VOGONS


First post, by FidoMike

User metadata
Rank Newbie
Rank
Newbie

Hello all, I am having an odd issue with two Pentium boards that I am totally stuck with. Really hope someone can help with this.

I have a DEC Venturis 575, from what I can tell the model number means 5 for pentium and the rest is the speed of the included processor, therefore it should be a Pentium 75 which the BIOS confirms.
On power-on, it complained of a dead CMOS battery, it has a socketed Dallas DS12887. It wanted me to replace and enter the BIOS and correct the values etc. Before replacing the RTC, I tried anyway to get it to do anything, I wanted to hook up a CD ROM and / or an 120GB HDD, on trying to detect either of these, the BIOS would just hang. I Initially thought it was just an old BIOS and those two drives might be too new, so I didn't worry much. I took those out and tried just a floppy drive, the drive that came with the PC seems to be faulty, so I swapped it out for another I know works.

I put in a Windows 98 boot disk that I use it for loading CD drivers etc for DOS games, the disk works in other PCs but on this board it seems to read the disk, I can hear 2 sets of the normal floppy access sounds, then a load of symbols and what looks like a part of exe file is dumped to screen and it stops. Very much like if you were to try to print out an exe file. The only words are 'FAT12' repeated several times. Oddly it hasn't frozen, a Ctrl + Alt + Del will reboot it. At this point I thought it was the disk gone bad, so tried an other, same problem. I then tried same drive and disk in an other PC (a DEC 486) and it was fine with it. I then assumed it was due to the Dallas RTC, and therefore ordered a new one from RS online. I replaced the DS12887 with a DS12887+, powered on, entered BIOS, loaded defaults with F9, saved it. Tried to boot, it no longer complains of dead CMOS battery, seems to hold the time fine, but it wont boot, same mess on screen as before. I also tried detecting the HDD or CD drive again, and it still hangs.

I tried lots of different things in the bios settings, disabled shadows/caches or anything it offered, I also discovered that the 'F9' load defaults only applies to the page currently displayed, so I even tried going to every page and every sub page loading defaults and saving. I even tried swapping the floppy drive to the end of the cable, it didn't like that, but when I told it to swap the drives, it seemed to accept it, but again none of these made any difference. I did also try the 'clear config data' under the advanced / plug n play area. Still nothing.

At this point I was starting to think the board was dead, so I took out the floppy drive, and its cable, put that into the DEC 486, and it worked fine. I then put the drive and cable back in the pentium, disconnected its power supply and used the DEC 486's one, still same mess on screen and no boot. I tried removing the risa card, swapping the CPU for a pentium 133 I had separate, still same.

I was now convinced it was a board problem, so I went on eBay and got a used but tested replacement, same part number as my existing board. I was pleasantly surprised it came with CPU and 40MB RAM. I decided to not do anything to it, and put the board in exactly as is with the CPU/RAM. It powers on, BIOS says its a Venturis 5120, and confirms a Pentium 120. It complains of dead CMOS battery, ok not unreasonable, and I have to load defaults in BIOS, and try to boot. Same output as before, corrupted mess with the FAT12 being only item recognisable. To be clear, I did not try replacing the Dallas chip on this board.

As this looks like the same issue as my original board, I then went back to my original and tried the password clear jumper, usually when I try these the boards wont post until it is removed again, but this started up with no change. I tried the jumper next to it which said something on the lines of recovery, this appears to be a BIOS recovery/flash feature, nothing on screen and tries to access the floppy. I stopped at that point as I don't have a BIOS floppy for it. I even tried the mysterious 'FACTORY SETTING' jumper near the AT power connector, seemed to do nothing, even after putting it back.

So in summery, two boards, both have this odd problem, and I think I have tired everything I can think of on the original board. I haven't done much on the replacement, as it seems to be an identical issue, and I don't want to do too much as it might get returned. I say might, as the price I paid for it, I might keep it for the RAM and CPU regardless of the board.

Any help or ideas would be very greatly appreciated.

