Is there EASY test for BAD GTIA in late model XE (China)

[Rybags]

There was one mod around (Polish?) that buffered the AN0-AN2 signals that go from Antic to GTIA. Not sure to what extent it fixed the problem though.

[+Stephen]

Thanks! Right after I posted, I found your reply on this thread: http://atariage.com/forums/topic/233858-testing-gtia/. I meant to come back here and say as much, but I forgot. So is there a common video mod that would help?

Yes - search for UAV (from member Bryan). It completely eliminates this issue, as well as giving the best video quality via luma/chroma that can be had. only way for better video is one of the RGB solutions (sopiha or VBXE2).

[santoscork]

I didn't read the entirety of this thread but there are apparently 2 types of GTIA bug/defects that are known, both of which are documented in the Source link provided under the four steps below.

 

This fix is fascinating, 4 in total. The second option is remarkable as it provides RGB out.

 

Repair 

There are four known methods to fix GTIA defects: 

1. replacement of the GTIA system with a fault-free one 
2. installing VBXE in the computer and connecting the monitor via RGB output
3. building a system compatible with Simius GTIA Fixer (it only eliminates the first type defect)
4. feeding the ANTIC pin 35 of the OSC signal from the GTIA system (foot 28) instead of FØ0 (this only eliminates the defect of the second type)

Source: http://atariki.krap.pl/index.php/GTIA, Google translated, https://translate.google.com/translate?sl=pl&tl=en&u=http://atariki.krap.pl/index.php/GTIA

 

On VBXE: http://atariki.krap.pl/index.php/VBXE, Google translated, https://atariki-krap-pl.translate.goog/index.php/VBXE?_x_tr_sl=pl&_x_tr_tl=en&_x_tr_hl=en-US&_x_tr_pto=ajax,elem&_x_tr_sch=http

 

VBXE (aka Video Board XE) summary:

In short, VBXE - a hardware expansion designed by Electron for the 8-bit Atari. It consists in the installation of a (quite sophisticated) system that, imitating the operation of the GTIA, generates an RGB video signal (with a horizontal frequency of 15 kHz, although promising work is carried out to generate a signal in the VGA standard). Contrary to the name, the system can be installed not only in XE but also XL computers .

Tidbit: VBXE emulation is built into the Altirra emulator.

 

By the way, the specs on the VBXE are ridiculous! You can find a PDF (English) on VBXE right here, https://lotharek.pl/files/vbxe/

Sold here: Lotharek https://lotharek.pl/productdetail.php?id=53 and the Brewing Academy: https://thebrewingacademy.com/products/vbxe-video-board-xe

 

This is definitely worth looking at for resolving the GTIA defect issue.

 

 

Edited July 12, 2021 by santoscork

[flashjazzcat]

Would Sophia 2 not be a better option? This would also fix the legacy video output.

 

[santoscork]

6 hours ago, flashjazzcat said:

Would Sophia 2 not be a better option? This would also fix the legacy video output.

 

I don't know if it would be better only because I don't have enough knowledge on the matter but if it is a full GTIA replacement and is compatible with the XE line then those faulty GTIAs found in the 130XE and 800XE made in China and Singapore would probably also work with a. Sophia 2board. Seems like a reasonable theory anyway. Thank you for raising the possibility. Sounds like it will actually just work too.

Edited July 12, 2021 by santoscork

[flashjazzcat]

1 minute ago, santoscork said:

I don't know if it would be better only because I don't have enough knowledge on the matter but if it is a full GTIA replacement

It is, yes.

[phaeron]

You can't rely on VBXE to completely fix this issue. It'll fix the visual glitch, but collisions are still handled by GTIA and at least some versions of this problem result in bogus playfield collisions in Graphics 10. On the other hand, Sophia 2 will fix the collision issue as well, provided that you're either using PAL or can tolerate the change in aspect ratio for NTSC.

 

[santoscork]

18 hours ago, phaeron said:

You can't rely on VBXE to completely fix this issue. It'll fix the visual glitch, but collisions are still handled by GTIA and at least some versions of this problem result in bogus playfield collisions in Graphics 10.

Why is that? I can't find anything on a bad GTIA causing collision issues in mode 10 but that doesn't mean it doesn't exist. Where is this documented? Maybe @electron can provide a word on this. I believe he created the VBXE. As far as I know the original GTIA stays in the computer and that the VBXE solves all the problems the bad GTIA exhibits. Anyone who read the statement I just made, don't take it as gospel. I know very little and only learned of this bad GTIA China XE machines.  

