VOGONS


First post, by NevilClavain

User metadata
Rank Newbie
Rank
Newbie

Ok, I apologize for the messy headline but I have a hard time putting this into a sentence.

A little over half a year ago I aquired one of my dream machines. A Siemens Nixdorf XPert tower from the late 90's (P200 mmx with Matrox onboard graphics and SBpro compatible sound), and I fear I have killed it.
I had a hard time setting it up with a CF and I had to do a lot of stupid stuff, and after about a week the PSU died. When I removed it I saw that it had no wires going into the +3.3vdc plugs on the ATX connector (astec SA146-3421). Strange,I thought, but connected a strd. ATX PSU and continued. After about another week of trying to set it up I finally got it working reliably with CF, onboard graphics, SCSI CDrom and a Voodoo 2. Really nice setup. After another reboot It stopped posting. The CPU fan and PSU fan spins up but nothing. I tried resetting the bios (taking out the battery, disconnecting the PSU and taking out the ram chips etc.) but nothing helped. I eventually gave up and just put it away. It's been a long time now and It's been bothering me.

It might be apparent by this point that I'm not hugely knowledgable about these things but I would like to learn as much as possible. I've got equipment to do some basic troubleshotting (oscilloscope and voltmeter, soldering station etc.) and I really want to get this machine up and running. I hope this community might be able to help steer me in the right direction.

1: Why were the +3.3vdc wires not present? Did I fry the motherboard by using a PSU that had them? Or the CPU? Any ideas?
2: Where do I begin when troubleshooting this? It seems the board is getting power (CPU fan) but there is no beep and no post.
3: What are the most likely problems? I haven't really started experimenting now, as I fear I got too careless the last time around. I've learnt my lesson and have decided to be more respectful in the future. 🙁

All help and suggestions are GREATLY appreciated!

Reply 1 of 16, by dionb

User metadata
Rank l33t
Rank
l33t

Not too familiar with this specific machine, but maybe can help with a few general things.

1) Some early ATX boards were really only AT boards with a different connector and form factor and didn't actually use the 3.3V for anything. It's possible that that was the case here, and that Siemens decided to cut costs by not providing for the rail or the wires in the PSU either. Does the sticker on the PSU mention any 3.3V? I looked up that model number and didn't find the SA146-3421, but did find the SA146-3481, and it certainly doesn't have any 3.3V:
3_454_104824_800_600.jpg

So my hunch is that the lack of 3.3V is by design and therefore not the cause of your problems.

2+3) Good question. The death of a PSU can frequently cause more damage further down the line, so it's entirely possible that killed your board. Then again, are you sure it's the PSU that's dead?

First up: if the board's still in the tower case, get it out and put it on a non-conductive, non-static surface. A book or newspaper works fine. Then hook up the absolute minimum of devices to power it up, so a known-good CPU, two known-good SIMMs, a keyboard and a known-good PSU. Use a screwdriver or similar to short the power switch pins on the board to turn it on. If it does boot now, you either have a short in your case, or one of the other components was the problem.

Then with hardware of this age: check capacitors. The types used <1999 were generally good, but 20 years later even good stuff starts to die. This applies both to the PSU and to the motherboard. A visual inspection finds the worst offenders, but even caps that look good can be dead. If you have any way of measuring ESR, do that. Failing that, just go by appearance. Bulging/leaking/otherwise not looking straight, flat&clean caps definitely need to go.

Next the board might be trying to tell you something. If you have a POST debug card, use that. If not, consider getting one (dirt cheap on well-known auction site and/or far-Eastern marketplace). The actual code is rarely that helpful, but distinguishing between a completely dead board and one well on the way to POSTing is useful.

Also, BIOS EEPROMs can get corrupted. Trying a BIOS recovery procedure may help, or even better: flashing a replacement chip.

