VOGONS


Reply 40 of 85, by Horun

User metadata
Rank l33t++
Rank
l33t++

I have a few other 486 VLB boards with cache and the caps for the cache are 100nF (or 0.1uF) radial multi-layer ceramic (or Tanatalum appearing types) on both of them. They both are stamped 104 MEE from what my magnify lens shows me. So any 100nF (o.1uF) SMT cap should work.

As for cache the board excepts 32kx8, 64kx8 or 128kx8 density main cache chips. If you decide to fix it and just go for 256k total cache you are better to use 9 x 32kx8, if you use 4x 64kx8 plus one 32kx8 for TAG (still 256k) OR 4x 128kx8 plus one 32kx8 TAG (for 512k) then you have to disable the "auto" setting in BIOS cache because the AUTO defaults to interleaved using both cache banks but you will only fill one bank, also you must set Cache Read Speed to 3-2-2-2. ---> from the Archived Amptron website, the DX-6900 v1.7 is the M912 v1.7.
Best to fill both bank0 and Bank1 with cache chips just in case you loose CMOS settings and did not make a note/remember the cache setting. It will boot but hang near end of POST if cache is not set right in BIOS using just one bank.
added: attached the archived webpage zipped.

Attachments

Hate posting a reply and then have to edit it because it made no sense 😁 First computer was an IBM 3270 workstation with CGA monitor. https://archive.org/details/@horun

Reply 41 of 85, by Nexxen

User metadata
Rank Oldbie
Rank
Oldbie
ykot wrote on 2020-08-12, 22:01:

By the way, you may also want to unsolder JP10-JP36 direct jumpers and put 2x6 headers there so you would be able to configure the cache.

Absolutely, already ordered some.

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

Reply 42 of 85, by Nexxen

User metadata
Rank Oldbie
Rank
Oldbie
Horun wrote on 2020-08-12, 23:13:
I have a few other 486 VLB boards with cache and the caps for the cache are 100nF (or 0.1uF) radial multi-layer ceramic (or Tana […]
Show full quote

I have a few other 486 VLB boards with cache and the caps for the cache are 100nF (or 0.1uF) radial multi-layer ceramic (or Tanatalum appearing types) on both of them. They both are stamped 104 MEE from what my magnify lens shows me. So any 100nF (o.1uF) SMT cap should work.

As for cache the board excepts 32kx8, 64kx8 or 128kx8 density main cache chips. If you decide to fix it and just go for 256k total cache you are better to use 9 x 32kx8, if you use 4x 64kx8 plus one 32kx8 for TAG (still 256k) OR 4x 128kx8 plus one 32kx8 TAG (for 512k) then you have to disable the "auto" setting in BIOS cache because the AUTO defaults to interleaved using both cache banks but you will only fill one bank, also you must set Cache Read Speed to 3-2-2-2. ---> from the Archived Amptron website, the DX-6900 v1.7 is the M912 v1.7.
Best to fill both bank0 and Bank1 with cache chips just in case you loose CMOS settings and did not make a note/remember the cache setting. It will boot but hang near end of POST if cache is not set right in BIOS using just one bank.
added: attached the archived webpage zipped.

Thank you man!!
😀

Someone suggested 10-16V, is that the correct range?

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

Reply 43 of 85, by Nexxen

User metadata
Rank Oldbie
Rank
Oldbie

I've been looking around and it looks like those are 0.1µF 50VDC (in fact they are 0805 smd), and keeping in mind Horun's radial 0.1 and type 104.
I can try that unless anyone has something to say.

@Horun, as I don't know much on capacitance I think that anything above 5 volts will do, being the voltage of the ram chips.
For the price I can give it a go (0.70€ for 100).

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

Reply 44 of 85, by Nexxen

User metadata
Rank Oldbie
Rank
Oldbie

Can't really tell, but there is a chance the board and cache will work without these capacitors. It would still be a good idea to populate them to reduce the noise and improve system stability.

By the way, you may also want to unsolder JP10-JP36 direct jumpers and put 2x6 headers there so you would be able to configure the cache.

