451 replies to this topic

[Frankie]

I'm trying to imagine it on my XL, with a RAM 320XL plugged into it. I don't think I can picture the angles of the two...
So I will just have to have one :-)

Curt Vendel, on Wed Mar 23, 2011 8:10 PM, said:

Thats just begging for an XL or XE styled plastic case to be made for it

drac030, on Wed Mar 23, 2011 1:09 PM, said:

Hi

For some time there have been some ongoing works on a new PBI interface allowing to attach an IDE device to an 8-bit Atari. Previous solutions of this type - KMK/JŻ IDE and IDEa - had the disadavantage of being rather large, being limited to XE computers only (this is KMK/JŻ IDE), and having problems to cooperate with certain computers due to the well known problems with the PHI2 signal not being generated the same across various Atari XL and XE motherboard revisions, and across computers with different hardware extensions installed.

These problems have been addressed. The new interface:

1) is not much greater than a 2,5-inch IDE disk

2) does not require an external power-supply

3) contains an XE (CART+ECI) connector, an XL (PBI) connector, a cartridge port, and a connector for a 2,5-inch IDE disk

4) contains 96 KB ISP Flash ROM and 32 KB RAM available to the internal PBI BIOS and flashable from the Atari

5) contains 512 KB or 1024 KB ISP Flash ROM for SpartaDOS X (or any other cartridge of such type), flashable from the Atari. This ROM can be physically disabled using a dedicated switch. The switch is not installed in the device shown on photos below.

6) contains a real time clock circuitry to be used with SpartaDOS X. If someone doesn't want to use it, the module is anyway fully compatible with an external R-Time 8 or internally mounted ARC.

7) last, but not least, it also contains a LED flashing during the I/O

Five prototypes exist, one of them was presented on the Forever party 2011 in Slovakia, few days ago (to those who came to our desk and asked what it was). A larger production batch is being perpared, a subscription topic will be opened as soon as the final price is calculated and (of course), if anyone wants the device.

Photos:


1) Top view, with the disk dismounted. CART+ECI plug at the left, cartridge port at the right.


2) Bottom view, PBI connector visible on the far end


3) The same stuff at different angle.


4) Top view with the disk mounted.

The PCB has holes for 2,5-inch disks, either "old" or "new" type (e.g. 2 GB Toshiba MK2103 or 30 GB Toshiba MK3018). You can see that one of the holes is misplaced on the prototype PCB It is most visible on the first photo. This obviously will be corrected in the final batch.

Estimated price apiece of the final batch is, according to the hardware designer, "not much more than 200 PLN" (plus shipping). It may get a bit lower or higher depending on the quantity of the batch and the final price asked by the factory guys for the PCBs and assembly (coz it will not be assembled manually, but at a regular electronic factory instead).

A PLN (Polish złoty) exchange rates are approximately these:

1 USD = ~2,8 PLN
1 EUR = ~4 PLN

[David_P]

Added bonus feature: I suspect you'll be able to plug your Ethernet cart into the cart port on the interface while using the 800xl cart port for something else.

[venom4728a]

I am interested in purchasing 2, let me know when preorders begin please!

[JohnBuell]

I definitely want to give one a try!

[drac030]

bob1200xl, on Wed Mar 23, 2011 6:45 PM, said:

How does this work with no external supply?

You're right, I forgot about this issue: it works without external power supply with XEs and with 600XL. In 800XL there is no 5 V on the PBI, so the power must be either supplied externally, or the computer must be modified by adding a wire to supply the 5 V to the PBI, like in 600XL.

[MEtalGuy66]

I have a question concerning the XL-PBI side of this interface.

Has the layout been CORRECTED from the original IdeA design, or is it the same?

The reason I ask this is because there are 2 ways to populate the board on the XL side. If you look at the FIRST PICTURE in Drac030's post above, you will see that there are 3-rows of pads. If you put a female edge connector on the outer 2 rows (as shown on the board in the picture,) you can plug the thing directly into an XL. If you use the center row, and one of the outside rows, you can install a 50 pin IDC header and use a normal flexible cable, with a 50 pin IDC connector on one end, and a female card edge connector on the other end.