Beyond that, you get into measuring voltage lines on the board for breaks etc, but don't do that until you've tried the quick&dirty things above. If these things don't fix it, it would be very useful to know exactly which board we're talking about (Siemens-Nixdorf used a Dxxxx naming scheme IIRC).

Finally, over on vcfed.org there's a user Overmann who posted he got one of these last year and had it up and running straight away. Possibly he could help, if only regarding the PSU and the missing power lines.

Reply 2 of 16, by NevilClavain

User metadata
Rank Newbie
Rank
Newbie

Hi dionb,

Thank you for replying so quickly! 😀
Yes, that picture of the PSU is very similar to mine. It has no mention of 3.3vdc, just like that one. My PSU no longer does anything. It made a dying rattle, and choughed and smoke came out of it. It didn't take the motherboard with it when it died as the the motherboard worked fine with another PSU for about a week before the board stopped posting. I did connect it just now with two different good PSU's and the same thing happened. The CPU fan starts, the floppy drive light is constantly on (perhaps because I have the floppy cable the wrong way around?) but it makes no sound or beeps. I have tried several different ram sticks without luck, all of which were good when I stored them, but I will test them in another machine! I have no known good CPU that is similar, but I do have another socket 7 CPU. An AMDk6 II-450, but can I use this on this board? Will it just clock down to whatever the FSB allows?

I have connected the board to a PSU, with only RAM attached, but no post. I did notice a faint etectrocution when I touched both the board tower and the PSU at the same time (the PSU was laying on the desk next to the tower. I will follow your directions and do it properly with tested RAM and with the board taken out of the tower.

The hardware looks good to me. The capacitors look great, if fact, but I've learnt that looks can be deceiving when it comes to caps. All of them look flat and clean! I don't know how to measure ESR, but I will read up on it!

I don't have a post debug card either, but will order one right away. That sounds useful 😀 It's one of those PCI cards that is hooked up to a small LCD screen, right?

I've had a look on the board and the bios does not seem to be socketed. I will look again. BIOS recovery procedure sounds like a good and easy step. 😀

Regarding Overmann. I am that same idiot, and this is that machine! 😀 I don't know why I have a different handle on this site. I was looking for this machine for years and when I finally get hold of one I screw it up after only two weeks. 🙁 Btw. the 486 Motherboard that I bought from you is working like a charm! 😀

I've uploaded some pictures of the Siemens Nixdorf board

Attachments

  • IMG_20180513_235825.jpg
    Filename
    IMG_20180513_235825.jpg
    File size
    97.18 KiB
    Views
    494 views
    File license
    Fair use/fair dealing exception
  • IMG_20180513_235636.jpg
    Filename
    IMG_20180513_235636.jpg
    File size
    385.25 KiB
    Views
    494 views
    File license
    Fair use/fair dealing exception
  • IMG_20180513_235806.jpg
    Filename
    IMG_20180513_235806.jpg
    File size
    143.41 KiB
    Views
    494 views
    File license
    Fair use/fair dealing exception
  • IMG_20180513_235716.jpg
    Filename
    IMG_20180513_235716.jpg
    File size
    900.67 KiB
    Views
    494 views
    File license
    Fair use/fair dealing exception

Reply 3 of 16, by dionb

User metadata
Rank l33t
Rank
l33t
NevilClavain wrote:

Hi dionb,

Thank you for replying so quickly! 😀
Yes, that picture of the PSU is very similar to mine. It has no mention of 3.3vdc, just like that one.

Then that's one worry less: the lack of 3.3V is by design.

My PSU no longer does anything. It made a dying rattle, and choughed and smoke came out of it.

OK, that's pretty clear then 😉

It didn't take the motherboard with it when it died as the the motherboard worked fine with another PSU for about a week before the board stopped posting. I did connect it just now with two different good PSU's and the same thing happened.

Maybe, but I'd say the PSU blowing and the motherboard dying so close together is unlikely to be coincidence.

The CPU fan starts, the floppy drive light is constantly on (perhaps because I have the floppy cable the wrong way around?)

