PC@LIVE wrote on 2025-01-24, 23:36:

Having the pin connection scheme, you can have the board made, and then you have to solder all the necessary components, maybe 🤔 who has a too damaged board, could make it work again with a new PCB.
For example, I kept an old 386DX from the late 80s, maybe 🤔 it could be repaired with a lot of (a lot of) work, but having a new PCB made, ultimately could be the easiest solution, because you don't have to go in search of traces interrupted by corrosion.

This board with the chipset can only handle 286 CPUs. But I have also created a 386SX board. This uses a 386SX CPU. A different chipset is required for DX. It also has more address lines.

Nelson68k wrote on 2025-01-25, 10:47:

Test the maximum speed. At the moment it is running stable at 24.58 MHz.

The attachment IMG_9480.jpg is no longer available

The attachment IMG_9479.jpg is no longer available

I saw that you have a working card, congratulations 🎊🎈🎉🍾
In the next few days as soon as I can, I will try a fairly similar card, with headland chipset and AMD 12 MHz cpu, and 1 MB of RAM.

both schematics from the 286 and 386 sx would be a awesome starting point for a ht18 Project ( i know chipset is different to HT12). A pdf would be cool for me and i won´t blame anyone for mistakes

I was finally able to run the last test. The CoPro is recognized by the BIOS and the test programs run without errors. I think the board is now complete.

I'm tinkering with an ATX version of the 286 board. ATX power on/off is already working. There is even room for more stuff.

Nelson68k wrote on 2025-09-05, 17:19:

The attachment IMG_0006.jpg is no longer available

I'm tinkering with an ATX version of the 286 board. ATX power on/off is already working. There is even room for more stuff.

ATX is then treated as AT, requiring to still power down manually?
And has a place for a -5/12 close by?

Kudos!

Nelson68k wrote on 2025-09-05, 17:19:

I'm tinkering with an ATX version of the 286 board. ATX power on/off is already working. There is even room for more stuff.

Sweet! I'm really happy you are doing this.

But .... it really ATX (or LPX) without a PS/2 Mouse?

And a PLL for the Crystal would be cool to allow multiple speeds.

Nexxen wrote on 2025-09-05, 17:33:

ATX is then treated as AT, requiring to still power down manually?
And has a place for a -5/12 close by?

A simple push button is sufficient for switching it on and off. Just like with any ATX case. -12V is provided by the ATX power supply. The -5V is generated by a voltage regulator. This is located above the ATX connector. Unfortunately, I don't have a suitable component at the moment.

douglar wrote on 2025-09-05, 17:37:

But .... it really ATX (or LPX) without a PS/2 Mouse?

And a PLL for the Crystal would be cool to allow multiple speeds.

It should be exactly ATX. Not uATX or LATX, LPX ....

I'll change that to PS/2. Along with a few other things.

A PLL would be possible. I'll have to take a look. There aren't that many options: 12/16/20/25 MHz.

Nelson68k wrote on 2025-09-05, 17:42:

I'll change that to PS/2. Along with a few other things.

You might need to swap out the 8042 with a newer controller (Holtek HT6542B, AMIKEY-2, VIA VT82C42N) and then map the mouse ports to the controller.

And then get something like MR Bios on the board that knows how to support a PS/2 mouse.

Re: Native PS/2 mouse implementation for 386/486 boards using the keyboard controller

I have Mr. BIOS for the 286. And also VIA VT82C42N from my 386 replica.
However, I had considered integrating a complete I/O controller with serial, parallel, floppy, and IDE.

Nelson68k wrote on 2025-09-05, 22:03:

I have Mr. BIOS for the 286. And also VIA VT82C42N from my 386 replica.
However, I had considered integrating a complete I/O controller with serial, parallel, floppy, and IDE.

Cool. I/O controller. Great idea. You got the space on the PCB. Seems like the thing to do.

But the PS/2 mouse really fills out the I/O shield

What settings (and VGA) were the current 3dbench results for the board achieved with?

Looking back through my HT12 board results (I have two of them) I get 7-8fps at 20MHz, 0ws.

At 24MHz (the max I have been able to push an HT12 stably) I'm up at 8.9-9.9 depending on VGA card (Trident 8900 on the low end, ET4000 on the upper).

I was not able to get 25MHz running well on any of my HT12 boards.

I wonder if you are able to push the design any higher?

megatron-uk wrote on Yesterday, 06:37:

What settings (and VGA) were the current 3dbench results for the board achieved with?

The board runs at 24.58 MHz and 0 WS. I can't get it to run stably at anything higher. I use my CL-GD5401 256kb graphics card.

megatron-uk wrote on Yesterday, 06:37:

I was not able to get 25MHz running well on any of my HT12 boards.

Same here.

megatron-uk wrote on Yesterday, 06:37:

I wonder if you are able to push the design any higher?

The chipset is designed for 16 MHz. I think it's overclocked to its maximum at 24 MHz. I don't think it can be improved any further.

Yes, seems like it's the absolute maximum the chip can run at. One of those later HT22 parts would be really interesting!

I do like the Headland chips - probably the most solid 286 design, IMO. Plus the ability to provide EMS, and supported by The Last Byte to give UMB. They're definitely one of the best choices for a modern board like this.

I've tested HT12/A at 30MHz. Memory doesn't like it very much 😉

Re: Kixs's 286 to the Max Too

It's alive. https://www.youtube.com/watch?v=ulXmGhlY0zg

kixs wrote on Yesterday, 07:21:

I've tested HT12/A at 30MHz. Memory doesn't like it very much 😉

Re: Kixs's 286 to the Max Too

I have some 45ns SIMMs somewhere. I should test them sometime.

Nelson68k wrote on Yesterday, 10:20:

It's alive. https://www.youtube.com/watch?v=ulXmGhlY0zg

excellent!
might build one of these - indeed HT12 is a very nice chipset and I think I have a perfect donor for this 😁

Could the HT12 be replaced by some "pie", or other project?
They aren't mainstream to find... just curious.