800XL and 1084S: Artifacting

[flashjazzcat]

I've just gotten hold of a Commodore 1084S monitor, and I'm absolutely delighted with it. VBXE output from my 130XE is stunning. I was expecting similar results from my UltraVideo modded 800XL via chroma/luma, but the picture is plagued with artifacts. This is the same 800XL whose s-video output is almost indistinguishable from VBXE when fed into my LG flat panel, where no artifacts are apparent. It doesn't seem to be the custom cable I've made, since the still-active, largely unmodified lagacy output of my VBXE 130XE is artifact free through the 1084S. It's really frustrating, since the artifacts completely spoil the pseudo 80 column display on the 800XL.

 

Any clues? Obviously chroma/luma share the same ground but why would the XL circuit behave differently to the XE in this way? The composite video circuit of the 800XL has been completely obliterated, to the best of my understanding.

 

This is the s-video from the 800XL through the LG flat panel:

 

 

Needless to say it's next to impossible to capture the image on the 1084S, but it is awash with artifacting.

Edited May 7, 2010 by flashjazzcat

[Rybags]

Can't replicate on mine, although my 800XLs default chroma circuit is kinda buggered due to one of those SuperVideo upgrades that I'll have to fix at some stage (chroma virtually non-existant)

 

One possibility that I think I found with one of my capture-cards - the monitor might very well combine Chroma/Luma and just process it as Composite.

Sounds utterly stupid for sure, but I think some monitors and capture devices do that.

 

Maybe you could verify it by just running luma only to the monitor. In theory, doing so with a black background in hires should give the same results as if chroma was there anyway.

Edited May 7, 2010 by Rybags

[flashjazzcat]

OK. No artifacts with just the luma connected. But what's extra puzzling is the absence of artifacts using an XE with no RF modulator. Surely that suggests the signal isn't mixed at destination? I'm thinking that despite removing the RF mod and a bank of resistors in the 800XL, the chroma and luma are still being mixed at source - just not to a degree to which the LG LCD display is sensitive.

Edited May 7, 2010 by flashjazzcat

[Rybags]

I'm no electronics expert, but would something like a diode on the GND for Luma output help in any way?

[Bryan]

Hmm... I was going to ask if it has a convergence problem, but you said it looks okay with just luma.

 

I can't imagine any Commodore monitor would degrade an S-Video signal intentionally.

[GroovyBee]

Looking at the Jerzy Sobola schematic of the 800XL have you lifted the leg on capacitor C54 at the junction of L7 and R56? There are two C54s on the schematic its the one connected to R67. Check resistors R41 to R45 are connected to +5 too. Where are you picking your chroma and luma signals up from?

[flashjazzcat]

I just tried a diode, but it didn't work: removing the CV circuit alleviated most of the problems the diode in chroma has attempted to fix in the past. Actually, I built a "clean" cable after removing the CV in the 800XL, because the cable with the diode was introducing colour noise.

 

Secondly, I tried different hook-up points for the chroma, progressively backing up the circuit from the original (junction of R56 and C54), to the leg of Q5 which is connected to R66. No difference. This is extremely puzzling.

 

Just to clarify, this 800XL has all 4 legs of the RF box disconnected, and components R56, C54, C55, R67, and R68 removed from the board.

[Kjmann]

One thing about commodore Monitors it that they suffer from "Signal Bleed over" in the chroma Luma Inputs. The chroma and Luma inputs inside the monitor are not separated well at all. the 1084's are better but it seems that they were never able to get rid of the "Bleed over" Completely. the earlier Commodore monitors were even worse with this problem.

 

This is the reason for needing a Capacitor on the Chroma Line output from the 800XL to filter out the DC Voltage, when using a Commodore Monitor. otherwise the luma will pick up the DC voltage that comes off the Chroma Line causing Artifacting problems and in some cases even a nasty puke Green Tint to the picture. this doesn't solve the problem completely. it just gets rid of some of the artifacting and guarantees that you wont see the Puke Green Tint. =)

[flashjazzcat]

One thing about commodore Monitors it that they suffer from "Signal Bleed over" in the chroma Luma Inputs. The chroma and Luma inputs inside the monitor are not separated well at all. the 1084's are better but it seems that they were never able to get rid of the "Bleed over" Completely. the earlier Commodore monitors were even worse with this problem.

