• Please review our updated Terms and Rules here

Unstable 286 board

I

iulianv

Experienced Member
Joined
Jan 5, 2012
Messages
75
Location
Romania
Hello everyone,

I need the help of experienced people to debug some issues with the motherboard in the attached photo. I couldn't find an exact manufacturer or model - the closest one I found is this one (except for the location of the JP5 battery jumper and an extra JP6 that I have next to the FPU, they pretty much look alike):

http://stason.org/TULARC/pc/motherboards/A/AUVA-COMPUTER-INC-286-BABY-286-BAM-12-G1.html

The problem is that my board is pretty unstable (I'm using plain DOS 6.22 for testing). At first FDD access would hang the system (tested with three I/O controllers, two floppy cables and two FDD units), then someone suggested that I should remove the FPU, which I did - FDD access worked after that, but after putting back the FPU it was still working. Then I pulled out the eight 80ns RAM chips and put four 70ns chips - the system still worked, but FDD access broke again. Then I put back the original RAM chips - FDD still dead, lots of instability (freezes in games like Xenon2 or Volfied, hangs while doing nothing with Norton Commander open).

I sure hope it's the RAM's fault - is there any memtest-like tool available for pre-386 computers? I'd be very disappointed if something else were broken on the board, it's the first 286 board that I could get my hands on and I was really hoping to be able to build a system around it...
 

Attachments

  • headland286.jpg
    headland286.jpg
    99.8 KB · Views: 3
Silly question but is whatever PSU you are using known-good? And, assuming this is being 'bench tested', is the board properly grounded via the mounting holes?
 
Well, I've never payed much attention to my AT PSUs (except for keeping them clean and free of bulged caps) - I have three or four of them and I guess they qualify as no-name, but I've done plenty of testing with them and any of them could hold any 386, 486, Pentium or Pentium Pro system that I built over time. Same goes for grounding - whatever board I get I first bench-test on the table, and no board has ever behaved like this on me so far...
 
Last edited:
Need to ID the board. Are there any numbers/names on the PCB? Who made the BIOS?

Do the switches for memory correspond with the diagram you referenced, everything counts out right with switch changes? It just doesn't quite look the same but close.

Tried a different video card?

Tried a different boot disk?

Lots of memtest programs available at http://cd.textfiles.com/
 
Last edited:
BIOS is Quadtel - I'll post the exact info later on today.

When I tried to only use half of meg of RAM I did notice some difference between the info on the site and what appeared to be the right dip-switch config for 512KB on my board, but I think any misconfiguration of those dip-switches should only cause the board to complain that the configured and installed amounts of RAM are different, but not generate instability...

Speaking of RAM, the maximum configurable amount of "system RAM" in the BIOS is 640KB (possible values are 256, 512 and 640); I assume this is conventional RAM and the remaining 384KB are reserved for other purposes. But what happens if I'm only using four RAM chips (512KB)? What (if any) is the amount of reserved RAM in this case? Also, how fast should the RAM be (are 70ns and 80ns valid values for the 16MHz CPU and 10MHz FPU)? And, in case I'd like to populate the four DIP16 sockets for parity, should the parity chips and the RAM be the same speed?

I tested with two video cards (TVGA8900C and TVGA9000i-3), but I only had DOS installed on one hard-drive to play with...
 
70 or 80ns would need a wait-state at 16MHz. Have you tried lowering the CPU clock speed to 8MHz (if you can)?
 
Yes, it's possible and I did that as well (Ctrl-Alt-<minus> and Ctrl-Alt-<plus> toggle between 16 and 8 MHz); I also changed JP17 which, if I'm to trust the link I found, toggles between 0 and 1 wait state, but I couldn't "measure" or see any effect...

Speaking of Ctlr-Alt- combinations, I find it fun that the BIOS can be accessed (Ctrl-Alt-s) from anywhere, anytime (no matter if the OS is running or not) :).
 
Have you tried reseating all socketed chips. Or perhaps some of the soldering joints may have failed...

But as mentioned, you should check that the voltages on the board are correct during runtime. Even if a PSU is good, failing capacitors/components on the motherboard itself can draw a lot of current.
 
