SIDE3 "Birdies" with 1200XL

[+DrVenkman]

So here's a weird thing I noticed a couple days ago and decided to record tonight:

 

 

Note that the weird graphical glitches ("birdies") are only on the second horizontal line of "a" characters, and are not consistent - they jump around from "a" to "a" on the screen. Some "a" characters exhibit this from moment to moment, some don't, and which ones are which cycle around. If I remove the SIDE3 cart, the glitches are gone completely.

 

For background, this is a 1200XL that I rescued about 4 years ago, then stuck an Ultimate 1MB into it, and then a UAV. Along with my SIDE2 cart, it has been rock solid and completely stable.  After getting my SIDE3 cart last fall, I started seeing some weirdness and instability. I replaced the stock vintage 74LS08 chip in the Phi2 clock circuit with a modern, faster 74F08, and added a small ceramic cap on the U1MB itself between the RESET and GND pins, per @tf_hh's posts elsewhere. The instability went away and the system returned to its reliable self.

 

In recent months, the 1200XL hasn't been used much - I've been building, testing and playing around with my 576NUC+ beta board, plus my Incognito 800 for the most part. So this week I returned the 1200XL to the rotation and noticed these weird graphical glitches. What's potentially telling is that I left the system running for a couple hours while I did my usual weekly virtual Happy Hour with some widely-scattered friends over Zoom, and when I checked the system again after a couple of pints, the birdies/glitches are completely gone. No sign of them at all.

My first thought is I might have an ANTIC that is starting to go bad, but works properly as the system warms up over time - the glitches appear to be related to one specific bit or something, for only one character (the lower-case "a"). However, as I noted above, if the SIDE3 is removed and the system boots to SDX without it, there are no glitches anymore. 

 

So ... is this an issue with my SIDE3 cart that goes away as it warms up? Is it related to my ANTIC or maybe my CPU when cold? Is it a power draw issue related to run time? I'm running the system off a vintage Atari 9VAC brick, though the two 7805's in the 1200XL are new as of 4 years ago, but the capacitors are vintage as well.

 

I welcome your thoughts and suggestions. 

[+bob1200xl]

Ya... I'm pretty sure that is the character set ROM falling off the buss too soon, leaving you with a $FF latched in ANTIC for that row of the character. I used to see those when I was fooling around with 14mhz upgrades. You want memory to drive the buss longer, not shorter, so high speed ICs may not be the answer. If I remember, the character that failed was the space character - get lines buzzing about in the empty regions of the display.

 

Just try the usual suspects...

 

Bob

 

[_The Doctor__]

Try an hct series chip instead of an f series, they are usually in between with good drive, also consider sometimes the cpu is marginal.

Edited July 31, 2021 by _The Doctor__

[Rybags]

I used to get that on my 800XL with the 32in1 and VBXE I think was the config at the time.

There's various fixes, from memory sometimes a different CPU will cure it.

[+DrVenkman]

16 hours ago, Rybags said:

I used to get that on my 800XL with the 32in1 and VBXE I think was the config at the time.

There's various fixes, from memory sometimes a different CPU will cure it.

Well, I swapped the SALLY that was in the machine (an NCR chip from 1984 pulled from a parts 130XE) with an MOS made chip from 1982 (pulled from a parts 5200), but the problem persists. So next I swapped the 74F08 with the original 74LS08. Now the "a" characters are fine, but I see the same effect with "(" characters. And as before, after an hour or so, the problem goes away. In all cases, if I remove the SIDE3, the problem disappears. 

 

Anyone else have any suggestions? I don't think I have any working spare ANTIC chips to try without pulling one from another working system. Could it be a power issue? 

[Mazzspeed]

When the problem disappears, you could spray some form of cool spray (usually available at electronics suppliers) on each chip in turn and see if the problem returns.

[_The Doctor__]

74HCT08