So when you say you keep it on the "back plane", do mean you keep the board mounted to a removable tray that is part of the case?

Do you really mean "random beep codes"? Or that you just don't recognize it? If it's reproducible, then it's not really random.

EMI is a thing. What is the motherboard, and do you have pictures of the tray and case where it doesn't boot?

There could also be some mechanical contortion forcing an IC or the video card to be slightly misaligned when fully mounted.

Agreed, my guess would be the devices going into the pci/agp/pcie slots. Most likely the video card not seating all the way. Put the motherboard back in the case, Take the back face off of the card so the only thing interfacing with the card is the slott. Be sure it is fully inserted and give it a shot.

Yes, the tray. Couldn't find the right word for it. Gonna edit original post.

I'm gonna also record the beeps and show the motherboard inside the case later today. When I say "random" I mean that don't match any pattern I've found but is always the same.

The motherboard is a Gigabyte GA-7ZX. But the manual doesn't mention the beep codes.

I'm really sure that the video card (an AGP tnt2 m64) is fully in, I tried several times. But I guess I can try remove the slot bracket to be sure it is.

mrfusion92 wrote on 2023-05-25, 06:17:

Yes, the tray. Couldn't find the right word for it. Gonna edit original post. […]
Show full quote

Yes, the tray. Couldn't find the right word for it. Gonna edit original post.

I'm gonna also record the beeps and show the motherboard inside the case later today. When I say "random" I mean that don't match any pattern I've found but is always the same.

The motherboard is a Gigabyte GA-7ZX. But the manual doesn't mention the beep codes.

I'm really sure that the video card (an AGP tnt2 m64) is fully in, I tried several times. But I guess I can try remove the slot bracket to be sure it is.

How is the cooler secured...socket tabs or thru holes? - is it possible that the tray + mb is warping somehow when fitted to the case, causing bad contacts somewhere (tho hopefully not solder cracks!)

mrfusion92 wrote on 2023-05-24, 21:39:

I have this full ATX case with detachable tray where the motherboard is screwed in. […]
Show full quote

I have this full ATX case with detachable tray where the motherboard is screwed in.

So I screwed the motherboard on the tray, and then placed inside the case. Connected the power supply, a video card and... no POST with random beep codes.
I'm not kidding the sequence of beeps doesn't match any manual I've found online. 9 short (I think at least, they are very short and quick) then 3 long. It is an AMI BIOS. No other devices connected, just the video card.

Removed the motherboard from the case, keeping it on the tray, tried with same power supply and video card... and it POSTs okay.... Back inside the case and I got again only the beeps.

I did inside\outside the case three times. Out works, inside not.

I don't understand at all what it can be. A short somewhere? But the motherboard is always on the tray even outside the case.

Googling "9 short beeps" gave me this:

Nine short beeps mean that there has been an AMIBIOS ROM checksum error.

If the board works fine outside of the case, but not when mounted inside the case, then the problem probably lies with the interaction between motherboard and case.
Does your tray have any standoffs or are they missing? (in some cases they are always missing)
Are the standoffs all of the same and of the correct length?
Is there a standoff that is shorting the board? (this can happen if a standoff is placed where the motherboard doesn't have a corresponding hole).

mrfusion92 wrote on 2023-05-24, 21:39:

I don't understand at all what it can be. A short somewhere? But the motherboard is always on the tray even outside the case.

Ok, I must have overread this bit.

At this point, I'd suggest to try another known working motherboard inside this case and see if it still doesn't work.

If the above 2 are not the issue, another thing you could try is to take the mobo+tray combo outside of the case, but then not lay it flat on your testbench but holding it up horizontally and then try to start it up.
It may be causes by some kind of mechanical failure somehow (cold solder joint, some kind of warping or mechanical stress or a short that's only afflicting the board when the board is sagging a bit when inside the case or something)

I'd suggest putting your mobo+tray on your workbench and having the combo lean against a wall or something and then powering it up and see if it still works or then doesn't work. Don't need to have any PCI/AGP cards in for this test, just the basics (CPU + mem).

Tetrium wrote on 2023-05-25, 08:39:

I'd suggest putting your mobo+tray on your workbench and having the combo lean against a wall or something and then powering it up and see if it still works or then doesn't work. Don't need to have any PCI/AGP cards in for this test, just the basics (CPU + mem).

I thought similarly, he can try with the case lying on its side as a first easy step to see if it's something mechanical when the board is positioned vertically.
Although I'm leaning more towards the "something shorting" idea, only because it's difficult to reproduce with accuracy a marginal contact.

konc wrote on 2023-05-25, 10:37:

Tetrium wrote on 2023-05-25, 08:39:

I'd suggest putting your mobo+tray on your workbench and having the combo lean against a wall or something and then powering it up and see if it still works or then doesn't work. Don't need to have any PCI/AGP cards in for this test, just the basics (CPU + mem).

