Bestec 250atx-12e partially fried emachines motherboard

[kleung]

Need help/opinion on whether this system is repairable.

I just picked up an emachines H2542. This is a pentium 4, socket 478 motherboard system. Unfortunately, I also have the bestec 250atx-12e power supply that is known to frequently fail on the 5v rail and fry motherboards.

Peering into the psu; I can see that there are some bad capacitors there so I may repair / discard the psu.

HOWEVER, the motherboard powers up but is partially damaged.

Symptoms

1) audio/usb ports show up in windows but don't work. I've tried switching psu's and still doesn't work. This has been mentioned before in youtube videos as one of the components that gets damaged by the bestec psu failure.

2) When the system is powered off; the fan on the power supply still runs and the cpu still feels hot to touch.

So my question is... Is this a fried chipset issue (non-repairable) or a capacitor/component repairable issue on the motherboard.


I'm not sure what would need to fry in order to have no audio/usb, and standby/off motherboard with power still going to cpu (cpu fan does not spin though).

Ironically, the system still runs for now.

I don't even know what component may fail due to 5v rail overload on the motherboard to cause these type of symptoms. Ideas/suggestions appreciated.


BTW - no visible badcaps on the motherboard

[everell]

Re: Bestec 250atx-12e partially fried emachines motherboard

The Bestec power supply can be fixed, especially if the 5vsb circuit is modified. Others who have experience with the Bestec fried mother boards will probably tell you that the mother board is fried. I cannot comment as to whether or not it is fixable.

If you would like advice on fixing the Bestec, I will be happy to comment on it. If you decide not to fix it, please don't trash it as I am interested in saving it.

[kleung]

Re: Bestec 250atx-12e partially fried emachines motherboard

Thx for the reply. I followed with great interest and horror at the complexity of the mods you made to the psu in the other thread

Kudos on your design mods.

As for me. I mIght just replace the bad caps and use it until it fries the rest of this old pentium bOard

[c_hegge]

Re: Bestec 250atx-12e partially fried emachines motherboard

The problem with the bestecs is on the 5vsb (purple wire on the ATX connector), not the standard 5v rail. The 5vsb rail powers the motherboard when the PC is off, so an overvolt will usually fry the chipset, which generally means the board is a write-off.

You can also fix the overvolt problem by replacing one of the really tiny caps (provided that the PSU hasn't destroyed itself).

[mockingbird]

Re: Bestec 250atx-12e partially fried emachines motherboard

How do you know that it fries the chipset and not just a fet?

[c_hegge]

Re: Bestec 250atx-12e partially fried emachines motherboard

On most boards that get Bestec'd or Antec'd, the southbridge will get red hot when you plug the ATX connector in, which indicates that it is shorted internally.

[kleung]

Re: Bestec 250atx-12e partially fried emachines motherboard

THanks for the quick and useful replies.

What I've done

1) Opened up PSU; saw that the 2 caps (c1&c15?) were bad on the esr meter. No bulging. area around c1 was charred/brown.

Replaced

10uf-50v => showed esr infinite, capacitance 6 uf
220uf-25v => showed esr 3, capacitance 160 uf


Question; I tested a 1uf-63v cap with a esr of ~ 25. Not sure if that's normal or not for 1uf capacitors???

2) I did not check the purple atx / 5v line before the recap. However, AFTER the recap; it shows 5v.

In fact, powered on; ATX psu lines show

Grey +5, Purple +5, yellow +12, Red +5, Blue -12, white -5
no surges noted.

I measured the voltages across the capacitors above (when replaced) and seems to show C1 (10uf) => 11-13 v; c15? (220uf) => 17-18 v approximately.

3) Tested ESR on all motherboard caps and none seem particularly bad.

Still have problem of no sound/no usb. When I have the computer powered "off"; psu fans still runs and the cpu heatsink gets hot to touch (short circuit somewhere).

Any suggestions => southbridge shorted/fried.

Funny thing is that the system still runs fine; just no audio/usb still and the power off still pushes power to the psu fan and cpu.

Luckily; pc ps/2 connectors still work.

[kleung]

Re: Bestec 250atx-12e partially fried emachines motherboard

duplicate post deleted

[c_hegge]

Re: Bestec 250atx-12e partially fried emachines motherboard

Question; I tested a 1uf-63v cap with a esr of ~ 25. Not sure if that's normal or not for 1uf capacitors???

