A tip for those designing new hardware for vintage machines...

[syzygy]

I like figuring things out, but I don't like figuring things out that get in the way of figuring out the things I actually want to figure out. If that makes any sense.

I've initiated a return, I'm done with this. I'm getting ready to put these back in storage for a while and move over to another project. Being ADD, I tend to be happier just having a bunch of unfinished projects that I can switch between as my interests change. Although it results in a lot of unfinished things cluttering up the place.

That is not unreasonable. Everyone finds their own way. Personally, I was hoping that you would stay with it until it was solved, but again, it is not an unreasonable decision in my view.

Hey, I just finished a project that I started in 1992...it feels good...but sounds crazy to my friends.

 

[JonnyGators]

That is not unreasonable. Everyone finds their own way. Personally, I was hoping that you would stay with it until it was solved, but again, it is not an unreasonable decision in my view.

Well, the original problem that needs solving is the blue lava card, which is past the point of returning. This was just an attempt at a simple replacement.

Since the blue lava card is in 2 pieces, the logical starting point for troubleshooting to me was determine which half has the problem. I'm still awaiting the arrival of a new CF to IDE adapter to see if that makes a difference. I suspect it won't. I made 2 other purchases that had quick shipping options - an IDE DOM, which required an IDE male to male adapter to connect to the XTIDE card, and an SD to IDE adapter.

The size of the SD to IDE adapter card didn't allow them to sandwich together properly. So I purchased a short IDE extension cable. This experiment didn't go well, I had to shut it down once I smelled burning, and found one of the strands in the middle of the IDE cable turned black. Not sure what went wrong there, but then again, I was pairing things that in theory should work to me - SD to IDE adapter ought to work as a standard IDE device. XTIDE CF card with removed CF to IDE adapter ought to work like an IDE controller. If they're both designed to work to proper standards, they should work together. But, I'm finding a lot of these amateur projects don't necessarily work to the standards you'd expect. Of course, a step along the way resulting in charred wires now calls into question if further damage has been done to devices I already am trying to troubleshoot without any idea how. So there's that.

Meanwhile, I moved on to the IDE DOM and adapter. No burning there. But also no device present. Not sure if that's confirmation that the IDE controller is not seeing devices, an effect of trying to use the IDE controller in a way it wasn't designed for, or a result of something burning from the previous failed experiment.

Not sure where to go next there, other than wait for the replacement CF to IDE portion to arrive, and see what comes from that.

 

[Malc]

I wanted to buy a storage device for my computer, not a rubik's puzzle.

I would expect that one selling devices like this ought to have a responsibility to provide the required tools, or at the very least detailed documentation on the requirements and how to obtain and use the required tools to make the device work. Texelec has not done that at all, they just took my money and sent me a device that is completely useless to me as is, and have provided no response nor direction as to how to use it.

I tend to agree, But to be fair the Original VCF XT-IDE Controllers were designed to work in the IBM PC/XT computers and they do that very well, Then came the Lo-Tech controllers, I actually own Revision 1 / 2 and 4 of the Original VCF / Glitch Works controllers and a Lo-Tech IDE-CF controller, All work perfectly fine in my old IBM 5150 / 5155 / 5160 / 5162 and 5170 computers. Texelec are now the official suppliers of the Lo-Tech adapters and i'm sure they do their best but it's just not possible to guarantee that these controllers will work in all computers.

Over the years various 3rd parties have come and gone taken the original designs and put together their own Mix, Moving parts around and even leaving parts off etc, Got the PCB's fabbed as cheap as possible and use parts sourced from China and sell them on Ebah, Basically clones of clones of clones, And we all know what happened to the Asgard's.

FWIW i only recently found out that my GW r4 XT-IDE does not work in my Amstrad PC1640, I thought it had died but after testing it in my 5150 and 5160 it works perfectly fine, In one of my 5160's i have a r2 XT-IDE and swapped that with the GW r4 and put the r2 XT-IDE in my Amstrad PC1640 and it works, I'm not going to go kicking Glitch in the balls as i know his PCB's are quality and my legs ain't long enough

 

[Zare]

If you're going to design new hardware to use in 30+ year old machines, maybe build it to last a bit longer than 2 years. I know I'd be extremely embarrassed if I designed, built, and sold a product, only to have it die in 2 years while a 30+ year old machine it was made for continues to work flawlessly.