Light constantly on sounds like cable the wrong way round. In any event, try with no floppy connected.

but it makes no sound or beeps. I have tried several different ram sticks without luck, all of which were good when I stored them, but I will test them in another machine! I have no known good CPU that is similar, but I do have another socket 7 CPU. An AMDk6 II-450, but can I use this on this board? Will it just clock down to whatever the FSB allows?

FSB isn't an issue, it's the motherboard that determines the clock, not the CPU, so it can always run slower than specced. Bigger issue is voltage. A K6-2 wants 2.2V but your Pentium MMX get 2.8V and I don't see any voltage settings on the board. It's possible that it might autodetect or something, but more likely it would feed that K6-2 2.8V as well. Not a good idea.

What I do see: you say that's a Pentium 200MMX in there, but if I look at the dipswitches it's set for 1.5x, which an MMX reads as 3.5x, so the CPU is set for 3.5x66=233MHz. Now usually you can overclock a Pentium MMX with ease (I had a 200MMX running at 350MHz once), but in case of doubt, set it back to original speed or - even better - lower. I'd recommend setting it at 120MHz for testing, so 2.0x 60MHz. See here, now switch 1,2,3,4 are set to on,off,off,off. Try off,on,on,off instead.

I have connected the board to a PSU, with only RAM attached, but no post. I did notice a faint etectrocution when I touched both the board tower and the PSU at the same time (the PSU was laying on the desk next to the tower. I will follow your directions and do it properly with tested RAM and with the board taken out of the tower.

That shock sounds like an earthing/grounding problem to me. Is the PSU connected to an earthed power outlet? And the monitor? In itself that's not likely to be the cause of your problems, but it's definitely something to look into.

Another thing to rule out: that COAST cache module. It's unlikely to be the cause of the problems, but it's easy enough to remove just in case.

The hardware looks good to me. The capacitors look great, if fact, but I've learnt that looks can be deceiving when it comes to caps. All of them look flat and clean! I don't know how to measure ESR, but I will read up on it!

I had forgotten the spectacular build quality of Siemens-Nixdorf stuff. Massively thick boards, huge heatsinks, high-quality components - and those silver coloured legends. Beautiful. No guarantee that it all still works, but I'd trust this stuff more than most.

It's a D990 rev E11 motherboard by the way, the name is in the little white field above the dipswitches.

I don't have a post debug card either, but will order one right away. That sounds useful 😀 It's one of those PCI cards that is hooked up to a small LCD screen, right?

Yes, although I suggest getting a PCI/ISA combo card. Frequently the POST stuff is only output on a boards' oldest bus - and of course you have at least one system with no PCI at all.

I've had a look on the board and the bios does not seem to be socketed. I will look again. BIOS recovery procedure sounds like a good and easy step. 😀

Yes, I'm surprised by that. Most Siemens-Nixdorf boards I've seen have nice socketed stuff. I have a BIOS EEPROM here from a D1192 board. Does make me wonder what that PLCC32 socket at the rear of the ISA slots is for...

Regarding Overmann. I am that same idiot, and this is that machine! 😀 I don't know why I have a different handle on this site. I was looking for this machine for years and when I finally get hold of one I screw it up after only two weeks. 🙁 Btw. the 486 Motherboard that I bought from you is working like a charm! 😀

Talking to yourself probably isn't going to help much no 😉

Good to hear the 486 board hasn't let you (or me) down 😀

Reply 4 of 16, by NevilClavain

User metadata
Rank Newbie
Rank
Newbie

PCI/ISA POST card is ordered and should arrive in the next couple of weeks. I just remembered that I do also have a pentium 133, but that also has a different voltage rating then the MMX (EDIT: AAAND wrong socket 😜 ).
There is a three pin header on the MB called VCORE. I think we have the voltage selector. 😀 It is currently OPEN, no pins are connected, but I can't find instructions regarding this header,
And, you are right about the CPU, it's a 233 mhz. I remembered it wrong. Made in USA in 1998 as far as I can tell.