 

This is the reason for needing a Capacitor on the Chroma Line output from the 800XL to filter out the DC Voltage, when using a Commodore Monitor. otherwise the luma will pick up the DC voltage that comes off the Chroma Line causing Artifacting problems and in some cases even a nasty puke Green Tint to the picture. this doesn't solve the problem completely. it just gets rid of some of the artifacting and guarantees that you wont see the Puke Green Tint. =)

I see. So the 130XE which exhibits zero artifacting simply happens to have less DC voltage on the chroma line? I'll give this a try. Is a wiring diagram required, or is one simple placing a cap across chroma/ground?

[Bryan]

I see. So the 130XE which exhibits zero artifacting simply happens to have less DC voltage on the chroma line? I'll give this a try. Is a wiring diagram required, or is one simple placing a cap across chroma/ground?

 

Just put a cap in series with the chroma line:

 

Chroma----------||------------>TV

 

Chroma is a fairly high frequency signal so you shouldn't need a very big cap, something like a .47uF or so should work well. If the cap is polarized, put positive toward the computer.

[flashjazzcat]

Thanks. Not having any success with this approach yet, however. The cap really make no discernible difference on this 800XL. Will look into it further tomorrow.

[+bob1200xl]

The components that you have removed delete ALL the CHROMA from the COMPOSITE pin. Put C54 (.001) and R67 (75ohms) back into the PCB and use COMPOSITE for CHROMA. This will put a capacitor (C54) in the circuit, as suggested.

 

 

 

One thing about commodore Monitors it that they suffer from "Signal Bleed over" in the chroma Luma Inputs. The chroma and Luma inputs inside the monitor are not separated well at all. the 1084's are better but it seems that they were never able to get rid of the "Bleed over" Completely. the earlier Commodore monitors were even worse with this problem.

 

This is the reason for needing a Capacitor on the Chroma Line output from the 800XL to filter out the DC Voltage, when using a Commodore Monitor. otherwise the luma will pick up the DC voltage that comes off the Chroma Line causing Artifacting problems and in some cases even a nasty puke Green Tint to the picture. this doesn't solve the problem completely. it just gets rid of some of the artifacting and guarantees that you wont see the Puke Green Tint. =)

I see. So the 130XE which exhibits zero artifacting simply happens to have less DC voltage on the chroma line? I'll give this a try. Is a wiring diagram required, or is one simple placing a cap across chroma/ground?

[Kjmann]

I see. So the 130XE which exhibits zero artifacting simply happens to have less DC voltage on the chroma line? I'll give this a try. Is a wiring diagram required, or is one simple placing a cap across chroma/ground?

 

Just put a cap in series with the chroma line:

 

Chroma----------||------------>TV

 

Chroma is a fairly high frequency signal so you shouldn't need a very big cap, something like a .47uF or so should work well. If the cap is polarized, put positive toward the computer.

 

I Personally use a 22uf 10volt Axial Cap.

[flashjazzcat]

I Personally use a 22uf 10volt Axial Cap.

I actually tried one last night and it made no difference. I'll take a photo of the board later and post it, just in case someone can spot something I've missed. It must be fixable, since that 130XE has zero artifacting.

[Jacques]

Flashjazzcat:

do you remember? I also had a case with 800XL and s-video mod where theoretically complete separation (RF modulator removed, removed connections between LUMA and CHROMA leading to composite video) didn't help and I was still left with some artifacts on this 800XL. Luma is perfectly clear, but together, the S-video was still artiffacted :/ It's weird, looks like there's still some connection between LUMA and CHROMA in 800XL (at least in my old 800XL, not XLF).

[flashjazzcat]

Flashjazzcat:

do you remember? I also had a case with 800XL and s-video mod where theoretically complete separation (RF modulator removed, removed connections between LUMA and CHROMA leading to composite video) didn't help and I was still left with some artifacts on this 800XL. Luma is perfectly clear, but together, the S-video was still artiffacted :/ It's weird, looks like there's still some connection between LUMA and CHROMA in 800XL (at least in my old 800XL, not XLF).

I do recall, and I tend to think - albeit from a less informed standpoint - that there's still a vestigial connection between luma and chroma at source. It's not enough to upset the LG M227WD, but it does spoil the party with the 1084S. Might be wise to study the 800XL video circuit schematic.

