ATX Power Supply Question

[MAXPAYNE]

Hi, i want to modify a pc power supply for 30 V, 15V output and i have rewinded the transformer. can you help me how can I calculate the feedback resistor  ?


why there are 3 feedback resistor from 3.3v, 5v and 12v output in the original power supply. will it be possible to use just one feedback path (let 12v) to stabilise all voltages.... ?

[tienngoc]

First: can you post the schematic of your circuit?
Second: the value of the feedback resistor from origin circuit?
Third: I think the circuit need three feedback line in order to make the accuracy of the voltage regulator.

(sory about my bad English)

[MAXPAYNE]

First: can you post the schematic of your circuit?
Second: the value of the feedback resistor from origin circuit?
Third: I think the circuit need three feedback line in order to make the accuracy of the voltage regulator.

(sory about my bad English)

It is not possible for me to post the ckt as i am modifying a readymade PC SMPS.

The feedback goes to pin 1 of TL494 and pin2 is referenced at 4.86V

For 12V: 27Kohms
For 5V: 4.7Kohms
For 3.3V: 15Kohms

from pin1 to ground :  24K || 470K || 220K (|| means parallal)

[Walkura]

This multiple regulation system is described in marty brown's book .
Some time ago i tryed to explain some one how to include a second rail into the feedback loop .
(he was also modifying a atx supply .)
http://www.sonsivri.com/forum/index.php?topic=12707.0
The book i mentioned i uploaded some time ago here on sonsivri .
http://www.sonsivri.com/forum/index.php?topic=12417.0
Have a look at page 26 and 113 of BROWN__M.__1990_._Practical_Switching_Power_Supply_Design.pdf
At page 26 he gives a basic explanation of the system on page 113 he explains the setup you describe .
With the information and examples he provides its a piece of cake .
(you should also download Pressman if it interests you those 2 i consider almost as the bible of switchmode supply's)

Good luck and be carefull

[MAXPAYNE]

please link of ""switch mode powersupply's by Pressman""

[Walkura]

Pressman ,Brown ,Power supply cookbook and basso,s Switch-Mode Power Supply Spice Cookbook are on the following link.
Some time ago i uploaded them on rapidshare and posted it on the following link.
http://www.sonsivri.com/forum/index.php?topic=12417.0

Also maybe this schema is helpfull to you .
http://www.pavouk.org/hw/en_atxps.html
Read the chapters in brown or pressman and get your calculator

Good luck .

[tienngoc]

Also maybe this schema is helpfull to you .
http://www.pavouk.org/hw/en_atxps.html
Read the chapters in brown or pressman and get your calculator

Good luck .

[/quote]

Your link is so great.
I have a switching powersupply, with input: AC100V, output 5V/10A. Using the TL494 IC.
How to modify --> input 220V, ouput 12V/6A. Calculate the feedback circuit is one of many problems need to do.

[MAXPAYNE]

Also maybe this schema is helpfull to you .
http://www.pavouk.org/hw/en_atxps.html
Read the chapters in brown or pressman and get your calculator

Good luck .



Your link is so great.
I have a switching powersupply, with input: AC100V, output 5V/10A. Using the TL494 IC.
How to modify --> input 220V, ouput 12V/6A. Calculate the feedback circuit is one of many problems need to do.

All the atx PS diagram I found on net use only one error amplifier. Does anybody have  any diagram with both error amplifer used (One for feedback and another for "I don't know") ?

[marty.markoh]

Also You can look out this site for useful information:
http://www.wikihow.com/Convert-a-Computer-ATX-Power-Supply-to-a-Lab-Power-Supply

[Walkura]

The other pin is usualy connected to the Vref with a resistor and to minus with a capacitor .
When you have a look at the datasheet of the SG3525 (it very simular in function but has totempole outputs for mosfets)
http://www.st.com/stonline/products/literature/ds/4286/sg3525.pdf
Have a look at the test setup the potentiometer will set the dutycycle .
By regulating the voltage on the non inverting pin 0.9 > 3.9 volts will make duty cycle 0 > 100 % (minus dead time)
By doing that with the inverting input 0.9 > 3.9 Volt will regulate duty from 100 % > 0%
Probably your *dont know* comes from the compensation pin or the Vref.
Tomorrow i will have a look for some more schema's .
@Tiengnoc.
Have a look at the schema i posted yesterday how it is arranged around the mains rectifier .
Usualy they have a voltage doubler like this for 115 Volt use .
Probably that is all you have to do .
Otherwise your in for calculating ,replacing parts and making a new transformer .

Have a nice eve Walkura

[MAXPAYNE]

Now i have made 30v from atx pc power supply. but there was a noise coming from the transformer. The noise increases as load increases. Why this is happening ? Is the core getting saturated ? what should i do to remove the noise ?


