1088XEL Alternative Mother-Board Project

[Panther]

36 minutes ago, mytek said:

Because as it turns out it doesn't matter and can be done either way with the same results (learning curve). And bottom line I believe the extra logic got nuked on the XLD mainly because I ran out of gates and didn't want to add yet another IC to implement something that wasn't necessary.

Got it, thank you for the info.  Then, I can free up one of the NAND gates to use for the disk swap.  🙂

 

[+mytek]

8 hours ago, Panther said:

Got it, thank you for the info.  Then, I can free up one of the NAND gates to use for the disk swap.  🙂

Aww so you're going to do what was done in the XLD on your XEL, so that you can use the PS/2 keyboard ALT+N command to initiate the disk swap feature. Good idea 👍

 

And that is exactly why the RAM R/W logic was altered on that unit

[_The Doctor__]

4 hours ago, mytek said:

Aww so you're going to do what was done in the XLD on your XEL, so that you can use the PS/2 keyboard ALT+N command to initiate the disk swap feature. Good idea 👍

 

And that is exactly why the RAM R/W logic was altered on that unit

might make a nice addendum to that build out as well

[+mytek]

22 hours ago, _The Doctor__ said:

might make a nice addendum to that build out as well

Well here's a quickie attempt to do that...

 

So the idea is to give the 1088XEL the ability to do a Disk Swap on the XEL-CF3 drive via a PS/2 (TK-II) ALT+N command. First of all it will require that the latest TK-II PIC firmware be flashed (from HERE).

 

Then if we look at a snippet from the 1088XLD schematic we can see where the pertinent connections are. These were designated as JOY1 and JOY2 on the XLD schematics, which represented their primary purpose of selecting which joystick port was designated for the PS/2 Mouse. However these served double duty, and via a NAND gate would also activate the disk swap feature on the XEL-CF3 board.

 

Now if we take a look at a marked up snippet from the 1088XEL schematics, we can see that the desired connections for JOY1 and JOY2 can be intercepted at J25 (PS2-AUX). All it takes to add the necessary logic is to make use of a single NAND gate, which can be stolen from the RAM R/W circuit with some creative bypassing, or perhaps it would be better to simply use an entirely new 74HCT00 IC instead. Power for the new chip would also be available from J25, with pin 1 supplying the GND (HCT00 pin 7) and pin 3 supplying the +5V (HCT pin 14). Diode D1 needs to be removed so as not to interfere with the PIC's new roll as a swap disk controller, and the PIC chip does need to be flashed with the latest firmware that supports this feature. Keep in mind that the possibility of connecting a 2nd PS/2 keyboard is no longer an option either through J25 or J11.

 

 

For reference as to where the /swap output of the NAND gate goes, here is the XEL-CF3 schematic: XEL-CF3_V1_1_schema.

 

74HCT00 Pin-Out Diagram

 

It's also possible to connect the JOY1 and JOY2 connections to the 1088XEL J14 header via pin 3 and pin 5 respectively to now control the mouse by pressing ALT+M on the PS/2 keyboard (this steps through the selection process). It would be good to still retain the LEDs to let you know which port has been designated as the mouse.

 

That's as far as I wish to dig into this at the moment, and good luck to anyone that attempts to do this

 

EDIT: I could see this being done as a small upgrade PCB with the NAND gate IC and a couple of pin headers -- anyone game to make that happen?

[Panther]

Okay, so simply remove D1 and use one tiny little board with a $0.40 TI SN74AHC1G00 that can plug into J14 (using a stacking header so as to not lose access to this) and have a single pin to connect a wire to pin 2 of J2 on the XEL-CF3.

[+mytek]

1 minute ago, Panther said:

Okay, so simply remove D1 and use one tiny little board with a $0.40 TI SN74AHC1G00 that can plug into J14 (using a stacking header so as to not lose access to this) and have a single pin to connect a wire to pin 2 of J2 on the XEL-CF3.

Yep that would do it 👍

 

[Panther]

Oops, J25, not J14.

[+mytek]

31 minutes ago, Panther said:

Oops, J25, not J14.

I knew that, and was just testing you   NOT - passed right by me.

[Panther]

38 minutes ago, mytek said:

I knew that, and was just testing you   NOT - passed right by me.

I hope I passed! 

A slim little board to squeeze between the PS/2 connector and the UAV...

 

 

Edited February 11, 2023 by Panther
Changed the image to a better view.

[+mytek]

55 minutes ago, Panther said:

I hope I passed! 

A slim little board to squeeze between the PS/2 connector and the UAV...

