VOGONS


First post, by gex85

User metadata
Rank Member
Rank
Member

Hi folks,

the other day I received this motherboard: MSI MS-5169 Ver. 2.1, Super Socket 7 ATX, ALi Aladdin V chipset. It came with an AMD K6-2 300 MHz (300AFR) and refused to start (powers on, but no beeps, no POST, nothing). Jumper positions are all correct.
Upon closer visual inspection I found this MOSFET next to the ATX power connector:

MS-5169 MOSFET.jpg
Filename
MS-5169 MOSFET.jpg
File size
607.25 KiB
Views
427 views
File comment
MOSFET
File license
CC-BY-4.0

To me this looks like it got too hot, cracked and even melted the solder so that it slid down on the PCB until it was stopped by the EC3 cap.
Just wanted to make sure you guys agree with me in that regard. Do you think it's worth trying to replace the MOSFET? Or is it likely that other components got damaged beyond repair when the MOSFET died?

Here's the datasheet:

Filename
CEB603AL Datasheet.pdf
File size
515.34 KiB
Downloads
8 downloads
File comment
Datasheet CEB603AL
File license
CC-BY-4.0

Cheers!

1992 - i486DX2-66 // 1997 - P1-233 MMX // 1998 - P2-350 // 2000 - P3-650 // 2001 - Athlon 1400 // 2003 - Athlon XP 3200+ // 2008 - Xeon E5450 // 2015 - Xeon E3-1240v5

Reply 1 of 13, by Socket3

User metadata
Rank Member
Rank
Member

Personally, I'd replace the mosfet. That's a nice motherboard, it's worth a trying to repair it in my opinion.

I doubt that's the original mosfet - the solder work tells me it was replaced at one point... I'm no expert of course, and after googling some pics of your board, the original mosfet is indeed made by CET (can't make out the model from the pics).

Mosfets are a common failure point on any motherboard. I believe that replacing it should fix the board.

Reply 2 of 13, by gex85

User metadata
Rank Member
Rank
Member

Thank you for your assessment! I think I'll try to replace it.
I thought that what looks like a sloppy solder job might indeed be the result of the MOSFET getting too hot and melting the solder by itself...

Since I'm not too experienced with this type of repair, do you think this would be a suitable replacement?
Vishay SUM90N03-2M2P-E3 (Datasheet: https://www.vishay.com/docs/74342/sum90n03.pdf)

Edit: I have now removed the defective MOSFET which was pretty straightforward and dug out an already dead donor board (A7V266-E). This one has a couple of GFB75N03 MOSFETS which I guess I could use. The voltages match (VDS = 30V, VGS = +-20V) and the max. drain current is higher (80A) while the RDS(on) is a bit lower. Datasheet: https://www.digchip.com/datasheets/download_d … number=GFB75N03

1992 - i486DX2-66 // 1997 - P1-233 MMX // 1998 - P2-350 // 2000 - P3-650 // 2001 - Athlon 1400 // 2003 - Athlon XP 3200+ // 2008 - Xeon E5450 // 2015 - Xeon E3-1240v5

Reply 3 of 13, by pyrogx

User metadata
Rank Newbie
Rank
Newbie

Be careful with that K6-2, chances are that it was also killed when powering on the board. Or it was the cause of the cracked/dead mosfet. If a K6 dies due to overvoltage, it can cause a short-circuit on its power supply pins Vcc2 and Vcc3 to GND. This in turn will also kill the voltage regulator on the board.

Reply 4 of 13, by kalohimal

User metadata
Rank Member
Rank
Member

Don't think the mosfet had become so hot the solder melted. The melting point of 60/40 solder is around 370 deg F and the lead free ones even higher at around 420 deg F. It looks more like a sloppy solder job.

The mosfet shouldn't be cracked. If you have a multimeter, you could measure the input and output voltage and check if it is working properly. The leg that is connected to the coil is the output (you can just measure the voltage at any leg of the coil against ground). Also measure all other mosfets on the board, including those for the CPU and RAM. The CPU mostfet output should match the Vcore of your CPU (with CPU in socket). The RAM mosfet output should also match the RAM voltage.

Slow down your CPU with CPUSPD for DOS retro gaming.

Reply 5 of 13, by gex85

User metadata
Rank Member
Rank
Member
pyrogx wrote on 2020-07-15, 08:38:

Be careful with that K6-2, chances are that it was also killed when powering on the board. Or it was the cause of the cracked/dead mosfet. If a K6 dies due to overvoltage, it can cause a short-circuit on its power supply pins Vcc2 and Vcc3 to GND. This in turn will also kill the voltage regulator on the board.

