VBXE and dimming screen

[damanloox]

So.. my fully loaded 600XL saga continues (unfortunately). After "fixing" jailbars on RGB output I've noticed another rather worrying issue. The screen seems to be "dimming" randomly (ie. it can happen 5 seconds after boot or it can be OK for 5 minutes and then it starts doing this strange thing). Here's a short video (it doesn't really show the issue as it's rather low quality): https://photos.app.goo.gl/dxkF2muUQm4Ao6rf7  

It seems like VBXE "turns off" the signal for a split second - both video and audio. Sometimes several times in a row in 2-3 seconds intervals, sometimes one or two times. Then it works OK for a few minutes and then starts again. I'm not sure if it's the TV (I only have one that has RGB input) but the same TV works perfectly when using s-video or composite signals. It could be just RGB/Scart input doing this (and not Atari/VBXE) - not sure. 

Has anybody ever noticed something like that with VBXE? Any ideas what can be wrong?

 

[flashjazzcat]

This the sync drifting out (of spec re: the sync voltage the monitor expects). I saw the exact same thing on my first VBXE machine a decade ago (a 130XE) and it was distracting as Hell. Only happened with the LG M227WD monitor, though; others had no issue with it. At the time, I used a variable resistor to attenuate the sync signal coming off the CD4050 a bit, but - since taking sync from composite is off the table in your situation - changing CD4050 is worth a shot as well (this makes sense, since not every VBXE machine here which takes sync from CD4050 pin 15 causes the LG monitor a problem). Failing that, try a variable resistor.

Edited May 12, 2021 by flashjazzcat

[tuf]

I've got some Sony LCD monitors that do that and there's no way to disable them.  Some sort of low light circuitry.  Could it be your monitor trying to do some ambient light adjustment?

 

[damanloox]

@flashjazzcat thanks, I'll try changing CD4050. I don't like the idea of variable resistor for some reason  But yes - it's the same monitor as yours (19 inch variant but everything else is exactly the same - m197wd). Worst case scenario - I'll use OSSC or something like that with mo other monitors (unfortunately that LG is the only RGB that I have). And yes - SYNC off composite is out of the question - I didn't pay for VBXE to have jailbars ;)). And - while on the subject - do you have any idea why SYNC/COMPOSITE generates jailbars (as opposed to "clean" sync from CD4050)? Perhaps there's a way to prevent that...?

 

@tuf - I don't think so. I don't think it's as "advanced"  There's also nothing in the menus to suggest it could be that...

 

But - why is it losing audio as well?

[flashjazzcat]

21 hours ago, damanloox said:

But yes - it's the same monitor as yours (19 inch variant but everything else is exactly the same - m197wd).

So we can immediately discount all other theories.

21 hours ago, damanloox said:

I don't like the idea of variable resistor for some reason

You can use it to find the optimial resistance, though, and then simply replace it with an inline resistor (or - I suppose - preferably a voltage divider) on the sync line.

21 hours ago, damanloox said:

And - while on the subject - do you have any idea why SYNC/COMPOSITE generates jailbars (as opposed to "clean" sync from CD4050)? Perhaps there's a way to prevent that...?

I don't, but I assume it is because of extraneous information in the signal (after all, it is carrying luma and colour data as well as sync).

21 hours ago, damanloox said:

But - why is it losing audio as well?

Because the monitor thinks the input source is going completely offline when it drifts out of spec (albeit momentarily).

Edited May 13, 2021 by flashjazzcat

[damanloox]

OK, quick update... 

Both ebay and radionics failed to deliver the chip in almost a week but.. I'm happy to report that 330 Ohm resistor on CSYNC fixed the problem and now I have uninterrupted clean RGB

Thanks!

[_The Doctor__]

sounds like the resistor fix should be a part of the VBXE kit going forward

[flashjazzcat]

10 minutes ago, _The Doctor__ said:

sounds like the resistor fix should be a part of the VBXE kit going forward

It's only required in a handful of cases.

[_The Doctor__]

even more of a reason to put it in the documents etc. these sort of issues get forgotten and buried and have to be answered time and time again over the years... Just thankful you were here to answer it. Hopefully you will live long and prosper... but what if there were no FJC at the time another needed the solution. Still good to document... maybe some experiments to see if it would adversely affect other monitors having it there, or actually be good across the board.

 

Edited July 10, 2021 by _The Doctor__

[flashjazzcat]

Oh - I have no objection to 'putting it in the documents'; you seemed to advocate providing a resistor of some aribtrary value in the box with the VBXE. This would be difficult, given that the resistance value required appears to be completely indeterminate. Of course there would be no harm in Lotharek at least alluding to the fix in the installation instructions, the extreme rarity of any issue at all notwithstanding.

 

In the meantime, even if FJC died tomorrow, we fortunately have the self-documenting phenomenon of forum threads. Googling 'Atari VBXE dimming' brings me directly to this topic.

 

[damanloox]

I've tested this issue (just thought I need to get some value form my oscilloscope from time to time ;)) and in three cases (ie. three Ataris I used for testing) 4050 generates TTL level CSYNC. That means we need 470 Ohm resistor to provide "SCART" level output (ie. 75 Ohm). That's what I'm doing on my new "daughter board" and it works perfectly fine.

"Resistor kit" for VBXE is not really necessary as most people will take CSYNC from composite video anyway (which is already "SCART" compliant). The only "issue" is with CSYNC coming directly from 4050 because it's TTL level (and even though some TVs can deal with that I'd still recommend putting 470 Ohm resistor).

I BTW - I sincerely hope FJC doesn't die for at least next several decades and outlives us all  

Edited July 10, 2021 by damanloox

[_The Doctor__]

hmmm AtariAge forever? too many dead links out there for me to count that I relied on thinking they'd never go away... not even a wayback machine crawl to pull them from. I hate to admit it... but the best we have is paper and a multitude of users keeping digital copies of everything. Putting it in the docs is certainly a great way to go. who knows what a goggles search will yield in 8 years...

I'm starting to think google might me renamed blinders... it seems to prefer keeping us in it's own little bubble idea of what we want to see and what it wants to sell us. Goggles or blinders seems fitting.

[_The Doctor__]

heh, FJC will live forever... he is powered by Jazz, and energy derived from purring cats... since there is no shortage of either... we will have Jon, a cat, a roach, and a bunch of rocks sitting at the end of the universe. Perhaps there will be a restaurant there as well.

[flashjazzcat]

For a new user to have access to someone's local copy of the revised documentation, there would have to be some centralised web resource (which, it seems, we have decided is unreliable), or some surviving member of the community in possession of said document or said knowledge. It appears we are covered for the moment, then.

51 minutes ago, damanloox said:

The only "issue" is with CSYNC coming directly from 4050 because it's TTL level (and even though some TVs can deal with that I'd still recommend putting 470 Ohm resistor).

That's a good point which was well worth clarifying.

51 minutes ago, damanloox said:

I BTW - I sincerely hope FJC doesn't die for at least next several decades and outlives us all

Eighteen months ago, I would have laughed at the very notion of outliving anyone, but as time goes on...