 

 

That looks sweet 👍

 

Are you going to get one made? Because of how small it is, I would suggest OSH Park which will probably come in the cheapest of everyone, considering that shipping is included.

[Panther]

After measuring fit on all sides and verifying connections, I think this is my final version.  I'll have a batch of them made.

 

[Panther]

I decided that it would be best if the entire board could fit between the PS/2 connector and the UAV board itself, since with the female header it will likely sit at about the same height.  Though an option would be to simply solder this board onto the J25 header, I think sticking with a connector would be better, so...I rotated the 74AHC1G00 and revised all the connections...got the board down to 3.9 mm wide!  Then I found out the minimum width for the PCB fabricator is 5mm.  🙄  So, I will make it exactly 5mm and this should fit the XEL, just a bit tightly.  Some slight filing of the (excessive) board edge might be needed.  At least I can probably fit the silkscreen back on.  The 3.9mm version shown...

 

[Panther]

Here's a question...you specified using RB2 and RB3 for the Alt-N output, but isn't it just RB3 that's used as the switch for this?  Why connect RB2 to the NAND gate/swap?

 

[+mytek]

8 hours ago, Panther said:

Here's a question...you specified using RB2 and RB3 for the Alt-N output, but isn't it just RB3 that's used as the switch for this?  Why connect RB2 to the NAND gate/swap?

Keep in mind that RB2 and RB3 I/O bits from the PIC serve double duty, and function both as PS/2 Mouse Assignment and as the Disk Swap Button. To see what I mean, look at the logic table below.

In normal usage as a mouse-to-joystick port assignment selection, each I/O bit serves as a 'high' enable directing the mouse to be assigned to a given joystick port via the Mousetari PIC chip. You can also think of RB2 as Joystick1 and RB3 as Joystick2, and when either is 'high' then that respective joystick port becomes a mouse port. In that usage there are 3 possible settings, with one of those being no mouse assignment. These are latched I/O bits when used for the mouse assignment function. ALT+M steps through the mouse assignment selection each time its pressed.

 

The forth state, with both RB2 and RB3 set to 'high' is not latched and only momentary in action -- reverting back to the previous latched state, pulsed by the PIC every time ALT+N (Next Disk) is pressed. The NAND gate responds to both I/O bits being 'high' and produces a corresponding 'low' output to momentarily send a swap signal to the U1MB via a specific XEL-CF3 address.

 

If you try to skip the NAND gate and control the swap with just one of the I/O bits, the logic is inverted with the swap button being held 'low' and only being pulsed 'high' momentarily. It might still seem to work, but puts the XEL-CF3's swap NAND Flip-Flop in a disallowed state, which could potentially compromise operation. EDIT: it would act the same as the swap button being permanently pressed, which if I recall correctly can cause a potential problem in FJC's U1MB BIOS's interpretation.

 

If you incorporate the NAND gate as part of your upgrade board, then all things are possible (e.g., Mouse Assignment and Swap), and there are no questionable disallowed states being created.

[Panther]

Thank you for the thorough explanation.  I never made use of the mouse capabilities, so I hadn't paid much attention to the operation.  I've ordered the TI gates and female headers, and will have the PCBs soon.

 

[Panther]

Thank you, mytek for the excellent suggestion and diagram.  This works perfectly.

 

[+mytek]

9 hours ago, Panther said:

Thank you, mytek for the excellent suggestion and diagram.  This works perfectly.

 

That's an extremely tiny and nice looking PCB you created 👍

 

At that small size, I bet OSH Park would be a good place to have it made as well, since their batch process pricing favors tiny boards. Coupled with their free international shipping, I think you would be hard pressed to beat this with either JLCPCB or PCBway. Any thoughts about maybe uploading the design and providing a link? Because I think others with a 1088XEL would like this.

[Panther]

Unfortunately, OSH Park's minimum PCB size is 0.25" x 0.25", or 6.35mm wide.  My expanded version at 5mm wide just barely fits as it is.  Of course, a taller header could be used to bring the board up above the UAV.  Gerber files attached.  I've also put them up on AtariVerse:

http://www.atariverse.com/files/hardware/xel-swap/

I have a number of boards that I can send out to people in the U.S.  I might even be willing to assemble some.

 

XEL-Swap.rar

[malachykidd]

14 hours ago, Panther said:

Unfortunately, OSH Park's minimum PCB size is 0.25" x 0.25", or 6.35mm wide.  My expanded version at 5mm wide just barely fits as it is.

Fab multiple boards onto a master board, with the sub-boards separated by drill slots?  Then you could snap the sub-boards apart.  I've never used OSH Park, so I don't know how much they'd charge for that, but it looks like they offer it.