Jump to content
Faicuai

SOLVED: Vertical Banding, Atari 800XL (<=RevC) @ sVideo.

Recommended Posts

Not long ago, and after performing some specific changes and tests to enable sVideo on my primary 800XL unit, I did notice something unusual going on with the Luma signal, when displaying full-frames with single color (at higher luminance levels):

[url="http://www.atariage.com/forums/topic/155245-vertical-lines-on-many-xe-computers/page__st__125#entry2366585"]http://www.atariage....25#entry2366585[/url]

Past that point, I conducted a few more tests, and did notice that (in the above conditions) the Luma signal should appear COMPLETELY clean and free of any disturbances, even if chroma-info was being generated.

Now that I got my twin / back-up 800XL unit (HongKong, RevC, Alps Keyboard, fully-socketed, [i]thick[/i]-plate shields, and ~7,000 units apart from my primary 1984 unit), I decided to fully trace down this rather elusive issue, and look for a SIMPLE and IMMEDIATE solution to it, to be implemented on a rather STOCK configuration.

Here is a sample of the results which partially disconnecting ONE (1) capacitor (C[i][b]54[/b][/i]) on the MoBo (absolutely nothing else):

The before (C54 in place):

[attachment=243849:_1D38222-web.jpg]


The after (lifting RIGHT-side of C54, nothing else):

[attachment=243850:_1D38223-web.jpg]
(Note: right-side of the image shows aliasing of LCD pixels and capture's camera sensor... on-screen image is CLEAN)


Please, take note that CONTRARY to well-known video mods out-there, the OPTIMAL reduction of vertical banding (on s-Video-enabled output) is achieved by de-coupling the [i]COMPOSITE CHROMINANCE [/i]signal at its exact Luma+Chroma juncture (RIGHT-SIDE end of C54, where final COMPOSITE VIDEO signal is generated), instead of decoupling the [i]COMPOSITE LUMA [/i]component (R56-end) at that same joint. I dertermined this, because I temporarely disabled ALL chroma-signal paths (there are TWO), as well as LUMA signal paths (there are THREE) subject to further combination, and then re-enabled one-by-one, in all possible / relevant combinations.

The results, on a stock <=RevC machine, boils down to ONE (1) culprit: [b][i]chroma signal output at end of C54, being combined with its Luma counterpart, in order to generate Composite Signal[/i][/b].

In other words: the 800XL video output circuitry IS NOT EXACTLY designed with high-quality composite-output signals in mind. It is, INSTEAD, much better suited (believe it or not) for {Y,C} signal / svideo output.

I will later post (on a separate topic) what, in summary, is truly needed to obtain a relatively high-quality output sVideo signal with virtually ZERO mods to the motherboard, other than soldering three wires and installing high-quality toggle switch, all which cost $5 or less.

Enjoy!

Share this post


Link to post
Share on other sites
you are a statisticians dream? nightmare?

a sample set of one... want me to send you a 130XE, and you can show me whats wrong with it?

just asking...
sloopy.

Share this post


Link to post
Share on other sites
Hi,

hmm.... lifting the right side of C54 is explained in nearly all supervideo or ultravideo mods. So what´s new on this post?

Sorry, should be NO affront, I want to understand it only...

Share this post


Link to post
Share on other sites
[quote name='tf_hh' timestamp='1336725050' post='2520014']
(...)

hmm.... lifting the right side of C54 is explained in nearly all supervideo or ultravideo mods

(...)

[/quote]

Feel free to post the [b](exact) link[/b] with the[b] (exact) explanation[/b]. That would certainly contribute with everyone's perspective...

Unfortunately, all I cound find during the last two (18) months were mass-murder removal/ butchering of resistors, extirpation of RF-modulators, piggy-backing of foreign resistors, and butchering (in general) of the video-circuit... Maybe you are certainly better at understanding those combined recipes, though.

Share this post


Link to post
Share on other sites
I was under the impression the A8 video circuit was butchered at the design stage (or such is the assertion in several respected mod-guides). Nevertheless, if all existing mods are so ill-conceived and destructive (assuming your search was more than cursory), thank God you're here! :)

Share this post


Link to post
Share on other sites
[quote name='Faicuai' timestamp='1336743661' post='2520102']
Feel free to post the [b](exact) link[/b] with the[b] (exact) explanation[/b]. That would certainly contribute with everyone's perspective...
[/quote]

