Checking for circuits, but then Copilot is telling me that an Ultra 3 / LVD U160 is not compatible, even with the cable to a SE mode interface on the 2940UW, you'd need a 19160, 29160, 29320 - unless that cable is a fancy active converter

I've gotten lucky and found old stock SCSI cables & terminators at my local microcenter. They don't have any of it in the online inventory, and sometimes I get a strange look from the cashiers when buying them.
I would not recommend trying to make your own 68-pin terminators, the pin spacing is brutal.

Small pin spacing should not be a problem.
Unless something weird, of course.
I can deal with electronic and pcb, it's been one of my jobs.
I live in a small town in Canada, not much Microcenter or whatever.
A local store sold me a Voodoo 2 and a Vortex 2 for 20$, he is fabulous but not magic 😀
I don't know much about SCSI compatibility. In the 2000 I had a 18 G with a also 2940U2W., a scanner and a cdrom. It worked perfectly.
I've read about nightmares but never experimented .
Let's say I'm a noob about SCSI.
I just suppose the electronic circuit for a terminator may be quite simple, and I'am sure i can replicate it if I get a schematic.
Passive or active.

Matth79 wrote on Today, 00:37:

Checking for circuits, but then Copilot is telling me that an Ultra 3 / LVD U160 is not compatible, even with the cable to a SE mode interface on the 2940UW, you'd need a 19160, 29160, 29320 - unless that cable is a fancy active converter

This is true. Your disk is U160 SCSI(LVD). Your SCSI card is Ultra-wide SCSI(SE). I also recommend getting a different scsi card as lvd scsi disks are much easier to get. Bear in mind some lvd scsi cards does not support optional boot rom support. Although i encountered some 68pin seagate drives support both se and lvd.

homemade scsi terminators are not easy as high speed scsi has strict requirements for proper termination. usually it as achieved by discrete ic such as UCC5640.

https://www.ti.com/lit/ds/symlink/ucc5640.pdf … www.ti.com%252F

In theory, one can reproduce the functionality of such ic with modern parts....

15MB/sec off an IDE disk is about expected for that system.

The 2940 might double that.

In the end, it will be the access time and not the throughput that makes the difference.

I once tested win98 on a k6-2 475 on an ali chipset board with a 2940UW / 7200rpm 4.5GB Seagate jobbie. The win98 screen just appears for a couple of seconds while the disk goes gggggk before it enters the gui. Shutdown is faster than the screen can mode switch. Quite mind blowing. I attributed it to the tagged command queuing improving seek efficiency.