 

Here is an excerpt (Google translated) for the link below as I tried to find something definitive on the collision issue you reported as they relate to XEs made in China and Singapore but no collision issues are reported. https://translate.google.com/translate?sl=pl&tl=en&u=http://atariki.krap.pl/index.php/GTIA. Everything else is just info that doesn't really mean too much I suppose.

 

Some GTIA chips are factory defective, with the main symptoms being that the multi-color modes are displayed incorrectly. There are at least two types of these drawbacks, in the first (and worst) case the multicolor modes work completely badly, in the second - the irregularities are of a slightly different kind and show themselves only under certain conditions. Generally, this is a problem eminently related to the timing within the GTIA, in the circuit that receives data from ANTIC's ANx bus.

 

This is a BASIC applet (by MacGyver ), thanks to which you can easily determine if your computer has a defective GTIA:

10 GRAPHICS 9 
20 FOR I = 0 TO 15: COLOR I 
30 FOR J = 0 TO 4 
40 PLOT I * 5 + J, 0: DRAWTO I * 5 + J, 191 
50 NEXT J 
60 NEXT I 
70 GOTO 70

 

A simple test for a type two defect

With the following program (written in Atari BASIC with fragments of machine code) we can check if the GTIA has a second type defect:

10 REM - GTIA LINE SPLIT TEST - 

11 REM ** done by Seban / Slight ** 

12 REM 

13 FOR I = 1536 TO 1556: READ A: POKE I, A: NEXT I: GRAPHICS 9 

14 FOR X = 0 TO 31: COLOR X / 2: PLOT X + 48.1: DRAWTO X + 48.96: NEXT X 

15 FOR Y = 96 TO 191: COLOR Y-96: PLOT 48, Y: DRAWTO 78, Y: NEXT Y 

16 POKE 87.8: COLOR 1: POKE 710.0: POKE 709.15 

17 FOR X = 0 TO 157 STEP 2: PLOT 79.0: DRAWTO X, 191: NEXT X 

18 POKE 623.0: X = USR (1536)

19 DATE 169,0,141,10,212,141,27,208,162,6,202,208,253,169.64,141.27.208.76.0.6

 

 

Repair

There are four known methods to fix GTIA defects:

1. replacement of the GTIA system with a fault-free one
2. installing VBXE in the computer and connecting the monitor via RGB output
3. building a system compatible with Simius GTIA Fixer (it only eliminates the first type defect)
4. feeding the ANTIC pin 35 of the OSC signal from the GTIA system (foot 28) instead of FØ0 (this only eliminates the defect of the second type)

 

 

 

At the last minute I did find this comment in a Sophia 2 thread, take a look at the excerpt.

 

 

 

For posterity, I found another fix to the bad GTIA but even that page doesn't describe collision issues.

 

https://atari-owner.com/club/articles/fix-a-800xe-65xe-130xe-with-bad-gtia-chip.57/

 

Image rendered by a bad GTIA on an ATARI 130XE.

 

The obvious fix is to replace the bad GTIA (CO14889) with a good one.

Here is a hack by adding a 74LS175 Chip to fix the Bad GTIA (This fix was originally designed my Simius from Poland)

 

1)

Cut the traces that connects Antic Pins 2,3,5 TO GTIA Pins 18,29,20 (see image below)
 

 

2)

• Piggy back the 74LS175 on top of the connecting only Pin 8 (GND) and pin 16 (vcc) Legs
• Connect Pin1 of 74LS175 to pin16 of 74LS175
• PIN 28 OF GTIA (CO14889) to pin 9 of 74LS175
• Pin 18 of GTIA (CO14889) to pin 10 of 74LS175
• Pin 19 of GTIA (CO14889) to pin 15 of 74LS175
• Pin 20 of GTIA (CO14889) to pin 6 of 74LS175
• Pin 2 of Antic (C021698) to pin 12 of 74LS175
• Pin 3 of Antic (C021698) to pin 13 of 74LS175
• Pin 5 of Antic (C021698) to pin 5 of 74LS175

 

3) The Result

 

Again for posterity only, see link for XE Field Service and search "Bad GTIA" in browser.

http://www.jsobola.atari8.info/derelit/xefsm.pdf 

 

Edited July 13, 2021 by santoscork

[hueyjones70]

Will the 14805 replace the 14889?

[Peri Noid]

Just one remark: it may happen, that when you put a faulty GTIA in another machine, it will work correctly. I had duch an issue - a colleegue had an 800 with a GTIA showing faulty behavior. I took it out and placed in my 800XL - and it was fine there. On the other hand side, I had several XE-originated GTIAs, which were also broken in an 800XL. So it may be worth checking.