I'd also be embarrassed if I were a rich corporation and I've put leaky parts on my several thousand dollar hardware that'll make huge physical damage in 10 years time.

I've burnt through XTCF by connecting it to a wrong riser. Fixed by just replacing socketed ICs which costs peanuts. If any of the passive elements burnt out it would be visible and I'd replace them easily too.

You can't expect DIY people to have full Quality Assurance teams, x-ray the boards, stress test each one, whatever big companies can do. But you can expect it to be documented, servicable, multiple implementations around, and using most common components.

 

[westveld]

Well, the original problem that needs solving is the blue lava card, which is past the point of returning. This was just an attempt at a simple replacement.

Since the blue lava card is in 2 pieces, the logical starting point for troubleshooting to me was determine which half has the problem. I'm still awaiting the arrival of a new CF to IDE adapter to see if that makes a difference. I suspect it won't. I made 2 other purchases that had quick shipping options - an IDE DOM, which required an IDE male to male adapter to connect to the XTIDE card, and an SD to IDE adapter.

The size of the SD to IDE adapter card didn't allow them to sandwich together properly. So I purchased a short IDE extension cable. This experiment didn't go well, I had to shut it down once I smelled burning, and found one of the strands in the middle of the IDE cable turned black. Not sure what went wrong there, but then again, I was pairing things that in theory should work to me - SD to IDE adapter ought to work as a standard IDE device. XTIDE CF card with removed CF to IDE adapter ought to work like an IDE controller. If they're both designed to work to proper standards, they should work together. But, I'm finding a lot of these amateur projects don't necessarily work to the standards you'd expect. Of course, a step along the way resulting in charred wires now calls into question if further damage has been done to devices I already am trying to troubleshoot without any idea how. So there's that.

Meanwhile, I moved on to the IDE DOM and adapter. No burning there. But also no device present. Not sure if that's confirmation that the IDE controller is not seeing devices, an effect of trying to use the IDE controller in a way it wasn't designed for, or a result of something burning from the previous failed experiment.

Not sure where to go next there, other than wait for the replacement CF to IDE portion to arrive, and see what comes from that.

So the system was working with the Blue Lava card, then stopped?

And the Lo-Tech card also did not work?

Kinda sounds like the cards aren't the issue

 

[JonnyGators]

So the system was working with the Blue Lava card, then stopped?

And the Lo-Tech card also did not work?

Kinda sounds like the cards aren't the issue

That's an over simplification.

The Blue Lava card stopped seeing any device present when trying to boot, and when booting to a floppy the OS continued to not see an HD present.

With the Lo-Tech card, it did see a master device, but was unable to boot to it. When booting from floppy, the OS would see an HD, and allowed partitioning, formatting, and use of it as a C: drive.

Very different symptoms for the 2 cards.

The evidence suggests that something has failed on the blue lava card, and the Lo-Tech card is incapable of properly formatting a CF card into a bootable device. If there is a way to make a CF card bootable for the Lo-Tech card, no one has provided a working method, plain and simple. I've tried every suggestion made, they've all failed, so I'm returning it, I find it ridiculous that someone would sell a device that is at best difficult and at worst impossible to use, without documentation, support, or necessary tools.

 

[Eudimorphodon]

The evidence suggests that something has failed on the blue lava card, and the Lo-Tech card is incapable of properly formatting a CF card into a bootable device. If there is a way to make a CF card bootable for the Lo-Tech card, no one has provided a working method, plain and simple. I've tried every suggestion made, they've all failed, so I'm returning it, I find it ridiculous that someone would sell a device that is at best difficult and at worst impossible to use, without documentation, support, or necessary tools.

Did the blue lava card *ever* work all the way? Or are you saying it always never worked, just the exact symptoms changed?

Most people don't have this level of difficulty with these cards. My initial suspicion was CF compatibility issues, and I'm still not sure that's entirely ruled out, but at this point I *am* coming around to the idea this is actually some kind of BIOS compatibility issue with the particular clone you're trying to use these on. It is *not* normal that wipedisk didn't work as expected.

