Make a power amp out of op-amp ICs in parallel?

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I was thinking about designing a transistor power amp, when the thought occurred to me that it would easier and cheaper to just buy a load of op-amp ICs such as LM481 and hook them up in parallel. If I put a diode in line with the DC power to each IC, then it should be safe.

Is this feasible or am I overlooking something obvious and important?
I'm not sure what you think the diode will do, don't use it.
All you need is to make sure the dc offset between channels is low, and to ensure stability, depending on topology current sharing output resistors may also be needed.

I can't remember the name, but I recall a Scandinavian designer on head-fi building such a thing.

Current is limited only by patience and budget, so the max supply safe supply voltage will tend to limit output power.
Wow! So it would work.. That's excellent. I'll leave the diode out and just build it, bearing in mind that each op-amp IC has to share the current from the power supply equally. Thanks a lot guys.

Just one more quick question: roughly what level of current in excess of the safe operating limit for an average opamp is required to destroy that opamp? I want to have a rough idea how far I can push it.
Kolloi, on 25v rails an LM3886 IC is easily good for 30W into 8 ohms, the parts count is low and the sound quality is rather good. This sounds like it fits your needs exactly :)

You can get LM3886 for about £5 each from Farnell, or maybe a bit cheaper on eBay if you can find them.
I've been read into this area for a while.
From what I've seen, the main idea of this topology lies in:
1. Extract "the best" 5w-10w
2. Gain of 1-5 possible
3. Extract ultra low distortion at high frequency

By using Composite AMP design with CFA high output OPAMP.
For a typical design , check the ADA4637+ADA4780 design in ADA4780 datasheet.
This kind of CFA inside a loop yields literally no phase shift at all.

However, the device must be chosen carefully.
The cost range can be 2-3 times based on the parts you pick.
You might look at this article.
The 5532 OpAmplifier, part 1
There is also a thread on this site about the amp.

I've read that already.
IMO, massive general purpose OPAMP is too costly.
Those high current CFA can easily achieve 10 times current output while the cost of 1 amp is not much higher than 5532.
Also this can reduce the parts count greatly and therefore reduce PCB size and overall cost.

However with high current CFA, thermal design is the key.
They need to be heat sinked.
Founder of XSA-Labs
Joined 2012
Paid Member
You might like this: 40 x OPA1602’s


Building a massively parallel op amp power amplifier
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.