crtemudriver losing sync, how to troubleshoot?

[feltz]

Hey all, really hoping someone can help me track this down as i'm pretty lost, and barely understand whats going on.

My setup, I have a trinitron consumer TV and CRTEmudriver.
Sometimes, it works great and no problems at all. Other times, when I boot it up, it is very sensitive and seems to get out of sync (image disappears/reappears)

For example, I was sitting in Windows, and simply moving a window to the corner of the screen caused it to get out of sync
See here: https://vimeo.com/manage/videos/644956835


I'm using NTSC preset without any adjustments, as i assume this was the safest for consumer NTSC tv

[abispac]

I'm using NTSC preset without any adjustments, as i assume this was the safest for consumer NTSC tv

Use generic 15khz instead.

[feltz]

What is the reasoning behind that? Seems like using that preset is less safe than NTSC, as it has a wider range.

[Zebidee]

Looks like a sync problem, possibly a complete absence of sync. Many TVs can still show a picture even with no sync signal, because they can guess where it should be from the picture information, but they can lose track as soon as you do something fancy (like move windows around on a desktop).

The important question is, how do you connect the PC to the TV? SCART, RGB, component, something else? Do you use an adapter or transcoder? What sync signal processing is there?

[feltz]

My hardware setup is a Sony Trinitron consumer TV, which has S-Video input.
I use a high end RGB/HV converter, detailed here: https://www.axunworks.com/RGB-to-Composite-S-Video-p341706.html
(Many others are using this in the same configuration, and was manufactured with CRTEmudriver in mind)
Here's another video, which might give more insight.


I'm just trying to pinpoint the culprit, video card, driver, rgb converter, or my ignorance 

[feltz]

Ok, i think I have some more info, and I think I know when its happing, just not why.
Every time i attempt to troubleshoot this problem, everything has been powered off for awhile, I power it on, and it gets out of sync.
I spend an hour or two attempting to change things (wipe modes from crtemudriver, reprogram them, try different resolutions, etc)
Previously, it would eventually start working and I had no idea what I did to fix it.

Now, i just realized something, i spent about an hour working on it today.
When i took the time to capture this video, uploaded it, sat down and made this post, i went back to launch a game, and the problem is gone 

So, what the heck? This sounds more like a hardware issue, maybe its my TV? cold solder joints? old capacitors? ugh..

Here's it working, with absolutely no changes other than re-running 30 mins later:

[Zebidee]

Mortal Kombat is a challenging game to be testing with, as it runs at ~55hz, which many consumer CRT TVs have trouble doiing. TVs tend to be better at ~60hz (NTSC) and/or ~50hz (PAL).

If TV is consistently better after warming up then may well be capacitors (or something else). First step is to have a look at the chassis to see if there is anything obviously wrong, like leaking/bulging capacitors.

[feltz]

Yeah i know what it natively runs at, but the TV doesn't support that, and Groovymame is adjusting to 59.940hz

[Zebidee]

I had another look at your videos. It definitely looks like a sync problem.

Using that transcoder you should continue using the NTSC preset in CRTEMU and flick the switch on the transcoder to NTSC.

Are you using H+V sync via VGA cable to the RGBS/HV jack on the RGB-2YC, or composite sync? You can setup H+V or composite sync via VMMaker, and could try both to see if there is some improvement. Looking at the Y/sync signal on an oscilloscope would explain a lot. You may be missing one or both of the sync signals.

Would could also test your setup on another Svideo TV, if available, to see if it is a TV issue. They are getting old   

[abispac]

Again, on my experience, on my trinitron, i had nothing but trouble using ntsc settings, i swithed to 15khz generic monitor settings and problems where gone. Ive also had problems with my cable, once i double checked and found about it, things were smooth ,but thats Just me.

[feltz]

I'm buying another CRT TV off Craigslist today, as I dont have another input source to test, and this seemed the easiest route.
I'll update my findings after some testing.

I do want to point out, that man this setup looks beautiful, even in an interlace resolution. Kudos to Calamity for the hard work on this.

[feltz]

Are you using H+V sync via VGA cable to the RGBS/HV jack on the RGB-2YC, or composite sync? You can setup H+V or composite sync via VMMaker, and could try both to see if there is some improvement. Looking at the Y/sync signal on an oscilloscope would explain a lot. You may be missing one or both of the sync signals.

Would could also test your setup on another Svideo TV, if available, to see if it is a TV issue. They are getting old   

First, yeah, i'm using VGA cable direct to the RGBS/HV jack on the RGB-2YC, it's supposed to handle the sync automatically from my understanding. Here's the manufacturer's response when i asked before i purchased:
"VGA only support 15K RGBS and RGBHV. Compatible with 15K RGBHV signals of PC, Raspberry Pi and MISTER.
It can also support the RGBS 15K signal input of home game consoles and JAMMA arcade machines."