[url="http://ftp.pigwa.net/stuff/collections/nir_dary_cds/Hardware%2520Projects/Super%2520Video%25202.1XL/supervid5.html"]http://ftp.pigwa.net.../supervid5.html[/url] (Version describes using "add-on" resistor, not very cute, but it works. Also named CV switch)

[url="http://www.wolfpup.net/atarimods/supervid.html"]http://www.wolfpup.n...s/supervid.html[/url] (Another one using replacement resistors and explaining the need for CV switch = lifting C54)

[url="http://www.atari8.co.uk/downloads/UltraVideo%20XL%201.0.pdf"]http://www.atari8.co...ideo XL 1.0.pdf[/url] (Extreme version removing all parts for CVBS, only leaving Y/C S-video signals and changes only the wrong coupling for monochrome transistor)

[url="http://www.atari8.co.uk/downloads/UltraVideo%20XE%201.0.pdf"]http://www.atari8.co...ideo XE 1.0.pdf[/url] (Same for XE models)

These are the main articles, there are a lot of more, sometimes translated, repainted and so on to other languages.

Personally I use the SuperVideo 2.1 XL mod with new resistors, but I didn´t change all parts, because after modding more then 50 systems I found not all parts are needed. For the XL I remove L5 by wire (shifting up voltage to ~5v instead of 4.3v for video circuit), R116 to 8.2 ohms, R56 & R53 to 75 ohms, R52 to 2.7k or 3.0k. Of course removing C56 if present.

If you want to preserve full usage of modulator, CVBS signal AND s-video, I use a shottky diode instead of a switch. Lift the right side of R56 and solder a diode (it must be shottly!) in serial. Cathode shows to the right, lifted pin. Now the monochrome signal is not disturbed by color informations, Y/C output is clear but CVBS and modulator also works fine. Just try it.

After modding a lot of systems on local retro events I found out, that only inserting the switch/diode or so isn´t always enough. Some older TFT displays or also tube monitors have a bad picture, no sharpness. So it´s not ok (IMHO) so say that all mods are butchering - it depends on the display and ALSO on the parts on the XL mainboard. Sometimes it´s better to change the CMOS 4050 to a 74HCT4050 - picture gets extremly better. On other systems and monitors it doesn´t...

And, last: On NTSC the world might be different to PAL :-)

Share this post


Link to post
Share on other sites
[quote name='tf_hh' timestamp='1336747767' post='2520139']
(...)

[url="http://ftp.pigwa.net/stuff/collections/nir_dary_cds/Hardware%2520Projects/Super%2520Video%25202.1XL/supervid5.html"]http://ftp.pigwa.net.../supervid5.html[/url] (Version describes using "add-on" resistor, not very cute, but it works. Also named CV switch)

[url="http://www.wolfpup.net/atarimods/supervid.html"]http://www.wolfpup.n...s/supervid.html[/url] (Another one using replacement resistors and explaining the need for CV switch = lifting C54)

[url="http://www.atari8.co...ideo"]http://www.atari8.co...ideo[/url] XL 1.0.pdf (Extreme version removing all parts for CVBS, only leaving Y/C S-video signals and changes only the wrong coupling for monochrome transistor)

[url="http://www.atari8.co...ideo"]http://www.atari8.co...ideo[/url] XE 1.0.pdf (Same for XE models)

(...)

[/quote]

THANKS!

The list of resources you list (above, which I had already read them) precisely make my point: they add to confusion, motivate trial-and-error, bring out "SuperVideo mod gone bad" posts, divergence between what a circuit diagram shows vs. what's really on the MoBo, further desoldering, cutting, piggy-backing, mass-murdering of resistors and passive components, etc., etc.

And contrary to what you write, in NONE of those links you will find direct, clear answer to this problem (vertical banding on Y/C output), and provided switch-installation directions [b][i]DO NOT [/i][/b]bring the best results out of the 800XL (at least on my two main units, and not in my Scope).

The important thing is that if you are happy with it (or it meets your objectives), then [b]PERFECT[/b].

As for me, most of such mods are [i]redundant[/i], [i]unnecessary[/i], and not properly backed-up with visual (step-by-step) evidence of the improvements... You end up NOT KNOWING exactly, what happened, and if you really had to perform them (in first place).

Below is a sample of my primary 800XL unit, with [i]ABSOLUTELY [/i]everything in place, all video-outputs fully operational, clean MoBo, and pretty clean sVideo output (without all the butchering and vertical banding), no pick-up resistor on Chroma line, no separate power-line & filtering, plus Ultra1MB installed, AND shields installed back:

[attachment=244123:_1D38184-web.jpg]

[attachment=244124:_1D38186-web.jpg]

Share this post


Link to post
Share on other sites

Sorry to revive an old thread, but can anyone suggest the appropriate capacitor on an Atari 800 that would correspond to this modification? I have exactly the same problem.

