can anyone recommnd a good tutorial for using a hex editor to split a bios odd/even ?

Vipersan wrote on 2023-03-29, 10:38:

I must conclude that the Compaq III 286 lunchbox has more in common with XT than AT.

Okay..

Can't judge. As I said, I don't have access to those systems. 🤷‍♂️

Vipersan wrote on 2023-03-29, 10:38:

All now run versions of XTIDE at boot...but only IDE_XT.bin has been successful in booting the original hard drive ...and sometimes the floppy drive.

But how? Did XTIDE Universal BIOS (XUB) manage the original hard disk drive by itself or did it give control back to the BIOS?

I'm asking bevause XUB has the ability to co-exist with the original HDD routine, I vaguely remember.
That's handy if an existing file card or MFM/RLL controller is installed, too.
XUB was meant to be used with those XTIDE cards (XT CF Lite etc), after all.
It was intended as an add-on to PC/XTs.

Vipersan wrote on 2023-03-29, 13:42:

can anyone recommnd a good tutorial for using a hex editor to split a bios odd/even ?

WinHex has such an ability. It can also merge odd/even files.

That being said, I don't know of any tutorial. Surely there are, I just haven't read/watched them.
Maybe someone can give a tip? Please? 🙂

Hi Jo
I certainly didn't press anything to allow control to be handed back to the bios.
It may well have happened automatically ?
All I observed is that the original 40mb C: drive was not 'identified' but was 'seen' and booted.

If I can learn how to use WinHex to split a bios...then in theory I can split IDE_XT to flash to the 2 ...27C256 chips the modified bios is expecting to find.
..or at least give it a go..
🤣

Not Sure what I have yet....but looks promising..
Found and downloaded Romwak.
Then activated the exe in CMD window..
using..
romwak_x64.exe /b IDE_XT.BIN OUT1.BIN OUT2.BIN
IDE_XT.BIN an 8k bin file was divided into two 4K .BIN files
Out1.BIN and OUT2.BIN
certainly looking promising.
regards
VS

Looking good...looking _very_ good
Wrote the 2 bins to 2x 27C256 eproms (not forgetting to fill the buffer with zeroes - dont want them pesky checksum errors) and popped them in the Compaq..
didn't boot at first but 50:50 chance I had them the wrong way round..
no way to know which bin is odd and which is even...other than to pull them again and swap them over..
here is the result.

Still have to change that boot order I think.

...and it does indeed look like having identified and found no hard drives ..it defaulted to the internal ide ports...
I must build an OS on a more modern drive or CF card...to check if XT_IDE is actually working.
rgds

It's a start. 😀 👍

In the screenshot, XUB scans port 300h - that's a weird address.
- It's being used for XT IDE card rev1, according to the web ? And some network cards (MPU-401 also often is on 300h or 330h) ?
https://forum.vcfed.org/index.php?threads/get … orking.1238012/

From what I know, an ordinary AT IDE port is at 1F0h (primary) or 170h (secondary).
- Some XT IDE ports are said to use 300h, 320h or 324h though.

The older Compaq Portable II is said to use a "standard 40 pin IDE hard disk controller card".
Because "Compaq were the first to use IDE standard".

https://www.angelfire.com/amiga/pc100/compaq2.html

Edit: Can't find information about the Compaq Portable III, except the manual (not helpful, no details).
However, this guys seems to has successfully installed an CF adapter in a Portable III:

https://www.oldsilicon.com/compaq-portable-iii

Maybe that person has some more information, dunno. 🤷‍♂️

Good news ..and some bad.
I reverted to the Leflash combination of modified bios which maps in the vacant sockets after moving jumpers E10 and E12 on the mainboard..
got this working yesterday..
ran the compaq setup floppy in order to set drive type to zero...I figured this a neccessary step to avoid conflicts..ie...between the compaq and xtide fighting for the IDE interface.
I'm using IDE_AT version of XTIDE.
although I have learned how to config the xtide bios before splitting and flashing...
useful info though not really neccessary since LeFlash provides binary dumps of all 4 needed chips.
All these binary files are written to 27C256 eproms.
The story so far..
XTIDE AT runs at boot though I might at some point change the boot order with the config tool.
I built 2 DOS operating systems ...
one on a flash card..
The other on a quantum 544 mb hard drive.
and of course I still have the original conner 40mb hard drive to test..which has windows 3.x on it and is bootable.
So...what actually works ?
The original conner is detected ...not identified ...and never boots.I'm not really surprised by this and is perhaps related to the inability to identify it.
Presumably when xtide interrugates the drive and gets no response it gives up and hangs.
The Quantum drive OS dos 6.22 is identified and boots just fine(prepared on windows 98se pc).
The 512mb compact flash ..also contains DOS 6.22 (prepared on windows 98se pc).and boots no problem in my windows 98se machine......is seen ...indentified...but also refuses to boot ?
Does this mean I am restricted to using mechanical hard drives ?
Maybe just this brand of CF is the problem...but yet it boots ok on another machine.
Once again I thank LeFlash for his work ..and for sharing the binaries.
Comments and help always welcome...particularly regarding the CF issue.
rgds

Been trying all day to get XTIDE to use CF cards..
tried various adapters ..ribbons...and CF cards..
It simply doesn't like them..
Sometimes I can get a card recognised...but no chance of booting.
On the other hand it appears to work just fine with mechanical drives....such as the one I just prepared.
DOS only on a 2gb partition ...on an 8gb drive.
the lack of CF compatibility has me totally stumped ?
rgds
...hopefully one day some enthusiast of these venerable old lunchbox computers will clone the ram expansion card...and I will get to add XMS.
'till then I guess 640k base memory it is..
Sad that I have larger simms but they cannot be used in the 2x simm slots.
256k per slot appears to be max.

The CF adapter in the picture.. Does it allow inserting the CF card in both ways?

In the past, I had encountered CF adapters that took CF cards in both ways.
Confusingly, one of them wanted the CF the wrong way around - the label had to face downwards.

So please, if you can, check the CF card orientation. Maybe the adapter is not correctly wired, not sure.

Edit: If the card's name/model number appears on screen, it's correctly installed.

Edit: The old mechanical drives perhaps pre-date ATA-2 specifications, that's why they can work with the old BIOS out-of-box.
Modern CF cards by comparison were made with ATA-2 and later in mind.
XTIDE Universal BIOS does support IDE DOMs and CF cards just fine.

Edit: My apologies for the poor wording here. I don't know how to put things into words right now.

I forgot to ask - did you use an 1:1, 40 line ribbon cable for IDE?
The modern 80 line cables perhaps cause trouble. While the extra grounding wires itself are no problem, the special wiring that allows the 80 line detection might be a problem.

Hi Jo..
I have tried 3 different cf adapters just in case one was faulty...etc..
..and at least 2 ribbon cables 1:1
I have also tried a CF adapter plugged directly into the ide socket ...one with a female connector....so as to eliminate the ribbon.
Also tried various brands of CF ...very occasionally one would be recognised and identified...and even though it contained a bootable copy of DOS...
(tested on my 98se PC)...it just hung and never booted.
I have read there is something else I can try..
booting to a floppy with Fdisk on it ...and running fdisk /mbr
not tried this yet..but nothing to lose I guess.
I dont suspect the problem is anything to do with the CF adapters I've tried...but could be CF card related.
...or maybe the old/original IDE interface actually on the motherboard is incompatible with CF ?
A timing issue perhaps ?
most folk would be using an XTIDE adapter pcb with a CF adapter connected directly to it ..plugged into an ISA slot..
I'm trying to avoid this route...therefore not having to use the ISA sidecar/backpack.
rgds
VS

