• Please review our updated Terms and Rules here

Forthcoming XT-IDE Board - Cast Your Vote

Forthcoming XT-IDE Board - Cast Your Vote

  • As original XT-IDE, with a 40-pin header only

    Votes: 4 10.5%
  • With a 44-pin header and board space to mount a 2.5" IDE HDD (i.e. a hard-card)

    Votes: 7 18.4%
  • With an optional Compact Flash socket (as master or slave) and a 40-pin header

    Votes: 26 68.4%
  • With a Compact Flash socket only

    Votes: 1 2.6%

  • Total voters
    38
  • Poll closed .
Update - I've been chipping away at this, and so has Ian and his team on Dangerous Prototypes (DP). The DP XT/IDEv2 board, which is a 100% CPLD implementation, has gone to the board house. It's main advantage is cost - the board is only a couple of dollars, and the BoM is probably only another $10-15. It is mostly SMD based but this isn't as much of a problem as it sounds, with zero experience I soldered only up no problem.

Anyway, for compact-flash, as mentioned previously it is difficult to produce this in a board small enough to be cost effective. However I've got a first-cut of a board with both CF socket and 40-pin IDE header(see attachments). This is a direct decendent of the DP board, itself depending on the amazing work of the original of course. Changes:

- added the CF socket
- removed IRQ header
- replaced through-hole 28C64 with SMD part
- extended DIP switch to 12-way and integrated the remaining jumper header functions on to it
- changed DMARQ PD to 5.6k
- changed IORDY to 1k PU

However in order to route this board the 3.3V lines are 6 mils only, ground VCC + GND are 12 mils. I'm not sure that's really enough.

The idea is that the CPLD code will be compatible with the DP versions (standard or 'Chuck-mod'), and BIOS wise it will be compatible with either the XT/IDE Universal BIOS or Alan's future memory-mapped IO version for enhanced read & write performance.

Cost wise it should be similar to the DP version. CF header isn't cheap but SMD 28C64 offsets some of that.

Any comments very gratefully received... this is my first attempt at anything in Eagle so it's very probably littered with issues!

Eagle files are here.
xt-cf-ide-v0.1.jpg
 
Last edited:
You know, you could do a CF-only version of the XTIDE that uses the 8-bit mode of a CF card and probably have to provide little more than PROM address decoding for the BIOS. With some clever memory address decoding, you could even have a card that beats the current XTIDE on throughput. In any case, the result would be an inexpensive card not requiring a CF adapter.
 
I agree with Chuck on this. A lot of us really only want the ability to use a DF card simply because we can add a CF card to the rear of our current system using an expansion slot cover plate adapter. That way, we can remove the card while the computer is powered down, take it to a modern computer to either reload it with a new OS, or add more programs etc, then simply plug it back in all without opening a case.

I would love to do something with this, even on an older version of XTIDE. I was even thinking about installing one version or another in the expansion unit I'm building, but that is a discussion for another time. Peronally though, I would love a CF only version that is already set up to mount in an open expansion bay on the back of the computer.

You know, you could do a CF-only version of the XTIDE that uses the 8-bit mode of a CF card and probably have to provide little more than PROM address decoding for the BIOS. With some clever memory address decoding, you could even have a card that beats the current XTIDE on throughput. In any case, the result would be an inexpensive card not requiring a CF adapter.
 
The thing is, we'd need an available ISA bracket with a CF slot in it.

But, you can achieve that already with any of the XT/IDE varients and a seperate CF socket-on-an-expansion-bracket as available from your chosen auction website :) Actually I have this setup in a 5155 (allbeit with an SD-card) and it works very nicely.

In any case, the spec of the card two posts up is derived from the poll result at the top of the thread. If we're now saying simply CF-Card is the prefered option, that makes board routing very much easier!
 
Last edited:
Oh, I understand that. Sometimes I get too many ideas.

For example, I was thinking that, given the speed of an XT, a CF card is an expensive waste. A small MCU with a couple of SDHC slots would provide storage and perhaps ADC and async comms all in one--and at a lower cost for storage with performance comparable to a hard disk.