We need to be clear about something: there is almost *nothing* to these XT-CF adapters. All they do is decode a small range of I/O port addresses into a pair of "select" signals for the IDE/CF port and run those select lines, three address lines, an inverted version of the RESET signal, the I/O read-write lines, and the 8 data lines straight to the IDE or CF plug(*). This is true no matter who you buy the card from. There are quality difference between them (better quality cards will have cleaner trace layouts, more competently implemented noise suppression capacitors, better soldering quality, etc), but bluntly speaking when you're using one of these things it's almost the same as just cabling your CF card straight to the ISA bus. If weird things are happening chances are extremely high it's either your CF card or it's a software issue. (Either the XTIDE BIOS is misconfigured or it's somehow incompatible with your system.)

(* Some designs put a buffer on the data lines between the CF/IDE port and the ISA data lines. Generally speaking that's a better idea than not, but it makes no *software* difference.)

Maybe you need to find someone who will sell you a tested pre-paired adapter/drive combo, because that's the only way you can get any assurance that the whole widget is *supposed* to work, at least in a PC that's compatible with the XTIDE BIOS. Maybe you should get back on the horn with TexElec and ask them if you can pay them an extra $10 and trade what you bought for SD version, because in this case you *will* essentially be buying the drive from them; it's that adapter it comes with that does the job of "looking like" an IDE device and by buying it from them they should be standing behind it being compatible with the XTIDE BIOS. I mean, I guess if your luck is in fact bad enough I suppose you might manage to find an SD card that's incompatible with the adapter, but I use these IDE/SD adapters myself (as do a lot of other people), and I've never seen a card incompatibility problem. (Old slow cards perform like garbage compared to new ones, and I'd personally recommend using an industrial or "video rated" card for higher rated durability, but I've *never* seen a card not work at all.)

 

[JonnyGators]

Did the blue lava card *ever* work all the way? Or are you saying it always never worked, just the exact symptoms changed?

Sorry if I've not been clear. It worked perfectly, until recently, when it stopped detecting any device, as shown in the picture I posted from the computer booting up.

 

[Zare]

If there is a way to make a CF card bootable for the Lo-Tech card, no one has provided a working method, plain and simple. I've tried every suggestion made, they've all failed, so I'm returning it, I find it ridiculous that someone would sell a device that is at best difficult and at worst impossible to use, without documentation, support, or necessary tools.

The operation of these devices is simple, known, documented, and there are all the tools you need (xtidecfg).

XT-CF is two things - CF card hooked onto ISA bus, and XUB universal BIOS. You can even plug them in separately. The XUB ROM is a driver between system BIOS and the plugged in CF. System BIOS understands CHS only, while CF cards are internally LBA. XUB translates in between.
Using a different addresing scheme than what is actually on the media, and you won't have important bits for boot on the places you expect them.

First, upgrade XUB to R625. Zero out that CF you're using, boot from DOS 5 floppy and do a fdisk /mbr. Then try to format /s and boot into it.
Doesn't boot? Set CHS translation to 'NORMAL'. Still doesn't boot? Enable user CHS and enter the values corresponding to your card.
Still doesn't work? You have gremlins in the machine. (or aren't using a supported compatible CF card after all)

 

[modem7]

Blue Lava Systems XT-IDE Deluxe (XT-IDE CF+)

All of sudden, it stopped working. One day, you flipped the power switch on the computer, to discover that your XT-IDE no longer showed the CF card.

Post #52 contains a summary of what you have tried.

Noted is that you have made no comment about the possibility described in post #53, but I assume that is because you did not add a card.

Latest status is in post #102. "Not sure where to go next there, other than wait for the replacement CF to IDE portion to arrive, and see what comes from that."


Lo-tech XT-CF-lite rev.2

Pictured at [here].

"With the Lo-Tech card, it did see a master device, but was unable to boot to it. When booting from floppy, the OS would see an HD, and allowed partitioning, formatting, and use of it as a C: drive."

That sentence, in its entirety, indicates to me that the hardware is fully functional.

A 'different CHS-to-LBA translation' problem ruled out because partitioning and formatting done on the ComputerLand BC88.

