1088XEL Alternative Mother-Board Project

[+mytek]

That is just beautiful! I'm really glad I ordered that Realan case now. I'll definitely be ordering those top and back panels!

 

What are you using for the Atari logo on the front of the case?

 

Hi Stephen

 

I got the label through a shop called Lukparts on Esty... Atari Label / Aufkleber / Sticker / Badge / Logo [134] ($2.99/ea)

 

It comes with both a front and rear peel off backing so that only the raised metallic logo and lettering remains, with nothing left between the individual parts. The end result looks like it was embossed from the anodized aluminum of the Realan case. And it really works well with the silver look of the engraved lettering on the top and rear panels.

 

EDIT: I trimmed off the Registered Trademark symbol to give it a more symmetrical look (I hope the Atari police don't come after me ).

 

 

I did have to make a few adjustments on the panels I had fabricated by Front Panel Express (not due to anything they did wrong). Luckily it was relatively minor stuff that I was able to fix such as the DIN-13 hole size being a bit too small, an incorrect offset on the holes for the CF Adapter mounting screws, and the need to do a bit of edge sanding to get the top panel to fit. So I've carried these changes into the panel files, which should insure that the next set works from the git go.

 

- Michael

Edited August 2, 2017 by mytekcontrols

[Van]

LOVE that case badge!!! Had to order a couple straight away Great find!!! They also do 'Atari800XL' and 'Atari600XL' badges that fit OEM cases, may have to get one or two of them too.

Yogi

[+mytek]

Well unfortunately it looks like we missed the Mid-Summer design release

 

I know some people were vested into this happening, and I am truly sorry that the projected schedule has slipped. But believe me when I say, I am not sitting on my butt playing video games over here while I'm suppose to be getting this project finished up (well perhaps I am playing a few games, but that's all part of the testing process ).

 

So why the slip? It's mainly related to getting the things that were planned from the beginning to all work as they should. Some of this required waiting on gathering other gear in order to test compatibility issues, some of it was attributable to getting all of the beta boards built and powered up, and then there were the inevitable bugs that weren't always easy to track down. And of course unlike simply cloning the original A8 motherboard, this one had all sorts of custom aspects incorporated right from the git go, which introduced just as many uncertainties. AA member exxosuk has also begun an alternative motherboard project of his own, only his is focused on the STFM (link). His approach appears to be a two part one, first with duplicating the basics and testing to be sure that works, and then the 2nd phase will be to add in features and expansion ports that didn't originally exist. Perhaps I would have been better off to take this approach as well, since it will obviously take at least two iterations of the 1088XEL to get it right. Should of, would of, could of

 

So when will the project release occur? Sorry to say with all things considered it's looking like early Fall. Still need to troubleshoot Stephen's board to find out if his PAL problems relate to a design issue or something else. Have to initiate a build of the next rev once all design issues have been rectified, and do a verification test of at least one assembled unit. And of course I have to finish up at least the minimal docs required for release, although that'll probably occur in parallel with the first two items. So that easily consumes the entire month of August.

 

- Michael

Edited August 5, 2017 by mytekcontrols

[Van]

Well, being involved in some of the testing I can say that the design is solid. The issues we have overcome are/were minor (once discovered) and a fair amount of 'human error'. All to be expected with a prototype Design errors and improvements have been incorporated into the release design so builders can avoid some of the pit-falls the beta boards had. Well worth the extra time spent!

 

I expect that whatever problems a builder faces will be due to a faulty part or faulty construction. This is not to say that building the XEL requires expert skills, it's just a matter of taking your time and double checking yourself. Any builder that can read schematics, identify parts, handle through-hole soldering and use a DMM should do well.

 

In fact it's remarkable that a new design, a complex design, could go from an idea, to a working prototype, to a refined release in under a year! JOB WELL DONE

Yogi

[Theoryman]

Well, better right than quick...

 

I'll keep collecting parts...

 

--

[+Stephen]

Yeah - wonderful design. Big reason for the delay is me. I tried everything in my power to get my PAL board working. 2 ANTICs, 2 GTIAs, 2 CPUs, every combo of the various chips, testing with NTSC clock, PAL clock, 2 different PAL crystals, etc. I just couldn't get it working.

 

According to the tracking #, Mike should receive my board this coming Tuesday. I don't know why the post office can't get a 19 ounce box to California in under 6 days, especially considering they charged me $18. I really think it's about time the USPS is allowed to go under, and become privatized. The government in this country cannot run shit, unless running something head first into the ground is considered a success. Anyhow - wrong forum for that.

