VOGONS


First post, by kotel

User metadata
Rank Member
Rank
Member

Hi

Finally decided to post about this MSI MS-7046 VER.1. It was working before I got my hands on it. My attempts to remove the BIOS landed with some caps being knocked off (I hate myself for destroying it, already replaced all I could find). Then I got some parts for it. At first boot it was sometimes displaying a black screen. Then I got an great idea to put an 1.8GHz celeron inside (it doesn't support it, did it because I thought if optiplex 330 supports celerons and C2D's, it will support them too (it was before I knew about searching stuff)). After plugging it in, I was greeted with an burning smell. All caps under the CPU look good, not sure about the VRM since its coolers are soldered onto the board. CPU is fine. After some time I finally got my POST card. Not to my surprise there weren't any codes executed. Poking around the NB I found that some of the caps had 0 ohms (will try to post pics of them later today). Some caps near the P4 connector had 10-8 ohms. Also found out that the NB is getting scorching hot after some seconds and the CPU gets hot (around 60-70 C while measuring by hand) after around 30-40 sec (yes, I know this is a LGA775 P4 3.06GHz, but I don't think it should get that hot). So now the question is, what is at fault? The NB or some other components? Already tried different CPU's (P4 3.06GHz and 2.80GHz, celeron 1.8GHz), PSU's and CMOS batteries.

"All my efforts were in vain...
Let that be my disappointment."
-Kotel

Reply 1 of 7, by Karbist

User metadata
Rank Member
Rank
Member

This board has 2 switching buck converter , the one near the pci-e generates 1.5v and the one under the ram slots near the atx connector generates 1.8v,
you need to measure the resistance between the output inductor and ground and see which rail is shorted.
there is a chance that the low side mosfet is shorted.

Reply 2 of 7, by kotel

User metadata
Rank Member
Rank
Member
Karbist wrote on 2024-07-30, 12:58:

This board has 2 switching buck converter , the one near the pci-e generates 1.5v and the one under the ram slots near the atx connector generates 1.8v,
you need to measure the resistance between the output inductor and ground and see which rail is shorted.
there is a chance that the low side mosfet is shorted.

I couldn't find those buck conventers. There are only 2 EEPROMS and an amplifier near PCI-E and under the ram there are only transistors. Measured all chokes to ground and only CHOKE2 has around 30 ohms to GND. Is there any way to check if the low side mosfet is shorted without desoldering the VRM coolers?

"All my efforts were in vain...
Let that be my disappointment."
-Kotel

Reply 3 of 7, by Karbist

User metadata
Rank Member
Rank
Member

I marked the inductors, if they are not shorted to the ground, turn on the board and make sure 1.8v and 1.5v are present on these inductors.

The attachment ms-7046.jpg is no longer available

Reply 4 of 7, by kotel

User metadata
Rank Member
Rank
Member
Karbist wrote on 2024-07-30, 18:10:

I marked the inductors, if they are not shorted to the ground, turn on the board and make sure 1.8v and 1.5v are present on these inductors.

The attachment ms-7046.jpg is no longer available

The 1.8v inductor has 2.7V while the 1.5v inductor has 1.7V. I have no idea what could cause this. Any ideas?

"All my efforts were in vain...
Let that be my disappointment."
-Kotel

Reply 5 of 7, by Karbist

User metadata
Rank Member
Rank
Member

I made a mistake, this is a DDR1 board, so 2.7v is correct for the memory rail. 1.7v on the other rail is fine too.
sorry I ran out of ideas.

Reply 6 of 7, by kotel

User metadata
Rank Member
Rank
Member

Do you think that 5.8 ohms on inductors near CPU (on an socket 478 based mainboard) is a short or not? Didn't want to make another post about my problem in this forum.

"All my efforts were in vain...
Let that be my disappointment."
-Kotel

Reply 7 of 7, by Karbist

User metadata
Rank Member
Rank
Member

If there's a cpu in the socket, 5.8 ohms is the cpu internal resistance. you should remove the cpu and measure again.