A lot more complicated than I first thought then..
but for what it's worth I pulled both of the chips one at a time and got lucky in that the chip type was printed on the underside..
27C256
Read them both with the TL866 and zipped the binaries up..
If nothing else you might find them interesting...and I have a backup now.
regards
VS

Well fwiw I flashed your xtide bin files xeven.bin and xodd.bin to 2x 27C256 eproms and popped these into the 2x vacant positions U32 and U40...
The 386 still booted of course as these were not mapped in presumably.
I then looked at what jumpers were available and relevant ?
Only 1 seemed relevant Jumper E12 ...default setting 1/2....which I moved to the other position 2/3.
No attempt to boot...nothing on screen.
moving this back to 1/2 ignored the 2 chips at u32/40..and the 386 booted as normal.
A bit out of my depth here as although I am a reasonable hardware technician...not so much re- software.
🤣

I recently aquired a Compaq III 286 version ...
After deducing that there was a faulty 256k simm in it...as parity errors occurred during the ram count.
I sort of have it working.
2 x replacement simms fitted I now get the full memory count...but had to hang a 1.2mb floppy outside the unit as the internal floppy drive is faulty.
I then created a setup floppy on my win 3.1 computer ...so allowing me to set the time and date...
Hard drive type 17...and rebooted.
The conner hard drive is still good..so I have a working system of sorts.
I really wanted to try the modified bios and xtide.
so UV erased 2x 27C256 and 2x 27C64 eproms..and wrote the U72 Odd.bin and U89 Even.bin to the 2 x 27C256 eproms verified both...then wrote XEven.bin and XOdd.bin to the 2 x 27C64...then verified...with my TL866 programmer.
Fitted the U72 and U89 ...and Xeven and Xodd ...but sadly the Compaq doesn't start up.
Moved jumper E12.
Have I got this wrong..?
Perhaps got the chip size/type wrong...or maybe in the wrong places...?
I'm attaching a photo.
and wondered if Leflash might take a quick look and perhaps tell me what I might have done wrong.
regards
VS

Just re-read instructions...and it does appear I've made a stoopid assumption in that I used 27C64 chips for the XTIDE odd/even chips..
I assumed wrongly that this would be ok..
Maybe not..
so I'm gona dig out a couple more 27C256...to erase and program.
Hopefully this will sort the problem.
fingers crossed.

Sadly ..no change...
Still no screen illumination...nor boot process...
So writing 2 replacement XTIDE odd even chips on 27C256 eproms made no difference.
I guess it's time to re-fit the original bios chips to make sure it still boots ok.
???

Left the 2 XTIDE chips fitted and just re-fitted the original bios odd even chips...and moved jumper E12 back..
boots no problem...so at least ..no damage done.
Phew..
I'm now wondering if my bios chips are diffent revisions to those modded by LeFlash..?
Could I have yet another version of the Compaq III 286 ..and his bios (modded) is incompatible with my hardware ?
Attaching a photo of my original bios chips ...
I shall read these and dump them here...
Compaq Portable III model 2660
There appears to be a lot of empty space on these 2 bios chips of mine..
FF from 0000 to 4460 on both chips.
rgds

So I guess LeFlash modded .bin bios files are variations on these REV 2 dumps I grabbed from here..
https://www.minuszerodegrees.net/rom/rom.htm
I wrote these rev2 .bin files to another 2 x 27C256 odd/even
and these also prevent my portableIII booting.
I opened each file with the TL866 programmer and ..yes they are substantially different to mine.
I have to assume there are more than one versions of the PortableIII 286.
..and when it comes to bios ...one size doesn't fit all.
The model number IS important.
rgds

Ok..none of this was making sense to me...ie ...why so much unused space in the bios chips.
Then it hit me..
these are not 27C256 chips..but I had been reading them as such.
Finally plucked up enough courage to peel one of the stickers ...gently...so as not to rip it.
Sure enough ...mine are 27C128...so 16K chips.
Now it makes sense.
Further digging makes this the earliest version ? rev K ?
So I'll read them again...but I suspect they are already archived somewhere.
but I guess this is why LeFlashes modified bios doesn't work for me.
I'm pretty useless beyond hardware issues..
perhaps he can mod this version ..or perhaps someone already has ?
rgds

My Portable III 286 is coming on nicely..
Having been successful with the modded bios and integrating xtide internally....I can now use CF cards.
Here is a photo of the physical mod with the side mounted CF reader.
Once I have aquired another 3.5" to 5.2" carrier/frame ...I will be fitting a 1.44mb floppy drive..instead of the 1.2mb drive.
Nice though it is ..simply not as practical since disks are difficult to come by.
The only thing driving me nuts now is that I only have access to 640k base memory.....and only the ram expansion card is the way to get more.
I live in hopes that one might turn up on ebay eventually...but unlikely imo.
rgds
VS

Vipersan, you have a rev K BIOS.
To replicate LeFlash work, you need to change two bytes of U72 - one at 0x3896 from 0xDF to 0xEF, another at 0x1B7E from 0x20 to 0x10.
If it won't work, revert the change at 0x1B7E, then also change two more bytes at 0x1D63 - from 0xE8, 0xE5 to 0x90, 0x90 and, finally, one byte of U89 at 0x1D64 from 0xCA to 0x90.

Ooooh. I have a recently-repaired Compaq P3, a set of 27C256 ROMS, a ROM programmer, and a 3D Printer, and I have been wanting to get the system working with XT-IDE and Compact Flash cards! 😁

I just need to wait for my cables, and some CF-IDE adapters, and I'm in business!

OH BABY!!!!!!

I ordered the electronic parts a few hours ago, but I wanted to see if XT-IDE would load from 27C64 EPROMS. Turns out they DO!

I may re-program these later with an up-to-date version, but JUST SEEING THIS MADE MY NIGHT.

All right. I've updated the ROMS with the latest version. I have had some problems where I can't seem to get multiple CF cards on the IDE cable, though. It only accepts one, and that's on the end of the cable. Even with the drive positioning jumpered, it considers both adapters as the same drive, and causes a conflict. Is there a way to avoid this?