 

I really hope the issue is something simple. At this point, it very well could be an incompatibility of sorts with the original black PCB candle Ultimate 1MB.

 

You guys are in for a real treat when this gets a release!

[+mytek]

According to the tracking #, Mike should receive my board this coming Tuesday. I don't know why the post office can't get a 19 ounce box to California in under 6 days, especially considering they charged me $18.

 

Such a deal, only $3 per day of travel

 

Stephen's not hanging up anything. As Yogi pointed out, a lot of stuff has been refined, improved upon, and some other issues besides PAL have gotten fixed. If anything Stephen has saved the day. Because if everything had gone smoothly for him the 2nd rev of the design would have already gone to the printers so to speak, and we'd still have issues.

 

-------------------------

Switching subjects

-------------------------

 

I got the LED status panel PCB today for the Realan case, stuffed one with components, and tested it for fit.

 

 

 

Looking good

 

- Michael

Edited August 6, 2017 by mytekcontrols

[+DrVenkman]

Take your time. I'm tapped out on hobby money at the moment anyway. Fall is good for me.

[Kyle22]

Better late (and perfect) than early and buggy!

 

[+mytek]

I got Stephen's board under test. Here are the symptoms he reported having...

Well, I seem to be the last one with a non-working (PAL) setup. I'll post my latest detailed findings. The initial problem was corrupted PMG when running in PAL mode.

Upon advice from Mike, I switched U11 from a 74HCT08 to a 74LS08 and reomved the jumper between pins 5&6. This made the system completely unstable. It would crash before booting most times, and from SDX, just doing a DIR CAR: would guarantee a lockup. Also, the PMG corruption was still there.
I tried running NTSC clock & crystal with the PAL chips installed. I still got PMG corruption and crashes.
Swapped CPU (original was UMC UM6502I out of an 800XE) for a 1200XL CPU. System seems stable now (before, DIR CAR: would immediately crash) but PMG issue is still there.

So I started swapping ANITC and GTIA chips (both NTSC and PAL) and discovered:
Everything works UNTIL a PAL ANTIC is installed.
AMI 9119MFU system is completely unstable with 74LS08
AMI 8730MDX system is more stable but still has borked PMG
When in SDX, doing a DIR CAR: will crash it in 2 lines.

PAL clock with NTSC ANTIC - everything works fine
PAL clock with NTSC ANTIC & GTIA - everything works fine

I'm at a loss for what to check next, but hopefully these symptoms can offer a clue as to what the problem is.

He provided me with the board having the 74LS08 chip in place (same as a stock A8) and PH0 buffering intact (stock no-mod setup on V1.0 board). However I'm not sure which CPU he actually gave me. Antic is AMI 8730MDX which was reported to be more stable (too bad I don't have the AMI 9119MFU to test with, which was reported as very unstable). Board is setup as PAL and has Candle 'BLACK' U1MB.

U1MB BIOS Settings

1st Menu
Extended RAM: 1088KB RAMBO
System: XL OS 1.3
BASIC slot: Altirra BASIC
XEGS slot: Disabled
SpartaDOS X: Enabled
Graphical OS: Disabled
Boot to loader: Disabled

2nd Menu
VBXE base: Disabled
Soundboard: Disabled
Stereo Pokey: Enabled
Covox: Enabled (note no Covox Device is actually present)

3rd Menu
PBI BIOS: Enabled
PBI device ID: 0
Hard disk: Enabled
ATR swap button: Enabled
Boot drive: D9
-All other setting are Disabled-

Note: I had to disable the Hard Disk since I was testing without it, otherwise it would hang on start-up at a blank blue screen.

4th menu
SIO driver: HSIO
SIO devices: D1-D4
SIO driver D1-D4: Enabled

Other Menus are unimportant.

Test Results

So on my initial tests on the board as it came to me, I didn't see any instability issues. However I did see some PMG corruption, which was corrected when I swapped the 74LS08 for a 74F08. Doing a DIR CAR in SDX with either the 74LS08 or the 74F08 didn't cause a lock-up as reported.

With SIO2PC connected and RespeQt active...
 
D1: DIR CAR

Volume:      DOS 2.0
Directory:   Root

    494 FREE SECTORS

Without any drive connected and/or turned on...

