13.8V 15A offline SMPS w/ PFC

not actually built from scratch but it's my first time messing with an offline SMPS. luckily, no fireworks and I'm still alive! :D

started off with a Dell desktop PSU. bought it from a surplus shop for about US$8.00

[IMGDEAD]https://i119.photobucket.com/albums/o128/djquan128/PCpsu/CIMG5318.jpg[/IMGDEAD]

when I took off the cover and checked what could be done, I realized it isn't the same as most PC PSU's with TL494 and LM339's etc. looked further and saw a chip on the live side, another on the output side, and a small daughter board.

about 2 days of thinking about what to do and learning around the circuit, figured out that the chip on the live side is a TDA16888 which is a PWM controller with PFC. nice! :D the daughter board supplies 15V to the TDA16888 and continuous 5V 2A for standby and for the watchdog logic (the unknown chip at the output side) and the supply also has temp controller for the fan.

so I proceeded to remove all components related to the watchdog circuit and the outputs that I didn't need. moved a few components, a jumper here and there and there you have it. :smash: we have a regulated PSU.

I don't know if I'm allowed to post what mods I did since it is really dangerous to modify these things if you don't know what you're doing.

pictures:

secondary side with most of the components removed, bare essentials are left.
[IMGDEAD]https://i119.photobucket.com/albums/o128/djquan128/PCpsu/CIMG5312.jpg[/IMGDEAD]

[IMGDEAD]https://i119.photobucket.com/albums/o128/djquan128/PCpsu/CIMG5314.jpg[/IMGDEAD]

pic of the PWM controller chip and PFC inductor
[IMGDEAD]https://i119.photobucket.com/albums/o128/djquan128/PCpsu/CIMG5320.jpg[/IMGDEAD]

leftovers
[IMGDEAD]https://i119.photobucket.com/albums/o128/djquan128/PCpsu/CIMG5313.jpg[/IMGDEAD]

waveform@ trafo 12V secondary, no load
[IMGDEAD]https://i119.photobucket.com/albums/o128/djquan128/PCpsu/CIMG5316.jpg[/IMGDEAD]


waveform@ trafo 12V secondary, 8A load (don't have any full load tests since I don't have a big enough dummy load)
[IMGDEAD]https://i119.photobucket.com/albums/o128/djquan128/PCpsu/CIMG5316.jpg[/IMGDEAD]

completed PSU
[IMGDEAD]https://i119.photobucket.com/albums/o128/djquan128/PCpsu/CIMG5321.jpg[/IMGDEAD]
 
I just did a 14.6A test at 3min continuous.

I did the finger test and the main switcher heatsink got hot (about 45degC) output rectifier sink got hotter (about 50degC, can keep my finger on it for a few sec) and the output inductor got burnin hot. (didn't give me a burn but it was really hot. can't keep my finger on it for more than a sec.) main primary cap got warm but not hot. same with main trafo.

I suspect output inductor heating is due to I2R losses as it doesn't get warm at idle.

will need to rewind it with thicker wire.

darn. I just realized that the waveform photos in the first post is wrong. the loaded waveform should be this one:
[IMGDEAD]https://i119.photobucket.com/albums/o128/djquan128/PCpsu/CIMG5309.jpg[/IMGDEAD]
 
luka said:
Hi

Please do post, what you did, with note that there are HIGH VOLTAGES ;)

Great work, now you can do one that has say 25A at 5v, so you will get supply that will do 25A at your output voltage


I have to find the notes where I put everything that I did to this PS. dunno where I left it. :D

I have a 5V 80A supply over here. :hot: but that is too much work. it also has a 12V 65A module in there too. both needs 300VDC and 24VDC input.


I have to figure out a way to put a overcurrent limit to this supply though. I was able to pull around 22A (output drops to only 13.77V!!!) out of it but scared of doing so for more than a few secs as the parts are pushed way above their limits.
 
djQUAN said:



I'm more concerned with the core 'cooking' and changing its properties and then BAM!!! :xeye:
The curie temp for magnetics is much higher so you can run them super hot. Try to find out what core is used and look at the specs.
It looks like a powered iron core and most likely it is since they are cheaper as compared other materials.
 
iron powder cores are color coded so you can figure what material they are. Then go to the magnetics inc website and download there software and plug in the appropriate core part # (measure core size to find part #), number or turns and wire size. You can play with the output current to then see core and wire losses and saturation effects by the drops in permeability. The software will also tell you the expected lifespan.
 
djQUAN said:
it's the long term effects that I know of. I have read here many times that a hot running core loses its capacity to store energy or something like that.

maybe Eva or others could chime in if I'm wrong?:xeye:

Yes running them hot reduces reliability so it's alway good to run them cool- just like anything else. I never said it's good to cook eggs on them.
 
switchmodepower said:
Yes running them hot reduces reliability so it's alway good to run them cool- just like anything else. I never said it's good to cook eggs on them.

but you said...

switchmodepower said:
The curie temp for magnetics is much higher so you can run them super hot. Try to find out what core is used and look at the specs.
It looks like a powered iron core and most likely it is since they are cheaper as compared other materials.

so I thought that's what you meant.

oh well, still at the bench. will probably keep the windings the same. just need to add a current limit at the output to keep it within 15A max.
 
It depends on your goals. There are trade-off. If you want an efficient ps then you optimize for low losses. If you just care about not going "BAM" then you can run them hot. I personally would never go about 125C but in certain circumstances you may need to run close to the max rating of the wire/core.

Check out http://www.micrometals.com/material/Coreloss_aging.html

They discuss core loss vs time, temp, and operational hrs. In one instance they state that in that given circumstance, core, etc you can operate at 100,000hrs at 125C. Now you need to know all operational parameters especially what core you have and the flux density since the core losses change with temp as shown.

Good luck
 
it isn't that easy. I had little difficulty because I used an existing winding and didn't use the others. rewinding the trafo is another issue because this is a forward converter (or I think it is, or it's either a flyback type? :xeye: ) and the transformer characteristics are critical.

if it is push pull, it is easier to change the output voltage by changing feedback circuit and trafo secondaries.