ediflorianUS wrote on 2025-03-17, 13:31:

I am sorry to pull up this topic , but I want confirm Similar Issue on (my) old ECS K7SOM+ 5.x /5.2 ...-forgot the version 11/19/2002
* After changeing my low PC2100 ddr to a SDR Kingsone 133Mhz module , system with a Socket Duron 1200 in it , powered/booted/posted up. (before that I got black screen of death) - I almost started a recap for no reason (caps seeme fine I checked them visually).

Nothing to be sorry about. If anything, thank you for adding more information / further confirmation that this is an actual issue.
FWIW, I haven't pulled mine out of storage yet to try and play with it to see if this issue can be remedied by a newer BIOS.

ediflorianUS wrote on 2025-03-17, 13:31:

Thank you. (so keep in mind the ECS stuff with DDR+SDR , with a duron , test with SDR first ,I am not even sure if a duron suports DDR, could not find refrence on CPU-world).

Socket A / 462 CPUs don't have the memory controller onboard like the next gen AMD did (socket 754/939 and onwards.) Therefore, the type of memory that can be supported by the motherboard is completely down to the chipset - particularly the Northbridge, as this is where the memory controller resides, thus being the link between CPU and RAM.
Looking at pictures of your specific board (v5.2 and not v7.x, which is a completely different animal with DDR slots only), I don't see any jumpers to have to put on / remove for selecting between DDR and SDRAM... so all of that appears to be done automatically at the time of BIOS POST.
As such, this is why I think an updated BIOS might solve the issue here.
Before recommending that, though, I really need to try that on my board first.

ediflorianUS wrote on 2025-03-17, 13:31:

- IF anyone knows exactly wtich cap I need to change , (for no issues in the future) , just let me know - I don't want to change all the caps if they are not bulged and system P.O.S.T.'s.

Probably the caps for the CPU VRM first, as that's the only buck-regulated rail on these ECS boards, it seems (and thus the rail to stress caps the most.) That would be the caps between CPU socket and rear I/O port area.
Looking at pictures particularly for the K7SOM+ v5.2 (as v7.x appears to be quite a different board), I don't see any other buck-regulated rails, so only the caps mentioned above would be the most critical to change. Of course, if you do have the caps, maybe change the ones around the RAM slots and Northbridge too... though that pretty much most of the caps on this board. 🤣
In terms of what's more likely to fail: the green G-Luxon caps are definitely more likely to fail and tend to be more unstable over time, IME. If you have these on the CPU VRM, definitely replace them. On the other hand, if you have purple OST RLX, those seem to do relatively OK on these older ECS boards (but on newer ECS boards from the socket 939/AM2/AM3 era - replace on sight!... especially anything in 8 mm diameter.)

kotel wrote on 2025-03-17, 16:24:

But this aside how exactly did the black screen of death appear? Like after what amount of time, what were you doing with it etc.

Well, for my K7S5A, I have to do exactly nothing - the board simply just does not POST with DDR RAM installed.
It's been a while since I tested it, but from what I remember, I was getting POST codes on my POST card with DDR RAM installed. It would just hang at POST code D-something (IIRC) and that was that.

I use this board regularly and it did have random no-POST (after restart/power-on) while I was first testing it (enough power cycles would boot it). The juices were leaking out the tops of the CPU VRM caps after I got Windows 98 installed (and I ran 3dmark99, successfully). It hasn't had any issues since re-capping (unless you count my PCI IDE failing)

I love sharing the pile. That board uses a LOT of caps...

I had been just reminded about my first msi 845pe which didn't work, then worked for a tad bit and then died. The whole story is just too complicated to say so I summarized it in one word.
That might have to do with the chipset since it doesn't get any warm.

Maybe here the VRM or whatever ecs used here for the NB/SB got screwed and there's no voltage going to it? Also if the board does come out of reset check if the CPU VRM is generating proper voltages.