The classic (not only) cause is the first item at [here]. Looking at earlier posts: FDISK /MBR is not possible because the OP does not have access to a 'boot floppy with FDISK' of a DOS version that supports the /MBR option. Alternately, a wipe of the first sector on the CF was attempted, but understandably, the OP expected the methods provided to wipe all of the CF (even the DISPART tool in Windows 10 does not wipe all sectors by default). Verification that first sector wiped not attempted.

Post #77 confirms that the partition was marked as 'active'.

All academic now because, quote, "I've initiated a return, I'm done with this."

To save further loss of hair, as Eudimorphodon wrote, purchase a {XT-IDE/XT-CF card with booting-to-DOS CF attached} combination.

 

[Eudimorphodon]

Sorry if I've not been clear. It worked perfectly, until recently, when it stopped detecting any device, as shown in the picture I posted from the computer booting up.

Pardon me for losing track. It's been many, many pages of bellyaching about how bad the vendors are at supporting their products... despite the fact that it doesn't seem like you've actually tried the number one thing they suggest ("fdisk /mbr") and also constantly discounting that your CF card at least could *possibly* be the problem when, again, download the schematics if you don't believe me, there's practically *nothing* to these things. The "full fat" XT-IDE devices that support "plain" 16 bit IDE devices at least have a few moving parts (the data latches) between the PC bus and the device they're supporting, whether an XT-CF works is basically *just* up to the CF card and whether it chooses to run properly in the oddball 8-bit ATA access mode that this implementation relies on.

 

[JonnyGators]

Pardon me for losing track. It's been many, many pages of bellyaching about how bad the vendors are at supporting their products... despite the fact that it doesn't seem like you've actually tried the number one thing they suggest ("fdisk /mbr") and also constantly discounting that your CF card at least could *possibly* be the problem when, again, download the schematics if you don't believe me, there's practically *nothing* to these things. The "full fat" XT-IDE devices that support "plain" 16 bit IDE devices at least have a few moving parts (the data latches) between the PC bus and the device they're supporting, whether an XT-CF works is basically *just* up to the CF card and whether it chooses to run properly in the oddball 8-bit ATA access mode that this implementation relies on.

I tried fdisk /mbr. My PC OS doesn't support that command. I do not have access to a DOS version for this PC on floppy disks that support this command. If the manufacturers would provide a simple tool for formatting or prepping CF cards to the required state for their product, that would make a whole lot more sense than the problematic guessing game and dead ends this experience turned into.

 

[JonnyGators]

Alternately, a wipe of the first sector on the CF was attempted, but understandably, the OP expected the methods provided to wipe all of the CF (even the DISPART tool in Windows 10 does not wipe all sectors by default). Verification that first sector wiped not attempted.

I do not know how to verify that the first sector was wiped with the wipe tool that was run, but it was definitely verified that I ran the command, the command returned that it completed the wipe, and further attempts to partition, format, install DOS on the CF card, had the same consistent results of being non-booting, but otherwise functional.

 

[cjs]

If the manufacturers would provide a simple tool for formatting or prepping CF cards to the required state for their product, that would make a whole lot more sense than the problematic guessing game and dead ends this experience turned into.

Sorry if I got this wrong, but isn't the issue you're having not about the required state for the CF card, but the required state of a block storage device for your OS and BIOS?

I have the feeling that if I plugged that device into a (suitably ancient) NetBSD or Linux box, it would work just fine.

 

[Svenska]

I tried fdisk /mbr. My PC OS doesn't support that command.

Then maybe, just maybe, you should be trying to figure out how to get access to that command to run it successfully. Once.

If you are unable to get MS-DOS 5 to boot once, you are probably unable to get a vendor-supplied utility to run once, either.

The doctor gave you medicine for your symptoms. You don't like it, so you don't take it. The symptoms remain. Now you complain about the doctor, the medicine and your symptoms. I don't think that's smart.

 

[Zare]

I tried fdisk /mbr. My PC OS doesn't support that command. I do not have access to a DOS version for this PC on floppy disks that support this command. If the manufacturers would provide a simple tool for formatting or prepping CF cards to the required state for their product, that would make a whole lot more sense than the problematic guessing game and dead ends this experience turned into.