The AMD chip next to the CPU is the FPU? I didn't know AMD made a 80287 FPU. Sorry off topic for a second. As Per said remove all socketed chips and reinstall. Connections tend to go bad over time. Don't know why but I guess expansion and contraction due to Temp changes.
 
Looks like it also has sipp(30 pin simms with pins attached to them if that makes sence) sockets between two ISA slots and on the outer edge of the board by the battery. So could possibly use up to at least 4 megs like I had on my 286/16 if you do get it sorted. To use them all dip memory would need to be removed.
 
Last edited:
framer said:
I'd start with, is that battery any good? I'd cut it off and use an external battery. Just my 2 cents...
Click to expand...

And even that it might sound obvious perhaps you have some damage from leaked NiCd battery? NiCd batteries (after some time) tend to leak electrolyte (alkaline substance) that corrode metal (PCB traces, components pins). Make sure you don't see any bluish salt or black stains on traces and components near the battery or white salt on the battery itself. If such damage exists, first make sure to unsolder the battery, remove remains of electrolyte, salt, etc. and check connectivity on damaged parts / traces. If needed re-connect damaged traces with wires. Then you can connect an external battery instead of one that was soldered on PCB (I think some smaller cordless phone 3-cell NiCd 3.6V batteries can be used).
 
Wow... lots of answers :). One by one:

- absolutely nothing is written / labeled on the board (front, back, slots, chips, nowhere)
- it has a "Quadtel Enhanced 286 BIOS Version 3.05.01" with "BIOS ID: 0011E03003"
- I could only reseat the FPU and the RAM chips; the CPU, BIOS and keyboard controller chips seem to firmly fit and I'm afraid I'd break something if I insist too much
- yes, the AMD chip next to the CPU is a 10MHz 287 - one of these: http://www.chipdb.org/cat-287-945.htm
- I'm looking forward to finding some SIPPs (I know I could "manufacture" them from 30p SIMMs, but I don't have the tools, skill and patience to do it); again, if I'm to trust the link I found, the board will take 2 or 4 megs of RAM (although there is an "Extended Memory" setting in the BIOS that goes up to almost 16 megs, but I doubt I'll ever find 4MB SIPPs to try it)
- there is absolutely no corrosion around the battery; since I started playing with old hardware I "identified" two kinds of barrel-type batteries: some (usually blue) engraved "CHARGE" on one end, that are very clean (no leakage) and can still hold BIOS settings at least over a few days of power-off and some (usually green) not engraved, that leak, corrode and can't hold anything at all; this board seems to have the first type...

Now the good news: yesterday I found yet another set of eight RAM chips (70ns, from a Trident VLB video card) and used those to fire up the board again (in order to collect whatever BIOS info available and post it here); well, it behaved flawlessly at full speed (FDD access OK, no resets or freezes while gaming). I'm just hoping it will do the same the next time (and the next, and the next...).

Two more questions:

- if I were to populate the four DIP16 sockets for parity RAM, do those chips have to be 70ns too? I have around thirty DIP16 chips, but they are all 120ns...
- the photo I uploaded was 1600x1200 in its original version, but somehow got resized when attached; is there any way to "fix" this?
 
As it turns out, I'm not out of the woods yet... A few days ago I tried to do a fresh install of MS-DOS 6.22 on this system, using some 1GB hard-drive I had around (previously "prepared" with Seagate's flavour of Ontrack Disk Manager).

If I floppy-boot through Ontrack (boot from HDD then "press space to boot from diskette"), the system either hangs or resets after choosing the install directory (just before copying files). If I boot straight from the FDD (so only ~500MB of the hard-drive are available, but that shouldn't be an issue), the setup process goes on into the file copying phase, but rejects ("read error") all compressed files on all three disks. Eventually a "minimal", incomplete OS is installed, and if I boot it from the hard-drive I can EXPAND all the compressed files on all three disks just fine.

I found the link below (I guess SETUP.EXE wasn't expected to work all the time) and will try it tonight, but I'm curious if this behaviour can be considered "normal" when installing MS-DOS 6.22 on a 286 with 640KB of RAM...

http://support.microsoft.com/kb/117245
 
You must log in or register to reply here.
Back
Top