No one knows anything about this?!

...wow. When I wanted to connect my HDDs in cable select mode, I actually thought this would be common knowledge, and I was the only one late to the "party" 😁 Not that i'm on the fronteer of discovering hidden knowledge that no one's ever even heard of 😆

I am using my VIA MVP3 board both (ide 1 and ide2 channels) at masters. HDD is master and IDE CD is master, no slaves here.
I have selected always master or slave and i have no racks, cant help much here. Cable select does not work how you want it?

B24Fox wrote on 2025-09-22, 21:45:

Unfortunately as long as it's single, and it's jumper is set to "Cable Select", it's always seen as "Slave" no matter where it's connected.
Only when I connect another HDD that is specifically jumpered to "Master" or "Slave", does the "C.Sel" HDD conform as opposite.

I'm not even sure how that switcheroo works, pin 28 should be grounded at the interface, and on a 40 pin CS cable, it IS wired to the middle, drive with CS grounded assumes master, and cut from the end, drive with CS floating assumes slave. The drives absolutely do not negotiate with each other. The 80 wire cable does it with a little more finesse, so the master (28 grounded) is at the end (better for impedance if there is only one drive) and the slave (28 open) is in the middle.
And 28 is only connected to one drive, so they cannot signal each other anyway.

I'd say maybe pin 28 is dry jointed at the pin header, but that still doesn't explain the behaviour with another drive

marxveix wrote on Yesterday, 14:57:

I am using my VIA MVP3 board both (ide 1 and ide2 channels) at masters. HDD is master and IDE CD is master, no slaves here.
I have selected always master or slave and i have no racks, cant help much here. Cable select does not work how you want it?

well, like I said: I've been struggling to make a single HDD that is jumpered as "Cable Select", be detected as "Master" or "Slave" only depending on it's placement on the PATA ribbon/cable.

Matth79 wrote on Yesterday, 16:31:

I'm not even sure how that switcheroo works, pin 28 should be grounded at the interface, and on a 40 pin CS cable, it IS wired t […]
Show full quote

B24Fox wrote on 2025-09-22, 21:45:

Unfortunately as long as it's single, and it's jumper is set to "Cable Select", it's always seen as "Slave" no matter where it's connected.
Only when I connect another HDD that is specifically jumpered to "Master" or "Slave", does the "C.Sel" HDD conform as opposite.

I'm not even sure how that switcheroo works, pin 28 should be grounded at the interface, and on a 40 pin CS cable, it IS wired to the middle, drive with CS grounded assumes master, and cut from the end, drive with CS floating assumes slave. The drives absolutely do not negotiate with each other. The 80 wire cable does it with a little more finesse, so the master (28 grounded) is at the end (better for impedance if there is only one drive) and the slave (28 open) is in the middle.
And 28 is only connected to one drive, so they cannot signal each other anyway.

I'd say maybe pin 28 is dry jointed at the pin header, but that still doesn't explain the behaviour with another drive

This behavior of a C.S. jumpered HDD, ignoring it's position on the PATA ribbon cable and defaulting to slave, happens on both headers of the mobo (channels 0 & 1). So it's not a bad solder joint, or anything like that.
And it happens regardless if I use an 80 or 40 pin/wire PATA ribbon cable. So poor connection due to faulty cables is also ruled out.
And just to put this out there; all the ATA-33 (40 pin) cables that I own and measured with my multimeter, are all "straight through"; no wire skips any of the connectors. (I dunno if there were multiple types produced, or what..)

B24Fox wrote on Today, 01:14:

This behavior of a C.S. jumpered HDD, ignoring it's position on the PATA ribbon cable and defaulting to slave, happens on both headers of the mobo (channels 0 & 1). So it's not a bad solder joint, or anything like that.
And it happens regardless if I use an 80 or 40 pin/wire PATA ribbon cable. So poor connection due to faulty cables is also ruled out.
And just to put this out there; all the ATA-33 (40 pin) cables that I own and measured with my multimeter, are all "straight through"; no wire skips any of the connectors. (I dunno if there were multiple types produced, or what..)

I don't want to over-generalize from my experience with the hand me down and other machines with IDE that I worked in, but:
while the cable select jumper was on the drives going far back, nobody (I'm referring to your local in-town system builder here that would use a motherboard like an MVP3) used it in the 40-conductor era. They manually jumpered to "master" or "slave". In contrast, cable select was ubiquitous in the 80-conductor era--think Pentium 4 and AthlonXP. I can't explain exactly why that is either. Possibly, because the 80-conductor ribbon was very specific to IDE (whereas 40-conductor ribbon had other purposes), they had to make a highly custom cable anyway, incorporated the cut needed on that pin for cable select to work.
It's not impossible that it has nothing to do with the chipset, but is just how that pin is wired on your board. You could use a multimeter to see where the cable select pin on the motherboard goes.

I looked at the back of the mobo, and a lot of pins (on the PATA headers) seem to have no traces going to them (including pin28). On the front of the board there are some traces going under the plastic housing of the connector, but it's very crowded in that area, and there's not nearly enough room for everything to be conected on the surface. So I can only deduce that most of them are connected via one of the inside layers of the PCB.
So unfortunately is's almost impossible for me to see/know where most of the pins go. I reccon they should all go to the southbridge, since the IDE controller should be integrated in there; but the contacts of the SB are underneath it.. so i can't even poke around there... 😒

It would be indeed a useful thing to at least know if pin28 is at least connected to anything on the mobo, but I really have no ideea onhow to check in this situation.