My New 12V SMPS w/Computer PSU Transformer

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
EWorkshop1708 said:

I also used a 2200uf for the smps remote 12V input to power chip+totem poles. It's the green cap in the picture.

You mean you use it so that chip +...have more stable voltage? I hope you seperated it from main voltage on main caps, this cap will be used same as your other input caps, which would mean current will go from that cap through narrow tracks to other caps. I used one low voltage drop, fast diode to seperate that cap from main board caps. Now cap can charge but it can't be used as part of supply
 
EWorkshop1708 said:


In my original posts, the CT terminal was on the transformer primary, but the lead was clipped off, so I just soldered a thick wire on it to get access to it. In the original ATX PSU it didn't use the center tap.



So its the center tap huh.. I forgot about it... :) yeah I recall a connection with the xformer clipped out...

I forgot how many turns there is to this winding.. 20+20 perhaps???

I think I still have some untouched ATX cores here..maybe I could play with it
 
luka said:


You mean you use it so that chip +...have more stable voltage? I hope you seperated it from main voltage on main caps, this cap will be used same as your other input caps, which would mean current will go from that cap through narrow tracks to other caps. I used one low voltage drop, fast diode to seperate that cap from main board caps. Now cap can charge but it can't be used as part of supply

Great idea. The cap and pwm is fed directly by the remote 12V wire, but to keep cap charge from going out of remote wire to main, I'll be adding a diode. Thanks :D
 
RX5 said:



So its the center tap huh.. I forgot about it... :) yeah I recall a connection with the xformer clipped out...

I forgot how many turns there is to this winding.. 20+20 perhaps???

I think I still have some untouched ATX cores here..maybe I could play with it

I've never disassembled the transformer, I have no idea how many turns there are. I just used it based on the voltages it was designed for, and because of the step-up ratio.

I figured I'd have to guess a lot when winding a Toroid, so the ATX transformer is a good start, and without the hassle of winding any coils! :D

I would recommend playing with a few ATX cores. I never had a use for them before, and now I wish I never threw away so many before in the past.

So, far, so good! The power supply is outputting good voltages, never gets even warm. Only thing that gets hot are the 10W 50 ohm input snubber resistors.

So far, I've been amazed how well it works. :D There's no heat, and lots of current. At first I assumed the heatsinks needed to be bigger, but since they are usually cold to the touch, I'm happy with what sinks I used!

Hardest part was not the power parts, but the PWM part with the TL594 and totem poles. I also learned that you MUST have a 1K pull-down resistor for the PNP side of the totem pole, or you WILL have cross-conduction. Luckily for me during testing, those big MOSFETS handled any abuse and hot temps from cross-conduction and anything else during testing. Now that it works right, they never get warm! :bigeyes:
 
well, I was just wondering if the ATX core/winding/s are really for only 50KHz switching freq(its what they are designed for).... 50KHz is (I think) low for use with switching power amplifiers??? SPS80 PS I recall, is at 90Khz?!? I dont know how 50Khz freq would interact to switching amps...

with 50KHz sw freq, you could probably use the output filter coil/s that came with the ATX....

anyways, it is a good start..It never came into my thought before, using them... AND I had thrown a way a lot of them cores before....for -OTHER- uses.. :D
 
dexter2003 said:
Hello,
with that transformer you won't get more than 300w ....The original psu had 550w peak (12x22=264 ).

Thanks. I had a feeling that's what it would do based on the amperage for that coil. That's still very sufficient for the planned usage. I wonder if you keep the core cool, and use a higher frequency, if you can squeeze a few more watts from it.

I wish I had two or more identical cores so I could make another one with more power. :D

Thanks for all the input so far. I hope others can learn from this and make their own SMPS as well easy and cheap.
 
hi !!

Thanks. I had a feeling that's what it would do based on the amperage for that coil. That's still very sufficient for the planned usage. I wonder if you keep the core cool, and use a higher frequency, if you can squeeze a few more watts from it.

I wish I had two or more identical cores so I could make another one with more power. :D

Thanks for all the input so far. I hope others can learn from this and make their own SMPS as well easy and cheap.

hi how are you, could you please post the schematic?? the one that appears is low quality, please , well greets and so long
 
i just found this.this is too cool!! i have thought of rewinding them but not flipping it around.and yes acording to transformer theory you should get more power with a higher frequency.2xfreq=2xpower handling.providing the core responds and the winding can handle the current.i'd say 50% more is worth shooting for.i have been looking everywhere for something like this to power abunch(like 8 to 16)of lm3886's in my car.please post any more progress you have made.jer
 
Mosfet heatsink isolation

With regards to the heatsink isolation, I use a .5mm (approx) thick cotton cloth soaked in heatsink compound. Make the hole in the heatsink bigger, and fix a hard wooden stick in it, make a small hole into it to accomodate the screw coming from mosfet's tab.

That's it, it can easily sustain the high temp. till everything burn out or melt.

- Shashank
(India)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.