Internals:

Motherboard:

It's almost small enough to be a single board computer on a backplane.

I can't help exactly but often these POS terminals were sold under all sorts of different names. you can try to google any and every part number on the motherboard to find its fellow brother and sisters.

Got the Dallas rtc replaced with socket and a coin cell holder. Also have two quiet fans on order to replace the really noise ones fitted to the case (one on that CPU heatsink/heatpipe, the other bolted to the chassis drawing air out of the (fanless) psu... But it's a crazy tiny 40mm job!

I'll see what it is like after it comes together, but think maybe an ess 1868f and possibly a Dreamblaster S2 for audio options. I'll almost certainly fit a PCI video card instead of the mobile Cirrus chip. I think I have a Diamond Stealth 2000/virge somewhere, as well as a Millennium II.

chinny22 wrote on 2023-01-05, 16:46:

I can't help exactly but often these POS terminals were sold under all sorts of different names. you can try to google any and every part number on the motherboard to find its fellow brother and sisters.

Unfortunately the motherboard is totally unhelpful - not a single version number or model number on it other than "V 4.0".

And literally as I just posted that....

It does list voltage support down to 2.3v, but no mention of MMX processors. But it also doesn't specifically say it supports the IDT C6, which is what is fitted...

cool find. how do those winchips compare to their amd/intel peers?

Can't really add much, other than to say those DigiPoS machines are pretty cool.

I've got 3 of the later DigiPoS "Power PoS" Socket 370 Tualatin-based Celeron systems, the internal layout and case design is very similar to yours, I guess DigiPoS didn't want to make any radical new designs over multiple generations.

drosse1meyer wrote on 2023-01-05, 23:55:

cool find. how do those winchips compare to their amd/intel peers?

The FPU architecture is supposed to be pretty weak, worse than even the Cyrix 6x86, but they are relatively low power compared to the P54/55, 6x86 and K5.

They are listed as having MMX, but also have a relatively large L1 cache (64kb Vs 32kb of the Pentium) but are a simpler pipeline design than the Cyrix, AMD and Pentium so are not really competitive clock for clock.

This one is playing funny games with me - it's really not happy with the PS/2 keyboard through my Belkin kvm; every so often the keyboard stops responding and the 'system selected' indicator on the KVM front panel (which usually stays on when a PS/2 keyboard is detected on the selected port) blinks (meaning 'port selected - but no keyboard detected').

Can't seem to get it to boot from a basic 512MB CF card which I tend to test all my systems on first (works in CHS mode, so it's an easy way of running diagnostics across pretty much all of my machines - whether disk-size limited, with an XT-IDE bios, or something newer with native LBA). The floppy drive also seems bust - it seeks on power on, but the post test says 'floppy drive(s) fail', so I'll need to try some others.

It's not looking great for this thing so far!

Definitely some keyboard issues with this thing - key repeat seems broken (i.e. hold down a key and it doesn't repeat... at all), and the keyboard still 'goes away' now and again - it's not the KVM, as I've tried direct connection and the same symptoms appear. There's no corrosion anywhere (since it had a Dallas rtc) and no damage that I saw whilst cleaning up the board.

Can't get any floppy drives working with it - every single one reports a "Floppy drive(s) fail - 40" error at the POST screen.

The standard BIOS won't detect anything I've attached to the IDE channel other than the 4.3GB Seagate it came with, I did manage to get an XT-IDE rom on a RTL8139C 10/100 NIC running (I was needing a NIC installed anyway, as there's not going to be any other easy way to get data/software on/off this thing) and thankfully it will detect and boot from those devices that the BIOS won't detect (including small disks that should be within the BIOS LBA limits).

I can live without the floppy, but the keyboard behaviour is pretty much a dealbreaker if I cannot resolve it.

What's the voltage on the PS/2 +5? Think there have been cases with similar symptoms of random disconnects where there was some resistance on the +5 line causing the keyboard to reset when drawing current.

megatron-uk wrote on 2023-01-06, 13:51:

Definitely some keyboard issues with this thing - key repeat seems broken (i.e. hold down a key and it doesn't repeat... at all) […]
Show full quote

Definitely some keyboard issues with this thing - key repeat seems broken (i.e. hold down a key and it doesn't repeat... at all), and the keyboard still 'goes away' now and again - it's not the KVM, as I've tried direct connection and the same symptoms appear. There's no corrosion anywhere (since it had a Dallas rtc) and no damage that I saw whilst cleaning up the board.