This is by far THE BEST solution for connecting a PBI device. Plugging directly into the ECI or PBI slot is only good if you have the interace "hard-fixed" to the machine in some way so that it cant move at all. Otherwise, the connections are going to be very succeptible to vibration. This has been proven time and time again, since the dawn of PCB edge connectors. This is why cards in a PC have a back-plate with a screw.

So, the problem with the original IdeA board layout is that they got the orientation of the 2 rows of pins reversed. This means that you have to either lay the interface top-side-down, or make a very "custom" cable in which each "pair" of wires individually reversed.

THIS NEEDS TO BE FIXED ON THE NEW LAYOUT BEFORE PRODUCTION if it has not already been fixed. If the layout for the XL-PBI connections is the same as the original IDEa then it is WRONG.

Candle is very well aquainted with this problem, based on discussions that he and I have had concerning this in the past. Perhaps he can explain this to the designers if they are unable to understand what I mean by this.

Trust me, this is VERY important. It is the NUMBER 1 COMPLAINT that XL users have with the original IDEA board.

here is the standard that should be followed for a CORRECT connection to a standard crimped flexible cable:

NOTICE THE PIN NUMBERS ARE REVERSED FROM A STANDARD IDC HEADER NUMBERING SCHEME! This is exactly how it should be. EVERY PBI device which uses a standard crimped flexible cable that plugs directly into an XL uses this EXACT LAYOUT for the 50 pin header.
If you don't believe me, then look at an MIO, Supra, or Black Box.

Edited by MEtalGuy66, Thu Mar 24, 2011 4:04 PM.

[drac030]

As I am not the hardware guy here, I can't answer this, so simply let's wait until the designer replies. I only want to clarify (maybe pedantically) some minor points here:

MEtalGuy66, on Thu Mar 24, 2011 3:29 PM, said:

Has the layout been CORRECTED from the original IdeA design, or is it the same?

This design is not the redesigned IDEa (hardware-wise). I know it may seem complicated, but there is some "family" of IDE host adapters, which use the same software, so it may seem that these are subsequent versions of the same design. In fact it is not exactly so:

1. first generation: KMK/JZ IDE (aka 1.0)
2. second generation: KMK/JZ IDE 1.2 (same as the above, just more memory)
3. third generation: IDE Plus 2.0 (designed anew)

"IDEa" is a compatible clone of the "second generation" thing. This clone has been made from scratch by another designer, who drew the PCB anew and used different parts (he did it because in early 2000s there were already problems to purchase parts for a 5 years earlier design).

So: there is little chance the IDE Plus would inherit any design errors from IDEa.

Quote

This is by far THE BEST solution for connecting a PBI device. Plugging directly into the ECI or PBI slot is only good if you have the interace "hard-fixed" to the machine in some way so that it cant move at all. Otherwise, the connections are going to be very succeptible to vibration. This has been proven time and time again, since the dawn of PCB edge connectors. This is why cards in a PC have a back-plate with a screw.

You're probably right, I can only judge from my 130XE where the device is plugged into: it sits very firmly and tighly (much more firlmy than an average cartridge) and (so far) there have been no issues with the connection.

[MEtalGuy66]

drac030, on Thu Mar 24, 2011 4:26 PM, said:

As I am not the hardware guy here, I can't answer this, so simply let's wait until the designer replies. I only want to clarify (maybe pedantically) some minor points here:

MEtalGuy66, on Thu Mar 24, 2011 3:29 PM, said:

Has the layout been CORRECTED from the original IdeA design, or is it the same?

This design is not the redesigned IDEa (hardware-wise). I know it may seem complicated, but there is some "family" of IDE host adapters, which use the same software, so it may seem that these are subsequent versions of the same design. In fact it is not exactly so:

1. first generation: KMK/JZ IDE (aka 1.0)
2. second generation: KMK/JZ IDE 1.2 (same as the above, just more memory)
3. third generation: IDE Plus 2.0 (designed anew)

"IDEa" is a compatible clone of the "second generation" thing. This clone has been made from scratch by another designer, who drew the PCB anew and used different parts (he did it because in early 2000s there were already problems to purchase parts for a 5 years earlier design).

So: there is little chance the IDE Plus would inherit any design errors from IDEa.