Reply 1 of 11, by dragonkn

User metadata
Rank Newbie
Rank
Newbie

Maybe there is a way to disable CMOS/Dallas,
some board got this option
if you got nothing to lose I would try to set different combinations on J16 jumpers
I know it's a low chance, but maybe worth a try

Reply 2 of 11, by FidoMike

User metadata
Rank Newbie
Rank
Newbie
dragonkn wrote on 2021-09-13, 20:02:
Maybe there is a way to disable CMOS/Dallas, some board got this option if you got nothing to lose I would try to set differen […]
Show full quote

Maybe there is a way to disable CMOS/Dallas,
some board got this option
if you got nothing to lose I would try to set different combinations on J16 jumpers
I know it's a low chance, but maybe worth a try

Thanks for the reply!
It seems the two boards I have don't have the pins for J16, it is marked on the board and has the solder pads but no jumper block. I did try shorting the pads by holding a pair of tweezers on the pads to simulate a jumper being there, but in either position it didn't seem to do anything. I also noticed that several other components near that jumper are missing, which might mean whatever circuit it would activate isn't present. According to the service manual I was reading, J16 seems to be only present on the V1 boards used on the Pentium 60 models, these also appear to use a 3v button battery instead of a dallas chip. The boards I have do have markings for a battery holder where the dallas socket is fitted, but I don't know if simply removing the socket and putting a batter holder on would work, I suspect other parts would be needed.

A bit more info I found out, it seems the symbols and letters displayed when it tries to boot, is actually the boot sector, I saw an image from some utility that shows the boot sector, and it the contents appears to match what I see at boot time. No idea if that helps. Why would the BIOS display the boot sector from the floppy instead of running it?

I also tried some more tests, I tried the new dallas chip in the replacement board, again no change. I also tried simply removing the dallas chip and seeing what happens, the bios comes up, starts the ram count, complains of an error then resets, and repeats again.

Reply 3 of 11, by FidoMike

User metadata
Rank Newbie
Rank
Newbie

An update!
I decided to try a third power supply, this time it is an ATX 300W from an old Athlon 64 Compaq PC, and I used one of those ATX to AT adaptors on it. I also took a new floppy disk (its actually a 720k DD disk from a pack I bought new some years ago for my Amigia). I put the disk in the old working 486, ran sys a: then write protected the disk and put it in the pentium 75 (original board), with the new dallas, and the 300w ATX PSU... and it booted fine. I then started trying several of the previously tried disks, and 3 out of 5 worked fine.

I then swapped back the old PSU, tried 2 out of the 3 working disks - old problem of showing the boot sector. OK seems like a result. Put the new PSU back on, tried again, and the 2 disks I tried with the old PSU no longer boot, one of them shows the boot sector, the other shows the Windows ME boot menu, but then freezes. The ME disk was write protected, the other was not. I tried the DD disk again made on the 486, and it works.

I then tried the replacement board, the pentium 120, with the 300W PSU and its original dead dallas. The 720k DD disk works, ME disk shows menu and freezes, other disk shows boot sector.

So what I think is happening, the old PSU seems to be corrupting the floppy disks, but what I don't understand, what happened to the ME boot disk, that was write protected with the little tab, yet it now freezes, which suggests it got damaged somehow. Is it even possible for a PSU to damage a floppy disk like this?

I opened up the old PSU and carefully had a look inside, but i can't see anything obviously wrong, no bulging or leaked capacitors, no burn marks or anything.

Reply 4 of 11, by Horun

User metadata
Rank l33t++
Rank
l33t++

My only guess is that the old PSU has very bad ripple which could be causing the floppy to corrupt the data while trying to read the data.... just a guess.

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. Stuff: https://archive.org/details/@horun

Reply 5 of 11, by weedeewee

User metadata
Rank l33t
Rank
l33t
FidoMike wrote on 2021-09-18, 18:31:

no bulging or leaked capacitors, no burn marks or anything.

fyi,
While bulging or leaking are tell tale signs of bad capacitors,
Capacitors do not need to be bulging or leaking to be bad.

Right to repair is fundamental. You own it, you're allowed to fix it.
How To Ask Questions The Smart Way
Do not ask Why !
https://www.vogonswiki.com/index.php/Serial_port

Reply 6 of 11, by FidoMike

User metadata
Rank Newbie
Rank
Newbie
weedeewee wrote on 2021-09-18, 23:12:

fyi,
While bulging or leaking are tell tale signs of bad capacitors,
Capacitors do not need to be bulging or leaking to be bad.

Yes they could still be bad, I just don't have anything I can test them with at the moment. I am looking at some ESR metres on ebay and considering if I should get one, they seem to range quite a bit in price, but it might be useful to have one in the long run.

Horun wrote on 2021-09-18, 22:50:

My only guess is that the old PSU has very bad ripple which could be causing the floppy to corrupt the data while trying to read the data.... just a guess.