It really is a very nice motherboard, with a very nice and extensive bios. I did try to upgrade the BIOS (fujitsu are still hosting drivers and bios for this machine) but was told that my machine was unsupported (D990 bios) when i initiated the process.

The socket near the read by the ISA slots reads WAVE. I don't know what it does.
But I do see that the DIP switches has a setting called "Flash BIOS recovery mode". The manual makes no mention of it, but it is very lacking. The one hosted by Fujitsu is probably just a stub.

It looks like I took the battery out without putting it back in the last time. I'll have to see if I can find a CR2032 in the batterydrawer.

Last edited by NevilClavain on 2018-05-14, 13:00. Edited 1 time in total.

Reply 5 of 16, by dionb

User metadata
Rank l33t
Rank
l33t
NevilClavain wrote:

PCI/ISA POST card is ordered and should arrive in the next couple of weeks. I just remembered that I do also have a pentium 133, but that also has a different voltage rating then the MMX. And as you mentioned the board has no selector. It seems strange to me then that it does not autodetect. That Siemens made this board to only support ONE type of CPU.

You're talking about a company here that left out a standard voltage line on their PSU because their one board didn't need it 😉

But the MMX/non-MMX situation is different, thats split voltage vs single voltage. If you insert a single voltage CPU such as the P54C (non-MMX) it gets all its voltage at the Vi/o setting (3.3V normally), whereas a dual-voltage CPU such as the P55C (or indeed K6-2) gets separate Vcore (at 2.8V) and Vi/o.

Once again, it *might* autodetect voltage, but I doubt it. Siemens was conservative, they almost certainly only support Intel CPUs, so that means 3.3V single voltage or 3.3v/2.8v split voltage.

Anyway, have you already tried downclocking that CPU from 233MHz to 120MHz like I suggested?

It really is a very nice motherboard, with a very nice and extensive bios. I did try to upgrade the BIOS (fujitsu are still hosting drivers and bios for this machine) but was told that my machine was unsupported (D990 bios).

Had a look and this BIOS looks hard to find...

The socket near the read by the ISA slots reads WAVE. I don't know what it does.

Almost sounds like an upgrade option for the sound chip (Wavetable).

But I do see that the DIP switches has a setting called "Flash BIOS recovery mode". The manual makes no mention of it, but it is very lacking. The one hosted by Fujitsu is probably just a stub.

Possibly. but OEM manuals are rarely complete.

Reply 6 of 16, by eisapc

User metadata
Rank Member
Rank
Member

I wonder if the board really supports MMX CPUs. Allmost all my MMX capable boards use switchmode power converters for the CPU voltage while yours features a linear converter.
The difference is easily visible by the presence of coils near the CPU socket in one case or heatsinks in the other case.
I also wonder why they didn´t use the 3.3V line to feed the CPU if one is present. Siemens even did AT power boxes with 3.3V Power like the PCD-5T (Dual P5 133MHz) with features a power supply daughter board and feeds the boards through a second pair of AT bus power connectors. btw. the PCD-5T is not MMX capable and is currently running 2x P133.
I suspect you fried the linear converter. Try to use a multimeter to check if the 3.3V CPU power is still there.

Reply 7 of 16, by NevilClavain

User metadata
Rank Newbie
Rank
Newbie
eisapc wrote:

I wonder if the board really supports MMX CPUs. Allmost all my MMX capable boards use switchmode power converters for the CPU voltage while yours features a linear converter.

I am pretty sure that the CPU is the original. All the other xpert 4320 systems I have found have an MMX CPU (https://www.ebay.de/itm/SIEMENS-NIXDORF-XPERT … =p2047675.l2557) and the xpert 4320 that we had when I was a kid was also a 233 MMX (or 200 MMX).

eisapc wrote:

I suspect you fried the linear converter. Try to use a multimeter to check if the 3.3V CPU power is still there.