What Im getting at is that if the designer used the previous design's connector pinout as a standard, then it is WRONG and will cause many people alot of frustration. I totally understand that the device is a ground-up redesign, where the logic portion of the PCB is concerned. That's obvious.

Quote

Quote

This is by far THE BEST solution for connecting a PBI device. Plugging directly into the ECI or PBI slot is only good if you have the interace "hard-fixed" to the machine in some way so that it cant move at all. Otherwise, the connections are going to be very succeptible to vibration. This has been proven time and time again, since the dawn of PCB edge connectors. This is why cards in a PC have a back-plate with a screw.

You're probably right, I can only judge from my 130XE where the device is plugged into: it sits very firmly and tighly (much more firlmy than an average cartridge) and (so far) there have been no issues with the connection.

Use it constantly for 5 years like that, and you will be in 100% agreement with what I said..

Please pass the post I made above on to the board designers if you would.

Edited by MEtalGuy66, Thu Mar 24, 2011 4:34 PM.

[flashjazzcat]

The two or more hours I spent making a twist cable to connect the KMK/JZ 1.2's pin header to an XL edge connector were bereft not only of laughter but any degree of enjoyment. I can't see how flipping the board upside down could have avoided this: surely that simply connects pin 1 to pin 50?

Anyway: although I'm sure the ability to make twist ribbon cables is a skill which will stand me in fine stead during the oncoming years, I don't fancy doing another SCSI one. As I said before: if the unit could be plugged "cable-free" and flat directly into the XL (turned by 180 degrees), it would be as stable as when attached via the XE connector. Oriented at 90 degrees to the machine, however, it will be susceptible to knocks, so a cable connection will be hugely preferable.

The existing KMK/JZ's XE adapter is a good compromise in as much as the interface isn't actually mounted directly to the back of the computer. My lovingly-prepared twist cable enables the same trick to be performed with the XL, with the obvious advantage that we then forego the additional clutter of the XE adapter.

[drac030]

MEtalGuy66, on Thu Mar 24, 2011 4:32 PM, said:

What Im getting at is that if the designer used the previous design's connector pinout as a standard

Knowing him, I doubt But I'll let him speak for himself in this respect, as I said, I don't feel competent here.

Quote

Quote

You're probably right, I can only judge from my 130XE where the device is plugged into: it sits very firmly and tighly (much more firlmy than an average cartridge) and (so far) there have been no issues with the connection.

Use it constantly for 5 years like that, and you will be in 100% agreement with what I said..

As I said, you're probably right. When such problems appear, I think it always can be connected via a cable to avoid the problems you mention (like in IDEa) or mounted internally. For now let me appreciate the decreased desk's space my 130XE needs and its increased movability (I don't have to unplug the hard disk when I want to move the computer to another place in my home).

Quote

Please pass the post I made above on to the board designers if you would.

I am sure he's reading this thread.

[MEtalGuy66]

flashjazzcat, on Thu Mar 24, 2011 4:37 PM, said:

The two or more hours I spent making a twist cable to connect the KMK/JZ 1.2's pin header to an XL edge connector were bereft not only of laughter but any degree of enjoyment. I can't see how flipping the board upside down could have avoided this: surely that simply connects pin 1 to pin 50?

You populate teh header on the OTHER side of the board. That's how it works without making the special cable, but unfortunately, the interface ends up upside-down. I think you can trust me in this case. Ive explored ALL possibilites, with the 3 people that Ive had to come up with "solutions" for their IdeA board.

Quote

Anyway: although I'm sure the ability to make twist ribbon cables is a skill which will stand me in fine stead during the oncoming years, I don't fancy doing another SCSI one. As I said before: if the unit could be plugged "cable-free" and flat directly into the XL (turned by 180 degrees), it would be as stable as when attached via the XE connector. Oriented at 90 degrees to the machine, however, it will be susceptible to knocks, so a cable connection will be hugely preferable.

Over time, either hard-plugged solution XL or XE will cause constant wear to the card edge and connector contacts as vibration is transferred from the keyboard, through the machine, into the interface, and the connector is the "weakest link" so the contacts will move in relation to eachother ever so slightly and eventually cause accelerated wear.. This is a VERY REAL PROBLEM and it's very apparent to anyone who has owned a blackbox and used it with an XE for any period of time. It also REALY IS the reason that cards are bolted down in machines.

