• Please review our updated Terms and Rules here
  • From now on we will require that a prefix is set for any items in the sales area. We have created regions and locations for this. We also require that you select a delivery option before posting your listing. This will hopefully help us streamline the things that get listed for sales here and help local people better advertise their items, especially for local only sales. New sales rules are also coming, so stay tuned.

WTB XT IDE V1 assembled or PCB

Plasma

Veteran Member
Joined
Nov 7, 2005
Messages
2,064
Will buy or trade TMC-850 with BIOS if interested.
 
Just the PCB or assembled? I am trying to avoid surface-mount soldering. I didn't know if V2 was through-hole design or not? If it's assembled I guess it doesn't matter.
 
V2 is currently undergoing testing and is through-hole. It also includes high-speed UART support. I think it might be prudent to hang on a bit until they're released. There are some nice features on V2.
 
Hi,
This really hasn't been discussed much here yet AFAIK. Jeff, myself, and a few others have been discussing a way ahead on XT-IDE V2 privately. It is not clear to me what the community is really interested in at the moment so I guess the plan is to get a small batch of the XT-IDE V2 PCBs and pitch them here.

If enough builders are interested to warrant continued support then we'll do that. However, I really don't want to inadvertently get in the way of the CPLD version of XT-IDE either. Frankly I am bit confused as to what is going on so I think the plan is to just get some PCBs and try it out. In other words, throw something against the wall and see what, if anything, sticks.

Peaceful and respectful discussion welcome as always. Thanks and have a nice day!

Andrew Lynch
 
Hi Andrew

Re DP's take on it, I have six "v1b" boards (3.3V CPLD but with 74xxx series ICs too) which should work (these are again SMT), I just need to build & test them hopefully over the Chistmas break, and offer them here at cost. However that's the end of that 'branch' I think, as there is a full CPLD "v2" board currently awaiting production for testing I believe which will dispense with the 74xxx chips, but keeps the 28C64 ROM (for now). That, like the v1b, will be just a port of the original though with some small fixes, seperate resistors for LEDs and CSEL grounded.

The greatly extended functionality in Alan's version I think is in test, hopefully he will post an update too!

BIOS wise I did submit some bugs for the updated universal BIOS a while back.

Re community interest, my own simple poll did show by far the greatest number of votes for CF card and 40-pin header, but there are problems making it since the card slot uses so much board area and the pins are all muddled up compared to the 40-pin header. However with a single CPLD implementation, the use of a card with a standoff, and an SMT flash chip in place of the socketed 28C64, maybe it might be possible. Sadly I just don't have the skills to layout such a board.

Whatever you produce I don't think could 'get in the way' - there is defintiely a slow but steady demand for these things, and you have the clear advantage of being 'original and best' :)

Hope that helps!
 
I haven't received any feedback from Jeff or Mike on my version. Tomi's board was DOA so I sent him a fully assembled and tested board early this week but it still has a few thousand miles to get to him. ATM I don't even have a working ISA card to test with. I'm ordering more PCBs this weekend. But there is a 3 week lead on delivery.
 
ok, here's a summary of all the things XTIDE related going on:

1) the original XTIDE boards, popular about 2 years ago, are sold out and we've stopped producing them.
2) In the past 6 months or so, THREE separate branches of XTIDE controllers have appeared: XTIDE rev2, Dangerous Prototype's CPLD version, and EEGuru's CPLD version
2a) XTIDE rev2 adds better performance than rev1, adds a larger EEPROM for more BIOS goodies, adds a high speed UART, adds some onboard power plugs. This version is still a through hole design, and is fairly easy to build at home. Fully populated, the cards will be expensive ($60+). Full functionality of XTIDE rev2 was verified just last week. Since this version is based off XTIDE rev1, we already have a fully functional BIOS available. Support for booting over UART is coming along shortly. We assumed that this card was going to be dropped with the rise of the CPLD designs, but now that we've verified that the card is fully functional, all it would take is a PCB order and we could start shipping anytime.
2b) As explained by pearce_jj above, their card is coming along after some false starts. They are UK based, and the parts are SMT so not exactly home build friendly, but I'm sure it's possible. This is probably the cheapest solution, even considering the additional cost of shipping overseas. I don't know that for certain, but am guessing that SMT parts are cheaper because they are common and in production. Not sure what level of BIOS support is available here.
2c) eeguru's boards are CPLD but use through hole connectors. I believe this card will have even better performance than XTIDE rev2. It does not include a UART. I've built up a card and they are SUPER easy to solder at home. Prices are probably going to less than XTIDE rev2, but more than DP's version. We're a little short on software at the moment, but that should come online very soon, I would think that by the time cards are available, the full software suite will be available too. I like this card very much, plus they're purple. ;)

So basically, the market will probably have several options available early next year.