Should I check the PSU that I'm using? I can do that, but what would cause the linear converter to fry?

dionb wrote:

Anyway, have you already tried downclocking that CPU from 233MHz to 120MHz like I suggested?

No, not yet. I've just started trying stuff today. I have removed the cache and have tried activating bios recovery. No change. Next up is checking the ram to see if it's good and downclocking the cpu 😀

dionb wrote:

Had a look and this BIOS looks hard to find...

Do you mean the chip or the software? The bios on their website should work according to their description, so I don't know why it didn't want to upgrade.
Flash BIOS Update for SCENIC Multimedia PD/ PT5xx (D990) V4.05 R1.04.990 4.05 - 1.04 (17/03/1998) 0.3 MB Language: English
Working with Xpert 41xx/42xx/43xx

Reply 8 of 16, by NevilClavain

User metadata
Rank Newbie
Rank
Newbie

Just tried downclocking the CPU, but that made no difference.

EDIT: I have found an old AMD K6-II 550AGR too. 2.3V core/3.3V I/O, but I suppose that helps little. I'll have to order a CPU I think. Thankfully they are plentyfull. 😀 On the other hand, If I managed to find out what the different settings of VCORE does i might be able to run the board using the K6-II. I'll have to keep searching for info.

Last edited by NevilClavain on 2018-05-14, 19:09. Edited 2 times in total.

Reply 9 of 16, by .legaCy

User metadata
Rank Oldbie
Rank
Oldbie
NevilClavain wrote:
I am pretty sure that the CPU is the original. All the other xpert 4320 systems I have found have an MMX CPU (https://www.ebay.d […]
Show full quote
eisapc wrote:

I wonder if the board really supports MMX CPUs. Allmost all my MMX capable boards use switchmode power converters for the CPU voltage while yours features a linear converter.

I am pretty sure that the CPU is the original. All the other xpert 4320 systems I have found have an MMX CPU (https://www.ebay.de/itm/SIEMENS-NIXDORF-XPERT … =p2047675.l2557) and the xpert 4320 that we had when I was a kid was also a 233 MMX (or 200 MMX).

eisapc wrote:

I suspect you fried the linear converter. Try to use a multimeter to check if the 3.3V CPU power is still there.

Should I check the PSU that I'm using? I can do that, but what would cause the linear converter to fry?

dionb wrote:

Anyway, have you already tried downclocking that CPU from 233MHz to 120MHz like I suggested?

No, not yet. I've just started trying stuff today. I have removed the cache and have tried activating bios recovery. No change. Next up is checking the ram to see if it's good and downclocking the cpu 😀

dionb wrote:

Had a look and this BIOS looks hard to find...

Do you mean the chip or the software? The bios on their website should work according to their description, so I don't know why it didn't want to upgrade.
Flash BIOS Update for SCENIC Multimedia PD/ PT5xx (D990) V4.05 R1.04.990 4.05 - 1.04 (17/03/1998) 0.3 MB Language: English
Working with Xpert 41xx/42xx/43xx

Linear regulators can fail due to thermal stress caused by a increased current demand on the circuit,with time they become "aged" and fail.

Useful links
[ How to Multiboot any OS | Rebuilding GOG Installer for Win9x | 486DX-4 Build | S7 Time machine Build |
twitch stream (in portuguese)

Reply 10 of 16, by dionb

User metadata
Rank l33t
Rank
l33t
NevilClavain wrote:

Just tried downclocking the CPU, but that made no difference.

EDIT: I have found an old AMD K6-II 550AGR too. 2.3V core/3.3V I/O, but I suppose that helps little. I'll have to order a CPU I think. Thankfully they are plentyfull. 😀 On the other hand, If I managed to find out what the different settings of VCORE does i might be able to run the board using the K6-II. I'll have to keep searching for info.

