Fun with fixing I-Micro power supplies

[momaka]

Re: Fun with fixing I-Micro power supplies

Still think it's a waste to resurrect such a crap, not to mention putting polymers into it.

I disagree.
What these cheapo PSU have that more expensive units don't is simplicity. Sure they may not be very efficient and offer almost no protections, but all they need is good caps + a few modifications/improvements (as long as it's not a completely gutless PSU with hopeless design) and they will work forever. And that's because their design is very tolerant to component variations. With high-efficiency APFC PSUs, something goes just slightly adrift and the whole PSU stops working. Not to mention they are a lot harder to troubleshoot.

So with the mods that everell did here on this PSU, I certainly wouldn't call the PSU a "polished turd", let alone a "turd".

[everell]

Re: Fun with fixing I-Micro power supplies

With all these mods, I was hopeful we could come up with something to call it besides "turd" or even "polished turd". How about debonaire, cavalier, or even dapper.

[momaka]

Re: Fun with fixing I-Micro power supplies

With all these mods, I was hopeful we could come up with something to call it besides "turd" or even "polished turd". How about debonaire, cavalier, or even dapper.

All fine with me.

Unfortunately, majority of people don't appreciate a PC these days unless it has at least a 750W PSU, it seems. Hence all the negative backlash.

But just ignore that and keep modding them PSUs.

[rugger]

Re: Fun with fixing I-Micro power supplies

What fun is there in modding or messing around with a high quality PSU with Japanese caps and all the good stuff. There really isn't much you can do to improve such PSUs.

These junky PSUs on the other hand, offer much more exciting opportunities. I didn't repair a truepower 380W PSU because it was cost effective, I repaired it because it is fun to tinker, learn, and now I got it fixed and didn't kill it, I have a nice PSU for 5V heavy systems compared to modern 12V heavy PSUs.

[everell]

Re: Fun with fixing I-Micro power supplies

Very interesting experiment, everell (as always ). Do you by any chance have an oscilloscope, though? I'm rather curious to see the noise and ripple on the outputs, as super-low ESR caps can sometimes cause excessive "ringing" on the PSU output filters and possibly even produce more ripple and noise than if the PSU had failing/inadequate filter caps.

That in itself is the only reason I don't do poly recaps on PSUs, especially these cheaper units that are based on the oldschool slow-switching speed half-bridge design.

Here are some pictures I made showing noise spikes, 160 millivolts on the 5 volt line and 200 millivolts on the 12 volt line, 15.5 microseconds pulse time (64.5 kHz) using a Biostar M5VLQ pentium 3 mother board for a load.

[everell]

Re: Fun with fixing I-Micro power supplies

I used another similar I-Micro with cheap Chinese caps to compare with the recapped poly cap I-Micro above. As can be seen, the noise between spikes on the Chinese caps was about 50 millivolts whereas with the poly caps was about 20 millivolts. The amplitude of the spikes was higher with poly caps than with the Chinese caps. I am not sure if the spikes were coming from the power supply outputs or if it was from spurious signals in the air coming from the switching transistors and transformer. Both power supplies were tested using a Biostar M6VLQ pentium 3 motherboard as a load. Is it an improvement? Can't say for sure yet.

[Per Hansson]

Re: Fun with fixing I-Micro power supplies

Did you try putting a 10uF electrolytic and 0.1uF ceramic cap in parallel with the load as the Intel ATX testing guidelines mandates?
I realize you're testing with a mainboard so maybe it makes no difference in this case. (Depends on how close it's caps are to your probe point).
Still curious though...

[Dan81]

Re: Fun with fixing I-Micro power supplies

What fun is there in modding or messing around with a high quality PSU with Japanese caps and all the good stuff. There really isn't much you can do to improve such PSUs.

These junky PSUs on the other hand, offer much more exciting opportunities. I didn't repair a truepower 380W PSU because it was cost effective, I repaired it because it is fun to tinker, learn, and now I got it fixed and didn't kill it, I have a nice PSU for 5V heavy systems compared to modern 12V heavy PSUs.

This, I have loads of recapped Deer PSUs ready to be put in 5v heavy systems with 30A 40V rectifiers on the 5v rail.

[everell]

Re: Fun with fixing I-Micro power supplies

Next I did a poly cap on the outputs of a Bestec ATX-250 12Z. Here are the scope pictures. The I-Micro is a half bridge design. The Bestec is a current forward design. The scope pictures of the I-Micro above did not look too bad. The Bestec with no load had spikes at the switching frequence, but very little noise between the spikes. The Bestec with the same Biostar M6VLQ pentium 3 mother board as a load looks terrible. I would have thought that the Bestec with the current forward design would show much less noise when loaded.

[momaka]

Re: Fun with fixing I-Micro power supplies

Did you try putting a 10uF electrolytic and 0.1uF ceramic cap in parallel with the load as the Intel ATX testing guidelines mandates?

I second this.

Everell, you should follow the testing procedure according to the Intel PSU spec. See page 23 of this PDF document:

I realize you're testing with a mainboard so maybe it makes no difference in this case. (Depends on how close it's caps are to your probe point).

Indeed.

It's best to test PSU with some kind of a linear load (i.e. non-switching).

Next I did a poly cap on the outputs of a Bestec ATX-250 12Z.
...
I would have thought that the Bestec with the current forward design would show much less noise when loaded.

Well, it matters quite a bit how the feedback loop of the output filter is designed. Some feedback loops simply aren't designed to handle the ultra-low ESR of polymer capacitors. In addition to that, if you lowered the output capacitance due to using those polymers, that too will affect the feedback loop.

In essence, many older PSUs like these are simply not designed for polymer capacitors. On that note, you may want to look up output filter ringing on switching power supplies.