I unsoldered the jumpers and put some headers (pic coming). Job done but not a pro job 😀
Waiting for the capacitors to get delivered (it took forever!) and give it a go (probably in a week).
I'm going to use 5 cache chips as I have them from a socket 4 mobo.
I'll have to go with some wires as I lifted some pads and cut a couple of traces (talking steady hand...).

To do list:
1. test the board without cache: (i)- if it still works at all (ii)- benchmarks
2. test with everything soldered "no cache": benchmarks
3. test with everything soldered "with cache": benchmarks

Fake chips are needed for "bios continuity" (sounds good to me 😀) to allow boot even if none is actually installed?
Or even without any installed and more than 0 is selected will work??

Last edited by Stiletto on 2020-09-24, 21:41. Edited 1 time in total.

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

Reply 45 of 85, by ykot

User metadata
Rank Newbie
Rank
Newbie
Nexxen wrote on 2020-09-24, 10:27:

Fake chips are needed for "bios continuity" (sounds good to me 😀) to allow boot even if none is actually installed?
Or even without any installed and more than 0 is selected will work??

I wouldn't be surprised if "fake chips" are just dummies, so if they are missing, no one will notice. 😀
But if you are still running modded BIOS for the fake chips, it would be a good idea to flash it with a proper image to be able to support real cache. I can't say for sure, but at least on some of my 486 motherboards (don't remember which ones), when I removed cache chips without changing any jumpers or BIOS settings, they would still boot and run, but obviously without cache.

Reply 46 of 85, by Nexxen

User metadata
Rank Oldbie
Rank
Oldbie
ykot wrote on 2020-09-24, 17:46:
Nexxen wrote on 2020-09-24, 10:27:

Fake chips are needed for "bios continuity" (sounds good to me 😀) to allow boot even if none is actually installed?
Or even without any installed and more than 0 is selected will work??

it would be a good idea to flash it with a proper image to be able to support real cache

Absolutely!

RAM traces to Cache are good; VCC is good; GND good.
I think it's definitely going to work. If not, nice ride.

I am led to believe that 28 and 32 pin chips are just different options for cache, the 4 extra pins are redundant VCC and gnd, the data is on other pins.
At least, that's why some have 32 pins with 28 chips, maybe need extra juice. I'm ignorant on the subject, so if wrong no offence if corrected.

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

Reply 48 of 85, by bjwil1991

User metadata
Rank l33t
Rank
l33t

And cache between 512KB and 1MB uses either all 32 pins (Bank 0, Bank 1, and TAG) or mostly 32 pins (Bank 0 and Bank 1) and only 1 uses 28 pins (TAG).

Discord: https://discord.gg/U5dJw7x
Systems from the Compaq Portable 1 to Ryzen 9 5950X
Twitch: https://twitch.tv/retropcuser

Reply 49 of 85, by Nexxen

User metadata
Rank Oldbie
Rank
Oldbie

Interesting info!!

I bought 28 pin sockets. Can cut another 4 pin and solder it...
Do I need more than 256kb of cache?

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

Reply 50 of 85, by bjwil1991

User metadata
Rank l33t
Rank
l33t

If you have 64MB RAM, yes. Less than that, 256KB is sufficient.

Discord: https://discord.gg/U5dJw7x
Systems from the Compaq Portable 1 to Ryzen 9 5950X
Twitch: https://twitch.tv/retropcuser

Reply 51 of 85, by Nexxen

User metadata
Rank Oldbie
Rank
Oldbie

Back on the board.
I soldered the 100nF 50V smd caps and the cache sockets.

Now the board won't boot. Diagnostic card reads "d3 d4" "d4 d3"; cache caps are shorted when power is on.

There is one smd cap that shorts when red lead of the multimeter is put on one side but not the other, when checking continuity.
Cap C45 at the bottom righ corner of the chipset.

Any ideas? 🙁

Edit: C48 also behaves strangely. Ground and +5V do not short at power connection. PSU is OK.
Edit 2: I use tag chip 1x 32k x8 + 4 x 32kx8 (tag + 4 chips) all same size Bank 0 is populated Bank 1 is empty . It doesn't output video.

Attachments

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

Reply 52 of 85, by cyclone3d

User metadata
Rank l33t++
Rank
l33t++

Looked at one of my M912 v1.7 boards. The caps are supposed to be there though I am not sure it matters or not since you are only using 28-pin cache chips.