I tried hooking up the power supply to an oscilloscope and seeing if I could see anything, unfortunately at the time I didn't realise the correct method includes limiting the bandwidth to 20mhz, therefore the numbers below are going to be off by a fair bit, but it did give some interesting results. Initially the PSU must have known it was under test, as it worked perfectly (and the PC booted fine to MS DOS). It measured 128mv vpp on the 12v and 140mv vpp on the 5v while the PC was working fine, see first picture (QuickPrint2). It stayed like this for hours, I kept it running, rebooting over and over, and it refused to have a problem. I then decided to hook up a second floppy drive as I was going to try to fix one of the corrupted disks, it was powered off maybe 2-5 minutes. Then when I powered it back on, it went a bit crazy, see second and onwards pictures, the BIOS hung before the ram count and the 12v rail went up to 372mv vpp, then 640mv, then 804, then 244, then 592. This wasn't just a single spike, this was constant spikes for maybe 2 minutes or more, I even had time to grab a multi metre to check the voltage. I tried removing the second floppy drive while it was on to see if that was causing it, it didn't seem to affect the ripple at all. I turned it off and powered back on again, and it seemed to behave as it had before. I even tried re-adding the second floppy, and it seemed fine again.

Would bad capacitors alone cause this? I can understand it taking time to get warm, but this took a very long time for it to happen maybe 2 hours or more, and it went back to normal right after. I considered the floppy drive to be the cause, but that was from the working 486 pc. Unless the extra floppy was just enough extra load to push it to error. But then why did it do it right away when testing the the other times. At the time of the fault, it was powering the motherboard, a 120GB HD and 2 floppy drives. The PSU says it is 100W.

Note I did adjust scale/time a bit between some of the images, but you can see the values on the bottom of each image.

The attachment DS1Z_QuickPrint2.png is no longer available
The attachment DS1Z_QuickPrint11.png is no longer available
The attachment DS1Z_QuickPrint12.png is no longer available
The attachment DS1Z_QuickPrint13.png is no longer available
The attachment DS1Z_QuickPrint14.png is no longer available

Reply 7 of 11, by 1ST1

User metadata
Rank Newbie
Rank
Newbie

Just for your reference, this PC was made by Olivetti, originally it should be a M4 Modulo P75. So you can swap the processor up to 200 Mhz (without MMX!) . The Olivetti own version has a slightly different frontpanel, that's it.

Can you please check the on mainboard printed BAxxx number and post here?

Currently I am trying to find out which DEC Venturis / Celebris / DECstation model made by Olivetti is which Olivetti model. DEC PC (DECstation/Celebris/Venturis) to Olivetti PC series translations (in this link you find a link to download customer test and driver diskettes in the DEC versions for these machines.

That BA-Number could help me to identify the exact Olivetti model name.

Reply 8 of 11, by FidoMike

User metadata
Rank Newbie
Rank
Newbie
1ST1 wrote on 2021-09-20, 05:32:
Just for your reference, this PC was made by Olivetti, originally it should be a M4 Modulo P75. So you can swap the processor up […]
Show full quote

Just for your reference, this PC was made by Olivetti, originally it should be a M4 Modulo P75. So you can swap the processor up to 200 Mhz (without MMX!) . The Olivetti own version has a slightly different frontpanel, that's it.

Can you please check the on mainboard printed BAxxx number and post here?

Currently I am trying to find out which DEC Venturis / Celebris / DECstation model made by Olivetti is which Olivetti model. DEC PC (DECstation/Celebris/Venturis) to Olivetti PC series translations (in this link you find a link to download customer test and driver diskettes in the DEC versions for these machines.

That BA-Number could help me to identify the exact Olivetti model name.

I have two boards and one case, none have any numbers starting with BA. Here are the numbers it has got, excluding the serial numbers (they don't have a BA in them either):

Board 1:
E02 (sticker)
54-24486-01 (DEC part number) (sticker)
30095 (on the bottom of the board in the silkscreen with the digital logo).

Other board:
E02 (sticker)
54-24486-01 (silkscreen / pcb)
50-24485-01 (silkscreen / pcb)
D01 (silkscreen / pcb)
54-24486 (sticker)

Case:
FR-920WW-CB (model) - note this is a Pentium 75, also marked as a Digital Venturis 575. It says made in GB, both in English and in French. (stickers for the numbers)
36-42664-10.A01 (sticker)

Not sure if any of that is going to help, but there it is just in case.

(edited to clarify what each label is)

Last edited by FidoMike on 2021-10-01, 15:11. Edited 1 time in total.

Reply 9 of 11, by 1ST1

User metadata
Rank Newbie
Rank
Newbie

Stickers or integrated in the PCB?

Reply 10 of 11, by FidoMike

User metadata
Rank Newbie
Rank
Newbie
1ST1 wrote on 2021-10-01, 11:06:

Stickers or integrated in the PCB?

I edited my post to indicate which are stickers and what is on the silkscreen of the PCB.

Reply 11 of 11, by 1ST1

User metadata
Rank Newbie
Rank
Newbie

Did you had a look under the stickers?