Reducing voltage of TL494 based supply?

[Eva]

The supply is unregulated, the output inductors are part of a pi filter, so it's not recommended to use duty cycle regulation to reduce output voltage (unless you turn this pi filter into a conventional LC output filter, but then bigger inductors would be required). According to the pictures, the turns ratio of the transformers seem to be 16:6 (correct me if I'm wrong) so the nominal output voltage should be 12*16/6-.5=31.5V (and up to 14.4*16/6-.5=37.9V) per rail. Note that the voltage increases a bit when there is no load and that's probably what you have been experiencing (as you discarded everything after the PSU). Add bleeder resistors to draw 1W or so from each rail and consider changing transformer turn ratios if the idle voltage is still too high. Anyway, note that the voltage will probably drop to +/-30V or less under heavy load, and on the other hand, most amplifier modules can handle higher voltages when they are asked to output little power.

[theChris]

eva, why would a pi filter be used in a buck-derived regulated converter? Is it just assumed that the extra filtering prevents the voltage on the first cap from changing too much? otherwise it would seem that you'd be turning on into a pure capacitive load that may have droped significantly in voltage.

anon, paralleling supplies may not be the hard part --getting them to share load is. an imbalence in the references might prevent current flow. if one PSU attempts to force 30V, and the other 30.2V, then one PSU will force 30.2V and the other PSU will idle becasue that is the best action it can take to reduce voltage -- nothing, there is no way for it to force the voltage negative, the voltage just has to fall on its own, which it won't. the supplies can be forced to share load using current feedback in the outer-control loop.

[theAnonymous1]

Eva, you are right on the money; I just counted the turns and it is indeed 16:6.

I now see the light.

Paralleling is not the answer, and neither is adjusting the "duty cycle regulation(?)". I thought adjusting R501 was the answer at first because the one supply did seem to be regulated, but at the cost of the other supply not working at all. Paralleling the two in this state was just masking the non functioning supply.

I put R501 back at 33k and did some load tests. With 13.8v input and a 1k load across each rail the output was around 38.5v. With a 20R load on the outputs they went down to 36v. I don't want to try and draw more current without a heatsink attached.

Well, I learned some valuable lessons here. I should always listen to my more knowledgeable peers, and when someone tells me to rewind a transformer, I should just do it.

I will rewind with 14 turns on the secondary.

[theAnonymous1]

OOOOOOOOOUUUUUUCCCCHHHH!!!!!!!

Man, my fingers HURT! I can't imagine having to do that more than once in a lifetime.

Anyways, I rewound the transformers and all is well. I took too many turns off so the voltage is lower than I wanted, but it's workable. The class-d amps it is going to power needs a minimum of +-20v.

It was a royal PITA soldering and de-soldering the transformers. I was even using a 180w solder gun.

Results....

13.8v In = 29.6v (no load)
13.8v In = 27.8v (20R load)

12.5v In = 26.7v (no load)
12.5v In = 24v (20R load)
note: battery voltage dropped closer to 12v with the 20R load.

Well thank you all for the help. I hope I never have to do that again as long as I live.

[luka]

Hi

If you use Class D why did you even lower voltage, can't it work at +/-40?

Yesterday I have redone my trafo too, used 2x two wires for each sec, each wire made of 4x 0.5mm. So I had first to cut to size 16 wires of 0.5mm. If you don't know what I made, I made 4 litz wires, each sec with 2 of those. Result? F.... beautiful No heating of wire, no saging of output voltage...worth the energy

Btw, nice job done and you didn't use 20 diodes

[theAnonymous1]

Luka, the amps are the TA2022 based AMP1-B from 41hz.com. Maximum voltage is +-35v. Current limiting kicks in at >7A, so even +-35v is too high for a 4R load.

I was hoping I could get the output of the supply closer to +-32v @13.8v input. I'll live with what I have now. No way am I redoing the trafos again.

[luka]

Hi

A ok then, even if you won't listen that loud, your voltages will be more then enough. Great job done again...

[theAnonymous1]

I'm sure it will be leaps and bounds better than whatever power my head unit is putting out, not to mention quality.

Now all I need to do is repaint the heatsink and do all the wiring.