I thought similarly, he can try with the case lying on its side as a first easy step to see if it's something mechanical when the board is positioned vertically.
Although I'm leaning more towards the "something shorting" idea, only because it's difficult to reproduce with accuracy a marginal contact.

Good idea. Btw can test both methods. That way (if it fails to unearth the root cause of these issues) we can at least get a better picture of how consistent this error is.

You all made the assumption (rightfully) that when is inside the case I tested it in vertical position... but actually not, It was always in horizontal.

Tetrium wrote on 2023-05-25, 08:30:

Googling "9 short beeps" gave me this:

Nine short beeps mean that there has been an AMIBIOS ROM checksum error.

Yes I have read that but it doesn't do only the 9 beeps, after a brief pause it makes 3 other beeps.

I was in hurry during lunch break so I could make only these three pictures. After work will do more.

The tray

I'm using three plastic standoff for the rightmost holes, here the motherboard POSTed correctly

From the back inside the case, didn't had time to test it again

I got defeated by a case

Johny Case wins!
FATALITY!

On more serious note, you should get postcode card.

There is also the off chance that the IO bracket (shield) might be shorting on one or more of the connectors.
I have had that happen when in haste on newer PCs(USB for example).
Granted this would be unlikely on an older motherboard and would be dependent on the IO bracket being molded or snapped into the case.
But it doesn't appear in your photo that the IO bracket is attached to the motherboard when it's out of the case.
Just a thought.

So I do wonder about the quality of the motherboard screw mounts, as I have not seen square ones myself. Do they move when you wiggle them? And with them being squarish, I'd check all of them if they with their screws their metal seem to come close in contact to any solder pads on either side, where smaller rounder mounts were expected.

I did say EMI can be an issue. When I remember this happening back in the day, my memory serves that it generally happened with CPU speeds were around 200-300 MHz, which is around the socket 7 era. You have a socket A, which would make me think you do not have the problem, as I had not seen it with socket A. But you do have a very similar motherboard chipset (686a) that was used in super socket 7 too. Perhaps my original thoughts blaming the CPU should have been placed on the FSB speed, and certain board layouts and cases of the era.

I was thinking like others the orientation could cause it. That is, if placing it horizontal while inside the case triggers the issue, and outside the case on the tray does not, and it wouldn't matter if you placed it vertically. Then as another idea you could also try to place the case cover underneath the tray, as if it is in the case, but without the whole frame. That could prove interesting.

Of course as other people say, the I/O shield or something about mounting / not mounting the AGP video card to the case frame with the slot cover could be a clue. I could suggest trying another known working PCI card if you have not already attempted.

I had something similar with MSI 6309 socket 370. The board lying on the table with the components both ran and functioned normally, so it was fitted to the case. Nothing happened after it was turned on - no post, no beep. I took it out of the case again, tried it on the table, and started running again. Finally, the solution was stick duct tape around the holes in the board and add insulating pads to screws .
The board mounted in this way works in case without any problems.
I don't understand it - the board is still connected to the case via the I/O backplate, and I've never insulated the distance standoffs on other boards, and the pads under the screws were last used by me about 15 years ago. Weird problem.

I don't know how relevant this is for your case (pun intended!), but I was puzzled once by a... broken switch, a stuck reset switch. Of course this can only be a possibility to consider if you are not connecting the switches when outside of the case but you are connecting them when inside it.

Yesterday I was too busy, I couldn't do any other tests. I hope today or this weekend.

konc wrote on 2023-05-26, 08:28:

I don't know how relevant this is for your case (pun intended!), but I was puzzled once by a... broken switch, a stuck reset switch. Of course this can only be a possibility to consider if you are not connecting the switches when outside of the case but you are connecting them when inside it.

And we have a winner!

Well it wasn't the switch but this pointed me in the right direction since it was.... the speaker! Yes when the speaker is connected the motherboard doesn't post. Tested three times because I really couldn't believe it.

And indeed outside the case I NEVER tested with the speaker but inside the case I ALWAYS plugged it.

I'm really having a good laugh out of it.

You got defeated by a PC speaker ))
jk

mrfusion92 wrote on 2023-05-26, 17:34:

And we have a winner! […]
Show full quote

konc wrote on 2023-05-26, 08:28:

I don't know how relevant this is for your case (pun intended!), but I was puzzled once by a... broken switch, a stuck reset switch. Of course this can only be a possibility to consider if you are not connecting the switches when outside of the case but you are connecting them when inside it.

And we have a winner!

Well it wasn't the switch but this pointed me in the right direction since it was.... the speaker! Yes when the speaker is connected the motherboard doesn't post. Tested three times because I really couldn't believe it.

And indeed outside the case I NEVER tested with the speaker but inside the case I ALWAYS plugged it.

I'm really having a good laugh out of it.

hehehe this is how you learn once and for all when troubleshooting to really remove everything from the equation no matter how insignificant or irrelevant it might seem at that moment. Pfff a tiny reset switch or a tiny speaker so I can hear the beeps, yes let me connect this it shouldn't affect anything 😀 Glad you solved it.