You are interpreting your own problems and issues as a flaw of the product.
First of all, we're dealing with old computers. Right now you're using a modern mindset, buy, plug in, maybe run a tool, and it should work. Absolutely wrong.

Not how upgrading or reconfiguring old computers went in their age. XTCF if it miraculously appeared on the market in 1992 would be the easiest hard disk controller to setup.

1. List your hardware specs here. Not prosaic, not with rants, not by just a computer model. What BIOS do you have, what XUB is the card running, do you have HDD controllers in, and what floppy drive does the machine use, and what exact CF model/size are you trying to use.
2. You need to have an USB floppy that you can hook to modern machine and create floppies via images downloaded from the internet.
3. You need to have a CF reader for same purpose too.

Until you resolve these three points it's meaningless to come back.

The way you're interpreting the "fdisk /mbr" tip shows that you actually did not get the gist of the problem. Don't get overwhelmed by defetism. One step at a time, answer the questions above then we can proceed.

I do not know how to verify that the first sector was wiped with the wipe tool that was run, but it was definitely verified that I ran the command, the command returned that it completed the wipe, and further attempts to partition, format, install DOS on the CF card, had the same consistent results of being non-booting, but otherwise functional.

Get w32diskimager for Windows, and image that CF card to file. Open the file in the hex editor, if it's not zeroes in the beginning you didn't do it correctly.

A 'different CHS-to-LBA translation' problem ruled out because partitioning and formatting done on the ComputerLand BC88.

Did he verify it was a blank card prior? I did not pick up that.

 

[Chuck(G)]

If the C drive is recognized when you boot from a floppy, but doesn't itself boot, that can be caused by the lack of a "55 AA" at the end of the boot sector. Can you dump both the MBR and the boot sector of the CF disk?

At least that's where I would start.

 

[JonnyGators]

The ability to create new floppies from images of OS's that I do not regularly use is not something I easily have at this time, my XP machine with a 5.25 floppy drive is in pieces due to a mobo replacement being needed, and the replacement mobo has a broken clip for the heat sync, so I need to come up with a solution to get a heat sync on the processor properly, basically yet another project on the pile.

It seems that the direction this is going is that it's a simple case of a CF card not formatted properly to be a boot device in a PC, and no one having a method of doing this from a modern Windows PC with a USB device, or with a PC that's on a version of dos that doesn't support the fdisk /mbr command, and me not having the time to jump through the hoops to resolve this for this device.

My point is rather simple - I would never sell a product in this way. Selling a product that has very specific requirements that can easily leave a customer in a situation where they can't make it work with the tools available to them, requiring hours of trial and error and chasing things down and still not have it working is just not what I would consider an acceptable way to do business. It's a guaranteed way to end up with customers that have a negative experience, return the product for a refund, and move along to never do business with you again. If I were going to sell this product, I would get all the answers needed to make this work, and include that on the sales listing or site for the product, perhaps include a download .pdf to go along with the product, and if it were possible to make a little app that can do a few basic things like verify compatibility of a CF, wipe it to the state it needs to be in to work, things like that, include that with the product. A good starting point to have what you need to make it work. Not send a product with no documentation, no support, and basically "well....it might work with your CF card....and you might need to jump through some hoops to maybe make it work, but maybe not....good luck". This is not the experience I signed on for. I expected a storage solution that would work.

Now then - I suppose the most helpful comments in this thread were the ones pointing to the suggestion of buying an SD card solution, or buying one that comes with a CF card already paired with it, as those are working out of the box solutions. May have to play with the idea of an SD card solution at a future date, but I'm going to swap some parts around and put these back in storage to play around with some other machines. My only objection to the idea of buying one with a CF card already paired is, although that gives a good starting point, being locked in to a single card isn't exactly ideal, ultimately I would want the option of fully working functionality, which would include the ability to setup and pair a new CF card to the device if the need arises.

Anyways, that's my final say on the matter of the texelec device. Working or not, it did not meet my requirements for a successful device, and is going back, I am done with texelec, they have failed to sell me something that works or is supported to my requirements for success, I'm moving along.

