XTIDE tech support thread

[Zombie]

I am trying to restore my Tandy 1000 TL in a similar method to the process shown here but the TL is a bit more flexible a machine than the HX. Currently my TL HAS a hard drive in it, but it is an ancient 21 MB MFM RLL hard drive that is failing. I have two such drives. I want to replace that hard drive with a 512 CF Card, but I need an XT IDE hard disk card made for the TL. Would anyone be willing to assist me?

 

[NobodyIsHere]

I need an XT IDE hard disk card made for the TL

Hi! What does that mean?

Thanks and have a nice day!

Andrew Lynch

 

[Maverick1978]

Yup, that sounds familiar. DOS can hang if the MBR is just trashed enough. Clobbering it is often the only way out.

It's also a good idea to keep several copies of your MBR on Cylinder 0, Head 0--unless you're using a DDO, the rest of the track is unused and a good place to stash things.

Chuck (or anyone) - how would one go about keeping a copy of the MBR on other cylinders and heads? And how would one go about retreiving them? This is the first I've heard of that trick... sounds useful

 

[Chuck(G)]

Use your old friend DEBUG to create a program to do it. Here's a little way to create a program that copies the MBR from its usual place at sector 1, head 0, cylinder 0 to the next slot at sector 2, head 0, cylinder 0:

C:\>debug savembr.com
File not found

-a100
13B7:0100 mov ax,0201
13B7:0103 mov bx,0200
13B7:0106 mov cx,0001
13B7:0109 mov dx,0080
13B7:010C int 13
13B7:010E mov ax,0301
13B7:0111 mov bx,0200
13B7:0114 mov cx,0002
13B7:0117 mov dx,0080
13B7:011A int 13
13B7:011C int 20
13B7:011E

-r cx
CX 0000
:1e
-w
Writing 0001E bytes
-q

You'll have your "savembr.com". Only 30 bytes, and probably oversized by half.

 

[Maverick1978]

Sounds great. I've typed in debug scripts to make short com files in the past (usually from magazine articles), but had never heard of copying the mbr to other sectors and whatnot (though I'd seen programs that would back it to a file on floppy and allow you to restore it from floppy)

How would one go about moving the backup from 2,0,0 to 1,0,0? - My understanding of assembly and machine programming is miniscule, at best. When I used to program, it was always in a high-level language - and it's been many years since I've done anything more major than text-based quick-sort routine in Basic or C (something to make my life easier at work with their large, comma-delimited files full of mostly information that isn't needed)

 

[RWallmow]

Ok, so I built and "half-tested" my XT-IDE a few months ago (plugged in without drives and BIOS ROM came up) so I figured I was golden, now come time I have pulled my PS/2 Model 25 out of storage to put this XT IDE to use, and it isn't recognizing ANY drive plugged into it (3 CF flash cards, 1 CF Microdrive, 20gb Seagate, 80gb WD, etc...).

I am wondering if theres something specific I should be looking for that I may have done wrong in assembling it, I know im not completely hosed since its booting the ROM, so my soldering skills are at least half-witted ;-)

I HAVE tried it in other machines thinking it might have been something Model 25 specific and it behaves the same way in my Compaq Portable II and Portable III (with ISA expander) as well as a generic ISA pentium box.

Thanks all,

 

[pearce_jj]

Check the ribbon cable is connected the correct way around on the card (red stripe at pin 1). Check with a magnifing glass the 'quality' of the soldering - dry joints should be visible like this. Mine only seem to work with 80-wire cables.

 

[hargle]

Also verify that the IO dipswitches are still set to 300h. If they are moved, the BIOS must be reflashed in order to let the software find the card. If there is a BIOS<->IO dipswitch mismatch, you won't find anything, ever.

 

[RWallmow]

Check the ribbon cable is connected the correct way around on the card (red stripe at pin 1). Check with a magnifing glass the 'quality' of the soldering - dry joints should be visible like this. Mine only seem to work with 80-wire cables.

Ribbon is the correct orientation, and I did pull all the chips and revise any of the questionable solder joints, but I have only been trying 40-wire cables, OR CF adapter direct connected to card, I will try a 80-wire cable tonight.