Quote

The existing KMK/JZ's XE adapter is a good compromise in as much as the interface isn't actually mounted directly to the back of the computer. My lovingly-prepared twist cable enables the same trick to be performed with the XL, with the obvious advantage that we then forego the additional clutter of the XE adapter.

It's a good compromise, given that IDEa boards are already made.. It's not a good solution by comparisson to just using the correct pinout in the first place. And if at all possible, this new design needs to have the correct pinout.

[BillC]

MEtalGuy66, on Thu Mar 24, 2011 5:04 PM, said:

It's a good compromise, given that IDEa boards are already made.. It's not a good solution by comparisson to just using the correct pinout in the first place. And if at all possible, this new design needs to have the correct pinout.

Ken: The pinout is correct for using the female card-edge connector, switching to having a female IDC ribbon cable connector to a male header flips the odd/even pins unless the header is installed on the opposite side of the circuit board. Extending the board so that the header didn't interfere with removal/replacement of the HDD is all that would be required, users could then be given the choice of card-edge connector as shown or ribbon cable header mounted on the HDD/component side.

[MEtalGuy66]

The pinout on what? The new board, or the original IDEa?

I have tried the IDC header on the original IDEa board in every possible orientation, and it is impossible to have the IDEa sitting face-up on the table.. And MANY people have commented on this. The two pin-rows are reversed in relation to ANY OTHER PBI DEVICE..

This needs to be correct on the IDEa 2.0...

It's a major problem and MANY PEOPLE have complained about it..

Edited by MEtalGuy66, Thu Mar 24, 2011 8:48 PM.

[BillC]

Ken:

You are complaining about the IDEa device which was a clone of the second generation KMK/JZ IDE device.

My comments are about the IDE Plus 2.0, the device this thread is about, third generation of the KMK/JZ IDE device.

What I am saying is that making the circuit board longer would allow either connector to be installed as desired, just that they would install on opposite sides of the circuit board to each have the correct pinout. The center row of holes between the card-edge mounting pads is essential to this as card-edge rows are spaced 0.2" apart while the header rows are 0.1" apart, so the center row needs to be connected to one of the outside rows. Since the card-edge connector already installs on the bottom of the circuit board, the header could be installed on the HDD/component side instead if desired. Would you object to a PBI compatible header being installed like this?

Extending the circuit board may not even be necessary, it is hard to tell whether there is enough clearance to plug/unplug a 2.5" HDD if the header is installed, especially since the drive mounts spaced above the circuit board, the pictures supplied don't show that part of the circuit board with the HDD installed. Even if a vertical header would interfere a right-angle header may not.

MEtalGuy66, on Thu Mar 24, 2011 8:44 PM, said:

The pinout on what? The new board, or the original IDEa?

I have tried the IDC header on the original IDEa board in every possible orientation, and it is impossible to have the IDEa sitting face-up on the table.. And MANY people have commented on this. The two pin-rows are reversed in relation to ANY OTHER PBI DEVICE..

This needs to be correct on the IDEa 2.0...

It's a major problem and MANY PEOPLE have complained about it..

[TXG/MNX]

I would like one.

Is it possible to add a second cartridge port on the final version ?

Is there a possibility to add a feature board to extend this PBI interface with other things like rs232c,printer ?

Where can I find the Memory map of this device.

Can I connect more then 1 IDE drive to this device ?

[TXG/MNX]

o last question, will the ROM support the 65c816 I mean will you make a special version for this processor would be somewhat faster with PBI routines.

[Simius]

BillC is right. There is possible to install a right-angle header on the component side of PCB. Clearence is 5mm, that is enough for header, but without plastic insulator. Is possible, but not recommended because of interference between adjacent wires, resistance and inductance on power lines, etc. It was never testet in this connection. Maybe with short (a few cm) cable it will be ok.

TXG/MNX, on Fri Mar 25, 2011 5:48 AM, said:

Is it possible to add a second cartridge port on the final version ?

Is there a possibility to add a feature board to extend this PBI interface with other things like rs232c,printer ?

Where can I find the Memory map of this device.

Can I connect more then 1 IDE drive to this device ?