If you mean 25 ohms, then yes, that's a problem. the ESR should be a bit higher than a large cap, but 25ohms is too high for any cap.

As for your problems with the motherboard, though, I'm thinking your chipset is toast.

[kleung]

Re: Bestec 250atx-12e partially fried emachines motherboard

If you mean 25 ohms, then yes, that's a problem. the ESR should be a bit higher than a large cap, but 25ohms is too high for any cap.

As for your problems with the motherboard, though, I'm thinking your chipset is toast.

Yes, the ESR was ~ 25 ohm on the 1 uf-63V cap.

Could that be the reason that the psu is still sending power to the motherboard even though the system is powered off?

When the system is "off"; the psu fan is still spinning; usb devices are still on and the cpu is still receiving power (heatsink is hot).

As for the chipset; it seems that audio/usb ports are fried. Installation of new PCI usb and PCI audio cards seems to have restored function of audio/usb.

Funny that the chipset would only fry partially like that. I would have expected a no POST situation.

[kleung]

Re: Bestec 250atx-12e partially fried emachines motherboard

Everell
Just going through your fantastic circuit diagram for the Bestec 250atx-12e psu.

I was wondering

C13 on your circuit diagram... you have listed as a 0.1 uf -63v capacitor. Maybe I misread it but I thought it was a 1 uf-63v capacitor (haven't cracked the psu open again to see).

Also, what would be the purpose of that capacitor/what would happen if that capacitor failed?

Thanks in advance,

[everell]

Re: Bestec 250atx-12e partially fried emachines motherboard

Capacitor C1 10uF is the "critical capacitor" which caused all of your problems. When it starts going high ESR, the voltage regulator tries to go to higher voltage to compensate due to oscillation.

The 220uF capacitor (C50) is on the larger of the feedback windings. When C1 goes high esr, voltage goes up on ALL three of the secondary windings. This means that the 5vsb can go up to 15 volts or more, and the voltage across the 220 uF capacitor can go over 55 volts before exploding. Many have found only pieces of the 220uF capacitor inside the metal box.

Transistor Q2 (C1815) acts as a switch between the 220 uF capacitor and the main PWM chip. It also functions as a voltage regulator to keep the voltage going to the PWM chip at about 18 volts.

Capacitor C13 is just a filter between the switch (Q2) and the main PWM chip.

Your voltages look pretty good, voltage across C1 should be about 11 volts, and voltage across C50 (220 uF) should be about 20 volts. Voltage across C13 should be about 17 to 18 volts.

That particular schematic was drawn more than three years ago, and was my FIRST attempt at drawing a power supply schematic (only a section covering a few circuits). There may be a few errors. Finding ANY schematics on these power supplies is hard. So I simply posted what I found in hopes of saving others the same heartaches.

I attempted to repair the 5vsb on my first Bestec ATX-250 12E twice. Each time it simply blew up again. Very discouraging! That is when I decided to try my hand at modifying the circuit to incorporate a pwm type chip rather than blowing up more transistors. The DM311 chip worked great. Once the 5vsb was working, I determined that the power supply was still dead. The main pwm chip was not working properly due to the reference voltage wrong, internally shorted. So I replaced it and the reference voltage came up correct, but power supply still dead. The 7510 supervisory chip gets its power from the 5vsb circuit, so I figured it might be fried also. And it was. I replaced it and the power supply came back to life.

When the 5vsb circuit on this Bestec model goes bad, it can do lots of major damage. On my power supply, it took out most of the 5vsb components, the main PWM chip, and the supervisory chip. I have no idea as to what it must have done to the original owner's mother board.

My repaired and modified Bestec now powers a socket 462 mother board that I use to play MP3 music files. It has given no further trouble since the modification.

[kleung]

Re: Bestec 250atx-12e partially fried emachines motherboard

Everell, thanks for the awesome and detailed reply. That really was perserverence at another level to get the psu working.

I regret not having measured the +5vsb (purple) atx line's voltage before the repair as I'm not sure what it was running at.

Suffice it to say; the fact that the components around c1 were charred shows that it had failed and the voltage/temperatures were indeed high.

The motherboard (emachines) that it's in is semi-salvageable but honestly; probably not going to bother. Pentium 4 cpu/motherboards are almost free at this point (just gave one away last month myself).

Just trying to repair for fun and giggles.