 

Thanks

Share this post


Link to post
Share on other sites

Sorry to revive an old thread, but can anyone suggest the appropriate capacitor on an Atari 800 that would correspond to this modification? I have exactly the same problem.

 

Thanks

 

Assuming your unit is NTSC, MoBo <= RevC (and you have CHROMA signal enabled on video port), please, locate C54 (near-by the RF box) and unsolder one of its legs (DO NOT over-bend, break or cut).

 

You should see your vertical banding substantially reduced or eliminated, altogether. For a more permanent solution, a small switch should be installed in-between one leg of C54 and its contact on the Motherboard, for complete and voluntary isolation of chroma path while using Y/C output (svideo).

 

Please, see above images of one of my XLs (all of them have this mod, and all yield on the results shown above).

Share this post


Link to post
Share on other sites

Hi,

 

Is C54 in the Atari 800 (not XL or XE) the same as in the XL? After looking through the 800 schematics, I'm not sure there are any capacitors numbered below 100. Thanks

Share this post


Link to post
Share on other sites

Hi,

 

Is C54 in the Atari 800 (not XL or XE) the same as in the XL?

 

Duh!

 

My apologies (just flew-by your post).

 

With respect to the older 800, I tried to address this problem on mine during Incognito install (with Service Manual complete schematics). The bottom-line is that it requires significantly more analysis (and maybe work) because I went right after the key Y+C combination points, and could not reduce vertical banding.

 

The good thing is that my 800's Y/C output looks GORGEOUS on my Viewsonic 930B + iScan DVDO HD processor (the latter eliminates vertical bands either at the A-to-D front-end decoding-stage or in the digital domain itself).

Share this post


Link to post
Share on other sites

Have you tried the Y/C (S-Video) signal from the 800 monitor port?

 

The 800 model has it built in and all the models ive seen have given a great clear picture, the only model version/variant that does im aware off.

Edited by Magic Knight

Share this post


Link to post
Share on other sites

Yes, this is the port from which I am getting the vertical banding. I do not use composite video. I have noted that I also get the same effect with the 7800 and 130XE with their S-video. The kjmann mods describe something that sounds like it could be the issue, but I haven't seen a picture of what he is describing.

 

"To Make the Chroma Work With Newer TV's (Fix The Vertical Banding issue) is as simple as adding one side of a 75ohm Resistor to the IN Leg (Side Connected to Pin 21 of GTIA) of the 22uf Cap on the Chroma line and the other end of the 75ohm resistor to Ground on the Mainboard. This will Get rid of the vertical Banding (On Most TV's) Which is Caused By DC Voltage in the Signal and an Overly Strong Chroma Signal."

 

The photos at the top of this thread are exactly what I am getting, but for either modification, I don't know the corresponding pin(s) on the 800 (or 7800 for that matter). I'm hesitant to dismantle the 130XE first since the MyIDE installation is a bit delicate.

 

Thanks

Share this post


Link to post
Share on other sites

Yes, this is the port from which I am getting the vertical banding. I do not use composite video. I have noted that I also get the same effect with the 7800 and 130XE with their S-video. The kjmann mods describe something that sounds like it could be the issue, but I haven't seen a picture of what he is describing.

 

"To Make the Chroma Work With Newer TV's (Fix The Vertical Banding issue) is as simple as adding one side of a 75ohm Resistor to the IN Leg (Side Connected to Pin 21 of GTIA) of the 22uf Cap on the Chroma line and the other end of the 75ohm resistor to Ground on the Mainboard. This will Get rid of the vertical Banding (On Most TV's) Which is Caused By DC Voltage in the Signal and an Overly Strong Chroma Signal."

 

The photos at the top of this thread are exactly what I am getting, but for either modification, I don't know the corresponding pin(s) on the 800 (or 7800 for that matter). I'm hesitant to dismantle the 130XE first since the MyIDE installation is a bit delicate.

 

Thanks

 

Try a 47 or 51 ohm resistor instead of the 75 ohm one on the luma side, this may help a slight increase from 75 ohm may alter this the other way too,, The 22uF cap should be ok for the chroma direct from pin 21 (not likely a culprit for vertical banding now! (this is for xl/xe models) -

 

On 800 models - you may need a resister ladder for the multiple Lum's comming from the Gtia chip if you are doing the more direct method on this (like the Atari 400 requires). I havent performed this so far as the 800 for me has no issues from the monitor output.