Also verify that the IO dipswitches are still set to 300h. If they are moved, the BIOS must be reflashed in order to let the software find the card. If there is a BIOS<->IO dipswitch mismatch, you won't find anything, ever.

Switches appear right, but maybe one isn't fully toggled, I will triple check them tonight.

 

[Chuck(G)]

Ok, so I built and "half-tested" my XT-IDE a few months ago (plugged in without drives and BIOS ROM came up) so I figured I was golden, now come time I have pulled my PS/2 Model 25 out of storage to put this XT IDE to use, and it isn't recognizing ANY drive plugged into it (3 CF flash cards, 1 CF Microdrive, 20gb Seagate, 80gb WD, etc...).

I've mentioned this before--go carefully over your soldered connections again. For some reason, this PCB seems to be more prone to cold solder joints than others--and I've been soldering PCBs for about as long as there have been PCBs--and it caught me. I've recommended to Andrew that he go with a slightly larger pad size on his next iteration.

 

[hargle]

OK, my turn for some XTIDE tech support.

I've just installed one of my cards into a Honeywell turbo XT. Landmark speed test tells me I'm running a 7.672MHz.
When the machine is in turbo, I get corrupt data when *writing* to the HDD. In non-turbo it writes fine. Reads work perfectly at either speed.

I have tried multiple cables, HDDs and even 2 XTIDE cards. No change.

The biggest caveat to this problem is that this computer is totally obscure, completely undocumented, and has lots of patch wires on the motherboard, so it may not be 100% compatible or even debugged from the factory.

This is the only turbo machine I have in my collection. I am pretty sure others have used this card in faster machines before, so I am currently blaming some incompatibility in the motherboard itself. Can any other turbo users sound off here just so I can assure myself that we don't have a design flaw in the card itself?

Any ideas where I should go next? I really like this machine, but not being able to use it in turbo mode rather wrecks the need to keep it.

 

[Agent Orange]

OK, my turn for some XTIDE tech support.

I've just installed one of my cards into a Honeywell turbo XT. Landmark speed test tells me I'm running a 7.672MHz.
When the machine is in turbo, I get corrupt data when *writing* to the HDD. In non-turbo it writes fine. Reads work perfectly at either speed.

I have tried multiple cables, HDDs and even 2 XTIDE cards. No change.

The biggest caveat to this problem is that this computer is totally obscure, completely undocumented, and has lots of patch wires on the motherboard, so it may not be 100% compatible or even debugged from the factory.

This is the only turbo machine I have in my collection. I am pretty sure others have used this card in faster machines before, so I am currently blaming some incompatibility in the motherboard itself. Can any other turbo users sound off here just so I can assure myself that we don't have a design flaw in the card itself?

Any ideas where I should go next? I really like this machine, but not being able to use it in turbo mode rather wrecks the need to keep it.

Well, I haven't checked the bus speed on the 1000SX but it does coexist with the NEC V20 running at 10 MHz w/the XT-IDE installed. Try that HD in your SX and see what happens.

 

[Chuck(G)]

Used it both with 8088 and V20 at 8 and 10 MHz respectively. No issues.

I doubt that your motherboard was made by Honeywell. Is there any board in th99 that matches it?

 

[modem7]

I have an XT-IDE in my 8088 based XT clone. The high speed is 8 Mhz (about) and the XT-IDE works flawlessly.

 

[modem7]

Any ideas where I should go next? I really like this machine, but not being able to use it in turbo mode rather wrecks the need to keep it.

I've identified a compatibility issue with the BIOS found in early model IBM 5170s (further investigatation on my to-do-list).

It turns out that the second revision 5170 BIOS has issues too, but not as bad. Whereas the first revision BIOS won't XT-IDE boot at all, fitting the second revision 5170 BIOS results in the XT-IDE sometimes not booting. When it does boot, CRC checks show the ocassional corrupted reads of files.

I haven't seen any problems when the third revision 5170 BIOS is then fitted to the motherboard.

So it might be worth you trying a different (turbo) XT BIOS in case your issue is BIOS related.

