How to determine necessary resistor wattage

[mth5044]

For example, in this power amp circuit:



How do you know the necessary resistor wattage? I am aware of power's relation to amps and volts, but I'm not sure how that would be applicable in this situation. In the forum amp, Tiny Giant, all resistors are 1/4W that a related to the TDA power amp chip, but I have seen other power amp circuits that require 1W or higher resistors. Any clues? Thanks.

[R.G.]

I am aware of power's relation to amps and volts, but I'm not sure how that would be applicable in this situation.

That's all there is.

P = I * V, where both I and V are the RMS values. Unless you like elementary calculus, you'll be very interested in two approximations. One is that for DC, both V and I in the RMS sense are the DC values. For sine wave AC, the RMS value is the peak value divided by the square root of 2, 1.414...

If you have a mixed AC and DC voltage and/or current, if the DC is less than 10% of the AC, ignore the DC and assume it's all AC. If the AC is less than 10% of the DC, ignore the AC and act like it's all DC.  If you can't ignore one or the other, do searches for equations giving you reasonable approximations. In most cases, pure AC or pure DC can be used for approximating. In this case, you can ignore the DC on R7 and R8, assuming that what's across them is the full RMS voltage coming out of the amplifiers.  Likewise for R6 - it's being driven differentially, and has the full AC rms that the speaker gets across it. R2, 3, 4 and 5 will have the worst case power dissipation when they're not being driven, so assume half of Vs is across R2/R3, and also R4/R5.

When you get those power numbers, double them for derating, and pick the nearest nominal power rating.

[PRR]

> How do you know the necessary resistor wattage?

What R.G. tells you is, of course, true.

Here's a dummy-way. Even though you are no dummy, it is always good to have a cross-check against errors.

TDA datasheet at ST Microelectronics

Figure 4 Page 7 is a ready-to-try PCB layout *with* parts outlines.

Knowing that the chip is 20mm, we can estimate that the lay-down resistors are 8mm long. A flip through Mouser/DigiKey listings for 1/8 1/4 1/2 Watt resistors will suggest *about* what wattage they intended. I'd start with low-price 1/4W resistors, because I _think_ 1/2W are bigger than 8mm, and 1/8 watt are no cheaper (and too small for my fat fingers). Or measure resistors on hand.

Less-dummy thoughts:

At 14V supply, your R5 10 ohms ""could"" take the full 14V and dissipate 14^2/10= 20 Watts. As much as the speaker??

No, because there's no path from +14V through R5 which would pass more than R4. And 1K in series with 10r, most of the heat will be in the 1K. So figure 14V^2/1K, that's 0.2W in R4, less in R5. On that thought, 1/4 is just-barely enuff, you oughta go 1/2W.

But if you understand how the amp works, you know that pin 8 never lingers at +14V, but closer to half of that (for symmetrical swing). Now 7V^2/1K is just 0.05W. A 1/4W is plenty.

R4 also has audio swing. Assuming almost 14V peak-to-peak Sine, that's another 5V RMS or 0.025W. Simple addition isn't quite right for DC+AC, but 0.05W DC plus 0.025W AC is so much less than 1/4W that we will probably be safe.

Uh, in car-sound and also in guitar, it is common for any soundz to be cranked-UP until the output looks like square-waves. Then the AC power can be taken from the Peak voltage, 7V, or 0.05W of AC power. This still looks OK for a 1/4W part.

The only utterly unknown is R7 R8. We see they could get a full 7V peak sine-wave, though only over 1r*0.1uFd or 1.7_MHz_. If the chip would actually do that, it comes to 49 Watts!! In fact there should never be anything like 1.7MHz coming out. But the transistors are inclined to try (read a dozen books on transistor audio amplifier complications). The 1r+0.1u tend to suppress this tendency, and there really should be "no" power in the 1r resistors. The main thing is: they should not be wire-wound. Simple carbon is fine.

[mth5044]

Thank you both for the very helpful info! Very reminiscent of my old physics class, which I enjoyed. Thanks.

[Jdansti]

Thanks for asking the question. I was wondering the same thing looking at the TG amp schematic.