D1: DIR CAR

138 Device does not respond

Not sure if that is what Stephen was referring to when he said DIR CAR

 

Thus far I have run a few different PAL demos, and other than seeing some messed up PMG issues with the 74LS08 in place, it seems to be working perfectly with the 74F08 being used in its place.

 

 

Stephen... I'm not done yet, and will likely update your U1MB BIOS to the XEL specific one that FJC created, and then plug-in a CF adapter to really give this a good test run (I have lots of xex demos in a CF card that I can try). Any other specific behaviors you remember and/or programs (demos, ect.) that caused problems?

 

Edit: I also ran your PMG Basic test program and all looks good using the 74F08.

 

- Michael

Edited August 8, 2017 by mytekcontrols

[Brentarian]

Try "dir car:"

Edited August 8, 2017 by Brentarian

[+mytek]

Try "dir car:"

 

Thanks Brent

 

Okay just tried that with the 74LS08 in use at least a dozen times in a row and it works without a problem (does quite a long listing).

 

I'm wondering if Stephen possibly had PSU issues, since I am using one of mine to power his board.

 

- Michael

[+Stephen]

Damn - I'm glad to hear it was that simple. All of that because I didn't specifically order and try the F08 you suggested

 

Thanks so much for looking at this. Can I be a pain and ask you to please update the TK-II, VGATE, and MouseTari PICs please? I haven't yet made an adapter cable which will plug into my 130XE so I am having to run JOY2PIC thusly:

 

I am looking forward to the final product! I'll start preparing for VBXE install

[+Stephen]

 

Thanks Brent

 

Okay just tried that with the 74LS08 in use at least a dozen times in a row and it works without a problem (does quite a long listing).

 

I'm wondering if Stephen possibly had PSU issues, since I am using one of mine to power his board.

 

- Michael

Interesting. In all of my testing, I did not try an alternate supply, or check the voltage rails with my o-scope. Wouldn't surprise me if the supply was sub-par like everything coming out of China.

[+mytek]

Interesting. In all of my testing, I did not try an alternate supply, or check the voltage rails with my o-scope. Wouldn't surprise me if the supply was sub-par like everything coming out of China.

 

Not saying that it is your power supply, but then again I'm not seeing any of the instabilities you reported with your board as you sent it to me. However it does need a 74F08 to really work properly with PMG's.

 

Any other tests you ran that I should as well?

 

- Michael

[+mytek]

Can I be a pain and ask you to please update the TK-II, VGATE, and MouseTari PICs please?

 

I was planning on doing that anyway .

 

So here's what I had in mind...

1. Install 74F08
2. Add 4.7K data line pull-ups.
3. Replace your crystal socket with one that actually fits tight.
4. Update U1MB with all the new versions, including the upcoming revised XEL-i3 PBI code.
5. Include an assembled XEL-i3 CF Adapter and one of those RGB-THRU fix-it boards.
6. Provide you with a known working PSU.

Edit: Most of the above also applies to Simon's board as well. Still waiting on some IC's to show up from back East (as in China) which are required to build a couple more XEL-i3 boards (that is why I have been holding off on returning Simon's XEL).

 

The fix-it board is probably optional if you go with a VBXE, although it does have a CSYNC buffer on it that the VBXE might need (depends upon how much the VBXE loads the circuit).

 

 

- Michael

Edited August 9, 2017 by mytekcontrols

[Brentarian]

I haven't yet made an adapter cable which will plug into my 130XE so I am having to run JOY2PIC thusly:

 

I programmed my first few pics just like you did, but I ended up making a cable soldered to a socket (one cable/socket each for the two pic types).

 

(Michael, if you ever have the need to make changes to joy2pic down the road, it would be awesome if you could squeeze in a socket onboard with jumpers to select different pic pinouts (like the jumpers on UAV for different 8bit models.)

[+mytek]

 

I programmed my first few pics just like you did, but I ended up making a cable soldered to a socket (one cable/socket each for the two pic types).

20170808_201056.jpg

 

(Michael, if you ever have the need to make changes to joy2pic down the road, it would be awesome if you could squeeze in a socket onboard with jumpers to select different pic pinouts (like the jumpers on UAV for different 8bit models.)

 

I started doing something like that, but then the XEL came along and I dropped it. Then I saw these for real cheap, and was thinking maybe getting one and copy it while also cloning in the JOY2PIC part of the circuit.