Posted on: July 03, 2008, 11:43:11 11:43 - Automerged
If i put my hand on the high side transistor heatsink the noise decreases rapidly.....

[Walkura]

When your transformer saturates thats usualy the end of your mosfets .
Not to mention since you didn't change the amount of primary turns the fieldstrength in Gauss shouldnt have changed .
If you have a scope you can look at the secundary side of the transformer ,
when you see a sharp rise on the end of the pulse you have a saturation problem .
Do keep in mindwhen you change from lets say 15 volt 2 Ampere to 30 Volt your max current will be 1 Ampere .
Maybe you hear a mechanical sound from too loose windings (just wild guess)
When you touch the high side transistor heatsink you increase the capacity of that point
There is a certain capacity between the mosfet and heatsink .
If you have a capacity measurement on your meter you can try to measure it (without voltage connected) without and with hand .
Also have a look if the heatsink is electricly connected to lets say minus in which case you could add the measured capacity change to that point .
In one of those books i uploaded they do describe the influence of the capacity caused by montage method of the mosfets c.q. transistors .
Afterall by placing your transistor on a silicon isolator you in effect create a capacitor.
Try to measure if possible
(always keep a isolation transformer on the primary side when you measure on the primary side in order to save your scope and yourself)


Good luck and be carefull .

[MAXPAYNE]

For switching MJE13007 is used , no MOSFET. and  I have run the PS abt 15 min with a load on the 30V side and all the time the hissing sound came. The transformer is so rugged that there isn't any chance of making sound itself.

Should I lower the diameter of primary winding coil (that means guage will be one step higher) ? The used 23 s.w.g. ..........

[Walkura]

Good evening Max .

Leave the primary winding as it is .
This winding dictates the fieldstrength in the core
(
.................. 10^8*V*Ton
B Gauss = -----------------
................... n*Ae cm^2 )
Making the diameter bigger isn't always usefull cause of the skin effect with higher frequency's .
Maybe its a stupid thought but since you went up from 12 to 30 Volt .
What voltage are the elko's of your output filter ?
With 12 Volt output i would place 25 Volts ,with commercial supply's there's usualy not to much over capacity to cut the cost .
I think that its Brown's book where you have a example half bridge 180 Watt's .
First walk through his example and make those calculations so you can be sure you did it right .
Second fill in your specs and make the same serie of calculations .
Just leave out the bits like current mode etc. cause its of no concern with a TL494.

Good luck

[MAXPAYNE]

Good evening Max .

Leave the primary winding as it is .
This winding dictates the fieldstrength in the core
(
.................. 10^8*V*Ton
B Gauss = -----------------
................... n*Ae cm^2 )
Making the diameter bigger isn't always usefull cause of the skin effect with higher frequency's .
Maybe its a stupid thought but since you went up from 12 to 30 Volt .
What voltage are the elko's of your output filter ?
With 12 Volt output i would place 25 Volts ,with commercial supply's there's usualy not to much over capacity to cut the cost .
I think that its Brown's book where you have a example half bridge 180 Watt's .
First walk through his example and make those calculations so you can be sure you did it right .
Second fill in your specs and make the same serie of calculations .
Just leave out the bits like current mode etc. cause its of no concern with a TL494.

Good luck

Thnks for uor nice info. Infact i was thinking abt lowering the diameter. What do u mean by 'elko' of output filter ? How can i design the toroidal filter ?

[Walkura]

With the elko i mean the electrolytic capacitor of the output filter .
Since you have to do with a square wave the amplitude is decided when you calculate the transformer .
With that i mean that except for line variations the peakvalue(voltage) is decided by the volt's / winding .
(as example) even if there is lets say 4,2 volt/winding
Then that will just mean that for normal functioning the duty cycle wont be above 93 % .
4,3 volt * 3 windings = 12,9 * max duty cycle .
But that also means that since the designer knows the absolute maximum voltage to be expected the capacitor doesnt need that much overhead voltage .
After all worst to expect is Npri/Nsec*Duty cycle +- line variation
For a 12 volt powersupply to make it jump to 16 volt (assuming 240 Volt supply) it needs to jump to 300 Volt AC .
And even then it will be just a few pulses till the feedback kicks in and lowers the duty cycle till the voltage drops .
So probably your output condensator is either 16 or 25 Volt (if you didn't change it anyway)
It might be the source of your hissing (then again i didn,t see your project so its a bit of a wild guess)
Maybe its just a stupid thought of me (for myself i could forget things like that and just fire the thing up after i changed it)
For the calculations of the toroidal filter (i guess you mean the output filter)
Open Pressman's switching powersupply design 2nd edition and on page 77 of the file you have the proper formula .
Also read from page 103 onwards the chapter about halfbridge topology
(sorry for the paperwork but i can asure you that your understanding of the circuit at hand will improve by doing so)

Good luck and above all be carefull .