Currently in the process of building a copy of dos6.22 on a 2gb Transcend CF....
I tried to run up disk 1 of dos 6.22 on a 3.5 inch floppy+drive...but it refused to complete the boot from A:
Ok...this might have something to do with the bios not being set up correctly...and by that I dont mean XTIDE.
So made yet another copy of Compaq setup disk...but this time on a 1.44mb floppy.
This I did manage to boot successfully.
Once in the Compaq setup ...I made sure 1.44mb floppy was selected...it was ..but I confirmed this and set the hard drive 'type' to zero...ie no hard drive...saved and rebooted...with the transcend CF attached and disk 1 of dos 6.22 in the drive...
This time it DID boot.
This meant I could format the transcend to FAT16 from within the dos6.22 setup.
It formatted the Transcend formatted and rebooted to allow the installation to proceed..
Now installing from disk 2
Fingers crossed it will complete...and reboot at completion to DOS/Transcend 2GB CF.
Some weird crap going on here I think..
..and I'm sure there is some sort of conflict in play with XTIDE and the Compaq bios fighting for control..
I'll report back if successful.

That was painfully slow...but then I only have 640k ram ...and 12mhz clock...
but it worked !!!
..and the first time I didn't use an OS buit on another PC..
This OS install should be better set-up for the Compaq.
I might try this again with another Transcend and windows 3
rgds
Time for a beer..
I think I earned it.
cheers

That specific CF adapter I see on the photo has the problem, it grounds IDE pin 28, which breaks the ISA bus logic on old PCs where the line 28 in the IDE cable exposes ALE signal from the bus.
And it might even work seemingly ok, but make hw misbehave depending on circumstances. The copper around that pin should be cut from the ground plane.

vstrakh wrote on 2023-03-31, 08:24:

That specific CF adapter I see on the photo has the problem, it grounds IDE pin 28, which breaks the ISA bus logic on old PCs where the line 28 in the IDE cable exposes ALE signal from the bus.
And it might even work seemingly ok, but make hw misbehave depending on circumstances. The copper around that pin should be cut from the ground plane.

That is seriously usuefull info..vstrakh
I'll get that sorted today..
cheers
VS

Hopefully I isolated the right pin...as sadly they are not numbered on these adapters..
I have checked the pin I isolated with a meter in continuity mode.
I have 5 of these which all receive the same treatment if confirmed correct.
I'll be giving it a try shortly..
Tea and Toast takes priority
🤣

Vipersan wrote on 2023-03-31, 09:16:

Hopefully I isolated the right pin...

Yeah, that pin: Re: Troubleshooting 386 motherboard, video memory writes clashes with PCM audio playback

vstrakh wrote on 2023-03-31, 09:25:

Vipersan wrote on 2023-03-31, 09:16:

Hopefully I isolated the right pin...

Yeah, that pin: Re: Troubleshooting 386 motherboard, video memory writes clashes with PCM audio playback

Many many thanks mate..
Today ..I learned something.
🤣

Tested and working...and hopefully this will fix the odd ..unpredictable behaviour I've seen going forward.

So ..looks like the hard drive issue is fixed...and I now have a working version of XTIDE in the basic unit...thus no need for the ISA sidecar in normal use..
These ISA slots could now be utilised for a sound card and/or a vga card.
What I do find frustrating is that I'm restricted to 640k base memory..and apparently it isn't possible to add more using the 2 simm slots.
Expanding the memory to include expanded and/extended memory was done by adding a 2mb card which plugs into a proprietary modem card...or a 4mb expansion card which again plugs into the modem card...
Non of which I have of course...
I can't even find pictures of these cards on the www.
I'm attaching a photo with the proprietary connector where the modem card is connected by ribbon...outlined in yellow.
The Blue box outlides the ISA backpack connector.
Both connect to the ISA bus in whole or in part.
As usual compaq made it this way in order to restrict upgrades to the parts they sold.
I can't help wondering if a conventional IBM type ISA with configurable switches could be used to add extra ram via the ISA slots ?
..not that I have one...and chances are it wouldn't physically fit anyway.
Perhaps there is a more modern version of such a card that would by using modern parts emulate an IBM ram card and be physically smaller...?
rgds
VS