What happens if you remove those cache chips you have installed? You could have a bad one. You also only have 128KB cache installed right now.
What do you have the cache jumpers set to?

What does the back look like? Are you 100% sure you don't have any solder connections that are jumpered to each other that aren't supposed to be?

Yamaha modified setupds and drivers
Yamaha XG repository
YMF7x4 Guide
Aopen AW744L II SB-LINK

Reply 53 of 85, by Nexxen

User metadata
Rank Oldbie
Rank
Oldbie
cyclone3d wrote on 2020-10-09, 15:47:
Looked at one of my M912 v1.7 boards. The caps are supposed to be there though I am not sure it matters or not since you are onl […]
Show full quote

Looked at one of my M912 v1.7 boards. The caps are supposed to be there though I am not sure it matters or not since you are only using 28-pin cache chips.

What happens if you remove those cache chips you have installed? You could have a bad one. You also only have 128KB cache installed right now.
What do you have the cache jumpers set to?

What does the back look like? Are you 100% sure you don't have any solder connections that are jumpered to each other that aren't supposed to be?

I'm going to desolder the smd caps, it could be that the valure is too low. ATM they are 100nF.
Cache chips are connected to 5V. As caps are @50V my guess is that it should be ok.

I have only 8 chips, so yes, 128KB. Jumpers accordingly. Cache chips are fine, taken from another board where they work perfectly.

Back pic is on the way. It all looks ok, the damaged pins have been connected and all works as expected. Even wasted 2 hours checking all connections for continuity.

One more thing, I have a tag chips that is 32x8 but the manual says it should be 8x8 in single bank setup. Is that an issue?

The error displayed is memory size, so I don't know what's going on. (blowing brains all over the place emoticon)

Attachments

Last edited by Nexxen on 2020-10-09, 16:24. Edited 1 time in total.

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

Reply 54 of 85, by cyclone3d

User metadata
Rank l33t++
Rank
l33t++

Larger TAG chip isn't a problem.

I didn't see a reply about it. Are you still running the BIOS that was meant for boards with fake cache or did you update it to the regular BIOS?

Yamaha modified setupds and drivers
Yamaha XG repository
YMF7x4 Guide
Aopen AW744L II SB-LINK

Reply 55 of 85, by Nexxen

User metadata
Rank Oldbie
Rank
Oldbie
cyclone3d wrote on 2020-10-09, 16:19:

Larger TAG chip isn't a problem.

I didn't see a reply about it. Are you still running the BIOS that was meant for boards with fake cache or did you update it to the regular BIOS?

Fake cache (facepalm hard as Krupp chaps)
Without cache and same setup (256kb) shouldn't it be working?

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

Reply 56 of 85, by cyclone3d

User metadata
Rank l33t++
Rank
l33t++

yes, probably. All the M912 boards I have came stock with real cache so I don't know for sure. If you can post a copy of your fake cache BIOS I can try it on one of my boards and see what happens.

Yamaha modified setupds and drivers
Yamaha XG repository
YMF7x4 Guide
Aopen AW744L II SB-LINK

Reply 57 of 85, by Nexxen

User metadata
Rank Oldbie
Rank
Oldbie
cyclone3d wrote on 2020-10-09, 16:39:

yes, probably. All the M912 boards I have came stock with real cache so I don't know for sure. If you can post a copy of your fake cache BIOS I can try it on one of my boards and see what happens.

I don't have a copy of it, unfortunately.

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

Reply 58 of 85, by Nexxen

User metadata
Rank Oldbie
Rank
Oldbie

Problem solved, one pin actually broke between the socket and the hole of the socket on the motherboard (fractured).
It was probably making contact when tested and broke when inserting the cache chip.

Tiredness doesn't help when checking things.

Board boots without cache and 256KB selected via jumpers. Tomorrow I'll do the test to see if I need a bios upgrade.

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

Reply 59 of 85, by Nexxen

User metadata
Rank Oldbie
Rank
Oldbie

I installed 256kb of cache in 9 chips.

If disabled message at post doesn't appear, if set to enable it does.
Chkcpu states there's no cache when enabled.

Tomorrow I'll check how to test if there's cache with another program. Hope it's not a bios limit because of fake cache.

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K