Is my 130XE OK (picture inside)

[Sdw]

So I bought myself a little christmas present - a 130XE! So at age 32 I own my first Atari computer ever (about time, eh?)

 

Now I've read quite a bit about that dreaded GTIA-bug, but as I understood it, it mostly concerned the 65XE model.

Anyway, I typed in the little BASIC test-code I found here:

 

(Polish Atari site I think, auto-translated to English)

 

My computer displays all different gradients correctly, however there are small (can't tell if it is exactly 1 pixel or what) dark regions between some of the gradients. Here's a photo of the screen:

 

 

However if I change the code to show the gradients as long horizontal bars instead, the artifact is not present.

Am I right in assuming this is some kind of TV-intereference or something, and not a faulty GTIA since it only appears between vertical bars?

[Rybags]

Looks normal - you get that kinda crap with RF or composite. Practically any machine should generate something like that on a TV.

 

That screen is good for adjusting contrast/brightness. Good way to judge is to get the lowest luma value distinguishable from the background.

 

Problem is, you'll probably find it's different from a comfortable setting for normal text viewing.

Edited December 19, 2008 by Rybags

[ClausB]

Notice that the worst artifact is between luma values 7 and 8, where all 4 luma bits change. Apparently the lower bits change faster than the upper bits, so the luma goes darker before it gets lighter. The effect is less noticable between lumas 3 and 4 and between 11 and 12, where only 3 bits change. To test this explanation, plot the bars in reverse order and see if the artifacts are light instead of dark.

[Sdw]

Notice that the worst artifact is between luma values 7 and 8, where all 4 luma bits change. Apparently the lower bits change faster than the upper bits, so the luma goes darker before it gets lighter. The effect is less noticable between lumas 3 and 4 and between 11 and 12, where only 3 bits change. To test this explanation, plot the bars in reverse order and see if the artifacts are light instead of dark.

 

That's an interesting theory! I did the test and here is the result:

 

 

Looks like the artifacts are still dark. However you are definitely onto something with the bit values, because even in this direction, the worst artifact is in the 7->8 transition.

Also note that the other interference-looking stuff is coming from taking the picture. Other than the darker lines, the image looks good in real-life.

[ClausB]

So much for that theory!

 

How'bout this one: luma bits turn off faster than they turn on? The GTIA luma outputs appear to be open-collector because they have pull-up resistors, so this theory makes sense.

Edited December 20, 2008 by ClausB

[Sdw]

So much for that theory!

 

How'bout this one: luma bits turn off faster than they turn on? The GTIA luma outputs appear to be open-collector because they have pull-up resistors, so this theory makes sense.

 

I'm not that knowledgeable about hardware, so what you said about open-collector and pull-up was a bit beyond me, but something is strange, that's for sure!

I just got my SIO2PC so I've been able to run some demos/games/programs now, and this artifact is visible and rather annoying in many things that use smooth gradients. Has anyone else seen anything like this on their machine, or is it me that has ended up with a faulty HW?

[charliecron]

I've seen that behavior on my 1200xl composite video, and was very thrilled that it was gone after I did the video upgrade. I remember this demo in particular, "EXE" showed the problem, a single dark line messing it up, really annoying.

 

http://atari.fandal.cz/detail.php?files_id=325

[ClausB]

Has anyone else seen anything like this on their machine, or is it me that has ended up with a faulty HW?

I've seen it on my Ataris. I think it's normal.

[Rybags]

Normal, yes. Desirable, not really.

 

Pretty sure all my machines have had the same thing - notice it also occurs between luma value 11 and 12 although not as pronounced.

[ClausB]

Normal, yes. Desirable, not really.

Back in the day, we were not spoiled by crisp VGA displays, so it was a small minus next to a huge list of Atari plusses.

[Sdw]

I've seen that behavior on my 1200xl composite video, and was very thrilled that it was gone after I did the video upgrade. I remember this demo in particular, "EXE" showed the problem, a single dark line messing it up, really annoying.

 

http://atari.fandal.cz/detail.php?files_id=325

 

Yes! I noticed it in that demo too. Also in for examle this:

 

http://atari.fandal.cz/detail.php?files_id=5198

and this:

http://atari.fandal.cz/detail.php?files_id=5545

it was very visible. Well, I guess it will be extra visible in any demo that uses those GTIA 16-gradient modes, as you often get the luma 7->8 transition there.

 

Anyway, thanks to all for the info, good to know at least that it is a common problem and not a hardware failure.

[Rybags]

Would something like a really small capacitor value on the luma output help with this? Or maybe just do nothing more than make for a blurred display?

[ClausB]

Would something like a really small capacitor value on the luma output help with this?

Or maybe stronger (smaller) pull-up resistors on the GTIA luma pins? Something for the video mod designers among us to ponder.

[_The Doctor__]

I've seen that behavior on my 1200xl composite video, and was very thrilled that it was gone after I did the video upgrade. I remember this demo in particular, "EXE" showed the problem, a single dark line messing it up, really annoying.

 

http://atari.fandal.cz/detail.php?files_id=325

 

what was the upgrade that got rid of it?

[charliecron]

what was the upgrade that got rid of it?

 

Oh it was the Clearpic 2002 for the 1200xl.

 

Charlie

[+CharlieChaplin]

Well,

since some of the mentioned demos are Gr.9 or GTIA demos, it might be you have a faulty GTIA chip. These faulty GTIA chips are very common in XE computers that were built in China. The Abbuc forum also shows some nice pics, even if you cannot read/understand german language scroll down to the pics - pic 1, 3 and 5 shows correct GTIA (how it should be) and pic 2, 4 and 6 shows the faulty GTIA...

 

http://www.abbuc.de/modules.php?name=Forum...opic&t=4275

 

greetings, Andreas Koch.

Edited January 14, 2009 by CharlieChaplin

[Sdw]

Well,

since some of the mentioned demos are Gr.9 or GTIA demos, it might be you have a faulty GTIA chip. These faulty GTIA chips are very common in XE computers that were built in China. The Abbuc forum also shows some nice pics, even if you cannot read/understand german language scroll down to the pics - pic 1, 3 and 5 shows correct GTIA (how it should be) and pic 2, 4 and 6 shows the faulty GTIA...

 

http://www.abbuc.de/modules.php?name=Forum...opic&t=4275

 

greetings, Andreas Koch.

 

Looking at those pictures it seems that a faulty GTIA is much more severe than what I have.

Actually, if you look closely at the first "good-GTIA" picture of Demo 1 in that forum thread, you can see this luma-line in the middle there as well, looking very similar to mine so it must be a common issue.

[Rybags]

Don't the faulty GTIAs not even represent Gr. 9 properly?

 

And, practically everyone else has reported the same phenonema. In fact the problem likely lies outisde the GTIA within the video circuitry (which we know is flawed on practically all machines).