Thank you for the heads-up! I took out the CPU and checked the continuity between some of the VCC2 (AN9..19) and VCC3 (AN21..29) pins and VSS (AM8..30). And yep, short. So I will definitely use a different CPU and throw this one away.

kalohimal wrote on 2020-07-15, 08:46:

Don't think the mosfet had become so hot the solder melted. The melting point of 60/40 solder is around 370 deg F and the lead free ones even higher at around 420 deg F. It looks more like a sloppy solder job.

The mosfet shouldn't be cracked. If you have a multimeter, you could measure the input and output voltage and check if it is working properly. The leg that is connected to the coil is the output (you can just measure the voltage at any leg of the coil against ground). Also measure all other mosfets on the board, including those for the CPU and RAM. The CPU mostfet output should match the Vcore of your CPU (with CPU in socket). The RAM mosfet output should also match the RAM voltage.

Okay so maybe the original MOSFET died and killed the CPU as well, then it was replaced (sloppy solder job) but due to the CPU pins being shorted, the replacement died immediately as well... 😉

I do have a multimeter but since I have already removed the dead MOSFET I don't think there's much to measure at the moment. But thank you for the detailed explanation, as soon as I have a suitable replacement component, I will check everything.
First, I will have a friend of mine (who is an electrical engineer) check the datasheets of the dead MOSFET and the one I sourced from the dead Asus board to see whether I can use it.

1992 - i486DX2-66 // 1997 - P1-233 MMX // 1998 - P2-350 // 2000 - P3-650 // 2001 - Athlon 1400 // 2003 - Athlon XP 3200+ // 2008 - Xeon E5450 // 2015 - Xeon E3-1240v5

Reply 6 of 13, by dionb

User metadata
Rank l33t
Rank
l33t

Tbh, I'd say it's unlikely that the MOSFET was primary cause here. Yes, it probably killed the CPU, but something else killed the MOSFET. I'd check for additional shorts. Sometimes bad caps might do that, but with electrolytics that's unlikely (but possible: I once saw an MS-6309 where some electrolytic caps failed short and the MOSFET that failed because of that actually melted its way through the PCB 😮 ).

Reply 7 of 13, by gex85

User metadata
Rank Member
Rank
Member

Okay, so my friend said that while not really meeting the specs, the replacement from the Asus board would be close enough to give it a try.
So I soldered it in place (slightly less sloppy solder job than before):

MS-5169 MOSFET 2.jpg
Filename
MS-5169 MOSFET 2.jpg
File size
478.95 KiB
Views
350 views
File comment
MS-5169 MOSFET replaced
File license
Fair use/fair dealing exception

I put in a known-good Pentium 166 MMX, fired it up, and yes, it's alive!

MS-5169 POST.jpg
Filename
MS-5169 POST.jpg
File size
809.57 KiB
Views
350 views
File comment
MS-5169 POST
File license
Fair use/fair dealing exception

So this is already a great success. Although I agree with dionb, it's hard to imagine that the MOSFET itself was the primary cause. I will do some tests under load later and see whether the system gets unstable or any other components get too hot. Also, I will probably hunt down a proper replacement for the MOSFET that matches the specs of the original one better.

Huge thanks to all of you, especially to @pyrogx who has pointed in the right direction with the faulty CPU!

1992 - i486DX2-66 // 1997 - P1-233 MMX // 1998 - P2-350 // 2000 - P3-650 // 2001 - Athlon 1400 // 2003 - Athlon XP 3200+ // 2008 - Xeon E5450 // 2015 - Xeon E3-1240v5

Reply 8 of 13, by macroexp

User metadata
Rank Newbie
Rank
Newbie

I have a Matsonic MS6260S (Aladdin V PCChips clone) that had a similar problem. It died while rebooting, running a K6-2 500. The MOSFET looked like it melted the solder and lifted off the board- but it’s possible it just wasn’t soldered down well in the first place, and the heat sink wasn’t touching the pad on the board - so it probably went into thermal runaway and died.

Anyways, I found an exact replacement mosfet online and replaced it - fired right up and the k6-2 500 that was there was completely unharmed, still running today.

Reply 9 of 13, by kepstin

User metadata
Rank Newbie
Rank
Newbie

At least your MS-5196 board appears to have better caps than my (rev 4.0) did, mine was straight out of the Taiwanese capacitor plague era, and had a few leaking 1000µF caps. I ended up recapping the board, and stability improved quite a bit.