ICSP PIC ICD2 Programming Adapter PICKIT2 PICKIT3 Universal Programmer ZIF

About $9 w/shipping

 

- Michael

[+mytek]

Update U1MB with all the new versions, including the upcoming revised XEL-i3 PBI code.

Actually what I really meant to say "is update the Main BIOS with FJC's latest adjustments". The XEL-i3 PBI appears to be fine as is.

 

- Michael

Edited August 9, 2017 by mytekcontrols

[flashjazzcat]

Actually what I really meant to say "is update the Main BIOS with FJC's latest adjustments".

Still mulling over the "L" key question and optimising code, but I'll do a few real hardware tests in the next day or two and then we'll be good.

[Brentarian]

ICSP PIC ICD2 Programming Adapter PICKIT2 PICKIT3 Universal Programmer ZIF

About $9 w/shipping

 

- Michael

That is exactly what I need. Thanks!

[+Stephen]

 

Not saying that it is your power supply, but then again I'm not seeing any of the instabilities you reported with your board as you sent it to me. However it does need a 74F08 to really work properly with PMG's.

 

Any other tests you ran that I should as well?

 

- Michael

The only other test I could possible think of would be the PMG test i SysInfo, but if my little test program was working, and the PAL demos were working, I am sure it's fixed.

 

 

I was planning on doing that anyway .

 

So here's what I had in mind...

1. Install 74F08
2. Add 4.7K data line pull-ups.
3. Replace your crystal socket with one that actually fits tight.
4. Update U1MB with all the new versions, including the upcoming revised XEL-i3 PBI code.
5. Include an assembled XEL-i3 CF Adapter and one of those RGB-THRU fix-it boards.
6. Provide you with a known working PSU.

Edit: Most of the above also applies to Simon's board as well. Still waiting on some IC's to show up from back East (as in China) which are required to build a couple more XEL-i3 boards (that is why I have been holding off on returning Simon's XEL).

 

The fix-it board is probably optional if you go with a VBXE, although it does have a CSYNC buffer on it that the VBXE might need (depends upon how much the VBXE loads the circuit).

 

 

- Michael

Thanks for the extras! Let me know if you need some cash to cover the parts or labor. I know one of the resistors controlling the Mouse LEDs is wrong. I think hey both need to be 390Ohm. I have the resistors, if you could just let me know if that value is correct, I can fix that when I get the board back.

[+mytek]

The only other test I could possible think of would be the PMG test i SysInfo, but if my little test program was working, and the PAL demos were working, I am sure it's fixed.

 

Thanks for the extras! Let me know if you need some cash to cover the parts or labor. I know one of the resistors controlling the Mouse LEDs is wrong. I think hey both need to be 390Ohm. I have the resistors, if you could just let me know if that value is correct, I can fix that when I get the board back.

 

Yep 390 ohm is what they both should be... I'll change it.

 

As for paying me... not something I'm worried about to tell you the truth. Maybe I'll get you to convert a couple of U1MB's over to straight headers for me

 

BTW, I was at the post office today, and I saw that they have what they call a Medium Flat Rate Box that costs about $13 no matter what the weight. It's pretty close to the same size as the box you used, just a bit wider and not quite as tall. that'll be the one I'm going to use to send your board back to you (delivers in 3 days or less). When you use your own box, they'll charge you more every time, best to stick with their own boxes when doing Priority mailing. Something to file away in the brain for future reference.

 

- Michael

[+Stephen]

 

Yep 390 ohm is what they both should be... I'll change it.

 

As for paying me... not something I'm worried about to tell you the truth. Maybe I'll get you to convert a couple of U1MB's over to straight headers for me

 

BTW, I was at the post office today, and I saw that they have what they call a Medium Flat Rate Box that costs about $13 no matter what the weight. It's pretty close to the same size as the box you used, just a bit wider and not quite as tall. that'll be the one I'm going to use to send your board back to you (delivers in 3 days or less). When you use your own box, they'll charge you more every time, best to stick with their own boxes when doing Priority mailing. Something to file away in the brain for future reference.

 

- Michael

Great - thanks for everything. If you have any U1MBs you want converted, send them to me with the board, and I'll get them back to you. I have a great system down.

[+remowilliams]

Yeah - wonderful design. Big reason for the delay is me

 

Hey - I'm the guy who always has the edge case crap happen that people doing whatever it is haven't seen or heard of ever in their decades of experience. You trying to encroach on my turf??