# Question about power supply requirements for three paralleled TDA7293 audio amplifiers

#### J-Erbes

Joined Jun 16, 2019
3
I need some help in specifying power supply volts/amps needed to run a TDA7293-based 250W RMS amplifier board consisting of three TDA7293 modules connected in parallel (described on the TDA7293 datasheet as a 'modular application').

A TDA7293 amplifier module is constrained in four different ways (from the datasheet);
1. Supply voltage is +/- 50V maximum
2. Built-in current limitation of 6.5A maximum
3. Module heat dissipation of 50 Watts per module (150W total)
4. Additionally, in modular (parallel) mode, the TDA7293 must have the supply voltage derated so that the power heat dissipation per device doesn't exceed 50 watts. The supply voltage derating depends on load resistance. The supply voltage has to be less than +/- 50 V for a 6-ohm or greater load, less than +/- 42 V for a 4-ohm load, less than +/- 30 V for a 2-ohm load, or less than +/- 21 V for a 1-ohm load.

I'm going to be driving a 4-ohm speaker (rated 300 W RMS), so my supply voltage will have to be less than +/- 42 V.

I have two 120 V power supplies on hand at the moment.
One gives +/- 36 VDC @ 3.0 Amps (216 VA)
The other is +/- 36 VDC @ 6.0 Amps (432 VA)

My question is:
Do you add up the maximum amplifier output (250 watts) plus the maximum power heat dissipation (150 watts) to determine the minimum VA requirements of the power supply?, so that 250 + 150 = 400 watts minimum?

If not, how are you supposed calculate the size of a power supply?

#### ronsimpson

Joined Oct 7, 2019
1,125
I have a very strange way to think about the power supply question.
The TDA7293 has a current limit at 6.5A. ( it also said 10A absolute max) I think it will do 6.5A and not the 10A.
If you input a large sign wave that draws 6.5A pk then the average would be 4.5amps.
That gives me the idea that your 3.0A supply could not do one channel but the 6.0A would be a little over sized.
Music is not like a 6.5A signal (average) but the power will be much down from that. The peaks of music will be at full power but the average is down but I don't know by how much.

#### Audioguru again

Joined Oct 21, 2019
2,449
That is correct, a high power linear amplifier like you want produces massive heating. The output power plus the heating power must be produced by the power supply. Modern amplifiers use class-D for much less heating.

You said a +/- 36V supply. The output into a 4 ohm speaker will be less, probably +/- 30V or less which is 21.2V RMS. Then the power into a 4 ohm speaker is 21.2V squared/4= 112.5W or less. The heating will be about 56W or less.

Amplifiers are supposed to produce no distortion, certainly not 10% distortion shown on some of the graphs (but acid rock plays lots of distortion).

Why did you think the output will be 250W? Then the signal on the 4 ohm speaker will need to be 31.6V RMS which is 89.5V peak-to-peak. There is no way the amplifier will produce 89.5Vp-p when it has a 72V supply. Paralleling the amplifiers allows it to drive a load that is 2 ohms or less for higher power.

#### J-Erbes

Joined Jun 16, 2019
3
Why did you think the output will be 250W? Then the signal on the 4 ohm speaker will need to be 31.6V RMS which is 89.5V peak-to-peak. There is no way the amplifier will produce 89.5Vp-p when it has a 72V supply. Paralleling the amplifiers allows it to drive a load that is 2 ohms or less for higher power.
This was based on the board maker's claim of 250W, which is probably based upon the manufacturer's datasheet.

The attached graph from the TDA7293 datasheet shows the RMS/Distortion output of a single TDA7293 driving a 4-ohm load with a Vs supply voltage of +/- 29VDC. At a frequency of 1KHz, the amplifier produces 70W RMS @ 0.009% THD, 75W RMS @ 0.1% THD, 80W RMS @ ~0.7% THD, and 90W RMS @ ~6% THD. The numbers driving an 8-ohm load are even better.

So, if you believe the manufacturer's datasheet, it is possible that three of the TDA7293s in parallel could produce around 250W RMS @ 1% THD.

For my application, in a bass guitar amplifier, this is more than acceptable. I can live with a lot more distortion than 1%.
If it's the right kind of distortion, more of it often improves the sound.

In your defense, I do understand the calculations you have presented, and I do scratch my head trying to figure out how they can get 90W RMS out of one of these modules into a 4-ohm load using a +/- 29VDC supply.

I wish someone would explain it to me.

#### Audioguru again

Joined Oct 21, 2019
2,449
Look at the datasheet. With a plus and minus 29V supply and a 4 ohms speaker the distortion is low when the output power is a little less than 72W. It begins clipping (making clipping distortion) a little above 72W and produces a horrible sounding 10% distortion at an output of almost 100W. Then the fundamental sound is 72W and the distortion sound is 28W.
Yeah, electric guitars are always producing lots of distortion. But not just clipping distortion. They want distortion at any output power.

250W into 4 ohms is a signal at the 4 ohms speaker that is 32V RMS which is 32 x 2.828= 90.5V peak-to-peak. Paralleling amplifiers does not increase the voltage, instead it increases the available current so that a speaker that is 1.33 ohms will produce 250W.

250W into 1.33 ohms is a signal at the 1.33 ohms speaker that is 18.4V RMS which is 18.4V x 2.828= 52V p-p. A power supply that is about plus and minus 29V might work. Where will you find a speaker that is only 1.33 ohms? Three 4 ohms speakers in parallel make 1.33 ohms.

#### J-Erbes

Joined Jun 16, 2019
3

Another fellow did an extensive test of a single TDA7293 analyzed on a Rigol scope in FFT mode and found that the TDA7293 just started clipping at ~86% of the VDC supply voltage, so his +/-32VDC supply was clipping at +/-27.1VDC or 19.2V RMS into a 4 ohm load, giving him a clean 92.2 watts RMS.

(See
)

One of the TDA7293 datasheet graphs shows that in parallel mode, the supply voltage has to be derated to less than ~43 volts with a 4 ohm load.
If I were to use the maximum possible +/-42VDC supply, that would supply about +/-36VDC or 25.5 V RMS into a 4 ohm load.

Squaring that and dividing by the load looks like I might be able to get a clean 163 watts RMS into 4 ohms right at the verge of clipping (as long as I have enough heat sink).

Does that sound about right to you?