 

Im familiar with the 7800 internals - what are you using to get s-video on this? (slightly off topic, but relevant to the above)

Edited by Magic Knight

Share this post


Link to post
Share on other sites

I can't seem to find the 22uF cap off of pin 21 where I am supposed to attach a resistor, so I'm not sure if the mod applies to the 800.

 

I did try putting a 75ohm resistor between pin 21 and ground, with no effect. I wish I knew which capacitor in the 800 corresponds to C54 from the initial post, since there is evidence that the mod works on at least the poster's model.

 

Since I'm getting exactly the same effect on three different systems, I am hoping that a similar mod will resolve them. I would imagine that these systems all have some design elements in common, since issue is described on 400, 800, 800XL, 130XE, 7800, and 5200 at least. When I do a google image search for the terms "vertical banding", guess what comes up? A bunch of blue Atari 8-bit screens.

 

The 7800 has an 8-bit domain S-Video board. I did find some posts suggesting that the rf modulator be removed on a 7800 or 5200 - don't know if that would help some or all of these systems.

 

Thanks!

Share this post


Link to post
Share on other sites

The mod with the 75 ohm resistor and the capacitor wouldnt work on a 800 without extra bits (like the resistor ladder likely) - The XL/XE circuitry has a built in amplification circuit which takes care of the luma strength value and so little is needed when you know where to tap the points from.

 

The C54 capacitor is from the 800XL mainboard and is a cut which de-connects the composite circuit where the y/c are joined - this helps as composite can produce various results and can also affect s-video when the composite interacts with the signals when connected (pal 600xl's improve when some of the composite resisters are removed in order to get better s-video)

 

The 800xl board has little resmeblance from the 800 as it is very different in nature - I wouldnt go hunting for this particular component or function as you may end up breaking the wrong part off and in addition the ladder is likely to be needed.

 

A 75 ohm resitor to ground is cited in some schematics ive seen, but ive seen no real effect with this (see a previous thread covering this all in some detail and someone else who has issues - http://atariage.com/forums/topic/209136-kjmanns-s-video-upgrade-revisited-atari-400/page-3?hl=+revisited )

 

 

Also can you post a screen shot of the 800 screen you get? - and also the 130xe so i can see what extent you are getting.

Edited by Magic Knight

Share this post


Link to post
Share on other sites

If you look at the vertical banding in the initial post in this thread, it is exactly what I am getting on these systems, like this from a section of the Ballblazer screen.

post-8517-0-35071900-1378842412_thumb.jpg

 

Perhaps it would help to disconnect the composite circuit on the 800, if anyone can suggest how to go about that.

 

Here are the 800 schematics I am looking at:

 

post-8517-0-81078000-1378843821_thumb.jpg

post-8517-0-15962400-1378843841_thumb.jpg

post-8517-0-69436500-1378843852_thumb.jpg

 

Thanks

Share this post


Link to post
Share on other sites
Perhaps it would help to disconnect the composite circuit on the 800, if anyone can suggest how to go about that.

Thanks

I have attached a schematic with the Chroma amplifier(A104), the link between the Chroma/Composite signals(C164/R202), and the link between Chroma signal/RF modulator input(C170/R206) circled and labeled.

post-26063-0-42923600-1378848747_thumb.jpg

Share this post


Link to post
Share on other sites

So, do you think I should break the link between the Chroma/Composite signals, or both (also link between Chroma signal/RF modulator input)?

 

Thanks!

Share this post


Link to post
Share on other sites

So, do you think I should break the link between the Chroma/Composite signals, or both (also link between Chroma signal/RF modulator input)?

 

Thanks!

I would think that disconnecting just the Chroma/Composite link would be enough.

Share this post


Link to post
Share on other sites

I lifted the top leg of R202 indicated in the picture - no effect :/ Any other suggestions?

 

attachicon.gifR202.jpg

 

My suggestion remains the same: the 800 chroma-circuitry (and its Y+C combining points) are simply different than those of the 800XL's.

I quickly realized this after disconnecting the most obvious combination-point shown on the schematics... just to absolutely no effect, at all.

 

The only thing that comes to my mind, at this point, is trying to sample with Y/C "test cable" directly from early Y and C signal-points of the 800's video circuitry (following the Schematics), and see how it looks on the screen. The idea is to keep sampling inwards, progressively, from each Y & C possible points. This will save from (conversely) de-soldering most of the 800's video circuitry in order to find the likely culprit.

 

Good luck!

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...

  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...