Fun with fixing I-Micro power supplies

[Pentium4]

Re: Fun with fixing I-Micro power supplies

If you count the manpower and material

But then I'm not using your RS caps

poor efficiency it will become very expensive in the long term no matter how much you payed or did nto pay) for that.

I have 7 computers running 24/7, and most of them have PFC or PPFC now. The half bridge unit with no PFC is powering a computer that pulls 52W DC. Is it really worth the trouble to reduce a few watts on one machine? Not in my opinion. And I also doubt its efficiency is in the 60's. It's probably 75% at worst.

On the other hand 80 PLUS Gold certified Seasonic SSP-350GT or similar quite cheap unit has 100% quality internals and you will definitelly notice 25-30 % higher efficiency as long as you don't steal electricity…

Good units, but I also get those Antec Earthwatts a lot for free that just have one bad CapXon or Ltec. Much better value Plus, isn't working on PSUs part of the fun? What's fun about buying an already perfect PSU?

[Behemot]

Re: Fun with fixing I-Micro power supplies

Earthwatts are not different league, they are different sports compared to this.

You will always have at least half decent units like these Antecs to use the caps for, it is really waste to do anything with these i-shits IMO.

[Pentium4]

Re: Fun with fixing I-Micro power supplies

Right. But so is the other HB unit I have, not all of them are bad

[ChaosLegionnaire]

Re: Fun with fixing I-Micro power supplies

As per the first post, a computer shop salesman asked the question....why are so many IMicro power supplies failing right out of the box? Based on what we have seen so far, what could we tell him?

1. Only one had emi line filter.
2. Main input capacitors have too low value AND measure only 60% of marked value.
3. Switching transistors inadequate for advertised wattage
4. Switching transformer too small for advertised wattage
5. No pi output filters on any of them.
6. Output filter capacitor values too small, inadequate for advertised wattage.

Two were "fixed" or "enhanced" and I am confident they will probably do half of their advertised wattage. I am working on one more.....I like to play with power supplies.

Anyone have anything else to say to the computer shop salesman?

you cant tell a salesman any of that - he wont know what your talking about.
just say that they are shit and the parts used are substandard junk.

lol! i'd just say that it looks like a one year old designed, built and made the power supply. why do they let one year olds play with power supplies, i have no idea. even that is an insult to a one year old.

i have also seen soldering on power supplies and speaker systems that look like they were soldered by a one year old. why do they let one year olds play with soldering irons, i have no idea.

[everell]

Re: Fun with fixing I-Micro power supplies

Well......I have inherited two more IMicros. Both 400 watters! The first had a badly burned 5vsb area, something I didn't want to mess with. It had no line input filter, and no place for output pi filter coils. Just the single capacitor outputs. So it is going to the scrap pile.

The second has the same pc board as the three I managed to fix. No input filter, but full pi filters on the outputs. So I decided to fire it up and see if it had any life at all. To my surprise, the 5vsb had an output, and with ps-on switch closed, all outputs were present.

So............I plugged it into my FIC motherboard (K7MNF-64) with AMD socket 462 K7 sempron SDA3000 (2 gig and 166 bus speed) along with a gig of memory and 40 gig ATA Maxtor hard drive. I was using an 8x AGP card with 256 meg of RAM. All three of my previous repairs would power it OK, so why not try it.

So what happened? Nothing. No fan movement on the mother board. No output whatsoever.

So I removed it from this motherboard and plugged it into my Biostar motherboard with Pentium 3 (1 gig) and 512 meg of memory. Onboard video. Maxtor 40 gig ATA hard drive. Powered up and it works fine. Output voltages are 12.62 Volts, 5.07 Volts, 3.34 Volts, and 4.40 volts for 5vsb. Although the 5vsb is low, it didn't prevent the power supply from working.

The 5vsb has a very small transformer and two transistor circuit. I could fix the existing circuit, or I could replace it with an more reliable integrated circuit. But to fix the power supply properly, I would need to add a line filter, replace the output capacitors with something bigger, and the output diodes would need to be improved.

