# Is there a difference between specialized audio amplifier and general op-amp silicon?

I've been doing some research into creating a DIY audio amplifier from line-level inputs to speaker level power. I found a guide centered around the LM386 Low Voltage Audio Power Amplifier on YouTube here.

I have a fairly large collection of DIP package TTL logic and alike that I inherited from my father, which I am hoping to make use of in this project. I did some research, and it appears that the LM386 is just a specialized op-amp with desirable characteristics for audio signal amplification.

I located an LM324N op-amp in my collection, which looks to my incredibly untrained eye to have similar characteristics to the LM386. However, seeing how I have a rudimentary understanding of circuit design, it is highly likely that I'm missing some critical information that would allow me to see that the LM324N wouldn't work for my application. Am I missing something, or would the LM324N work fine?

• The 386 has built-in biasing circuitry, if I remember correctly, and can provide higher output current than the average op amp. Oct 17, 2018 at 0:09
• Specifically the LM324 isn't good for audio by default electronics.stackexchange.com/questions/341843/… Oct 17, 2018 at 0:30
• I am not sure about the validity of this statement, but the video mentioned that 1W is often more than enough to drive most medium-small speaker elements. The target device I have in mind is an old shelf speaker from a Sharp tape player/CD changer in our house. Oct 17, 2018 at 0:32
• Is such a 1W application within the capabilities of the LM324N? I believe the speaker was rated for an 8 ohm impedence. Oct 17, 2018 at 0:33
• I think an audio amplifier is a current amplifier, and an opamp is a voltage amplifier. Oct 17, 2018 at 1:42

The first clue about the LM324N comes from the datasheet title where it says "Low Power" in it. Your second clue about the LM324N is that it is a quad provided in a 14-pin TSSOP and plastic DIP. This will be a thermal resistance of about $$\90-100\:\frac{^\circ\text{C}}{W}\$$. Even if there was only $$\250\:\text{mW}\$$ dissipation per section, the part would be very hot. (Granted, the LM380 also comes in a 14-pin plastic DIP with similar thermal resistance. So it is only a clue. But also note there is only one section in the LM380.)

The actual answer comes from the following table. (The last column is for the LM324 device):

Here, you can see it only sinks or sources perhaps a few dozens of milliamps. This pretty much eliminates any idea that this is, or can be used as, an audio amplifier of any kind. (Note that the microamps shown in that snapshot from the datasheet you provided, as a sinking current compliance, is likely an error in the datasheet as glen_geek points out in a comment below this answer.)

The schematic itself on first blush might not be as clear, since it has a lot of the elements of a power amplifier. However, it also provides some added clues. Let's look at it in a side by side comparison with the LM380:

On the left is the LM324 schematic. You can see the green-circled current sources and sinks shown and you can easily notice the very small values for these currents. This suggests low-power. Especially the $$\100\:\mu\text{A}\$$ one that is providing base drive sourcing for the Darlington pair on the upper quadrant of the output, $$\Q_5\$$ and $$\Q_6\$$. Given something like an effective useful combined $$\\beta\$$ of about 500 or so, you can't expect more than about $$\50\:\text{mA}\$$ for sourcing. (Though $$\Q_7\$$ and $$\R_\text{SC}\$$ provide some kind of output current limiting circuit, too.)

On the right side, you can spot a few clues about it NOT being a general purpose opamp. $$\R_4\$$ and $$\R_5\$$ would be "parasitic" for the normal use of an opamp. You can see that they are not present on the left side, where the differential amplifier pair (Darlington structured) directly expose their bases to the (+) and (-) input pins. But they are added in the LM380 to provide some built-in DC path to ground and biasing. But most importantly, you can see $$\R_2\$$ providing direct negative feedback (NFB) from the output backwards to the input stage, where $$\R_3\$$ plays an additional role, as well. You don't build NFB like this into a general purpose opamp. You expect a designer to do that, externally. So you leave that option open to the designer in those cases. But for an audio amplifier IC? There, you can provide NFB -- especially in this case where the gain is fixed at $$\34\:\text{dB}\$$!

So there are a few clues:

1. "Low Power" in the title description for the LM324 vs "2.5W Audio Power Amplifier" in the title description for the LM380.
2. Built-in NFB for the LM380 and no built-in NFB in the LM324.
3. Fixed gain for the LM380. No fixed gain for the LM324.
4. Very low output current compliance specifications for the LM324, but where the LM380 claims up to $$\1.3\:\text{A}\$$ for the output on the first page of its datasheet.
5. Built-in DC bias path for the (+) and (-) inputs of the LM380 vs no such added bias path in the LM324.
6. Very low recombination currents being provided in the LM324, for use as base drive at the output stage (if nowhere else) also strongly suggests further limitations with respect to its being used as an audio output amplifier without added external circuitry.

Of course, the LM324 can be used as part of a larger external circuit that may be a useful audio amplifier. But it will at the very least require some kind of external output stage added to it in order to provide much higher output currents. Still, even then this particular LM324 opamp type has such small current sink compliances that I'd be very careful to only use it in a design where the sourcing capacity was in active use and where the sinking capacity isn't important (since it effectively has no ability to sink current.)

I also discuss the LM380 here and here, where you can find some expanded discussions about the LM380 audio amplifier IC.

There are similar substructures in the above schematics, though. So it's not always entirely obvious at a single glance. But if you see NFB in the schematic, that's one thing that's almost always a give-away. Of course, you could also just read the text in the schematic. If it is designed for audio, it's likely to specify operation into an $$\8\:\Omega\$$ load somewhere and to discuss "power" and likely even to have some discussion about how to design the "copper pour" that will help handle dissipated power from the IC. And general purpose opamps rarely specify an output current compliance much higher than about $$\25-35\:\text{mA}\$$. It's just not usually needed for general purpose use and the focus is often more on other important benefits.

• Thank you very much, @jonk. A little over my head, but I get the gist. Oct 17, 2018 at 3:30
• @ifconfig You can see a case where a low power opamp (general purpose) is surrounded with added components to turn it into an audio amplifier here. Just as an example so that you know what else might be needed.
– jonk
Oct 17, 2018 at 3:36
• @Jonk Checked a few LM324 data sheets. That "sink current" spec of 20 or 50 uA looks like a typo. Can't be that low. Another sheet says 20 mA which seems much more reasonable. Oct 17, 2018 at 3:47
• @glen_geek Thanks!! Good catch. (Yes, it did seem a little low and I didn't go check other datasheets and probably should have done so.) I'll update my response. Thanks.
– jonk
Oct 17, 2018 at 4:00

A simple answer is to understand that an audio amp tries to be a true ‘voltage source’ having 0 ohm source and unlimited current, yet none are.

All Op Amp have active current limiting on the output, for short circuit protection. With 8 ohm load at +/-25 mA that limits you to +/-200mV or +/-2mW. Thus a simple 1W power amp uses a complementary “Emitter Follower” output with negative feedback to the speaker. This becomes a 3rd stage to the Quasi-complimentary 2 stage (like Darlington) outputs already in Op Amps. Bigger PA’s may have 4 stages with transistors ganged on heatsinks in the final stage to spread the heat. These are all voltage buffers with current gain.

The power lost must consider the thermal resistance of the transistors and heat rise in ‘C/W for Rja may not be necessarily short circuit proof , but could be with extra work.