Going on the to do pile - trying to make the blue lava device work again. I have a CF card that was working as a boot device with it, and is showing that it still has it's data, so in theory if the blue lava device were made to detect a device again, it should work. Then I could move on to figuring out if I could setup a new CF card to work with it. Perhaps by then I'll have a working XP machine with 5.25 drive to make a newer DOS disk to add to my tool kit. A replacement CF to IDE adapter is on it's way here on a slow boat from China, which will likely confirm that isn't the issue, leaving me to figure out how to start troubleshooting the parts on the board. Perhaps I might start with some continuity testing with my multimeter to see if there's a failed solder connection. Not sure how to test the chips on the board, but these are the directions this thread could go in next, and move away from this CF formatting noise.

 

[Eudimorphodon]

My point is rather simple - I would never sell a product in this way. Selling a product that has very specific requirements that can easily leave a customer in a situation where they can't make it work with the tools available to them, requiring hours of trial and error and chasing things down and still not have it working is just not what I would consider an acceptable way to do business. It's a guaranteed way to end up with customers that have a negative experience, return the product for a refund, and move along to never do business with you again. If I were going to sell this product, I would get all the answers needed to make this work, and include that on the sales listing or site for the product, perhaps include a download .pdf to go along with the product, and if it were possible to make a little app that can do a few basic things like verify compatibility of a CF, wipe it to the state it needs to be in to work, things like that, include that with the product.

It’s interesting that you prescribe all these things while at the same time demonstrating why nothing anyone could do would ever be good enough for some customers.

Regarding the possibility of an “app” someone could use to “verify the compatibility” of a CF card, sorry, that’s not a thing you can do with the CF card mounted in a card reader in a newer computer. XT-CF relies on an oddball transfer mode that not all CF cards reliably support and, critically, is not something you could test in a card reader. (You miiiight be able to simulate it using the IDE port in an old AT-bus computer with low-level programming of the ATA registers. But definitely not on a USB dingus.) The only place you could definitively run this “app” is on the computer with the XT-CF adapter slotted into it… and then you discover some of your customers refuse to go through the effort to make floppy disks. So then I guess you’re expecting everyone selling these to package them with this software on floppy disk? How much more are you going to pay for that?

There are a million ways to zero a CF card in a reader; several have been suggested to you in this thread, as have methods of seeing what’s going on via hex/sector editors, and you’ve consistently refused to do them in favor of bellyaching about how something intended for geeky hobbyists isn’t guaranteed plug-and-play. You could also solve your floppy problem super easily by just buying a Gotek floppy emulator with FlashFloppy on it; someone on eBay will sell you a pre-hacked one if you can’t be bothered to go through doing that work yourself, and once you have it getting DOS 5’s FDISK onto your XT is literally 30 seconds work downloading a boot disk image onto a USB disk. But, no, instead of having any interest in doing the hacking it seems like you’d rather complain.

I dunno, but it kind of seems like maybe some of the expectations here might not be completely aligned with reality.

 

[Eudimorphodon]

… I guess a thing I would add is if you’re technically savvy enough you *can* completely generate a working raw disk image using tools on a modern computer that you can dump onto a memory card and boot up an XT-CF-equipped XT with having never touched a physical floppy (or emulator) on the XT, but doing it requires specific knowledge of variables like the size of the target card and how its geometry is going to be translated by the XTIDE BIOS. I did just this myself on a Macintosh using QEMU with a config file specifically set up to present a disk image with a BIOS geometry the same as what an 8+ gigabyte SD card in an IDE->SD adapter looks like to my Tandy 1000. Why did I do this?

• I found a bootable CD image of IBM PC-DOS 7 and it seemed like an interesting way to translate that into a working XT install, and,
• Doing stupid hacker things just to prove you can is most of the fun of this hobby.

In theory *maaybe* this technique could be translated into a magic installer that, given the size of a CF card, could make an educated guess how it’s going to look to the XTIDE BIOS, create a QEMU disk image with apropos geometry, launch a VM to do an automated DOS install from a source image, export the QEMU image to a RAW format, and DD it onto your CF card. But are you going to pay for that? And what happens when that all works but your CF card is still electrically incompatible/doesn’t actually support 8-bit mode?

Hard cases make bad law.