But again, I have too many ideas... :sigh:
 
I know, I've seen several of those adapters. I like the ideea of using an SD card too. In fact that would be even better. You know what I want right now? Any version of the XTIDE that I can install and start using to both use as additioanl storage for the 5000 plus games, and the ability to boot from several different partitions on the same SD card. I have to set up a couple of my DOS boot disks right now to add functionality for parallel and serial ports, plus I want to boot different versions of DOS from version 1.0 all the way through DOS 2000 with thise configurations added in. If I could get that functionality, I wouldn't even worry about the new version of the XTIDE for now.

Would the old XTIDE do that for me? And, are the instructions in VCF good enough to get me up and running? If so, anybody got one for sale? How much? I could almost guarantee purchase within 2 weeks the way things are right now. I am out of work and I need to pay someone for some work I had done. So if I know how much I need, I can get it budgetted in by the end of the month.



The thing is, we'd need an available ISA bracket with a CF slot in it.

But, you can achieve that already with any of the XT/IDE varients and a seperate CF socket-on-an-expansion-bracket as available from your chosen auction website :) Actually I have this setup in a 5155 (allbeit with an SD-card) and it works very nicely.

In any case, the spec of the card two posts up is derived from the poll result at the top of the thread. If we're now saying simply CF-Card is the prefered option, that makes board routing very much easier!
 
You and I both Chuck. You and I both. I guess it comes with age and intelligence. Keep coming up with ideas.
So far I have learned a lot from your thoughts on things like these.


Oh, I understand that. Sometimes I get too many ideas.

For example, I was thinking that, given the speed of an XT, a CF card is an expensive waste. A small MCU with a couple of SDHC slots would provide storage and perhaps ADC and async comms all in one--and at a lower cost for storage with performance comparable to a hard disk.

But again, I have too many ideas... :sigh:
 
Ground path - ground path - ground path.

As the selected supply voltage drops (3.3 in this case), a clean wide low impedance ground return becomes more and more important to ensure good coupling between components and lower transmission loss on the return. Traces are resistors. Small long traces are pretty significant resistors. As current flows through resistance, you lose voltage. And transmission losses are compounded each way - supply and return.

I still contend trying to drive legacy IDE hard drives (as people will do with this board) with 3.3V I/O CMOS output drivers over long ribbon cables will result in signal integrity errors in a lot of cases. Especially since IDE is TTL and no one really cared about impedance control on drives until the speeds went up to UDMA levels. Any significant leakage current in the drive and it could drop the effective supply voltage enough to add up with transmission losses and transients to effect digital signal levels.

I suspect DoMs, CF cards and modern drives with very short cable runs will work. Anything else will likely have very marginal signal integrity. No one at DP is prepared for a support thread twice as long as the current XT-IDE one. Until someone actually builds the DP board and proves it works reliably, I would hold off on basing anything else off of it.

You also have a few critical errors in the ERC. For example pins 77 and 99 aren't actually connected.
 
I still contend trying to drive legacy IDE hard drives (as people will do with this board) with 3.3V I/O CMOS output drivers over long ribbon cables will result in signal integrity errors in a lot of cases. Especially since IDE is TTL and no one really cared about impedance control on drives until the speeds went up to UDMA levels. Any significant leakage current in the drive and it could drop the effective supply voltage enough to add up with transmission losses and transients to effect digital signal levels.

Good point. Put a crufty old CDC 5¼" ATA drive at the end of a meter-long cable and you're likely to be surprised. It always surprised me that the ATA standard didn't call for parity checking over the signal cable. Since there's no way, other than catching the occasional command or status corruption, to tell if what you got was what was sent, it's always seemed to me to be a very optimistic way to do things.

Even SD cards CRC-check their conversation with the host.
 
Many thanks indeed for all the excellent feedback on this.

I wondered if the CF slot would fit sidewards, and it does! We'd need to mill the slot in the keystone bracket, but I should be able to arrange that.

