Switching power supply

[vainajala]

Hi all,

I thought about building a power supply for a thermoelectric (peltier) cooler element, which maximum power is 122W @ 15,7V (~7,8A).
I have a toroidal transformer whose ratings are 300VA and 2x35VAC, so dropping the voltage from about 40VDC (or something, I don't know the  exact voltage which comes from the transformer after rectifying&filtering) to around 12-14 volts would most propably produce tremendous amount of heat to the linear pass transistor(s)...  (40V-13V) * 6,5A = 175,5 W of heat if i calculate correctly. That could make good heater though

So I started planning a switching power supply. TL494 might be good choice as a controller IC, so i downloaded the datasheet and started designing with EAGLE...

Only problem is that the circuit presented in the datasheet has output voltage of +5VDC and current capability of 200mA.
I understand that the output voltage and current are both "sensed" by the TL494's on-board operational amplifiers, which control the PWM "chopping" process... changing resistor values might do the trick!

I think that by changing the resistor values, switching transistor, output coil, capacitors and the current sense resistor would solve my problem.

Would someone give me some advice ???

TL494's page: http://www.onsemi.com/site/products/summary/0,4450,TL494,00.html

View attachment 35449

 

[ante1]

vainajala,

Welcome to this community.

When rectifying and filtering 32VAC you end up with about 50VDC and this is to much for the TL494.

Ante :

 

[vainajala]

vainajala,

Welcome to this community.

When rectifying and filtering 32VAC you end up with about 50VDC and this is to much for the TL494.

Ante :

Thank you very much!

Yes, i was aware that after rectifying and filtering the voltage might be too high for the IC to handle. However I found this picture in the datasheet, which is the solution:View attachment 35452

 

[ante1]

vainajala,

Then I would change the TIP 32A for a MJ 4502 or similar. The inductor and diode should be higher rated (10A+) and the voltage and current feedback resistors changed to give you the desired output. A bit larger caps is necessary and remember heavy wiring for the input / output path.

Ante :

 

[vainajala]

I made the circuit with Eagle and modified it a bit.

• I have some MJ15004 transistors, so I'll propably use one of those
  
• Diode is ultrafast type, rated at 200V/8A
  
• Most propably I'll take the inductor from an old PC power supply
  

I only hope that the Hfe of MJ15004 is high enough (about 25)... maybe I should try darlington connection?

If you notice any errors, please tell me and I'll try to fix View attachment 35456

 

[ante1]

Vainajala,

Can you simulate it ?

Ante :

 

[vainajala]

Ante:

I thought that too, but I don't have the program.
Do you know some good simulation program?

 

[ante1]

vainajala,

I have Proteus but no model for the TL 494.

Ante :

 

[vainajala]

Ante:

I'm downloading TINA Pro demo version, I'll have to check if it has the model...

 

[ante1]

Vainajala,

Looks good to me, I would hook it up and try it!

Ante :

 

[vainajala]

TINA Pro didn't have it
But I'll try one more program: 5Spice

 

[vainajala]

OK, time to flood the forum again with my pictures ;D

Well, actually it's the same circuit, with a soft-start feature that should limit inrush current when starting. The board measures about 10,5 x 7,5 cm, and I designed it with the soft-start components in place.

If I ever get this thing to work, could this be added to the "projects" -section....? :
If you will take just one more Powersupply circuit!

View attachment 35466

 

[mixos1]

Hello vainajala

I encourage you to finish it and then i will surrelly add it under the project section with your name and e-mail. You have done a great work so far. When you finish it contact me using the contact page and it will be shown under projects section right away!

Thanks

 

[ante1]

vainajala,

Looks good to me, If you can fit a heatsink on the TO3 your home free. I can

 

[vainajala]

vainajala,

If you can fit a heatsink on the TO3 your home free.

Ante :

Heatsinking of the MJ15004 (in TO-3) is a problem, should I do a mofidied version with, for example, parallel-connected two or three BD912's (in TO-220)? Because TO-220 is way more easier to heatsink than the TO-3 in my opinion...

If I end up using the BD912's, I don't think that the TL494 has current capability to drive paralleled transistors. So driver transistors might be a mandatory addition in that case...

[glow=red,2,300]OR[/glow]

Should I just stick with the original design?

 

[MP1]

vainaja, I agree with the others. It is a nice looking layout. Perhaps you could leave it as it is and later make a new REV?

MP

 

[vainajala]

Perhaps you could leave it as it is and later make a new REV?

MP

That's OK, if you say so!
You know, I'm just trying to make it "even better" and easier to use...

I posted a picture about a thing which confuses me. I hope you can answer.

What do you think about the R12's value? Is it correct? The TL494 needs supply current of 15mA MAX. My transformer gives about 45VDC, so can I calculate resistor value like this?:
(45V-39V)/0,015A=400 ohms?

Would be glad for help! View attachment 35471

 

[ante1]

Vainajala,

I think you should stick to the TO3 and either move it from the pcb or fit a heatsink like this:

Ante :

View attachment 35472

 

[vainajala]

I have now built the powersupply and tested it a little bit. I unfortunately can't test it thoroughly, because I don't have a variable DC-supply. I'll try it tomorrow with two car batteries in series, resulting 24V.

I did some tests with a car battery charger as a mains supply, trying some 12V bulbs and small fans as a load. It did work quite nicely, and it lowered the output voltage to about 8-9V (charger's voltage is about 18V when not loaded).

I also measured volts and amperes & calculated effiency "n":

• With 21W car bulb: n=74%
  
• small 12V/100mA fan: n=68%
  
• 55W car halogen: n=46%
  

(charger's max. current sets the limit here, it's only 4A)

It definitely needs more testing, fine-tuning and of course heatsinking, but if it works after that, I think it will be worth of it View attachment 35496