Thermalwrong wrote on 2023-04-01, 00:50:

At last I've sort-of managed to troubleshoot something using my oscilloscope. I posted a picture of this board the other day wit […]
Show full quote

At last I've sort-of managed to troubleshoot something using my oscilloscope. I posted a picture of this board the other day with the front and back traces overlaid - I got this board a couple of years back as someone else had attempted to repair it and not got it working. I couldn't do much better although I did repair a bunch of the vias it still wouldn't do anything.

This is a WTC POP-3254 and the example shown on the retroweb appears to be in a worse state than my one. It's a rather old and basic 386dx board with no cache and the battery had destroyed some traces and vias by the 82c206 chip. It was coming out of reset and the clocks were working, but it didn't do anything.
Checking with the scope on the BIOS EEPROM I could see that there were signals on both the address and data lines, but some of the address lines seemed to have the same signal patterns and went from 5v to 2.5v which didn't seem right, most signals went to a lower voltage for a low signal.

Washed it up in the sink because this board has clearly been exposed to the elements and there were some white marks around most of the pins where I think the lead was leeching out. It cleaned up nicely but that didn't fix anything.
One of the 74LS373 chips had been replaced and socketed by the previous owner so I decided to pull everything in that area off, test it and check the connections of the traces, including the 82C206:
IMG_1156 (Custom).JPG
The pads/legs of the 82C206 were a bit corroded so that had to be checked too - there were bad VIAs hiding underneath it. Laboriously tracing and checking each trace that the 74LS373s attached to. The 74LS373 that wasn't replaced was bad after all, but its traces were in better shape.

WTC-POP-3254-tracing.jpg
A front and back picture lets me trace all the lines in Krita and see all the points they pop up - this is doable because this is a comparatively simple 4 layer board with VCC and GND on the inner layers, with signal routing on the outer layers only.
You can see how it looks for tracing here: Re: What retro activity did you get up to today?
This has the underside as the bottom layer with a filter mask layer to adjust the tint. The topside layer goes above that with the Blending Mode set to Lighten - which gives a really nice kind of overlay and the layers are different colours so it's easy to tell what I'm following. Some more layers on top with the set of pins I'm tracing. I'll definitely use this method again in future, I was able to find a bunch of broken connections by following the traces on the screen.
Then drawing a marking by the pin so I can tell it's been checked - only one wire had to be run on the outside, that's the one marked X.

Put it back together and put the 82C206 back on, it powered right up! Yay 😀
IMG_1159 (Custom).JPG
IMG_1158 (Medium).JPG
For some reason it has a 33MHz Intel 386, but the clock crystal is 50MHz for 25MHz CPU speed. The BIOS doesn't appear to be online and it came to me with no sticker on the EEPROM window, there was a big gap in the ROM file so I was worried the ROM was damaged, but it's fine after all. It's an AMI BIOS with a BIOS string of EOX3-6069-083090-KM
This thing is old enough that it has bios settings configured either with easy OPTI setup which has few options like wait state, or advanced OPTI setup where settings are configured by changing register bits, it's guided but it shows which bits are being altered for 11H, 14H etc. The board's from mid 1991 going by the chip markings but the BIOS and design both say '90, so it's probably the oldest computer I've got along with the Tandon motherboard.

edit: ergh okay so it's not working entirely. Stuff like keyboard works and I can change settings / save settings in the BIOS, but the computer can't boot from a floppy disk. Could that be a problem with DMA? That's handled by the 82C206 and there's a cluster of damaged traces there that were repaired.
It just says insert BOOT diskette and I've tried two known good drives, it does the track seek on boot so I think it's solely down to being unable to read the data.

As a follow up to this, I've been trying to figure out what's wrong with this POP-3254 / OPti 381/382 chipset 386 motherboard. The board posts, runs from a hard drive and can run programs, but it couldn't read floppies and PCM audio was coming out as a garbled mess in any test - that's a DMA problem. The 82C206 chip handles DMA and I fixed some traces that were broken underneath it as well as some of its traces that run over to the KBC - that's how it came back to functioning.
I could see that DMA was being processed like a DRQ was happening and there was a DACK following from what I could see but still garbled noise. I was getting ready to buy a new 82C206 and replace that but on a whim and following some traces from the 82C206 data pins, I pulled off the two 74F245 chips here, which from what I can tell connect it all to the bus - looks like they were in the path of the corrosion that this board was subjected to as the solder for them was tough to remove:

Both of them tested okay out of circuit but they had some dirt under them and the legs were pretty discoloured. I cleaned it all up a bit and put them back to get the board back to operational, again on a whim because I didn't think this would have fixed anything, tried a game and now sound works!

