Can i use a computer power supply to power audio amplifiers?

[kartino]

Dear Kambule,

No, it is not change the 3.3V and 5V to 12V. But replace wire size for overal secondary winding with same number of turn and direction and pin connections as existing. Hence the new modification is still have 3.3V and 5V and 12V but at low amps, it may approx 2A for self control and reference.

The desired power for amps come from new windings that have no connection with existings, totally separate winding. We calculate the desired refer to existing secondary number of turn.

Please read carefully, not so fast buddy,

Regards,
Kartino

[kartino]

Hi, Cpemma,

About http://www.qrp4u.de/docs/en/powersupply/ I have read that page several time. It is more complex even just produce plus rail only.

Until now I didn't find other modification for +/- rails that suitable for most amps. Hence what I did is the simplest way because no need to calculate reference, no need switching knowledge, no need rearrange control, still get +/- 12V for your fan, DC and short amps protections, to feed your pre amp, tone control etc.

For the same power level (200-300W), I am wondered if people coose that complex way.

Best regards,
kartino

[Eva]

Originally posted by kambule


Kartino,

I want to modify a ATX PSU to supply +/- 24Volts.
I see the post from Kartino of how to modify the transformer, but I need to be more specific. Refering to ATX schematics from EVA, I think :
(1) Rewind T3 secondary to twice more turns.
(2) Use the existing rectifier diodes.

As for the feedback (voltage stabilization) can I just keep R25 and raise the value to 56K ??

If 5volt is not required, can I just remove R26 ??

The 12V feedback path also goes to a diode D. Wot is the function of this path ?? Any modification necessary ??

Wot should I do with the 3.3v circuits, if I dont need it ?? Should I just remove the parts after L6 and D29 ??

thanks in advance for your advice.

kb

(1) Right. Turn count of the secondaries has to be almost doubled, altough there is some margin for improving the output current rating by using less secondary turns at the expense of increasing the minimum operating mains voltage required.

(2) The existing diodes will be probably schottkys rated at 40 volts, but for 24V output 100V diodes are required because they are required to block more than three times the output voltage. Standard ultra-fast ones like MURxxx series from On-Semi are fine). Please don't try to use the existing diodes if you add more turns to the secondaries, as this may cause a catastrophic failure.

(3) To obtain 24V from the feedback network shown in the ATX schematic, you can double the value of R25 and remove R26 as you have mentioned. Note that due to the regulated nature of these PSUs and due to the gentle turn ratio usually employed in the main transformer, changing the feedback network may allow to get up to 20V output without adding more turns (again, this depends on the minimum expected operating mains voltage).

(4) The components related to +3.3V, +5V, -5V and -12V outputs may be safely removed when they are not required

(5) Diode "D" is very important because it provides approx. 16V to 24V DC power to the control circuit as soon as the PSU starts oscillating. Some ATX power supplies may work without it if the standby circuit already provides enough voltage (see D30), otherwise a little auxiliary winding has to be added in order to continue powering the control circuit with the same voltage as before.

In AT power supplies this diode (D19 in AT schematic) is even more critical, because the primary side of these circuits is designed to self-oscillate and provide start-up power to the control circuit in the secondary side through diode "D". Since the self-oscillation frequency is almost always too low, the control circuit is required to turn off the primary side switches soon after it has received each power pulse in order to prevent transformer saturation, while each pulse increases the voltage across C19 until the TL494 starts working. Never remove D19 in AT PSUs, also, if the transformer is modified, a little auxiliary winding will be required in order to feed D19 and C19 with the same voltage as before (16V to 24V).

EDIT: The 13.8V PSU also employs self-oscillating startup, see D17 and C32.

[kambule]

EVA,

Eva,

Thanks for your helpful reply.

I need DC output of +/- 24v 8A.

1) please look at the attached diagram. The blur-ed component will be removed.

2) understand that the rectified diode rating to be 3xPIV.

3) a 12V zener is added to stabilized/limit the supply to the control circuit. As C21 has only 16v rating, a 12v zener is chosen.

4) my mains power socket is 230v. If I could find a Power supply with autoranging from 240 to 100v, I might be able to get +24v DC output by adjusting R25. I dont even have to rewind the transformers (If I need more current from -24v, rewinding might be necessary). The -12v rectifier diodes need to be upgraded as well to withstand the current.

Here is the orginal specification of a 220W ATX PSU (2theMax):
+5V - 13A
+3.3V - 17A
+12V - 15A
-12V - 0.3A
5VSB - 2.0A

5) Is the PS-ON pin for ON/OFF the PSU ?? Is a permanent ground required as long as the PSU is ON (I mean it is NOT a toggle type).

Grateful if could give some more advice.

kb

[kambule]

sorry ...here is the modified atx diagram

[Eva]

The modifications look right to me. However, don't trust so much that schematic, it's not drawn by me, I just found it in some webpage and it may have errors (indeed the AT has some mistaken connections). You should look at the particular schematic and component values of your PSU instead.

Concerning the control circuit, it may work better powered from +24V, but not more than that. That's because Q3 and Q4 control the (critical) turn-off process of the main switches and they require gentle base drive through R13 and R14 in order to do it properly, but these resistors are usually of a quite high value due to dissipation constraints. If you plan to put considerable stress on the main switches, you may consider replacing Q3 and Q4 by BD139 or some *fast* transistor rated at 1A and reducing the values of R13 and R14 in order to provide 50mA base drive or so.

[cpemma]

Originally posted by kambule
5) Is the PS-ON pin for ON/OFF the PSU ?? Is a permanent ground required as long as the PSU is ON (I mean it is NOT a toggle type).

The PS_ON (green wire) turns the psu on if low (grounded) so a light-duty toggle (not momentary) switch between green & black can be used for on/off control.

Note ATX supplies need a minimum load for regulation; this is commonly done with a 1A loading on the +5V line.

[destroyer X]

supplies.

Thank you all, by the gentle cooperation.

regards,

Carlos

[kambule]

Eva, cpemma, many thanks.

I shall try someday when I can find a cheap psu.


kb

[kartino]

Hi,

I checked again to many cheap ATX PSU. It is good for you that you must choose 450W version. The 350W (or 200W?) have too small transformer. For 450W version have similar to ETD34 size.

About modification we have two option:

1. Rewinding and make new secondary circuit as I mentioned before. Hi... you are right Eva, all 3.3V and 5.5V circuit can be removed for all the component. But I have checked only the 450W, that the voltage reference and control switch taken from +12V. I don't know for other version, but according to the posted schematic it can be too.

The good things is that no need to pull the 3.3V and 5.5V winding to transformer pins, hence make the winding simpler. Ahhaaa... I am happy too with this. You will have up to +/-42V, may be more.

2. Make new voltage reference as Eva mentioned. We will have max voltage up to twice from before (24V) max. But you also must consider to caps working voltag and diodes breakdown and remember that you must create new circuit for negative supply too.

OK, Carlos, and Kambule you have two choice. It based on your need only.

Eva, thank you for sharing your knowledge with us.

Kambule, don't wait so long.... I got "Made in China" 450W ATX PSU with only 75.000,- rupiahs, that is 8.0 in USD. I believe it will be only approx 6.0 USD at Hong Kong... not expensive. Compared to conventional power supply, it it really-really cheap....

Best Regards,
Kartino