U1MB, Incognito, SIDE2, SIDE3, 1088XEL/XLD Firmware 4.0 Released

[Gryfon]

On 6/6/2021 at 7:34 PM, flashjazzcat said:

I have, and unfortunately it often requires solutions such as replacing the CPU or changing out 74LS08 for a 74F08 (as Herb explains). Candle is aware of this phenomenon, but it appears that every machine which exhibits the issue is somehow situated outside of Poland. At some point, it would be useful to ship a machine which glitches with SIDE3 over to him for inspection. Unfortunately I encounter none but customer machines here which have problems, and they're usually fixed by the methods described above.

 

Quick update:  Been away from home for several weeks but now back and my 800XL has returned from Poland and investigation by Candle.  Judging from this and other posts, it looks like I also had a similar problem.

Candle managed to re-flash firmware on the SIDE3 whilst in a different machine.  Once back in my 800XL it worked again, but not with large (1MB cart) files.  He swapped the 74ls08 chip to a cd74hct08, tested it and its now behaving normally again.  I've only had a couple of hours to play with it ,but so far it's playing nice and passing all my tests.

[_The Doctor__]

cd 74hct08 almost always a good choice. decent drive and usually fast enough to get the job done

Edited October 20, 2021 by _The Doctor__

[flashjazzcat]

2 hours ago, Gryfon said:

Quick update:  Been away from home for several weeks but now back and my 800XL has returned from Poland and investigation by Candle.  Judging from this and other posts, it looks like I also had a similar problem.

Candle managed to re-flash firmware on the SIDE3 whilst in a different machine.  Once back in my 800XL it worked again, but not with large (1MB cart) files.  He swapped the 74ls08 chip to a cd74hct08, tested it and its now behaving normally again.  I've only had a couple of hours to play with it ,but so far it's playing nice and passing all my tests.

But it seems Candle learned nothing about the compatibility issue at hand, since he told me that he thought the machine itself had problems disconnected from SIDE3, and went down some different rabbit-hole.

 

This is unfortunate, since the 130XE I just upgraded with VBXE and U1MB for a customer black screened when I tested it with SIDE3 (the owner having suggested he might purchase SIDE3 at some point in the future). Replacing 74LS08 with 74F08 appears to have completely fixed the SIDE3 compatibility issue as per usual, but we are apparently no closer to understanding why.

 

Thanks for sending the machine to Candle, anyway: this is sincerely appreciated.

Edited October 20, 2021 by flashjazzcat

[_The Doctor__]

The problem would have to be timing or drive possibly timing and drive... side 3 might need to tighten it's window or move it forward or backward in time... possibly change it's trigger point levels. Who knows maybe a calibration solution might be needed where each Atari is tested with two or three presets for timing and and or three presets for level / drive... then save the best resulting calibrated adjustments.

Truly seems the best we have is to update to hct or f chips, but the monkey in the factories had fun randomizing the combinations with ls chips... so here we are.

cartridge timing has been an issue since ras and phi were changed on the 800 cart port as a fix and of course later models - pbi eci being driven by those chips. other fixes attempted to combine that window. And some cart solutions picked an abbreviated window to keep it within both ranges.

Edited October 20, 2021 by _The Doctor__

[Beeblebrox]

Having experienced issues with various A8's this year, (including my own badly unstable 65XE U1MB issue), most of which in part were cured by replacing 74LS08 with 74F08. I now have a batch of 74F08s on hand and tend to replace 74LS08 by default as soon as I open up any A8 I am working on. Thne the fun begins trying to work put any other issues.

[flashjazzcat]

Who knows indeed. The primary obstacle to finding a solution was supposedly not having hands on a machine which exhibited the common problem.

[_The Doctor__]

If I can ever get serious and have the time again, I'll see what I can find with the logic analyzer... I suspect it will still work when I dig it out. Perhaps washing years of mouse messes off of it. Though I suspect people are light years ahead with equipment and know how these days. I am sure they'll solve it all before most of us can eat breakfast.

Edited October 20, 2021 by _The Doctor__

[flashjazzcat]

You would think so.

[Level42]

The F08 did nothing to fix my, apparently, unique problem.

I have long installed back the original LS.

 

As mentioned before, no hardware add-on should require any modification to the original hardware. 

Edited October 21, 2021 by Level42

[+Stephen]

18 minutes ago, Level42 said:

As mentioned before, no hardware add-on should require any modification to the original hardware.

If the original hardware wasn't all over the place regarding timing perhaps that would be a fair statement.  These machines have always been picky with add ons - even back when ICD's MIO was new, many had to be tweaked.

[Mazzspeed]

30 minutes ago, Level42 said:

As mentioned before, no hardware add-on should require any modification to the original hardware. 

Bear in mind, you're talking about a 30yo machine interfacing with modern hardware...

 

Your particular issue is more than a little odd, I don't know if I've ever read of anyone else having such problems before? Did you end up trying the cart in another machine (I can't remember)?

[+ZuluGula]

How is your power supply and voltage under load? Some pheripherials are acting when the voltage is less than 5V. 

