VOGONS


First post, by GigAHerZ

User metadata
Rank Oldbie
Rank
Oldbie

So i have a Compaq machine, that unfortunately doesn't come to life.

It's a Compaq ProLinea 4/25s with a 486sx-25 cpu and ET4000/w32 graphics adapter integrated. 4MB ram on board.
I've cleaned up the components fully to make sure i have no bad contact issues with jumpers or anything else.

The CPU and GPU chips get warm.
POST card shows all volatages to be OK.
Based on POST card, it does properly come out from initial RESET.
CLK and Frame on POST card are lit, IORDY is "half-lit".
The POST debug card shows 4241, last code therefore is 42, the previous was 41.
Keyboard is properly reset - all lights come up an then go off.
There are no beeps, but i do hear a tiny pop from pc speaker when booting machine up. Probably artefact of RESET cycle.
And of course there's no picture nor even a signal produced for the monitor.

When i shut it down and in few seconds start it again, i get no BIOS codes at all on POST card.

I'm using known good PSU with all voltages (including -5V) present.
I've ordered adapters to be able to use my TL866 to read the bios. What type of chip the bios is? I would like to order some just in case as well.

I've looked through the board closely but it looks to be in pristine condition. Most specifically i tried to find maybe some bridges between the smallest legs of chips, but all looks good.
I've tried an additional, 486sx2-50 cpu on it with appropriate jumper changes, but the behavior didn't change. I also tried disabling the graphics card and changing the IO base address jumpers just-in-case, but it didn't change anything. I also tried adding known good additional RAM to see any difference, but nothing.

What else i could try? Unfortunately i don't have an oscilloscope to look into some signals, but maybe there are some other ways and things to check?

Attachments

"640K ought to be enough for anybody." - And i intend to get every last bit out of it even after loading every damn driver!

Reply 2 of 23, by GigAHerZ

User metadata
Rank Oldbie
Rank
Oldbie
chrismeyer6 wrote on 2022-05-26, 15:57:

Silly question is the bios battery good and have you tried a clear cmos?

Battery is good, it does hold the voltage well enough that it shouldn't be a problem. (2,7V, while rated at 3V)
I did try to clean the cmos - no effect.

"640K ought to be enough for anybody." - And i intend to get every last bit out of it even after loading every damn driver!

Reply 4 of 23, by Strahssis

User metadata
Rank Member
Rank
Member

That no life + pop in the PC speaker sounds like a similar issue I have had on an Olivetti before, although I can't confirm since I didn't have a POST card to work with at the time. Have you tried checking the oscillators and the capacitors yet?

Mimi: AMD K6-2/266, S3 Trio64, Diamond Monster 3D II, Sound Blaster CT2800, 32MB RAM
Satellite 220CS: Pentium 133, SVGA DSTN, Sound Blaster Pro, 64MB RAM
Contura 420CX: 486DX4 75, VGA TFT, Roland Serial MIDI, 16MB RAM

Reply 5 of 23, by GigAHerZ

User metadata
Rank Oldbie
Rank
Oldbie
tabm0de wrote on 2022-05-26, 21:48:

maybe try remove cmos chip, computer should work without it, at least does it on presario 433.

Woah, really? You mean BIOS ROM? I cannot think how could that ever work, but it wouldn't hurt either, so i tried it out - dead. No post codes or anything. Reset and pc speaker pop does happen properly.

Strahssis wrote on 2022-05-26, 22:00:

That no life + pop in the PC speaker sounds like a similar issue I have had on an Olivetti before, although I can't confirm since I didn't have a POST card to work with at the time. Have you tried checking the oscillators and the capacitors yet?

Thankfully this board has no electrolytics at all. Everything is ceramic. And if there would be any bad tantalums, they would blow into my face, which has not happened, so seems to be all good.
How could i check the oscillators without oscilloscope? Only thing i can currently say is that the CLK led is properly lit on POST card, but the board has 2 oscillators and maybe one of them is not good...

"640K ought to be enough for anybody." - And i intend to get every last bit out of it even after loading every damn driver!

Reply 6 of 23, by Strahssis

User metadata
Rank Member
Rank
Member

Ah that's very nice! I should've known that, since most of my Compaqs don't have electrolytics either. I don't have an oscilloscope, so I can't tell you how to do that. I have sadly always been limited to just replacing the oscillators and see what happens.

Mimi: AMD K6-2/266, S3 Trio64, Diamond Monster 3D II, Sound Blaster CT2800, 32MB RAM
Satellite 220CS: Pentium 133, SVGA DSTN, Sound Blaster Pro, 64MB RAM
Contura 420CX: 486DX4 75, VGA TFT, Roland Serial MIDI, 16MB RAM

Reply 8 of 23, by GigAHerZ

User metadata
Rank Oldbie
Rank
Oldbie
Strahssis wrote on 2022-05-28, 19:08:

Ah that's very nice! I should've known that, since most of my Compaqs don't have electrolytics either. I don't have an oscilloscope, so I can't tell you how to do that. I have sadly always been limited to just replacing the oscillators and see what happens.

I'm really tempted to go on that route - replace just to see what happens and if something changes... I might do that at some point if we get truly stuck.

Joakim wrote on 2022-05-28, 20:27:

Checked the power rails (when th computer is on) and see if something is shorted?

Haven't done that in the way you described. As the POST card shows all voltages with fully lit LEDs, i assumed there can not be a problem with that. Otherwise one of the LEDs would be dimmed a bit. I would also hope that with shorted rails either of those two would happen: 1 - something becomes so hot it explodes (like tantalum caps) or 2 - my relatively modern ATX PSU (with AT adapter and all voltages truly present, checked) would trigger some protection and would not run.