 

[hargle]

Used it both with 8088 and V20 at 8 and 10 MHz respectively. No issues.

I doubt that your motherboard was made by Honeywell. Is there any board in th99 that matches it?

I agree. Probably made by NCR.
I posted about this machine earlier, with a link to the th99 closest match I could find:
http://www.vintage-computer.com/vcforum/showthread.php?21167

Since this is now the 2nd device that doesn't seem to work quite right (the above linked thread talks about my failure to use a VGA card in it that works in all my other 8088 machines) I am truly suspecting the ISA bus is just wonky.
edit: interesting. after re-reading my own post about this machine, I mention that the XTIDE works just fine. I may have been only doing reads then (just booting).

 

[Minerva10210]

Hi Guys

I would appreciate some input on the error "timeout error while polling EEPROM" when trying to flash the "universal" BIOS v1.1.5 onto my card.

What I have done so far:
Tried the flashing process in 3 different machines (IBM 5155, HP P1x200, Clone PIIIx450)
Double checked all the solder work on the card.
Checked the DIPS and Jumpers with a DMM.
Changed ALL the IC's and EEPROM chips.

Still no joy.

One thing I noticed right now, is that when the card is in my 5155, and I have JP1 capped, the machine hangs, but it boots fine with JP1 uncapped. This means that the machine is in fact trying to read the BIOS off the card, but due to there not being anything on it, it just hangs??

Your thoughts please?

 

[hargle]

what EEPROM brand/type are you using?
The hang may be from corrupt data in the eeprom, although the mainboard BIOS is supposed to checkum the data before trying to execute it. (did the 5155 have that feature, or will it blindly jump into a corrupt option ROM?)

Try this:
boot with JP1 off to a floppy disk with debug.com available.
re-attach JP1 after booting
load debug
at the - prompt, type "d d000:0"
This is assuming your card is dipswitched to the default values of D000 for memory.

The dumped out data from debug *should* have some data in it. The first 2 bytes should be 55 AA. If you see FF FF's then either your eeprom is blank, or your eeprom is not decoding at D000. If you see corrupt data, like the 55 AA is actually 51 A8 or similar, then you have either a corrupt eerpom or some address lines are floating/stuck high or low.
If you dump out the same data more than once, and it changes, then you have a bad solder joint and a bit (or two) are floating.

 

[Minerva10210]

Hi Hargle, thanks for the reply!

OK, in order then

EEPROM Brand, from what I can gather from the markings on the chip is as follows:
AMTEL
AT28C64B
15PU
0929

I only have access to one machine currently, will double check the debug stuff at home on other machines as well, but currently I get the following:

-d d000:0
D000:0000 55 AA 10 E9 06 0A 00 00-31 31 2F 32 38 2F 31 30   U.......11/28/10
D000:0010 58 54 49 44 45 31 31 30-2D 3D 58 54 49 44 45 20   XTIDE110-=XTIDE 
D000:0020 55 6E 69 76 65 72 73 61-6C 20 42 49 4F 53 20 28   Universal BIOS (
D000:0030 58 54 29 3D 2D 00 00 76-31 2E 31 2E 35 20 28 31   XT)=-..v1.1.5 (1
D000:0040 FF FF FF FF FF FF FF FF-FF FF FF FF FF FF FF FF   ................
D000:0050 FF FF FF FF FF FF FF FF-FF FF FF FF FF FF FF FF   ................
D000:0060 FF FF FF FF FF FF FF FF-FF FF FF FF FF FF FF FF   ................
D000:0070 FF FF FF FF FF FF FF FF-FF FF FF FF FF FF FF FF   ................
-

So looking at the BIOS dump, it would seem like it started the flash process, but only got a few bytes written and then gave up?

Was I doing something wrong?

Cheers!

 

[modem7]

... although the mainboard BIOS is supposed to checkum the data before trying to execute it. (did the 5155 have that feature, or will it blindly jump into a corrupt option ROM?)

The 5155 contains a 5160 motherboard. All three revisions of the 5160 BIOS/POST do the checksum check.
It could be by coincidence that the 5160 found the ROM contents to checksum to 00.