You can try the DOS version of www.hwinfo.com to see if there are any IRQ conflicts.
This tool also fits on a floppy disk.

If the serial port is not directly mounted to the I/O card, make sure you have the right breakout cable layout for the serial port ribbon cable.

Different I/O cards can sometimes use different ribbon cable layouts ... it's a pain.

canthearu wrote on 2021-09-23, 06:23:

If the serial port is not directly mounted to the I/O card, make sure you have the right breakout cable layout for the serial port ribbon cable.

Different I/O cards can sometimes use different ribbon cable layouts ... it's a pain.

Oh I know this pain - I have tested all my cards and the ones that have the right cable are in a separate pile. One of my known good cards has the RS-232 directly mounted to the PCB so that eliminates any potential cabling issues anyhow.

You can try the DOS version of www.hwinfo.com to see if there are any IRQ conflicts.
This tool also fits on a floppy disk.

I have a few tools I use to check IRQs, but it never hurts to have more - thanks for the link 😁

I would check out the serial port in the problem machine with another serial connection, null modem cable maybe, and if that also has problems, suspect a corrupt BIOS.

Jed118 wrote on 2021-09-23, 03:39:

Here's where I got somewhere - I unplugged the IDE channel entirely from the controller card and booted from a diskette, then loaded the Microsoft mouse driver - it found a mouse and loaded it on COM1 - I did not have a program to test it (I'll get QBASIC and EDIT on the boot diskette in the next coming days). I thought it might be the CDROM, so I disconnected it from the chain (IDE cable). No change. Swapped the hard disk with one that had only COMMAND.COM on it (Still with CDROM disconnected), manually loaded mouse drivers - no dice.

So, booting from the HD and trying to load the MS mouse driver fails when it can't find the mouse. But disconnecting the HD, booting from a floppy and loading the MS driver seems to work (or at least doesn't give an error)? When happens if you connect the HD, but still boot from the same floppy and try loading the same driver (i.e. don't do anything that accesses the HD)? If you can, might also be worth checking if you've got -12V on any of the serial port pins.

I didn't catch that, maybe we're looking at a disk controller problem... or the annoying thing that tended to happen more on pre-Pentium systems... a hard disk doesn't work right on a particular system unless it's partitioned and formatted on that system. (BIOS geometry translation differences were assumed to be the culprit)

BitWrangler wrote on 2021-09-23, 15:20:

I didn't catch that, maybe we're looking at a disk controller problem... or the annoying thing that tended to happen more on pre-Pentium systems... a hard disk doesn't work right on a particular system unless it's partitioned and formatted on that system. (BIOS geometry translation differences were assumed to be the culprit)

You might be onto something (albeit this is an IDE system) - I plug a 270 Mb drive into this computer, which has been shown to boot on several other machines - and it says MISSING OPERATING SYSTEM. This BIOS has an autodetect feature though, and I know the 270 Mb drive is detected correctly (it has the CHS right on it, it's a WD Caviar unit)

However, the 850 Mb drive that's in there now for testing purposes was formatted on another 486 and it boots OK. However, that controller (ISA multi I/O) AND that hard disk (the 270 Mb), when transplanted into another computer, works and boots.

I'll try Snufkin's approach, but I think I may have tried it late last night (or something similar to it)

Jed118 wrote on 2021-09-23, 18:49:

However, the 850 Mb drive that's in there now for testing purposes was formatted on another 486 and it boots OK. However, that controller (ISA multi I/O) AND that hard disk (the 270 Mb), when transplanted into another computer, works and boots.

I had some weird issues with 386/486 mobos and CF cards, some would work, many would not - but it's not just down to BIOS, I've tried several multi I/O cards and some were more tolerant than others. The weird thing was with some combinations I was getting a HDC failure error and also FDC as well, even though the floppy drive worked perfectly with different CF card and/or I/O adapter. So maybe you have something like that but with HDDs?

One more thing the HDD and serial ports have in common is 12V line. I suppose if the PSU has dried-out caps and the line is noisy, more load on it would cause all kind of issues. So try another PSU and/or that HDD that isn't working with serial ports but connect it to power only (not the IDE cable - boot from floppy). See if that changes anything.

Thanks guys - here's what I tried next:

-Rail voltage was 5.19v, 12.39v

-No IRQ conflicts (handy little utility the hwinfo!)