So now the DMA for both floppy and sound is working perfectly again and I don't really know what I fixed. Maybe the 2x 74F245 chips were used to make up 16 bits and one wasn't working because of dirt under the chips causing connections where there shouldn't be, causing the garbled-ness? But the lower one may have worked to allow everything else to function? I don't know enough about this stuff yet.

I'm really glad this is fixed since I was going to get an 82C206 but couldn't see how a board could be so otherwise functional with just some badness on the DMA. The more I use this thing I see it's a pretty odd board, the original BIOS was from like 1990 and would get angry if I installed 1MB SIMMs, limiting the board to 2MB of RAM. That was fixed by putting another board's BIOS on it - thank goodness for the retroweb's comprehensive search functions and BIOS repository: https://theretroweb.com/bios/?chipsetId=366
I put the 1991 BIOS from a Biostar MB-1325PD on it - and it works! Now the BIOS is the much more 'modern' green/purple/brown AMI BIOS I'm more familiar with.
Normally I wouldn't try random BIOS files but did you know that many of the Opti 381/382 chipset boards are essentially identical?

There must have been a reference design, these ones are all exactly the same down to the placement of the 74' logic chips and the RTC circuitry.

Tempted to borrow a crystal socket from another dead board along with its 66MHz crystal, the chipsets are marked as 25MHz but now I want to see if it can do 33MHz since this is a 386DX-33 chip.

Finished my restoration and explore of my toshiba 4000CDT as well as remove a seriously leaky battery:

https://youtu.be/wf286ovgYRM

Can you cut away some of the plastic covering the batteries? It's hard to tell, but they look like they are in suspiciously good condition. If they leaked, you should see corrosion on the bartery cells.

I've experienced this same plague with laptops that I've repaired. It's really bizarre how it attacks the composite case material. I wonder what the chemical reaction is.

I would also suggest leaving a note of the battery information and tape it where the cells used to be. Cut the leads and leave the there for future repair as well. It's nice to retain the real-time clock, and I would appreciate these things if I ever purchased one of these second hand.

Kahenraz wrote on 2023-04-14, 13:41:

Can you cut away some of the plastic covering the batteries? It's hard to tell, but they look like they are in suspiciously good condition. If they leaked, you should see corrosion on the bartery cells.

I've experienced this same plague with laptops that I've repaired. It's really bizarre how it attacks the composite case material. I wonder what the chemical reaction is.

I would also suggest leaving a note of the battery information and tape it where the cells used to be. Cut the leads and leave the there for future repair as well. It's nice to retain the real-time clock, and I would appreciate these things if I ever purchased one of these second hand.

It was quite interesting that the batteries themselves look so clean. I have owned this laptop for over 15 years and there is no chance they were changed in this time!

The RTC and bios settings are interestingly held also by the laptops own battery, which even now still lasts around 20minutes for use, perhaps a very long time while idle. I will see.

Here are some photos of the battery I replaced on my Toshiba laptop. I also had Sony laptops with similar batteries.

Replacing the NiMH CMOS battery in my Toshiba laptop

My Compaq LTE is the worst. It has a NiCd battery, for which there is no direct replacement.

Argh... I forgot how annoying it is to install Windows 2000 in dual boot without pre-partitioning the drive. I remembered to keep the USB flash drive disconnected, but I only created a C: partition. I installed 98, but then I wasn't paying attention during the Windows 2000 install as I was doing other things - I ended up with Windows 2000 installed on an E: drive.

I wiped out all 2K boot files on C: (except bootsect.dos), reformatted the D: drive (formerly E:), added bootsect.dos (for Windows 98 SE) back to the newly created boot.ini during reinstallation and everything is back to ham and eggs.

Dug out a 2nd newer Maxtor 120gb HDD to go in my P4. Dual 120gb and a 400gb for IDE plus a 32gb SSD for XP should be plenty for XP/Linux.