I think that what I have done so far is good documentation for anyone who gets a IMicro power supply and wants to fix it.

So...........lets take a poll: how many think I should go ahead and fix this IMicro.......and how many think I shouldn't..............

[ChaosLegionnaire]

Re: Fun with fixing I-Micro power supplies

well, u'd have to fix that 12v line going out of spec too tho its only 0.02v out of atx specs. the fact that it also only works on one mobo but not another means there is a problem in the voltage regulatory circuitry and/or chip preventing the psu from working correctly on certain types of loading patterns.

personally, being a perfectionist, i would wanna fix the psu so it functions correctly across a wide range of loading patterns with a minimal "blind zone" whereby a certain type of cross-loading pattern makes the psu go out of spec. i imagine this might be time consuming. so if u have the time and wanna gain more experience in fixing (or even re-designing) a psu so it has full atx certification across a wide range of different loads, then by all means go ahead.

however, if it wont function right within atx specs for a certain type of realistic load pattern due to an inherant or unfixable design flaw in the psu circuit then just f-it and chuck it in the junk/scrap pile.

[Per Hansson]

Re: Fun with fixing I-Micro power supplies

I vote for blowing it up, but you know I'm biased

[momaka]

Re: Fun with fixing I-Micro power supplies

So...........lets take a poll: how many think I should go ahead and fix this IMicro.......and how many think I shouldn't..............

FIX IT.

I think it is better to have the PSU in working order, even if you never intend to use it to power anything. After all, if you do later decide to take parts from it, at least you know everything should be working because the whole PSU was working when you put it in storage. When you chuck stuff on the parts pile, you always have to check the parts you pull and wonder what works and what doesn't. Or at least, that's the way I think. I even fixed this TV power supply board despite the fact that I don't have the right TV to install it in.
Why? Because if I need to pull anything from that PSU, I know it must be 100% working, otherwise that PSU wouldn't be working either (or at least there is a good chance that it won't).

So............I plugged it into my FIC motherboard (K7MNF-64) with AMD socket 462...

Be careful there! I have a FIC AM37 motherboard that came out of an eMachines w2040. It was given to me because it got the high 5 VSB treatment from a Bestec ATX-250 12E. When I started troubleshooting it, I found some interesting things about the motherboard's design. Namely, the motherboard has several logic gate ICs directly connected to the 5 VSB rail. Those ICs have an absolute maximum voltage rating of 5.5 Volts. They were completely shorted input-to-output. I have not gotten replacements for them yet, but I did try making my own logic circuit on a breadboard and wired it to the motherboard. It still didn't work. Thus I am hoping that either my circuit was not up to the task (it was wired right, but it was done with TTL and not CMOS chips) and that the SB and NB are not directly connected to the inputs and outputs of the logic gates that blew... because if they are, that would make that motherboard a paper weight.

Now, I don't know how close your FIC motherboard's design is to mine, but I do know that manufacturers do tend to reuse the same design when it comes to system power. Therefore, I don't advise you use that motherboard with PSUs that have 5VSB circuits that you don't know if they work properly or not. For that matter, newer motherboards tend to be more forgiving, because they usually have regulators on everything. At worst, you will just blow an LPC I/O chip - which is still a nuisance to replace... but at least the motherboard is still repairable.

When I power up an unknown PSU for the first time, I always test the 5 VSB with a load first. I use a 12 V, 20 W halogen bulb. On 5V, it pulls about 1 Amp. That's good enough to test the functionality of the 5 VSB. If it passes, then I test with several more bulbs on each output. If it passes that, then I *might* try it on a motherboard .

[everell]

Re: Fun with fixing I-Micro power supplies

OK.....time to fix it! First problem, one of the 5vsb capacitors was bulging, esr measuring 61. So I replaced both 47 Mfd/16 volt suscon caps with 1500 Mfd/6.3 volt caps. The input thermistor looked burnt, so I removed it. It crumbled as it came out. Obviously bad. So it was replaced. Then I replaced the output filters on the 12V, 5V, and 3.3V rails with 2200Mfd capacitors. Each rail has 2 caps and a coil in pi configuration. Then I added a line filter coil and X cap.

