Confirmed the memory is 3v https://www.datasheets360.com/part/detail/tc5 … 65173905008436/ so not sure what the problem could be.

Looks like other people have my exact issue;

M919 Cache stick not working
PCChips m919 cache stick not working

Maybe the M919 needs a very specific type of FPM ram for the cache to work?

Both of my 919s have displayed this behavior at one point or another while messing with them. The 919 first detects a module, and reports the capacity. Not untill AFTER that point does it do a check on that module. If it fails for whatever reason, it often seems to"silently " disable it in subsequent boots without indicating this in any way

Try: boot with module installed. Enter bios, disable external cache. Save and exit. Shut system down, remove module. Start system without module, let it complete post. Shut down. Reinstall module. Power on, go in to bios and enable external cache. Disable any automatic cache speed setting and just try the slowest. Save and exit. Let it complete post, and see if it works now. If it is stubborn, try to combine the above with swapping ram in to a different slot.

Note: if cache speed is set "too slow" compared to ram the board may ignore it. If this is the case then bumping up the cache speed should make it show up in tests.

I currently own three M919 v3.4 B/F motherboards and one of the boards exhibits the issue noted above. My tests confirmed that the issue was unrelated to the cache module, the RAM, or the CPU because I switched over the same parts on all 3 boards. My two primary theories are,

1) some contacts on the cache socket are bad/intermittent, although when I peer inside with 10x magnification and a torchlight, the contacts look fine.

2) there's a via at, near, or around the cache socket which has an intermittent connection.

These are my two main theories because if I remove the cache socket, then move it in and out vigorously for 10 seconds, this seems to get the job done. Until the next day, after letting it rest a day or two, I have to do the ole in-and-out all over again.

Given how flimsy and thin PC Chips mainboards are, and how much force is required to move the cache module in and out, I'm a more inclined to believe there's a bad/intermittent via or joint somewhere near the cache slot. To further backup this claim, this motherboard also had an issue whereby the jumper that controls L1: WB/WT wasn't sending the WB/WT' signal to the CPU. Am5x86 was always in L1:WT mode. All the traces looked fine and I reflowed some joints, but I ultimately had to jumper a wire from the appropriate motherboard jumper to the WB/WT' on the CPU to fix the issue.

I have tried reflowing the entire cache slot, but this did not help the issue. The problem being intermittent makes testing more challenging. It was for this reason that I ultimately acquired a 3rd M919 (its a spare part for my Am5x86-180 system).

Thank you to both bertrammatrix and feipoa.

I removed the cache stick multiple times and used contact cleaner. No luck. I then did the following;

1. Removed the cache stick
2. Removed the battery
3. Set the CMOS power discharge jumper for 20 seconds the set it back again.
4. Reinstalled the battery
5. Powered on the machine, CMOS error appeared (because of previous CMOS discharge / no battery)
6. Entered BIOS setup
7. Disabled the External cache
8. Disabled the auto config for memory settings and set slowest settings
9. Saved settings and booted into DOS
10. Power off the machine
11. Installed cache stick
12. Power on the machine and entered BIOS settings immediately
13. Enabled External cache
14. Saved settings and booted into DOS

BINGO, cache is now working. Woohoo!!!

I suspect there is some hidden cache configuration that is only set on a total BIOS reset.

I then maxed the memory settings and did some benchmarks.

I'm now in the Quake 17 FPS club 😉

Without removing the cache stick, wait a few days and see if the cache is still working. I have a strong feeling that we haven't heard the last of this issue.

zuldan wrote on 2025-04-22, 10:15:

Thank you to both bertrammatrix and feipoa. […]
Show full quote

Thank you to both bertrammatrix and feipoa.

I removed the cache stick multiple times and used contact cleaner. No luck. I then did the following;

1. Removed the cache stick
2. Removed the battery
3. Set the CMOS power discharge jumper for 20 seconds the set it back again.
4. Reinstalled the battery
5. Powered on the machine, CMOS error appeared (because of previous CMOS discharge / no battery)
6. Entered BIOS setup
7. Disabled the External cache
8. Disabled the auto config for memory settings and set slowest settings
9. Saved settings and booted into DOS
10. Power off the machine
11. Installed cache stick
12. Power on the machine and entered BIOS settings immediately
13. Enabled External cache
14. Saved settings and booted into DOS

BINGO, cache is now working. Woohoo!!!

I suspect there is some hidden cache configuration that is only set on a total BIOS reset.

I then maxed the memory settings and did some benchmarks.

The attachment Memory.JPG is no longer available

I'm now in the Quake 17 FPS club 😉

The attachment Qauke.JPG is no longer available

The attachment Speedsys.JPG is no longer available

The attachment CTCM.JPG is no longer available

The attachment CACHECHK.JPG is no longer available

Good to hear you got it working. These 919s often require a good dose of alchemy, persistence and LUCK to get working properly

feipoa wrote on 2025-04-22, 13:27:

Without removing the cache stick, wait a few days and see if the cache is still working. I have a strong feeling that we haven't heard the last of this issue.

It's been a week. Does it still work (without removing/reinserting cache stick)?

feipoa wrote on Yesterday, 06:48:

feipoa wrote on 2025-04-22, 13:27:

Without removing the cache stick, wait a few days and see if the cache is still working. I have a strong feeling that we haven't heard the last of this issue.

It's been a week. Does it still work (without removing/reinserting cache stick)?

Let me test, give me 5 min

feipoa wrote on Yesterday, 06:48:

feipoa wrote on 2025-04-22, 13:27:

Without removing the cache stick, wait a few days and see if the cache is still working. I have a strong feeling that we haven't heard the last of this issue.

It's been a week. Does it still work (without removing/reinserting cache stick)?

Working like just out of the factory!

Maybe revision 3.2 of this motherboard behaves differently?

Hopefully it stays fixed. Clearing the CMOS seems too simple.