Is it performing as if there is no cache? What are your 3dbench and pcpbench scores? Also doom (use Phil's DOS benchmark kit at www.philscomputerlab.com).

Sounds like the cache jumpers are incorrectly configured or of the chips is damaged or missing. Make sure the chips are correctly oriented and pushed all the way into the sockets. It would also help to recheck if all of the pins made it into the sockets - more then once I noticed one leg of the cache chips did not make it into the socket.

Hi,

clueless1 wrote:

Is it performing as if there is no cache?

Yes it is exactly performing as if there is no L2 cache installed, even though on the BIOS post screen, it shows the amount that i set up with the jumper block.

On cachechk it reports "it seems theres no L2 cache installed - check for fake cache"
On speedsys there is only one bump, in the region below 8kb (internal cache) after that there is a flat line. Also it mentions that there is no L2 cache.

Regarding the jumper settings- i triple checked them by now.
Regarding the IC pins - i checked them, the ICs are seated and all pins are in.

Here i took a picture of my configuration which is 256kb, 4x 64kb chips.

I got 2 sets of chips (i originally wanted to upgrade the board to 512kb), both sets get me the same result - that is no L2 cache is recognized.

here are some more pictures

i hope anyone got another idea what could be wrong here.

thanks in advance,
srm.

Move JS2 to 1-2?

I agree with Tetrium. JS2 should be in 1-2 according to the manual.

Tetrium wrote:

Move JS2 to 1-2?

oops... yes that jumper was indeed wrong. 🙁
But it still wont work

i saw something really interesting.

on the pic from the mainboard you can see a free smd space where it reads "555". There is a smd resistor soldered between pins 1-2 that really looks odd. iam really not sure if this is some kind of manufacturing mistake or not...

srmeister wrote:

i saw something really interesting.

on the pic from the mainboard you can see a free smd space where it reads "555". There is a smd resistor soldered between pins 1-2 that really looks odd. iam really not sure if this is some kind of manufacturing mistake or not...

I was about to mention this...but I don't know.

I don't think I have this board so I can't check myself. But iirc it was a reasonably common 486 board so another Vogons member may have one

here is a close up of the odd smd part.

looking forward to hear your opinions

srmeister wrote:

here is a close up of the odd smd part.

looking forward to hear your opinions

I don't know for sure, but I wouldn't exclude the possibility that this was a factory fix. SMD has the same color as the ones near it and repairs of this sort were more common in those days.

.
LM555 is a timer and it looks like the SMD is across what would be pins 3 and 4 which would be Output and Reset.
http://www.ti.com/lit/ds/symlink/lm555.pdf
.
Pic of a Ver 1.1 showing LM555 populated. http://www.elhvb.com/mobokive/archive/Microst … MSI_MS-4144.jpg
.
Pic of a Ver 1.5 with no LM555 and no SMD. http://www.amoretro.de/wp-content/uploads/msi … motherboard.jpg
.

Try dropping the RAM to 32MB - if the cache is in writeback, that would be the maximum cacheable

From the pic you posted it seems like the cache chips are set to bank 2? From what I can tell, the cache sockets are numbered V9, V11, V13, V15 and V19, V24, V28 etc. I see you have the cache chips plugged into the sockets with higher values. Try moving the cache chips in the other sockets, or try using 8 cache chips if you have them.

kanecvr wrote:

From the pic you posted it seems like the cache chips are set to bank 2? From what I can tell, the cache sockets are numbered V9, V11, V13, V15 and V19, V24, V28 etc. I see you have the cache chips plugged into the sockets with higher values. Try moving the cache chips in the other sockets, or try using 8 cache chips if you have them.

I checked this yesterday and according to the manual the cache chips are seated in bank 1 (but I do agree with you it is odd practice to have the lower bank indicated with higher numbers, but strange things happen more often so it might still be the correct one). But as there seem to be problems remaining (and as sometimes manuals simply lie, especially the ones from those dark ages) he might as well give it a try in the other bank provided it won't hurt any of the components.

Trying won't hurt. At this point it's either that, or one of the chips is damaged. Or the chips size is smaller then advertised.

Thanks for your tips, problem is solved.
I reduced the RAM to 8MB - nothing.
I put the cache chips to bank 2 - nothing.

Then i thought i should read the f*ing manual page by page and check EVERY damn jumper on the board.
So i came across the CPU jumper settings page (see below) and the mainboard was jumpered as in the photo

so i put in the 3 more jumpers to JC5 and JC6 and now it works (see screenshot of cachechk).
JC8 is for the 1x or 2x CPU frequency (DX or DX2) see right page in the manual below. This at least was set correctly 😁

Thanks for all the help. Should have read every page of the damn manual earlier 🙁
i bought the machine with 256kb cache installed and with the CPU installed. I didnt change anything cpu related. So the mainboard+cpu did never run on full speed as it does now.

Next thing im gonna do is go for 512kb cache and maybe 40mhz FSB 😀

THANKS again!!

Great! Glad you got it solved. These experiences not only help you but all of us who read along. I think mistakes likes this are wonderful learning opportunities. 😀

Nice to see it's been solved and the board is working, yw 😀