We have an old CNC machine at my place of employment. It was retroffited with a controller back in the 90's. The operationg system on the machine is OS/2 warp. Well just recently, the CMOS battery failed and now I cant save the Date and time. A new battery has been installed and still the same issue. So at this point, I believe the MB is bad. My question is, can I find a replacement motherboard or cross reference one? I've attached a couple photos with the Motherboard part numbers and the error im getting at boot!
That's awesome! :-)
If you replace the CMOS battery it should work. Make sure you've replaced it correctly by testing voltages at various points downstream from where you soldered it back up.
If the software running on the machine is using a standard communication output like serial or parallel to drive the CNC device, then you should be able to get it to work with any piece of hardware that OS/2 supports so long as the hardware ports are mapped the same. I would try to mate up the motherboard MHz as close as possible, as well as memory and graphics card.
If, however, it is running a dedicated type of motherboard with explicit input/output ports, you'll have to find the exact duplicate machine. I think it's unlikely it will have dedicated input/output ports. You may also be able to find from your CNC machine online what kind of software and/or codes it uses to operate. They may have updated drivers that will run on newer equipment, even from another OS or would allow you to run OS/2 Warp in a VM in a more modern OS.
Pretty awesome. In 1996 I developed the software for a part positioner system for an electron-beam welding device. It was used to weld jet engine intake and exhaust manifolds halves together, as the top/bottom were machined separately. I used a regular DOS-based system for that controller, coupled to an interface card which received commands and values from my program. I basically gave it data on where I wanted it to go, and it reported back in real-time where it was. I then adjusted my target (multi-axis velocities) based on how it was tracking. I had planned one more stage before the project was pulled, and that was to map the known quirks of the device and create a compensatory system which took those quirks into account so that I would not be quite as reactionary to the real-time feedback for adjusting my position, but instead would already know where it lags behind or leaps ahead, and then have those already factored in as by data from the mapping analysis. Would've been awesome.
If you can find out the language for the CNC, and it's simple enough, you might be able to write a new controller for it. Would be a fun project (unless it's in a production environment, then it needs fixed now I assume). :-)
Best regards,
Rick C. Hodgin