So i ended up getting another TV off Craigslist for free, which has Composite and S-Video inputs (an Emerson 20 inch), and I do not get a S-Video signal at all out of the RGB-2YC transcoder to the Emerson TV (*edit* turns out this TV's S-video input doesnt work, as i tried another source). I'm wondering if the Trinitron i have is just handling the lack of sync and still displaying an image most of the time? Is that possible, while the Emerson doesnt know what to do without it?

The composite output from RGB-2YC to both TVs work fine.

[Zebidee]

So i ended up getting another TV off Craigslist for free, which has Composite and S-Video inputs (a ---smurfy--- Emerson 20 inch), and I do not get a S-Video signal at all out of the RGB-2YC transcoder to the Emerson TV. I'm wondering if the Trinitron i have is just handling the lack of sync and still displaying an image most of the time? Is that possible, while the Emerson doesnt know what to do without it?

This is exactly what I've been suspecting all along. The way the TV loses sync when you move windows around on the desktop is a telltale sign. CRT TVs are often quite good at picking out the signal, even with poor or missing sync. Some TVs are better than others.

I suspect the RGB-2YC is not reading the HV sync inputs properly, so it can't create the composite sync needed for your svideo signal. This can happen when a PC's GPU outputs sync pulses at a relatively low level, say 3.5v - 4v instead of the 4.5v-5.0v used by some others.

This voltage level is important because the sync combining logic on the RGB-2YC would almost certainly use TTL logic, which has a voltage threshold to recognise the sync pulses. If the sync pulses don't go above/below a certain threshold voltage level, and strong/long enough, then they won't be recognised by the TTL logic chips. The threshold voltage is often around 70-80% of the operating voltage. So, if the RGB-2YC is running on 5v, sync pulses peaking below about 4v might be missed.

All this is easier to identify if you have an oscilloscope or a good digital multimeter.

I bet you can tell I've encountered this issue before (when designing the GreenAntz transcoders). Good news is that we worked out how to get around it by adapting the design slightly.

TL;DR: Ideas/Ways around the issue (without hacking your RGB-2YC) could include:
- try a different GPU/video card, it may output the sync at higher level
- use/make a VGA breakout cable and connect the RGB+composite sync via the RCA jacks. You can use CRT_emudriver to output composite sync
- amplify the sync signal

I have some other ideas, but mostly they involve some hacking the sync signal input and I'd need to have a close look at the RGB-2YC first.

The composite output from RGB-2YC to both TVs work fine.

This would be because composite VIDEO (vs composite sync) is usually in strictly NTSC or PAL format, and the sync and color burst timings required would be generated and set within the RGB-2YC unit by crystal oscillators attuned to the required frequencies. So it won't really need the sync input anyway, you just set it to NTSC or PAL via a switch.

[feltz]

All this is easier to identify if you have an oscilloscope or a good digital multimeter.

I know just enough electronics to be dangerous, no oscilloscope, but curious what a "good" digital multimeter means. I have a Fluke 77 II multimeter.  If you dont mind and its not too much to ask, mind walking me through what signal to test?

TL;DR: Ideas/Ways around the issue (without hacking your RGB-2YC) could include:
- try a different GPU/video card, it may output the sync at higher level
- use/make a VGA breakout cable and connect the RGB+composite sync via the RCA jacks. You can use CRT_emudriver to output composite sync
- amplify the sync signal

These are great suggestions. I'm gonna see if i cant find a pre-made VGA->RGBS cable, this seems like the easiest route. I'm really hoping to keep the card as other than this issue, its a pretty good one and theyre hard to find.

One additional side note: The RGB-2YC has a RGB-Adj potentiometer, documentation says this is for brightness, but it does seem to affect the sync issues (still doesnt solve it when on max)

[Zebidee]

All this is easier to identify if you have an oscilloscope or a good digital multimeter.

I know just enough electronics to be dangerous, no oscilloscope, but curious what a "good" digital multimeter means. I have a Fluke 77 II multimeter.  If you dont mind and its not too much to ask, mind walking me through what signal to test?

If your DMM can measure DC voltage with a MAX/MIN function, or a basic scope function, it might be fast enough to detect the sync waveform pulses peaks/troughs. What we want to know is the peak voltage. If you simply measure DC voltage, your DMM may just give you the average, which will of course be a little lower than peak voltage. This is all much easier (and less vague) if you can use an oscilloscope to actually see the waveform.

I've tried doing this with my Fluke 17B+, which has the voltage min/max function, but with only limited success. I guess it doesn't work fast enough (the sync pulses are only 4-5 microseconds long).