[flashjazzcat]

I've studied the schematic and other than the confusing dual occurance of C54, everything is as expected: i.e. the luma and chroma are totally isolated on my machine. The only idea I had was backward leakage through a +5v supply to the colour circuit, but lifting a resistor made no difference. I also tried re-instating C54 and R67, to no discernible effect. Obviously some voodoo at work here.

[+bob1200xl]

A picture of what you are calling 'artifacting' would probably help. As I understand it, you get decent video with just the LUMA plugged in, but the picture is faulty when you plug in the CHROMA cable? This same monitor looks good when you plug in a 130XE?

 

A good demodulating circuit (the one in the monitor/TV) will de-couple the DC component and normalize the signal before it decodes it. Other, older circuits may just key off of the sync level or some such and decode that. Hard to know what your monitors are doing, but this is typical. The Atari is non-standard so some things work and some don't.

 

Although, it sounds like your 800XL is broken...

 

 

Bob

 

 

 

I've studied the schematic and other than the confusing dual occurance of C54, everything is as expected: i.e. the luma and chroma are totally isolated on my machine. The only idea I had was backward leakage through a +5v supply to the colour circuit, but lifting a resistor made no difference. I also tried re-instating C54 and R67, to no discernible effect. Obviously some voodoo at work here.

[flashjazzcat]

I'll take a picture later (camera is flat), but basically I'm talking about classic Atari artifacting. Chequerboard black and white patterns are either green or pink depending on their location, half colour clock vertical lines have green/yellow tinges, etc. It's the kind of effect games used to make use of.

 

Another, stock 800XL looks exactly the same. This seems to be something peculiar to the XL line. I'd be intrigued to hear from any other 1084S owners and their findings using chroma/luma inputs. This particular modified 800XL has the best, cleanest, sharpest s-video output I've ever seen from an Atari when plugged into a flat panel LCD. Artifacting in that scenario was apparent until I made modifications to the video circuit (i.e. severing connection between luma and chroma, at which point artifacting vanished and video quality attained excellence). I can accept the fact that the 1084S and the 800XL may just not get along well this way, but I'd heard great things about this monitor, hence my slight surprise that the 800XL's picture isn't great.

Edited May 9, 2010 by flashjazzcat

[mimo]

I'll dig my 1084 (not S:sad:) out of the garage. I have a couple of modded XLs and XEs around. Any particular program you want me to test with?

[flashjazzcat]

I'll dig my 1084 (not S:sad:) out of the garage. I have a couple of modded XLs and XEs around. Any particular program you want me to test with?

That'd be great, thanks. I usually test with The Last Word 3.x, not for egotistical reasons, but because the title screen and pseudo 80 column text mode test the mettle of any display. What's the difference with the non-S version of the 1084?

Edited May 9, 2010 by flashjazzcat

[mimo]

I'll dig my 1084 (not S:sad:) out of the garage. I have a couple of modded XLs and XEs around. Any particular program you want me to test with?

That'd be great, thanks. I usually test with The Last Word 3.x, not for egotistical reasons, but because the title screen and pseudo 80 column text mode test the mettle of any display. What's the difference with the non-S version of the 1084?

 

Ok, I'll d/l The last word.

There are so many differences with these C= monitors from manufacturer to manufacturer but the main difference is that the S is Stereo

[mimo]

Well I can't get LW working, just loads and then goes straight to Dos 2.5:?

Anyway, I ran the Drunk Chessboard demo, plenty of black and white blocks in that one. No artifacting at all with 65XE (1 meg RAM, SDX,Stereo pokey,etc,etc) 800XL (quick and easy video mod) and 800XL with Freddy (just hooked up the chroma line, no other mods).

I have just got a very simple cable that I made myself, no resistors, diodes etc used.

 

Just need to try FW again, the dir says 360 sectors free so something is amiss

[flashjazzcat]

Type "L", then "LW.EXE" at the DOS menu.

 

OK: managed to get some pictures which show results here; 130XE vs 800XL:

 

130XE (no RF mod, no other changes from stock):

 

 

 

800XL (No RF mod, chroma/luma circuits isolated):

 

 

 

It should be noted that my other, unmodded 800XL looks exactly the same.

[mimo]

Ok, still no problems here with 800xl, quick and easy video mod