1. No. What for?
2. No
3. Nowhere at the moment, but it will be published soon.
4. Yes, if not mounted a narrow, right-angle PCB with 2,5" HD header.

Edited by Simius, Fri Mar 25, 2011 7:04 AM.

[MEtalGuy66]

I don't want alternative suggestions.. I've already explained why DIRECT-PLUGGING of the interface into a machine is not a decent long-term arrangement where reliability/stability of operation is concerned. This is not just my preference, this is a known fact, and an industry-wide design convention of the most basic/rudimentary nature, and has been FOREVER. If you direct-plug a sizeable card with an edge connector, you hard-mount that card to the frame of the device.

What I want is a STRAIGHT ANSWER.. "YES" or "NO"... Can a 50 pin header be installed on the top side of the board, with any/all signals used by the interface exactly corresponding the the pin orientation shown in the diagram I posted?

If the answer is "NO" then that is unacceptable, and I DO NOT RECCOMEND ANYONE BUY THIS INTERFACE UNTIL IT HAS BEEN CORRECTED

Why can't you guys include the standard header layout that everyone has known works for the last 28+ years? (1090XL,MIO,SUPRA,BLACKBOX ALL USE IT) I do not understand why this is such a a major problem for you.

Quote

interference between adjacent wires, resistance and inductance on power lines, etc. It was never testet in this connection. Maybe with short (a few cm) cable it will be ok.

Dude.. I don't know what planet you come from.. Or if you're just trying to blow smoke up my ass, as an excuse not to do things right.. But that's total bullsh*t.. NOTHING on the ATARI talks fast enough for that to be an issue. Both ICD and CSS reccomended PBI cable lengths of less than 14 inches (this is in addition to an XE cartrdige/PBI adaptor board). SLOR is using a PBI cable of at least 3 feet to put his IDEA on a shelf above his ATARI, with no issues. I dont think a few inches is going to cause a problem. It never has for ANY OTHER DEVICE in the past.




BillC: I'm sorry but you do not get it.. I have tried headers on BOTH SIDES of the existing IDEA BOARD (repeating myself now) and there is no connectior population/cable orientation possibility that results in the cable directly exiting that side of the IDEA board and going straight to an XL PBI port, with the IDEA laying flat, face-up. Period.

I HAVE EXHAUSTIVELY EXPLORED ALL POSSIBLE OPTIONS HERE..
THIS IS A 100% CERTAIN FACT.
Please do not respond to this, as you have proven to me that your oppinion on this matter is useless to me. Please allow the hardware designer to respond to this issue.

Edited by MEtalGuy66, Fri Mar 25, 2011 8:04 AM.

[MEtalGuy66]

Sorry about that, BillC. I made that last post right after I woke up, and I somehow missed your last response. I apologoize for the irritated tone of my message. I do value your input, and I do realize that what you said regarding the possibilities of changing the layout (and the physical/mechanical considerations therof) do make sense. It wont let me edit my post at this point, or I would change it.



Here is my position: Either the current configuration matches the STANDARD which DOES WORK, or it matches the configuration used on the original IDEA, which does not work.

I am still looking for a simple "YES" or "NO" answer (regarding whether or not it does indeed match the STANDARD) from the hardware designer.

Perhaps someone who has a firm understanding of the nature of this issue, and is fluent in both english and polish needs to explain this to him.

Edited by MEtalGuy66, Fri Mar 25, 2011 9:45 AM.

[drac030]

TXG/MNX, on Fri Mar 25, 2011 5:50 AM, said:

o last question, will the ROM support the 65c816 I mean will you make a special version for this processor would be somewhat faster with PBI routines.

The device has 96 KB ROM, so I guess such support code is possible to fit into, if it proves to be useful.

[flashjazzcat]

I take it this is getting brand new, speed-record breaking firmware.

[BillC]

MEtalGuy66, on Fri Mar 25, 2011 9:09 AM, said:

Sorry about that, BillC. I made that last post right after I woke up, and I somehow missed your last response. I apologoize for the irritated tone of my message. I do value your input, and I do realize that what you said regarding the possibilities of changing the layout (and the physical/mechanical considerations therof) do make sense. It wont let me edit my post at this point, or I would change it.



Here is my position: Either the current configuration matches the STANDARD which DOES WORK, or it matches the configuration used on the original IDEA, which does not work.