I'm not sure how those pins came disconnected, now sorted (thanks for the pointer on the Design Rule Check function btw) and I've added ground polygons to both sides and beefed up the supply and gnd rails significantly. Also the CF socket was lacking any power in the first cut!

Hence revised board, CF only for now, bracket facing socket, files here.
xt-cf.jpg
 
Now that is truly awesome. So compact! With the CF card holder integrated onto the board, you will not need an extra slot or bay. In many 8-bit systems, those are precious. Even better, this design allows easy access to the card.
 
Many thanks, that is my thinking (and reasoning) too.

After many, many hours messing about with component placement and some manual routing tweaks, ZIP attached is another attempt at a combined CF & 40-pin header card. This time there are thicker power supply lines, ground polygons on both sides (with a some manually placed vias to provide as much coverage as possible), and of course the CF socket is mouned so the card will stick out the back of the machine.

As before, any and all feedback (however harsh!) greatfully received. Signal lines remain 6 mils, it won't route otherwise.

Design files here.
xt-cf-ide.jpg
 
Sorry to be a PITA, but whenever I try to download an save the attached zip file, I get the message"This file cannot be downloaded". There is no other work around, or error number. I would like to see what the design files have in them. Just saying...

Many thanks, that is my thinking (and reasoning) too.

After many, many hours messing about with component placement and some manual routing tweaks, ZIP attached is another attempt at a combined CF & 40-pin header card. This time there are thicker power supply lines, ground polygons on both sides (with a some manually placed vias to provide as much coverage as possible), and of course the CF socket is mouned so the card will stick out the back of the machine.

As before, any and all feedback (however harsh!) greatfully received. Signal lines remain 6 mils, it won't route otherwise.

Design files here.
View attachment 7616
 
Sorry to be a PITA, but whenever I try to download an save the attached zip file, I get the message"This file cannot be downloaded". There is no other work around, or error number. I would like to see what the design files have in them. Just saying...

looks to me like there was a URL mix-up.... this one seems to work (FYI pearce_jj). :)
 
That did the trick... Thanks el_VR

All of you who are currently working on the new revision of the XTIDE are just plain amazing. I know I'm a geek (hell, most people who know me will tell you I'm the biggest nerd they know) and am very proud of that, but you guys are GODs. Geeks of the Day!
I only wish I could have stayed on track with my education to learn half the things you're doing. My hat's off to all of you.

Now, as far as the zip file is concerned, I was able to download it and open it with no problem. The new link works fine. I can see a few things I understand, and I'm not going to bother you with educating me. Suffice it to say, that I can't wait for the new revision so we can all enjoy it as much as you all do in making it happen.




looks to me like there was a URL mix-up.... this one seems to work (FYI pearce_jj). :)
 
Sorry about the URLs, I just typed in the wrong domain name but now the edit button isn't available it seems.

@Betablue - I can see a few things I understand, and I'm not going to bother you with educating me - download Eagle PCB Designer - the free version is limited to two-layer boards of a certain size, but that is enough for this project :) I'm glad you're interested in this, yes the underlying work from the original board plus all the subsequent tweaks and of course the BIOS coding is quite something indeed. Of course the bit I've been working on is the easy bit, just jiggling components around and drawing on the tracks!

Updated design pack here, for anyone interested. I think that's pretty much done (CF only) - unless anyone has any comments otherwise. Next step is to try and add either a 40-pin IDE header or a CH375B USB interface.
 
Many thanks indeed for all the excellent feedback on this.

I wondered if the CF slot would fit sidewards, and it does! We'd need to mill the slot in the keystone bracket, but I should be able to arrange that.

Hi,

Not sure why you say you have to mill the brackets? You can buy standard brackets w/ PC Card holes in them which should work nicely with CF. Here is one. Price falls rapidly with a large order (i.e. at 50 they are ~ $6.50 each). Milling them by hand will cost at least that much and won't look as professional. I am sure there are cheaper producers as well... Or am I missing something?
 
Back
Top