Don't waste any money on new CPUs. CPUs are pretty near indestructible when safely inserted into a socket. If some surge killed the CPU, it will definitely have taken the motherboard with it. Also don't try any out-of-spec stuff (like K6-2 CPUs) before you have it working in-spec. Otherwise you're just inserting new unknowns.

If you really suspect the CPU, try that Pentium 133. That is supported by the board and is a nice simple bulletproof design.

For now I'd say the suggestion that the linear voltage regulator is toast (probably triggered by the PSU blowing) sounds most likely.

Reply 11 of 16, by NevilClavain

User metadata
Rank Newbie
Rank
Newbie

Ok, thanks to everyone for the help! Can I simply replace the linear voltage regulator? Would it work as a failsafe and prevent other components of getting kiled?
Wouldn't the PSU dying and taking the linear voltage regulator with it cause an immediate reaction rather then a week of proper function?

Reply 12 of 16, by NevilClavain

User metadata
Rank Newbie
Rank
Newbie

I've managed to find a replacement motherboard and thoight id test my PSU's before fireing it up and I've encountered somthing unexpected. This is the first time i've measured psu's, by the way.

The one that came with the machine (that died) is actually supplying voltages that are within spec on some rails, but not at all on others, so ill throw that away. The two old atx-supplies that I have and that I expected would work are showing healthy reading on all rails except the -12vdv (blue) where one reads about -9 and one about -8. Both of these supplies were in until fairly recently.

Is this reading a problem? Could i be doing somthing wrong, or have i misunderstood something?

Reply 13 of 16, by ninkeo

User metadata
Rank Member
Rank
Member

I'll ask the obvious but were your measurements taken under some sort of load on the PSU? Reason I ask is because PSUs can act funny without the load of a motherboard / disk drives on them.

Reply 14 of 16, by NevilClavain

User metadata
Rank Newbie
Rank
Newbie

Oh, no, I did not know that. Aboslutely no load. Just jumpered to be operational. Should I assume the PSU's are fine as all the other measurements were very good or is there some particular load I can put on the PSU that should make the measurements accurate?

Reply 15 of 16, by dionb

User metadata
Rank l33t
Rank
l33t
NevilClavain wrote:

Oh, no, I did not know that. Aboslutely no load. Just jumpered to be operational. Should I assume the PSU's are fine as all the other measurements were very good or is there some particular load I can put on the PSU that should make the measurements accurate?

For the 'normal' +5V and +12V lines, a hard disk (even a broken one) is a perfect test load.

-5V and -12V are more difficult. Most systems won't actually have anything that draws current on these lines. A small resistor could work as dummy load. But really these lines are utterly irrelevant to basic functioning and stability of a board, so you can ignore them unless doing exotic things with ISA sound cards with built-in amplifiers. Probably your PSU is working fine on these lines, just not stable without any load. But even if these lines are missing or completely wrong, it doesn't matter. They are not causing your current problems.

Reply 16 of 16, by perhenden

User metadata
Rank Newbie
Rank
Newbie

I've got two of these boards, revision GS2 and GS2:2. The first came with a p150 non-mmx, and the other with a p166 mmx. Both had the vcore jumpers (3 pins) open. The CPU is identified as MMX on boot, when I set the multiplier and clock to 166 and moved the MMX CPU to the other board. The GS2 is marked "72002", and the other one "73602". Both are marked "S SZ64". The boards look completely identical visually otherwise, with the GS2:2 having a larger heatsink next to the socket 7 and the cache slot. The MMX CPU keeps really cool, so I it's likely receiving the right p55c voltage.
The PSU died here, too, and I now see hangs and memory errors to a large degree, even when using another PSU.

Siemens used this motherboard in different revisions for both Xpert 52xx/53xx and Xpert 41xx/42xx/43xx, I guess that's why there's a number of locations on the motherboard prepared for extra headers, but with nothing soldered in to them. That could explain why the vcore jumpers are not used in our case.