I am still looking for a simple "YES" or "NO" answer (regarding whether or not it does indeed match the STANDARD) from the hardware designer.

Perhaps someone who has a firm understanding of the nature of this issue, and is fluent in both english and polish needs to explain this to him.

I have a solution, maybe not ideal, that would allow this device to sit HDD side up behind an 800XL using a ribbon cable. I estimate about 5" to 6" would be the minimum cable length.

The device sits in the same orientation as with the 130XE, with the cable attached to an angle header on the right end, then twisting/folding to plug into the PBI port as shown in the picture below, the red stripe is pin1.


It would require testing to be certain that the cable didn't add enough capacitance to cause problems, but the MIO always used ribbon cable and do did the Black Box when used with an XL.

[MEtalGuy66]

I don't understand what's so hard about using the correct layout to begin with..

A) I didn;t come up with the STANDARD, ATARI DID.. Every device since the 1090XL has used it.

B) WHo in their right mind would EVER think that having it plugged directly into the back of a 600/800XL, sticking up in the air, with everything on the board facing AWAY FROM YOU, including the CART CONNECTOR would be acceptable? I have often asked this question about the original IDEa board, and the only answer I am able to arrive at is that the designers gave ZERO CONSIDERATION to the physical/mechanical design of the device.

C) Even if you DID use this insanely inconvenient mating option, you would have to build some kind of framework to support it or it WOULD become extremely sensitive to vibration and card-edge wear over time..

The layout currently on the board makes no sense for anything anyone would actually prefer. WHy not use the correct STANDARD?

Edited by MEtalGuy66, Fri Mar 25, 2011 7:00 PM.

[MEtalGuy66]

BillC, on Fri Mar 25, 2011 6:17 PM, said:

I have a solution, maybe not ideal, that would allow this device to sit HDD side up behind an 800XL using a ribbon cable. I estimate about 5" to 6" would be the minimum cable length.

The device sits in the same orientation as with the 130XE, with the cable attached to an angle header on the right end, then twisting/folding to plug into the PBI port as shown in the picture below, the red stripe is pin1.
PBI cabling.jpg

It would require testing to be certain that the cable didn't add enough capacitance to cause problems, but the MIO always used ribbon cable and do did the Black Box when used with an XL.

Billc, That would make the connector have the wrong orientation, left-to-right. There is no solution that can be done by folding, or I (and many other people) would have already done it, rather than making painstakingly "pair-swapped" cables.

The "capacitance/inductive feedback" is not going to be an issue. None of the signals being carried by the cable are any different in respect to other PBI devices. I will say AGAIN.. This is the WAY ATARI CONNECTED the 1090XL...

Look at the difference in the length of the traces that go from the ECI port on an XE, to the CPU and other chips, and the length of the traces that go from the PBI port on an 800XL to it's CPU and respective chips. QUITE A BIT. You expect any sane person to think that a couple of inches are going to matter on the length of the cable?

What I am talking about here is plain truth, and common sense. And it's a major consideration when it comes to how this device can efectively be used. There is alot more to designing a PCB than just arranging the components in a manner that all needed circuits can be connected by traces in a given number of layers. The very FIRST CONSIDERATION should be given to the physical/mechanical design characteristics of the device. You guys should get behind me on this, and do not order it unless the design is corrected. We already put up with these well-known issues on the original IDEa. Theres no reason at all that this new design can not be decent. It's a very simple matter of moving/reorienting a few pads & traces by a small fraction of an inch.

Edited by MEtalGuy66, Fri Mar 25, 2011 7:28 PM.

[drac030]

MEtalGuy66, on Fri Mar 25, 2011 7:20 PM, said:

Billc, That would make the connector have the wrong orientation, left-to-right.

Not going into the hardware/standards discussion, I'd however like to say that it seems to me that BillC is right. You're complaining that connecting the device via a cable from the PBI connector in the interface to the PBI connector in the XL will make the interface lay turned upside down on the desk. That's right, it would. I just took a ribbon cable and checked that out: folding the cable like that (to the left or to the right, it doesn't matter) makes the interface to be turned side up.

Edited by drac030, Fri Mar 25, 2011 8:28 PM.