I'm surprised by how different the mosfet and capacitor layout around the cpu socket is between the two board revisions.

Reply 10 of 13, by gex85

User metadata
Rank Member
Rank
Member
kepstin wrote on 2020-07-16, 04:58:

At least your MS-5196 board appears to have better caps than my (rev 4.0) did, mine was straight out of the Taiwanese capacitor plague era, and had a few leaking 1000µF caps. I ended up recapping the board, and stability improved quite a bit.

I'm surprised by how different the mosfet and capacitor layout around the cpu socket is between the two board revisions.

Yes, caps are all looking fine to me.

I have now flashed the latest beta BIOS (v. 3.83) and successfully tested with a K6-2+ 550 MHz, which gets detected correctly 😀
I was a bit hesitant because the later BIOS versions for this board are provided by MSI for PCB revisions 3.0 and 4.0 only, but multiple sources stated that 3.83 beta should be working on revision 2.1 as well - and I can confirm it does. However, this PCB revision won't allow for VCore below 2.1V, which is 0.1V more than the K6-2+ would need, but in my experience that's no problem at all. I would have raised the VCore anyways so the CPU will run stable at 600 MHz. But I will take out the multimeter again and measure if the voltages are actually correct after swapping out the MOSFET, just to be sure. Don't want to ruin a rare K6-2+...

1992 - i486DX2-66 // 1997 - P1-233 MMX // 1998 - P2-350 // 2000 - P3-650 // 2001 - Athlon 1400 // 2003 - Athlon XP 3200+ // 2008 - Xeon E5450 // 2015 - Xeon E3-1240v5

Reply 11 of 13, by bloodem

User metadata
Rank Member
Rank
Member

"Yes, caps are all looking fine to me."

It should be noted that caps can look fine, but still be completely damaged 😀 In fact, this is the case most of the time. An ESR meter is the single best way to test them (especially in-circuit).

4 x Socket 3 / 3 x Socket 7 / 4 x Super Socket 7 / 4 x Slot 1 / 2 x Slot A / 5 x Socket 370
3 x Socket A / 1 x Socket 478 / 2 x Socket 754 / 3 x Socket 939 / 2 x LGA775 / 1 x LGA1155
Current rig: AM4 - Ryzen 5 3600X
Backup rig: LGA1151 - Core i7 7700k

Reply 12 of 13, by Tetrium

User metadata
Rank l33t++
Rank
l33t++
kepstin wrote on 2020-07-16, 04:58:

At least your MS-5196 board appears to have better caps than my (rev 4.0) did, mine was straight out of the Taiwanese capacitor plague era, and had a few leaking 1000µF caps. I ended up recapping the board, and stability improved quite a bit.

I'm surprised by how different the mosfet and capacitor layout around the cpu socket is between the two board revisions.

My MS-5169 also had at least 1 bulging cap when I handled it a couple weeks ago (moving stuff) and I already suspected a bad batch of caps because I had bought that board a long time ago and ended up never using it because while installing an MII I caused too much stress on one of those socket tabs and it broke.
I was kinda surprised as the caps plague only really started getting noticeable at the sA/P4 era.

Whats missing in your collections?
My retro rigs (old topic)
Interesting Vogons threads (links to Vogonswiki)
Report spammers here!

Reply 13 of 13, by kepstin

User metadata
Rank Newbie
Rank
Newbie
gex85 wrote on 2020-07-16, 07:02:

I have now flashed the latest beta BIOS (v. 3.83) and successfully tested with a K6-2+ 550 MHz, which gets detected correctly 😀

Interesting! I ended up not using the V3.83 BIOS on my Rev 4.0 board, since it did not correctly detect my K6-III+ CPU. (I've tried a bunch of BIOS versions, so my memory isn't entirely clear, but I think it ended up running with the on-die cache disabled, so it was extremely slow). Maybe I should try it again with a K6-2+ just to see.

But since everything is more or less working with the V3.50 BIOS (most recent non-beta with K6+ support), I'm reluctant to mess with it 😀

Tetrium wrote on 2020-07-22, 03:08:

I was kinda surprised as the caps plague only really started getting noticeable at the sA/P4 era.

There was actually a bit of overlap between the K6 and Athlon eras - Late K6 stuff was produced into the early 2000s as a low end/value option to the Athlon, which was introduced in 1999 (and Athlon XP in 2001). I guess they ended up being replaced by the Duron lineup over time.