Checked the voltages right now with multimeter and all is perfect while the machine/PSU is turned on.

"640K ought to be enough for anybody." - And i intend to get every last bit out of it even after loading every damn driver!

Reply 9 of 23, by GigAHerZ

User metadata
Rank Oldbie
Rank
Oldbie

So i took out the large oscillator and replaced it with another one that was laying around in my parts bin. Unfortunately it didn't make any difference - the machine still behaves as described before.
There is another one, near the left end of riser card slot. It doesn't have any markings on it, so i was not able to try something different there.

I once more went through the board inch-by-inch and tried to find any traces with any damage. But this board is truly in pristine condition... except it doesn't work. 😜

Attachments

"640K ought to be enough for anybody." - And i intend to get every last bit out of it even after loading every damn driver!

Reply 11 of 23, by weedeewee

User metadata
Rank l33t
Rank
l33t

does removing the jumper from P7 change anything ?

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 13 of 23, by GigAHerZ

User metadata
Rank Oldbie
Rank
Oldbie

So, I've received the PLCC to DIP adapters for my TL866II. Here's the bios from the compaq board.

Just quickly looking into it, it seems okay. I can read some human readable strings out from it like "COMPAQ" and the date of "07/02/93" plus bunch of error messages the bios can show.
Just in case, i also tried to just reflash the same bios image back to the eeprom to make sure it succeeds and has a "fresh copy" in it. It did succeed, but the machine still doesn't boot.

Unfortunately i can't find any other usable bios images from the internet. I did find some kind of file, which made a bootable floppy that uses rompaq to flash the bios it has internally, but my machine does not boot, so i need the raw image. And i couldn't find it from the image. The floppy image is attached - maybe some of you can extract the bios rom from it?

To have this information archived as well: The bios rom chip is TMS27c010A. 😀

weedeewee wrote on 2022-05-29, 09:22:

does removing the jumper from P7 change anything ?

Nothing... 🙁

Attachments

"640K ought to be enough for anybody." - And i intend to get every last bit out of it even after loading every damn driver!

Reply 14 of 23, by rasz_pl

User metadata
Rank l33t
Rank
l33t
GigAHerZ wrote on 2022-05-27, 06:53:
tabm0de wrote on 2022-05-26, 21:48:

maybe try remove cmos chip, computer should work without it, at least does it on presario 433.

Woah, really? You mean BIOS ROM? I cannot think how could that ever work

it wouldnt, OP meant old style RTC/CMOS/battery combo chip which your board doesnt have because its more modern integrated design

GigAHerZ wrote on 2022-05-27, 06:53:

The POST debug card shows 4241, last code therefore is 42, the previous was 41.

does it use standard port 80h? http://mrbios.com/techsupport/award/postcodes.htm lists non standard ports 85 86
http://www.bioscentral.com/postcodes/compaqbios.htm same codes

41 Check RAM refresh
42 Start write of 128K RAM test
43 Reset parity checks

stop at 42 would indicate bad ram? no idea if desoldering onboard will automagically switch to simms

Open Source AT&T Globalyst/NCR/FIC 486-GAC-2 proprietary Cache Module reproduction

Reply 15 of 23, by Joakim

User metadata
Rank Oldbie
Rank
Oldbie

The ram is a highly suspected indeed. you need an oscillator to test them and the soldering skills you can probably replace them.

I watch Adrian Black doing this all the time but it's way out of my own league.

Reply 16 of 23, by GigAHerZ

User metadata
Rank Oldbie
Rank
Oldbie

Unfortunately the closest thing i have to oscilloscope is a salea logic analyzer clone. 🙁

But i do have a reasonably good soldering station with hot air gun and some kapton tape, aluminium foil, etc. Probably will need to desolder the existing ram and see, if it starts behaving differently...
Good to know that there are others with same suspicion, gives me a bit of comfort to go into it. 😀

NB! I also watch Adrian's channel amongst other things. Currently is playing an Amiga 1000 repair video. 😁

"640K ought to be enough for anybody." - And i intend to get every last bit out of it even after loading every damn driver!

Reply 17 of 23, by GigAHerZ

User metadata
Rank Oldbie
Rank
Oldbie

So i desoldered the ram chips. It went amazingly well - the kapton tape held beautifully. I recently just bought another roll as the previous one melted very easily and you never know really, what is coming from aliexpress. This one was way better now. Probably my best ever desoldering job i've done. 😁

But unfortunately the behavior didn't change. It seems completely dead.
The debug card doesn't show anything and didn't show anything before desoldering either. I'm afraid it was a hiccup of the debug card itself earlier.

I also noticed the U101 (near floppy connector) becoming a bit warmer than other similar chips around. I desoldered that, too, but still - dead.

Only warm things are VGA chip and CPU. And if i try another chip in the socket, then the soldered cpu stays cool and the added cpu becomes hot. (With P7 jumper also changed when additional cpu is used)

Anything else i could try?

"640K ought to be enough for anybody." - And i intend to get every last bit out of it even after loading every damn driver!

Reply 19 of 23, by weedeewee

User metadata
Rank l33t
Rank
l33t
GigAHerZ wrote on 2022-05-26, 17:44:
chrismeyer6 wrote on 2022-05-26, 15:57:

Silly question is the bios battery good and have you tried a clear cmos?

Battery is good, it does hold the voltage well enough that it shouldn't be a problem. (2,7V, while rated at 3V)
I did try to clean the cmos - no effect.

FYI, 2.7V is close to empty for a button cell. Though I'm guessing that an empty battery likely isn't your problem.

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