L272A based guitar amplifier?

[Brisance]

While component shopping I picked up the only dual op-amps they had in stock, the T272A. Now checking the datasheet, I found it can handle 0.7A inductive loads, now that made me think of building a simple Op-amp amplifier with it,  feasible? Has anyone done anything with it?

[tonyharker]

Dont know this one have you got a link to the datasheet?

[PRR]

I do not think that is a right part number.

[Brisance]

My mistake, its a L272, remembered wrong 

https://www.fairchildsemi.com/datasheets/L2/L272.pdf

[R.G.]

Ah, time for some thermodynamics. Well, very little.

For that chip, the maximum power is specified at 1W, and the thermal resistance from junction ambient as 100C/W, and the maximum junction temperature as 150C.  All those are consistent, since when you have 1W being generated inside, the heat causes 100C temperature difference getting out, leading to a 150C chip temp in a 50C external environment.

For a class B power amp, assuming all of the heat comes from the output devices, the maximum Pd is

Pd = Vs²/20*Rl

We know Pd in this case is 1W, and Vs can be anything less than 40V, so we can pick Rl by the size of Vs: Rl (for max power out) = Vs²/20*1W

So for Vs = 40V, the Rl that lets you have the max power out is
Rl = 40*40/20*1 = 80 ohms. 

Not a good speaker load,

OK, if we have 8 ohms, what Vs makes the best use of our 1W?

Vs² = Pd*20*Rl = 1W*20*8 = 160, and Vs = 12V.  OK, now we're getting somewhere. We can use 12V and dissipate 1W while driving 8 ohms.

But what's the max audio power we can get out?

Well, the equations show that the max pd for a class B amp is at an output p-p voltage about 0.637 times the power supply total. So we can generate 12V*0.637 = 7.644V p-p, 3.8V peak, or 2.7Vrms, and then Pout = 2.7²/8 = 0.91W into 8 ohms at the max temperature range of the chip.

As it happens, the chip actually cools off when you go over this to the full output. The chip datasheet says it can do 22.5V peak to peak with a 24V supply and a 0.5A output; guessing at the 12V range, it may do 10.5V peak to peak at 12V, and that's about 1.7W out into 8 ohms. 

Realistically, you're getting maybe 1.5W of audio out of it when it's really cooking, and you need to use a 12V supply to drive 8 ohms. Changing the power supply means you have to change the loading to keep the maximum power coming out.

Comparing to the ubiquitous LM386, it looks like it'll do about twice the audio power out if you use it right.

[Brisance]

Oh well, R.G.:

Seemed like a great idea, but oh well when building me a small amp, will probably resort to IRFZ44 power mosfets.

[R.G.]

That was only step one.

The issue driving all that was the power generated inside the chip, which is of course proportional to I²R. The chip can work with lots of voltage, but overheats on speaker loads.

The thermodynamics say that you can still get good voltage output, you just can't use as much current as you need for big voltages. Or that you can't let those currents flow inside the chip.

So you can move the currents outside the chip by using an external current buffer - like a Class AB output stage. This makes the currents flow in external power devices with thermal resistances of perhaps 2C/W to 5C/W for TO-220 devices.  There are many such circuits on the net; look up "boosted opamp".

[Brisance]

Ah, as long as I need to use some other stage to actually drive the speaker, I'll ditch the idea and use them as regular dual opamps, altho they do not have standard pinouts. Just the "high power, designed for low impedance loads" part in the datasheet made me think about using it for amps.

[PRR]

While there are 0.7A and 1A "max" numbers on the sheet, performance is only specified to 0.5A.

The LM386 can dump practically 0.5A. And may be easier to use. Especially in the far future (2019?) when the obscure L272 is forgotten but the LM386 won't go away.

L272 gain-bandwidth is stunningly mediocre. 1/3rd MHz. I really think (and the app-list suggests) this is aimed at "mechanical" loads. The little motors that spin CDs and VCRs, that scan the lenses up and down. Yes, a loudspeaker is in that class but we like it to move well at far faster speeds than typical servos.

[Brisance]

Thanks again for wonderful input I picked them up because they were the only dual opamps in stock and the TL072s I ordered aint coming for another week. Guess I'll find some use for them.

[Brisance]

I did find an use for them though, I will use em to create virtual ground