Have a Deep Cool Tessaract Iv been debating on putting my 8 core rig into off and on, currently in a 2010? Antec case. However I just found this PSU NIB someone gave me ages ago that is literally perfect for my P4 rig that is in a beige case. . .i guess the question is if a red GPU would clash with the blue?

ODwilly wrote on 2023-04-16, 07:42:

Have a Deep Cool Tessaract Iv been debating on putting my 8 core rig into off and on, currently in a 2010? Antec case. However I just found this PSU NIB someone gave me ages ago that is literally perfect for my P4 rig that is in a beige case. . .i guess the question is if a red GPU would clash with the blue?

I have that exact same power supply in my sons socket A system. It's a great unit I've owned it since new they are solidly built.

ODwilly wrote on 2023-04-16, 07:42:

Have a Deep Cool Tessaract Iv been debating on putting my 8 core rig into off and on, currently in a 2010? Antec case. However I just found this PSU NIB someone gave me ages ago that is literally perfect for my P4 rig that is in a beige case. . .i guess the question is if a red GPU would clash with the blue?

I would recommend a visual inspection of the capacitors inside of this power supply before you put it into service, even if it is NIB. There is always a chance of capacitor plague.

Also, brownie points for the Ultra brand. I miss TigerDirect.

Installed this Aztec sound card (AZT2320 chipset) on my 386SX build.

Initially, it was not detected (so I tried different fixes) but then I realized I did not insert it properly in the ISA slot 😓

Tested with Wolf3D: it works.

creepingnet wrote on 2021-12-20, 18:23:

LOL, yeah, I'm really winning the PC Chips "lottery" at this point, I've got 3 - a M912 in my second 486, the Moondog Tower, tha […]
Show full quote

dionb wrote on 2021-12-19, 14:51:

Ah, that old train-wreck of a board - SiS 5591 AGP chipset ruined by adding an onboard SiS 6326 instead of just giving you an AG […]
Show full quote

creepingnet wrote on 2021-12-18, 20:17:

Decided to go through the Super Socket 7 system and see if it works.....not really getting anything, not sure if it's because of […]
Show full quote

Decided to go through the Super Socket 7 system and see if it works.....not really getting anything, not sure if it's because of the on-board video not having a connector and ignoring the PCI Video card, or if it's because the board has problems. Starting to wonder

[...]

PcChips M590
Super Socket 7, 3xPCI, 2x ISA, 2x DIMM, 2x 72 pin SIMM, on-board sound and video. I have the same unknown PCI Video card attached to this one as well as I don't have a connector for the on-board video header. It has an AMD K6-2 333MHz CPU in the socket, and 64MB of DIMM memory in both slots.

No Post, no beeps either. Not a lot of jumper settings either. AMIBIOS586D1

Ah, that old train-wreck of a board - SiS 5591 AGP chipset ruined by adding an onboard SiS 6326 instead of just giving you an AGP slot.

Two potential issues:
1) 5V/3.3V jumpers by the DIMMs. Silly because the DIMM slots are keyed for 3.3V only, but if these jumpers (J2A+B) are set incorrectly, bad things happen. Both need to be on 2-3
2) despite "PC100" labels all over the board, 5591 doesn't support 100MHz operation.

Assuming the VDIMM jumper isn't wrong, I'd suggest assuming BIOS EEPROM was dead/corrupt and flash a new one to see if that helps.

🤣, yeah, I'm really winning the PC Chips "lottery" at this point, I've got 3 - a M912 in my second 486, the Moondog Tower, that runs great and has a real Cache believe it or not. Then there's the dead M919 that never worked, that has a DX2 in it....tempted to buy a third DX4-100 AMD CPU since that seems to be what that board came with and I've read about voltage adapters acting up on those. Then there's this hot piece of crap - M590 - which I remember I bought one on E-bay years ago and sold it to someone because I just was not a fan of it TBH.

After some testing, I'm starting to think the board may be ok, but the problem is that I don't have the VGA Graphics connector for it. I only have an external PCI card to see with. The lights on the Diagnostic card seem to be doing what they are supposed to - Reset blinks on, then clears, and IRDY blinks on and clears on the first boot-up.