Can't get any floppy drives working with it - every single one reports a "Floppy drive(s) fail - 40" error at the POST screen.

Screenshot 2023-01-06 13-27-28.png

The standard BIOS won't detect anything I've attached to the IDE channel other than the 4.3GB Seagate it came with, I did manage to get an XT-IDE rom on a RTL8139C 10/100 NIC running (I was needing a NIC installed anyway, as there's not going to be any other easy way to get data/software on/off this thing) and thankfully it will detect and boot from those devices that the BIOS won't detect (including small disks that should be within the BIOS LBA limits).

I can live without the floppy, but the keyboard behaviour is pretty much a dealbreaker if I cannot resolve it.

hmm. dodgy solder joint? maybe try a little deoxit in the port?

It reads +4.94v, which I guess is within tolerance.

In edit I get key repeats on cursors and delete, but not on backspace or any other keys.

If I type at normal speed at the command line more than half the characters are missing, I have t o t y p e l i k e t h i s a nd p a u s e b e t w e e n e a c h k e y p r e s s. If I type slowly enough at the command prompt, all the characters are fine. Type fast, and they get dropped.

It feels more like something which is being set by the bios/keyboard controller, than anything else. I've tried with typematic delay on/off, minimum/maximum delays, etc.

Well it's running and it seems to be going through the various benchmarks tools without issue, but the keyboard is still playing up.

Here's the out-of-box results for Speedsys:

I did have to take out the Millenium II that I had fitted, as my Magewell HDMI capture card didn't like the frequency that the MGA was outputting various high res modes. It's now got a Voodoo 3000, but I'll go back to the onboard chipset once I've ran through the full suite of benchmarks.

Quite a few options in the BIOS to tweak for a point of sale system!

Reading the manual, it does have split voltage support for the 6x86L, so does that mean it would also support a P200MMX, I wonder?

Okay, so this is interesting; I tried another keyboard (my go-to keyboard for all the retro stuff on this KVM is basic Digital Equipment Corp PS/2 keyboard (part number PCXLA-KE). Nothing fancy, no multimedia keys, just a UK/ISO layout.

I thought maybe there was something in the keyboard that was just not happy with the Digipos, so I pulled out a common-as-dirt Dell PS/2 keyboard, US layout, from my spares collection.... and the keyboard won't even power up when connected directly to the Digipos, it just sits flashing the keyboard LED's constantly:

https://www.youtube.com/watch?v=UuBwfvK_TIo

It's almost as though there isn't enough power going to the keyboard port, or the clock signal isn't stable.

Video of the Digipos in action showing keyboard input being dropped as the typing gets quicker:

https://youtube.com/shorts/pXgD0HT0VLU

Stick to hunt-and-peck input and it would be (more or less) usable, apart from the lack of key repeat function.

Not sure there's much more I can do to resolve this one; as with most boards beyond the 486 era there isn't a discrete keyboard controller any more, and there isn't a plethora of bios images to try - the one option would be to try a vanilla Aladdin III bios (if I can find an attachment for my TL866 that will accept one of the fancy qfp flash devices this board uses).

It's not the BIOS / settings that are doing this, as I tried the various M1521 images available here: https://theretroweb.com/bios/?biosManufacture … 5&chipsetId=178

A couple of them work and the system boots up (not with the onboard Cirrus vga, as none of those bios files have an onboard CL7528), everything looks normal-ish, but key repeat and the 'fast typing drops keyboard input' symptoms still remain. It must be hardware.

The only thing I can think of now is to desolder the AT keyboard header and replace it with a new part (there's a PS/2 Pinout beneath the solder points for the AT header, so I may as well go straight with that).

One final thing would be to replace the power supply with either ATX (won't fit inside the case, but good enough to test) or a PicoPSU - but the power header on the motherboard is non standard, so I'd have to make up a small hardness to go from an ATX plug (or ATX to AT adapter) to the one this board uses.