[Level42]

1 hour ago, Mazzspeed said:

Bear in mind, you're talking about a 30yo machine interfacing with modern hardware...

 

Your particular issue is more than a little odd, I don't know if I've ever read of anyone else having such problems before? Did you end up trying the cart in another machine (I can't remember)?

Nope, I’m the only one, but I don’t care anymore as it works just fine with my old 4GB SD card.

 

I’ve tried everything, always same result.

 

if you design modern hardware that should work with 30 (40) year old hardware, (because you sell it like it should work on it) it is -your- task as a designer to make sure it works with that old hardware…..not the other way around.

 

of course there is always the possibility of unique set-ups causing problems, but then you should be open to feedback and take the issue seriously and do some effort to solve it….

Edited October 21, 2021 by Level42

[Level42]

1 hour ago, Stephen said:

If the original hardware wasn't all over the place regarding timing perhaps that would be a fair statement.  These machines have always been picky with add ons - even back when ICD's MIO was new, many had to be tweaked.

You’re turning things exactly the other way. The add-ons should adapt to the range of original hardware, how on earth can one expect it to be the other way around ?

 

[Level42]

But……let’s get back on-topic. We’ve been through all this already. Can’t wait for the next update of FJCs firmware to be released !

[+Stephen]

1 hour ago, Level42 said:

You’re turning things exactly the other way. The add-ons should adapt to the range of original hardware, how on earth can one expect it to be the other way around ?

If that's the case, then we need to re-evaluate every single mod made excluding the IDE Plus 2 because they all exhibit random issues on random machines.  For what it's worth, I've been using this hardware since 1982, and I own at least one of every machine made and multiple copies of every single upgrade released for the past 25 years.  I must be really lucky because I have never had to modify a single machine of mine to get any upgrade to work.

 

What can we surmise from this?  I don't know.  I could say that all upgrades are perfect because "Well, they work for me".  Or we can say all upgrades suck because they don't work for someone else.

[Mazzspeed]

To put things into perspective....

 

As an IT tech, it doesn't take long to work out that HP print drivers often don't work correctly under Windows 10 for no good reason whatsoever either.

[flashjazzcat]

11 hours ago, Level42 said:

As mentioned before, no hardware add-on should require any modification to the original hardware. 

You need to travel back in time and tell that to the folks at ICD who would ask users to ship their 800XLs out to them in order to fix timing issues on a bespoke basis and make MIO work with the computer.

 

Having said that, if technicians and users alike get side-tracked by red-herring problems and/or introduce new problems owing to unknown environmental variables (PSUs, etc), there is absolutely no hope of ever getting to the bottom of the issue. I had just better stock up on more F08s and forget about it.

[Mazzspeed]

43 minutes ago, flashjazzcat said:

You need to travel back in time and tell that to the folks at ICD who would ask users to ship their 800XLs out to them in order to fix timing issues on a bespoke basis and make MIO work with the computer.

 

Having said that, if technicians and users alike get side-tracked by red-herring problems and/or introduce new problems owing to unknown environmental variables (PSUs, etc), there is absolutely no hope of ever getting to the bottom of the issue. I had just better stock up on more F08s and forget about it.

Furthermore, this wasn't limited to Atari or even the A8. Certain Amiga 1200's needed timing fixes to work with certain accelerators.

[flashjazzcat]

Exactly. No-one said it was gonna be easy.

 

[flashjazzcat]

This took a bit of trial and error for some reason, but MEM.SAV is now robust enough to context switch between BASIC, CAR and an executable like Turbo BASIC XL:

 

 

[+CharlieChaplin]

6 hours ago, Mazzspeed said:

Furthermore, this wasn't limited to Atari or even the A8. Certain Amiga 1200's needed timing fixes to work with certain accelerators.

Even some PC hardware needs to be fixed to work with WIN 11. 

[flashjazzcat]

2 minutes ago, CharlieChaplin said:

Even some PC hardware needs to be fixed to work with WIN 11.

Or rather, the installer must be fixed to work on the older PC.

[Faicuai]

On 10/21/2021 at 6:48 PM, Level42 said:

The add-ons should adapt to the range of original hardware, how on earth can one expect it to be the other way around ?

Ding, ding, ding! CORRECT answer!

 

Only that it comes with a rather significant catch: this approach would, in turn, require to deeply research and understand the host-system's bus-timing, logic, tolerances, delays, quirks, skews, etc. on its unmodified state first, so (in turn) HW upgrades model or take into account such properties during their design stage or when implementing bus-logic / control on CPLDs.

 

And that, my friend, is actually part of the crux or mystery itself.... and not limited to XL-series or U1MB, let me tell you. 

 

It may be more challenging that what today's modern-tools and prowess suggest... ??

[_The Doctor__]

with the caveat that sometimes a substitution was done by a factory or penny pinching buy of chips... slightly out of spec for the machine but good enough to get the things out the door and the order filled. Sure they worked and barely passed the basic tests, but it did cause issues for PBI/ECI/and expansion card/card edge pick offs. We aren't talking a re work of the board or architecture here. Just a replacement of an out of spec chip for our dear machine.