Basic Question on Reading Datasheets

[schrectacular]

Hi all,

Last night I fried a couple TA7222s and now I guess I need help. When I'm building an amp and looking at an amp chip's datasheet, how do I calculate how much current it will need?

So take the looking at an LM386N-1, the datasheet https://www.futurlec.com/Linear/LM386N-1.shtml says it can take 9v and that it will output at most 1.25W. If I know I want to drive a 4ohm speaker, then how do I go about calculating how many amps to feed the chip at different voltage/ohm combinations? Isn't 9v into 4ohms mean 2.25A at 20.25w? But that can't be right...

I tried searching around the 'net and this forum but I must be searching for the wrong terms. My guess is it's rather simple but I just don't get it yet. If anyone could help clue me in or point me toward applicable readings I would be very thankful.

Oh and I fried the TA7222s with a 12v, 5A power supply. I read this datasheet http://pdf.datasheetcatalog.com/datasheets/400/502310_DS.pdf and thought, "peak output current is 4.5A, so I'll plug 5A into it and it will have all it needs!". I had a working circuit on the breadboard which worked at 9v 100ma, 9v 500ma, 12v 100ma just fine. I plugged the new power supply in and the sound was really faint like it wasn't connected right. Messed around reverting to previous versions, then switched out to the second chip. Worked fine with old power supplies! Great, chip must be busted! Plugged the 5A supply to the breadboard and "ZZzzzt!! - pooof..." a zapping noise and then a tiny little plume of smoke emanated from the chip. Actually now telling the story I wonder if the power supply I had inverted tip and I plugged it in backwards... Would that fry the chip?

[11-90-an]

Please take my advice with a pinch of salt...
Rated amps in power supplies state the maximum amout of current it can supply. So basically, using a higher amp PSU shouldn't fry your chip. BUT as far as i know, (which isn't much ) the chip can be fried because it ends up as power hungry and consumes the extra current and burn out.

I think you could have accidentally swapped polarity... have you tried checking with multimeter?

[anotherjim]

You may have things slightly the wrong way around. Most power amp types don't set or limit the power output -  they try to do what they are told within the limits of power supply and load.

It's a technical "mistake" to use DC formula to calculate audio amp power, but it's a useful mistake nonetheless. The DC figure will be an absolute maximum if a fault occurs or the amp gets told to make full swing square waves. The full square wave is nearly equal to the DC figures.  Don't forget single supply amps bias the output voltage 1/2 of supply, so it isn't the supply voltage you need to calculate with.

Calc for 9v 4ohm is 4.5x4.5/4=5W
For audio, you need RMS, not Peak, so multiply by 0.7.
3.5W It's still too much.
Will the output voltage swing be 4.5V? Probably not. 4.5V assumes the chip output devices can switch to near-zero output impedance. Real ones won't. Datasheet might be hiding the output swing figures. Power required will still be too high.
Is that chip capable of driving 4ohm with 9v power? Try 6v or raise R to 8ohm. 6v with 4ohm, a 386 is chillin' at 1.5W if working flat out. At 9v it will just about handle 8ohm.

Even if you have reasonable power supply "power" available for the datasheet requirements, it can go haywire if stability isn't assured - fast ringing/oscillation that is too high frequency to be audible can cause power dissipation in the chip to rise too much. The TA7222 data doesn't show a proper circuit. What is the circuit you have them in?

Have to say I just don't trust these chip amps that need extra "bootstrap" cap connections. All the burnt-out commercial ones I've seen are this type. More recent chip amps with simpler requirements seem almost bomb proof.

[schrectacular]

Ok, that's very helpful.

To answer your question, I believe it was this: https://www.circuitstoday.com/6w-amplifier-using-ta7222ap, but I burned both the chips I have and won't be getting any more, so that's that for those.

I do have a pair of TDA2030s, but I haven't put them on a board yet. I was looking at trying out the General Guitar Gadgets circuit http://www.generalguitargadgets.com/effects-projects/amps/tda2030-amp/, but, among other things, I don't have a bipolar power supply. Plenty of other projects to keep me busy... but I have a 6" 15w 4ohm proper Jensen amp speaker and a leftover 6" 2ohm car speaker that I eventually want to turn into amps. I have LM386N-1, LM386N-3, and some TDA2822's. The internet is rife with circuit examples for these so I should be fine.

So what happens when my amp doesn't have enough current? Will I be fine building little silly practice amps with normal pedal plugs (100mA)? Do I even need to worry about this?

[PRR]

...............Isn't 9v into 4ohms mean 2.25A at 20.25w?....

That's DC. It also seems to assume all heat is in the 4 Ohms, so the chip must be cold.

Audio amps swing both ways. Which means on average the load voltage is 1/2 and the load heat is 1/4. (The chip heat is rather more complicated...)

The real factor from supply voltage to load Sine power is more like 1/6th.

[anotherjim]

The TDA2030 is a fairly good amp, and easily used single supply...
https://www.elprocus.com/subwoofer-amplifier-circuit-using-ic-tda2030/

If you only have 100mA PSU, you won't be able to get much power, whatever chip you have.
But with 9v and low ohm speakers, you will always have trouble - either the puny PSU gets overheated or the chip does.
If you insist on using 4ohm or less, you will have to find a lower voltage PSU or get more speakers and wire them in series.
More speakers is good - if you can get the total coil resistance up to 8ohm or more, it will lower the current consumption, run the chip cooler and yet give more apparent volume thanks to having more speaker cone area moving air.

[schrectacular]

Ok, this is again very helpful. My PSU has a 9v 500ma option, so I can run stuff on that. Not opposed to purchasing a lower voltage supply if it would be helpful. I guess I should buy a proper 8ohm speaker while I'm at it...

I have a stripped-down "Ruby" on the the breadboard right now and the lm386 does get pretty hot on 12v... I have a little heat sink on the top so it's probably being more forgiving of me pushing the limits.

I have a couple other speakers laying around, the 2 ohm and a tiny 1.5" from a portable bluetooth giveaway. I'll try wiring them all up.

Oh and thanks for the link to the TDA schematic, that is bookmarked and I'll be sure to report back when I get to it.

[schrectacular]

Oooh even at 6ohms on the 500ma it sounds so nice. Especially with some oompf in front. I like it better without the gain engaged, maybe it's because it's on the breadboard and it picks up a lot of noise, but I think it sounds better overdriven by a booster pedal.

I guess my question remains, which is how should I back-of-the-envelope calculate what sort of amperage I want to have given an amplifier chip + schematic, a known number of volts, and a known number of ohms?

[PRR]

For full power sine-wave testing, 8 ohm load, assume the amplifier acts-like 50 Ohms to the DC supply.
9V/50r = 0.18 Amps

Proportional to load impedance: for 4 Ohms, figure 25 Ohms on the DC supply: 9V/25r = 0.36A

This assumes there is enough capacitance on the rail to carry the peak current.

Clean speech/music will average MUCH less (but still peak high when cranked).

[anotherjim]

A tip for the small DIL package chip amps. Although it's nice to have chip sockets for easy changing, in this case, the chip pins do provide some path for heat to dissipate into. So where you have good confidence in the circuit, solder them in the board. With stripboard builds, each track becomes a heatsink fin.