Next order of business is circuit correction. The mfg. put the power switch on the neutral side instead of line side. So I rewired it so that the power switch is now in the line side.

With no load, the power supply came up working OK. Next I plugged it into my FIC motherboard. The computer came up working fine.

Another IMicro saved from the landfill............................

[kaboom]

Re: Fun with fixing I-Micro power supplies

Be careful using that on a modern motherboard. I'd not load the 5VSB more than 1A.

That almost sounds like an early-mid P3 era supply (or one feigning as such) with that small aux transformer.

They'd usually burn up after any time keeping the chipset and memory hot with the system in suspend-to-RAM.

If you never use STR and kill AC power when the computer is "off," that should be okay. Unless it draws current from 5VSB while +5 is present, as on certain motherboards, or ones with +5/+5VSB jumpers incorrectly set.

[ChaosLegionnaire]

Re: Fun with fixing I-Micro power supplies

lol!!! well done! at this rate everell is going to be the dragon ball of repairing and saving psus! how many power supplies did everell save from the landfill? over 9000!!!

[momaka]

Re: Fun with fixing I-Micro power supplies

^ It's true!!!
See this:
https://www.badcaps.net/forum/showpo...6&postcount=22

[everell]

Re: Fun with fixing I-Micro power supplies

OK, here are the pictures of the completed restoration. It now has full pi filters with 2200 uF on each side. The 5vsb has pi filter with 1500 uF on each side. A line filter has been installed. And the power switch was moved from the neutral side to the line side. I will run it a while on one of my motherboards, then it goes up in the attic to join its many friends!

[everell]

Re: Fun with fixing I-Micro power supplies

No activity with IMicro power supplies for a while, so decided to "upgrade" this last IMicro I fixed. I had a Coolmax with pc board delaminating, so removed the nice large heat sinks from it and used them to replace the IMicro heat sinks. While I was at it, I also replaced the main switching transformer. I also replaced the no-name switching transistors with a pair of 13009 TO-220 transistors. Also replaced the 3.3 volt output diode with a more sustantial schottky diode.

On power up, it shorted the 12 volt output diode, a F12C20C high speed diode. Replaced it with a MBR20100CT schottky diode. No more problem. The 12 volt line came up low and the 5 volt line came up high, so I replaced the group regulation coil, and all voltages look nice now: 12.37v, 5.20v, 3.35v, and 5.25vsb. I was going to replace the 3.3v mag amp coil, but there was not enough room to install the bigger coil.

I removed the cheap fan and replaced it with a ball bearing fan out of a junked UPS unit. Now when I turn on this power supply, it sounds like a jet engine revving up! It does move a LOT of air.

[Behemot]

Re: Fun with fixing I-Micro power supplies

Those 5V voltages are almost ouf of spec, what exactly looks nice about that?

[everell]

Re: Fun with fixing I-Micro power supplies

With better availability and dropping prices on the non-liquid electrolyte poly capacitors, I decided to have some fun installing poly caps on the first I-Micro I fixed. It had Seacon 1000 uF capacitors in the 12, 5, and 3.3 volt outputs. Also Seacon 470 uFD in the 5VSB output and the -12 volt output.

The Seacon capacitors were all surprisingly good. All 1000 uF measured esr .02, and all 470 uF measured .06.

I used Nichicon FP caps for both values, all rated for 16 volts.

I first measured the DC output voltage on 12,5,3.3,-12. and 5vsb. Changed the 9 capacitors. Then measured the DC output voltages. Output voltages were all identical. Next, I connected it to my Pentium 3 Biostar M6VLQ mother board. Worked just as good as before the Polycap conversion.

Just think, 20 years from now this power supply will probably still be working fine.

[momaka]

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.

[Behemot]

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.

[Dan81]

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.

Just because it looks cheap that doesn't mean a turd can't be polished.

[everell]

Re: Fun with fixing I-Micro power supplies

Since this power supply has had several modifications, what would you call a "modified turd" that has been polished?