The M919 seems to have a stuck Reset line, so maybe if I can get it unstuck it will work. Other option is to burn a new Flash ROM for it, but I need to figure out what options I have to do that with at this point since I don't have an EEPROM burner (at least, not yet).

I'm going to see which board I can get to boot first and then probably sell the one that does not work to whoever can replace them. Also, I'm tempted to recap the M919. What I find odd about that one is it was sold at RE-PC to me as a working board but it never worked.

Did you ever get the m919 to work?

Kahenraz wrote on 2023-04-16, 16:01:

ODwilly wrote on 2023-04-16, 07:42:

Have a Deep Cool Tessaract Iv been debating on putting my 8 core rig into off and on, currently in a 2010? Antec case. However I just found this PSU NIB someone gave me ages ago that is literally perfect for my P4 rig that is in a beige case. . .i guess the question is if a red GPU would clash with the blue?

I would recommend a visual inspection of the capacitors inside of this power supply before you put it into service, even if it is NIB. There is always a chance of capacitor plague.

Also, brownie points for the Ultra brand. I miss TigerDirect.

They seem to have used different components on different versions, I popped this one open when I 1st got it and it is full of chemi-cons and good caps! My very 1sr run in with a used Ultra was full of blown Fuhjyuus

I've got an Ultra 550 I ran at max load (Crossfire and OCed X2, kinda system the experts would say 700W minimum for) for 2 years, it was gonna be temporary, first I was surprised it actually booted the hardware, then I was surprised it didn't crash on desktop, then I was surprised it didn't die benchmarking, then I was surprised it passed a 48 hour torture test, then I just abused the hell out of it for a couple of years, didn't kill it yet, caps don't look bulgy. I trust it more than some lower end Corsair/Coolermaster 550s I've got now. With that and another Ultra brand item, a cooler, holding up real well, I am missing Tiger Direct too now... though I still curse them for somehow never having K6-III in stock when I tried to order. (In those days though it was a couple of day lag between me seeing something I wanted, paying some cash into account, and moving a little from my other savings, making sure funds available, then trying to order by which time they sold out.)

edit: I particularly miss their boxing day sales, still using monitors etc I got then.

I still have several Ultra branded items, and run into them with other people's equipment periodically. Their house branded gear was often surprisingly good quality for the price.

Hello, i got two vintage motherboards. One of two have oily residues mixed with dust basically everywhere, so the brush and air compressed is not sufficient. This mobo needs some proper "shower" but i never did that, i have some WD40, but i read that even dish soap and a brush is sufficient ( even diliited alcohol with normal water then let the board dry for 48 hours ). Is the WD40 a proper solution ? Thanks

Mild dish soap like Joy Ultra or Dawn if it's really stuck on. Alcohol can also help, but you need the soap and water for rinsing anyways.

It sounds like flux residue from the wave soldering machine during manufacture. These chemicals are soluble in alcohol with some scrubbing, but try soap and water first.

WD40 will leave a really gross film of its own. It's meant for mechanical parts to prevent corrosion. I don't know why it's abused so much for other things that it's not even good at. Just use soap.

I somewhat unexpectedly fixed a GeForce3 Ti500 (or is it 550?) which was previously artifacting like crazy.

Still not sure what exactly did the trick, but I'm glad that it works now. 😁

Kahenraz wrote on 2023-04-19, 14:11:

WD40 will leave a really gross film of its own. It's meant for mechanical parts to prevent corrosion. I don't know why it's abused so much for other things that it's not even good at. Just use soap.

For the record, you might want to specify the "original" WD-40 as the brand also has "specialist contact cleaner" designed for PCB and other electronic devices.

Kahenraz wrote on 2023-04-19, 14:11:

Mild dish soap like Joy Ultra or Dawn if it's really stuck on. Alcohol can also help, but you need the soap and water for rinsing anyways.

It sounds like flux residue from the wave soldering machine during manufacture. These chemicals are soluble in alcohol with some scrubbing, but try soap and water first.

WD40 will leave a really gross film of its own. It's meant for mechanical parts to prevent corrosion. I don't know why it's abused so much for other things that it's not even good at. Just use soap.

I tried, even with a brush....this misture it's sticky. I can rub it away with my finger, there is a film of grease all over the motherboard, i also tried with alcohol but nothing. I will leave everything as it is to avoid damages, but never mind!