-I took the multimeter to the pins of COM2 as it was dangling on a ribbon and got some voltages on several pins (-8v, some were -5v) and I made a note of it to turn the PC around and do a proper mapping of what comes out of where on COM1, but before that I skimmed this page for more suggestions and saw Deunan's post - I swapped in an old ATX power supply through a converter, and wouldn't you know it, the damn thing picked up the mouse. Tried COM2 through a DB25-DB9 adapter and a different (MS) mouse, and it worked too. Tried CuteMouse on both ports and it worked too! I also made some changes in Autoexec and Config - I thought that the invocation of the CDROM (both the boot diskette and the hard disk invoke the driver through CONFIG.SYS) had something to do with it so I reconnected the original PSU and put in the boot diskette with the hard drive connected and in CMOS correctly, booted to A:\, swapped the MS diskette to A:\, ran the MS mouse driver and... it worked.

I rebooted several times, took a break, made tomorrow's lunch for work, came back and... mouse does not work.

(I'm saving this for future edit as the laptop I'm writing this on bluescreened twice)

TBC...

Put the ATX power supply back in and we get signal!

Goddamn, I need to get myself a scope already.

Anyone have any suggestions on where to get a good scope or other power supply testing hardware?

Thank you all for your suggestions BTW!

Yeah, I had the thought of the PSU being cantankerous as a possible reason for your problems. Someone else beat me to the idea before I could say it. Seems that they were right.

Yeah, I want a good o-sillyscope as well, but I'm not sure if I'd ever save enough computer hardware with it to be worth the expense.

Jed118 wrote on 2021-09-24, 02:03:

Anyone have any suggestions on where to get a good scope or other power supply testing hardware?

For new cheap scopes look for Rigol or Siglent brands, the entry-level series that start at about 50MHz of BW will probably do everything you need. No point in getting used analog scope unless you get it free (and working) or you need properly working XY mode. While getting a scope to test a PSU every now and then is probably not very cost effective, it is a nice a tool that can come in handy in other cases as well (looking at clock and other signals on semi-dead mobos for example). In the end you have to decide if you want/can afford a new toy - but as I've mentioned, these days there is a lot of cheap and decent stuff available. I do not recommend so-called USB scopes. Very limited, usually pretty poor in general, and the somewhat decent ones are expensive enough to consider a standalone model. Plus you'll fry whatever is connected to that USB if you mess up, which would most likely be your primary PC.

Just keep in mind that while it's safe to scope the secondary side of switching-mode PSUs, you SHALL NOT connect it to the primary side. Or it's bye-bye new toy, and the PSU possibly as well. Not to mention potential hazards to health and life. The only proper way to ever look at the primary side is via differential probe. I know there are some "workarounds" like messing witht the grounding and/or isolation transformers but in reality none of that is ever fully safe.

And do note that many PSUs need a minimum load to properly regulate output, or even start at all, so if you are really into testing those you will also need some dummy load. For now some old mobos and HDDs might do, or even some 12V lightbulbs (that's what I use).

Deunan wrote on 2021-09-24, 08:35:

Jed118 wrote on 2021-09-24, 02:03:

Anyone have any suggestions on where to get a good scope or other power supply testing hardware?

it is a nice a tool that can come in handy in other cases as well (looking at clock and other signals on semi-dead mobos for example).

These are starting to pile up, and I know I have the skills now to tackle them, just not the tools.

Deunan wrote on 2021-09-24, 08:35:

Just keep in mind that while it's safe to scope the secondary side of switching-mode PSUs, you SHALL NOT connect it to the primary side. Or it's bye-bye new toy, and the PSU possibly as well. Not to mention potential hazards to health and life. The only proper way to ever look at the primary side is via differential probe. I know there are some "workarounds" like messing witht the grounding and/or isolation transformers but in reality none of that is ever fully safe.

And do note that many PSUs need a minimum load to properly regulate output, or even start at all, so if you are really into testing those you will also need some dummy load. For now some old mobos and HDDs might do, or even some 12V lightbulbs (that's what I use).

I did not consider the USB ones for that very reason.

Of course, I recall the primary side training I got at Agilent over two decades ago 😉

I have a few dead hard disks that spin up only to make an awful racket - I also have a ceramic resistor from an LED flasher "adapter" for car LEDS (I modified the flasher directly, screw adding resistors, 🤣 what's the point!)

Thanks 😁