Let's talk audio power amplifier ICs

[cvanc]

Hi all-

Looking to retrofit something more modern into older stereo gear that originally used the (not very good, obsolete) STK amplifier modules.

It occurs to me one way to go here is to strip out all the old parts, retain the heat sinks and power supply, and pop in one (or two if they're mono) newer design amp chip.  I've done some searching - mostly on the T.I. web site and Digi-Key - and I'm not finding a lot of parts that will tolerate +/-45V supplies and have a single ended output.

Best candidate I've found so far is the TDA7293, which looks pretty good actually.  But I wondered if the brain trust here had suggestions on other parts I should be looking at?

Here's the 7293 datasheet:

http://www.st.com/content/ccc/resource/technical/document/datasheet/4f/18/a6/c8/21/33/41/8d/CD00001887.pdf/files/CD00001887.pdf/jcr:content/translations/en.CD00001887.pdf

Thanks!

[b_force]

A good candidate is the LM3886.
The only issue is that most of the chip/power amps are (getting) absolete.

[Lee Leduc]

Would you consider a class-D module? Parts Express has quite a few available, For example:

www.parts-express.com/pedocs/manuals/320-309--sure-aa-ab32191-2x300W-taS5630-class-d-amplifier-board-new-manual.pdf

https://www.parts-express.com/sure-electronics-aa-ab32191-2x300w-tas5630-class-d-amplifier-board--320-309

[jaycee]

The LM3886 is capable of excellent quality, and indeed is being used in a number of lower priced "hifi seperate" amplifiers.

The STK circuit however is fairly reproducable with modern parts

[b_force]

The LM3886 is capable of excellent quality, and indeed is being used in a number of lower priced "hifi seperate" amplifiers.

The STK circuit however is fairly reproducable with modern parts

As well as over priced high expensive hifi amps

[macboy]

LM3886 is (can be) quite good, but +-45 V is pushing it. For 8 ohm loads, it's fine but often 8 ohm "nominal" speakers can dip to 4 ohms or less at some frequencies.

TDA7293 has similar cousins: 7294, 7295.

[rsjsouza]

LM3886 is (can be) quite good, but +-45 V is pushing it. For 8 ohm loads, it's fine but often 8 ohm "nominal" speakers can dip to 4 ohms or less at some frequencies.

TDA7293 has similar cousins: 7294, 7295.

I was looking around and saw that its cousin LM4766 is a similar story... ±45V is a stretch.

The class D TPA3255 and TAS5630B could potentially be fed from a single rail.

[Audioguru]

The Chinese class-D amplifiers produce hundreds of Watts with extreme distortion but have low distortion only with an output of 10W or less. Don't you want it for audio and not to drive a light bulb?

[b_force]

The Chinese class-D amplifiers produce hundreds of Watts with extreme distortion but have low distortion only with an output of 10W or less. Don't you want it for audio and not to drive a light bulb?

Depends a little bit how bad or good they have done their PCB layout and parts they choose.
But in principle, the fundamental blocks of the "chinese amplifiers" can make an amplifier that is as good as or even (much) better than audio amp IC's or even most discrete designs.

[Gyro]

Here's a nice analysis of the LM3886 circuit - particularly the effects of trace and lead inductance on stability and phase margin under capacitive load...

https://www.neurochrome.com/taming-the-lm3886-chip-amplifier/

[Audioguru]

Here are the spec's for the "high power" class-D cheap Chinese amplifiers:

[ivan747]

A good candidate is the LM3886.
The only issue is that most of the chip/power amps are (getting) absolete.

Unless you go class D.

[jaycee]

Class D has its place.. but I would argue a hifi amplifier is not one of them. I'm sure there ARE Class D implementations that are very very good, but the cheap Chinese kits are not amongst those.

It's easy for the big boys to get caught out on this as well - Rotel released a Class D amplifier. It turned out to be useless if you tried to use it with an FM tuner, because the Class D PWM noise radiated out of the speaker leads and swamped out the FM reception!

There's plenty of kits for Class AB, though again watch out for crap Chinese kits on eBay - usually poorly designed and with fake output transistors.

A good start is Rod Elliot's P3A Amplifier - http://sound.whsites.net/project3a.htm. I personally like to use MJE350 for Q4, and MJE243/253 for Q5/Q6. You can even get away with TIP2955/TIP3055 outputs if money is tight

[MWisBest]

Here are the spec's for the "high power" class-D cheap Chinese amplifiers:

1% THD is not audible in actual sound reproduction. It's pretty much industry standard for maximum clean output rating in this sort of application. You might hear 1% THD with a sine wave, but that's about it. Oh, and at 100W you probably can't hear even 3% THD. At high volumes you can't make out distortion that well. Not that I would like to push an amplifier like that, just saying if you really tested it you wouldn't be able to tell.  I think it's safe to say that amp could at least do an acceptable 100W.

If you're using a chip amp you've already thrown "golden ear" kinda stuff out the window and are just concerned with real-world results.

[Henrik_V]

Search for "gainclone"

[Siwastaja]

It's easy for the big boys to get caught out on this as well - Rotel released a Class D amplifier. It turned out to be useless if you tried to use it with an FM tuner, because the Class D PWM noise radiated out of the speaker leads and swamped out the FM reception!

Class D makes most sense in integrated active speaker design. Then, wiring length to the speaker can be minimized, and the actual speaker can be simulated and tested for performance and EMI, making the output filter design easier.

[Zero999]

It's easy for the big boys to get caught out on this as well - Rotel released a Class D amplifier. It turned out to be useless if you tried to use it with an FM tuner, because the Class D PWM noise radiated out of the speaker leads and swamped out the FM reception!

Class D makes most sense in integrated active speaker design. Then, wiring length to the speaker can be minimized, and the actual speaker can be simulated and tested for performance and EMI, making the output filter design easier.

The filter is often omitted or is as simple as just a ferrite bead, when the amplifier is integrated into the speaker. This should theoretically make it sound better, but won't make any difference in practise, as the distortion due to the speaker and acoustics will dominate the filter.

[David Hess]

A quick search turns up the inexpensive TDA7293 which operates at up to a nominal +/-50 volts and has an absolute maximum rating of +/-60 volts.

It is too bad that TI discontinued all of their non-class-D power amplifier drivers.