Fun with fixing I-Micro power supplies

[kc8adu]

Re: Fun with fixing I-Micro power supplies

only if your time is worth nothing....

I guess if you have a large pack-ratted stash of salvaged good components and heatsinks you could re-manufacture almost any POS into decent 300-350 watt P/Ss. The original heatsinks would have put the realistic (= no shrapnel) power limit around 200W.

[PeteS in CA]

Re: Fun with fixing I-Micro power supplies

I guess if you have a large pack-ratted stash of salvaged good components and heatsinks you could re-manufacture almost any POS into decent 300-350 watt P/Ss. The original heatsinks would have put the realistic (= no shrapnel) power limit around 200W.

[everell]

Re: Fun with fixing I-Micro power supplies

The IMicron pc board from my scrap pile (shown earlier) does have room for line input filter and output pi filters. Maybe there is some hope for this one! So lets see if I can find enough parts to put it together properly! Found line filter components from one of the less desirable IMicron units. Found three pi filter coils in my pull out parts collection. From my pull out capacitors, found a pair of 2200 uF/16 volt capacitors for the 12 volt output, a pair of 3300 uF/10 volt for the 5 volt output, and a pair of 2200 uF/6.3 volt for the 3.3 volt output. Found a pair of 1000 uF/200 volt caps for the main input capacitors. Installed a pullout 7500 pwm chip, and a few other missing parts. Now it looks much better.

[everell]

Re: Fun with fixing I-Micro power supplies

Found three more IMicro's. One had a pc board like the ones in my first post. The other two had yet a fourth version of pc board layout. All three of these had one capacitor for each output voltage, no room for a coil and second capacitor to form a pi filter. So all will go to the scrap pile. The thing that caught my attention to this "fourth" pc board layout is that not only did it NOT have a line filter, it has no place to put it. Take a close look at the pictures. Nothing on the pc board layout for a line filter! How can you possibly make this power supply better.....when the pc board can not accomodate pi filters on the output and line filter on the input??? Cheap.....no wonder these things fail quickly.

[everell]

Re: Fun with fixing I-Micro power supplies

I had my doubts, but I tested both of these "rebuilt" power supplies by powering up a Gigabyte motherboard with 3.2 Gig Prescott Pentium 4 cpu and 2 GB of PC3200 memory and a PCI express video card. I was expecting them to either not turn on at all or at least give a few hick-ups. Both power supplies ran the motherboard without giving any trouble.

Both also powered up fine on my Pentium 3 Biostar M6VLQ test mother board.

[everell]

Re: Fun with fixing I-Micro power supplies

The second "fixer-upper" is also rated for 450 watts according to the case. I used a second set of the "better" heat sinks for this one. Also removed the ZWCON main capacitors and replaced them with Fuhjyyu 680 uF/200 volt capacitors. Added the three coils in the output pi filters, and replaced all output capacitors with "better" measured output pull-out capacitors. I really didn't want to spend money to put super good capacitors in a power supply which I know is not going to do its rated 450 watts. Added line filter transformer and X and Y capacitors. Gutted the two transistor 5vsb circuit and installed a Viper 22A smd chip. My thanks to ldsishere for making the circuit board. I installed two D209L pull-out switching transistors. They were singing to me, so replaced them with a pair of pull-out 2625 transistors. Now this power supply works.

[everell]

Re: Fun with fixing I-Micro power supplies

The first "fixer-upper" is rated for 450 watts as shown in the picture. I used the best of the heat sinks I had from the five units. I removed the ZWCON main capacitors and replaced them with Fuhjyyu pullouts from a Antec Smart Power. At least their capacitance was in spec! I used the larger sized switching transformer, and a pair of pull-out D209L switching transistors. I installed three coils in the pi output filters, and used some pull-out capacitors to finish off the output filter circuits. I gutted the two transistor 5vsb circuit and replaced it with a DM311 chip and circuit. Made a nice little heat sink from a piece of copper flashing. And....I added a line filter with X and Y capacitors. Now it looks like a "half" decent power supply.

[stj]

Re: Fun with fixing I-Micro power supplies

the word "shit" seems insuficient.

[everell]

Re: Fun with fixing I-Micro power supplies

The remaining two power supplies and "junked" pc board all had ATX 12V2 VER 2.0 marked on the pc board. This pc board has room for full pi filter on the output circuits. But all three boards had one output filter capacitor, the pi filter coil space was jumpered, and one capacitor space was not used. Even more curious, output filter capacitors for all three outputs was 680 uF/25 volts. I don't think that is enough filtering to meet ripple requirements for a 400 watt or 450 watt power supply!

Seems to me that I have seen this ATX 12V2 VER 2.0 pc board in other off brand power supplies. Has anyone else seen one?

[everell]

I decided to start another thread for a "cheap" power supply, the I-Micro series. A salesman at a computer shop asked me if I knew why The I-Micros had a problem of being dead out of the box. I had five "bad" ones and a pc board from one I had gutted some time ago. They varied from 350 watts, to 400 watts, to 450 watts. They also used three different power supply boards. Pictures of the circuit boards from three of them are in this post. All had capacitor outputs with no room for coils to form pi output filters. This was a key issue for me. I don't want to try improving a power supply with pc board not even having room for output filter coils. All but one board used main capacitors branded ZWCON. Measured capacity was only 60% of marked value on all of them. Output capacitors (bloated and otherwise) were branded HEN. Of the five pc boards, only one had a line input filter. For switching transistors they used 13007 or 4124 switching transistors. Two boards had larger switching transformers, four had smaller switching transformers. Size didn't seem to match wattage rating.