Apologies to eeguru here publicly that I've been unavailable for testing/development as of late, being distracted by the recent developments in XTIDE rev2 and holidays and real life stuff.

I think having another poll to judge demand for features may help us decide how many new PCBs to order. Since not everyone would need the ability to boot over serial port, there may not be a big demand for XTIDE rev2 for example.
 
DP's PLDs are cheaper because they are a more common package, more common vendor, and are 3.3V chips. I have doubts whether a 3.3V unbuffered drive will be sufficient for every hard drive people will throw at the board. However pearce_jj has done extensive testing with v1 boards with positive results (data lines buffered, control lines are not).

And the production boards will not be purple, sadly. That's a Dorkbot PDX mask color who are cheaper only for small board runs where you buy panel space by the square inch. A non-green mask might be only marginally more expensive though at a larger board house.

An no apologies. I'm going to put all the blame on Mike B. (/ducks from the moderator ban stick).
 
The DP board is bright red, if that helps! But lacks gold contacts.

I haven't been able to get my hands on an old 5V logic IDE drive still (admitedly I've not tried particularly hard) so can't comment on the overall compatibility of the 3.3V CPLD; hopefully I can get some out 'in the field' and that (very valid) point can be answered soon.

However it does work fine with later 3.3v drives (ATA/100+) and CF cards, which has been my focus.

The DP v2 board should be quite cheap to build - by far the most expensive bit being the 28C64. Soldering SMT chips is very easy (blob it on thick over the entire row of contacts, then go over it wil a sucker and "ta-da!"). Howver the small bits (caps/resistors/led) do need tweezers, patience and flux.
 
I think it's fair to say that if you want longevity, the V2 card is probably your best bet, although it may not be the least expensive. Everything about it is out in the open and parts should be comparatively easy to obtain and, other than the BIOS in the PROM, nothing is programmable and problems are easy to troubleshoot with nothing more than an DMM and a logic probe.

If you want something less expensive, one of the others may be a better choice.
 
Hi,
OK that's a pretty good summary of where we are at. So I guess we'll proceed with a small batch of XT-IDE V2 PCBs in the near term. I am guessing to make the minor tweaks and spin up the PCBs this weekend and make a manufacturing order. The development of the two CPLD boards can continue in parallel and when they take off early next year they can support the demand for XT-IDE PCBs from then on out.

Does this sound like an acceptable plan? I've had enough of my own projects crush that I certainly don't want to do the same to someone elses project! If there is a problem with what I am describing please speak up!

Thanks and have a nice day!

Andrew Lynch

PS, the XT-IDE V2 does not require the ROM or the UART to function as an IDE controller. Leaving those sections DNP will significantly reduce the cost to build the unit.
 
Basically the plan on my CPLD board is to offer it (boards or kit) in the same fashon you and Jeff have done in the past. We keep hitting delays with personal time and schedules so that's why things have been delayed. So there are no surprises, I need to restate it will be significantly faster - which may enhance demand when I do get the first batch of production boards in. Though I'm not sure how much people will care about that since it is going in vintage XT class machines after all.

I was now targeting early Jan but with Christmas looming a couple weeks away, that may get pushed out yet again.

It seems the DP CPLD version is completely abandoned. There haven't been any significant posts in months and no one is actively pushing the effort forward.
 
I need to restate it will be significantly faster - which may enhance demand when I do get the first batch of production boards in. Though I'm not sure how much people will care about that since it is going in vintage XT class machines after all.

I'm not certain what you mean by "significantly faster". Andrew's V2 boards have what some call the "Chuck Mod", which enables read access by "IN AX,DX - STOSW" instruction pars on an 8088 or "INSW" on a system with a V20 or better. So, how much faster can one get?
 
Ever considered doing a busmaster DMA version for the XT to eliminate the extra move? I think that might be a significant improvement over either controller.
 
'busmaster' and XT don't go together. I considered doing one for the PCjr as the raw hold and hold ack lines are on the connector. 8-bit ISA slots are at the mercy of the 8237. It wouldn't speed raw transfer performance up any, however it would allow the processor to consume the prefetch queue while the transfer was in progress. Would also require an interrupt when done too. Could be done, I'm just not convinced there's much benefit.
 
Back
Top