick with the old Class AB amps for a while yet, but I enjoyed learning a little more about Ds from your reference.

[Bob Scott]

I did find IR's cache of application notes for class D. An-1135 has nice basic drawings showing where the current goes and has layout guidance:

Class D Audio
AN-1070: Class D Amplifier Peformance Relationship to MOSFET Parameters PDF
AN-1071: Class D Audio Amplifier Basics PDF
AN-1097: IRS2011 and IR2011 Comparison PDF
AN-1129: IRS20954S and IRS20955S Comparison PDF
AN-1135: PCB Layout with IR Class D Audio Gate Drivers PDF
AN-1138: IRS2092 and IRS2092S Functional Description PDF
AN-1141: IRS20955S and IRS20957S Comparison PDF
AN-1146: IRS2093M Functional Description PDF

I found those here: http://www.irf.com/technical-info/appnotes.htm

EDIT: See http://www.irf.com and search for part number "IRAUDAMP" for a list of (very limited supply) available class D amp boards, fully assembled. Although these assemblies are available from DigiKey and Arrow, they are cheaper if ordered direct from IR. eg: IRAUDAMP7D, 2 X 250W, 91% efficiency, $199. Just add power supply and case.

They have over/undervoltage, overcurrent and thermal protection. Full documentation including schematics, BOM and gerber files are online at IR in the datasheets for each board type.

Man, this abundance of support and information from IR makes a guy feel all warm and fuzzy inside!

[Bob Scott]

UPDATE: IR "Class D"

I noticed a problem with the IR class D basic design. The first clue I got was that these amps are immune to clicks and pops. It's like they have a built-in click and pop filter.

The second clue was that the modulation frequency drifts with modulation, and is not constant.

Reading one of the manuals, I found out that these amps use a 1 bit DAC, a Sigma-Delta modulator to generate class D. For those of you who aren't familiar with the workings of a Sigma Delta (translates as "sum of all the changes") modulator, I have an analogy. It's as if you have a blind crane operator. You have to signal to him to raise the hook 30 feet. All you are allowed to signal to him is "a little higher" or "a little lower" so he can move the hook in increments of only an inch. So, it will take you 360 instructions of "a little higher" to get the hook up 30 feet. That's how a Sigma-Delta modulator works, slowly. No wonder it can't respond quickly to clicks and pops, and that makes me doubtflul that these class D amps have the slewing rate necessary to respond to high power 20,000Hz bursts. Any high amplitude audio will be subject to slewing rate limitations.

When the Sigma-Delta modulator is exactly at the right level, it does not stop. It has no rest point. It keeps signalling "a little higher" and "a little lower" over and over as a square wave.

[ringo47stars]

I wasn't saying it was a battery charger I was saying that it looks like the same design as a battery charger because it modulates a power supply (the 60-100 volts@10amps). Then I thought that sounds like an idea that could be done. Use an amp for a battery charger. The audio source and modulator-amp would emulate a memristor or the usual modulator. And the value for the charging volts would have to be more practical because how many people need to charge up a 60 volt@10amp battery. So if the modulated volts changes to the 12v that matches the input to the amp or 386 why not modulate that power and with the same auio source maybe that would increase the decible level without the huge power source to the speaker because the pulse or audio would be in the power of one of the amps. Then that modulated output could be used as a battery charger if the audio source to do that could be found. Or an amp too.

[CeaSaR]

Looking through my cache of old mags and found the April 1996 edition of Popular Electronics.
Starting on page 30 is an article on how to build a 60 watt switching amplifier by Rolando R. Burbon.
If anyone is interested, I could send some scans or maybe Photobucket or Postimage them next week.
My kind of build, TL074, 2N4401/03's and IRF9530/530's. Nothing fancy.

CeaSaR

[CeaSaR]

Ah, finally I have been published. The October '09 N&V Q&A column replies.
Thanks Russ!

CeaSaR

[MicroRem]

To quote one of my favorite philosophers...."This is the kind of spontaneous publicity I need! My name in print! That really makes somebody! Things are going to start happening to me now!"


Congrats on being published, Cesear, I'll wait for my new issue and check it out.

Tom

[CeaSaR]

Thanks MicroRem/Tom,

I really wasn't expecting to be put in there, I just wanted him to look at the board and comment.
It wasn't even a dissection or correction, just a heads up. I'll take the printed acknowledgement
any way it comes, especially spontaneously.

Again, thanks to Russ Kincaid for the mention.

CeaSaR

[Bob Scott]

Looking through my cache of old mags and found the April 1996 edition of Popular Electronics. Starting on page 30 is an article on how to build a 60 watt switching amplifier by Rolando R. Burbon. If anyone is interested, I could send some scans or maybe Photobucket or Postimage them next week.

Me please! I'd like to get a copy, and maybe one of your publishment too.

My kind of build, TL074, 2N4401/03's and IRF9530/530's. Nothing fancy.
CeaSaR

AHA! You DO use op-amps! I'll remember, next time you insist on transistor schematics.

[CeaSaR]

Bob,

I should be able to get to the scan this weekend sometime. I'll post a link here.
As for the "publication", here it is from the N&V Oct. 2009 Q&A Mailbag section:

"Dear Russell:
Re: Class D amplifier; June
2009, pages 31-32. There has been
a thread on the N&V forum about
your Class D Amp that needs your
clarification.
— Charles Ryberg

Response: I understand the
confusion; the design only does the
positive alternation of the audio. It
does not work; I don’t know what I
was thinking. Perhaps the design
could be salvaged by repeating
and paralleling the circuit with the
output diodes reversed to give the
negative alternation of audio —
much like a class AB linear amp."

Nothing special, but still exciting.

AHA! You DO use op-amps! I'll remember, next time you insist on transistor schematics.

DRAT! You caught me! Actually, I never said that I was opposed to them, just that I wanted to go
with discrete construction for those few projects I have asked about so far. Maybe once things get
better situated here I'll be able to get on the "band"wagon.

CeaSaR