The good news is that even if your DMM only gives you an average sync voltage, this should be within ~0.3v of peak sync voltage (based on a sync pulse interval of 4.6us per line, each line 64us long). So you can get a close and useful reading even with a cheap DMM, just you may need to take the above drop in voltage readings into account.

An easy way to access the VGA pins for testing is to use a breakout cable, a male-male VGA cable, or one of those VGA male header housings with screw terminals for attaching your wires. All numbered and easily accessible.

These are great suggestions. I'm gonna see if i cant find a pre-made VGA->RGBS cable, this seems like the easiest route. I'm really hoping to keep the card as other than this issue, its a pretty good one and theyre hard to find.

One additional side note: The RGB-2YC has a RGB-Adj potentiometer, documentation says this is for brightness, but it does seem to affect the sync issues (still doesnt solve it when on max)

That's right, the pot should only affect RGB signals, not sync. RGB has intensity, but sync is binary - either on or off

[feltz]

SO I picked up a VGA-RBGS cable, and got it "kinda" working with CRTEmudrivers csync enabled.
When I first hooked it up, i was getting vertical rolling. Digging deeper, i tried flipping the h/v sync polarities, and that seemed to solve that problem.

Now, i'm back where to my original issue, where i'm guessing the sync signal is just not strong enough.. Its losing sync.

- amplify the sync signal

Do you have any suggestions on a simple way to do this? I kinda feel like this RGB-2YC might have bad some false advertising if it cannot handle the sync voltages from VGA

[B2K24]

Ideally you want to go VGA out from your PC into a component transcoder then component into your consumer set. That's how I use my consumer Sony Trinitron and phillips CRTs. I only use component and a transcoder and don't have any sync problems.

[Zebidee]

SO I picked up a VGA-RBGS cable, and got it "kinda" working with CRTEmudrivers csync enabled.
When I first hooked it up, i was getting vertical rolling. Digging deeper, i tried flipping the h/v sync polarities, and that seemed to solve that problem.

Now, i'm back where to my original issue, where i'm guessing the sync signal is just not strong enough.. Its losing sync.

Don't forget to regenerate and reinstall the video modes after enabling csync.

- amplify the sync signal

Do you have any suggestions on a simple way to do this? I kinda feel like this RGB-2YC might have bad some false advertising if it cannot handle the sync voltages from VGA

I can suggest some ways that involve soldering, but I'm guessing you don't want to do that. Depending on the issue with the RGB-2YC another solution might be better/simpler. Oscilloscope on would be ideal, but failing that, if you can measure the sync input voltage with a multimeter we will have a better idea if what is going on.

Some video cards output sync at lower levels than others. You might get better results with another video card or PC. I know this is a crap solution, but is one I've encountered before with other transcoders.

Ideally you want to go VGA out from your PC into a component transcoder then component into your consumer set. That's how I use my consumer Sony Trinitron and phillips CRTs. I only use component and a transcoder and don't have any sync problems.

He's using svideo input. If you're using a component TV for retrogaming may I humbly suggest a GreenAntz VGA/RGB to YPbPr transcoder.
http://forum.arcadecontrols.com/index.php/topic,164245.0.html

[feltz]

Don't forget to regenerate and reinstall the video modes after enabling csync.

Yeah I made sure to do this.

I can suggest some ways that involve soldering, but I'm guessing you don't want to do that. Depending on the issue with the RGB-2YC another solution might be better/simpler. Oscilloscope on would be ideal, but failing that, if you can measure the sync input voltage with a multimeter we will have a better idea if what is going on.

Some video cards output sync at lower levels than others. You might get better results with another video card or PC. I know this is a crap solution, but is one I've encountered before with other transcoders.

Ok I just ordered an osciliscope, i think that is best in the long run. I'm not afraid of doing some soldering, but i think its better to know the issue before attempting a solution. Should be here in a couple days, now figuring out how to read it and interpret the results 

[Zebidee]

What scope did you order? You can get some ok usb scopes for around $100-150

[feltz]

I got the Owon VDS1022I PC oscilloscope.
I saw this video and it was a suggested budget scope:

[Zebidee]

Yeah good, it does the job.

[feltz]

Whee! Got the oscilloscope up and running, and tested the composite sync out of the video card. Its sitting around 4.9v so i think thats good.

As far as measuring the sync on the y/c luminance channel coming out of my encoder, i have no idea what im looking for 

Any suggestions?

[Zebidee]

Maybe you can post some screenshots of your sync with the scope. You can do a screenshot from windows by holding down the windows key and pressing "print screen". Screenshots are saved by default under  \Users\<username>\Pictures\Screenshots

Your scope can analyse 2 channels at a time. I've attached a screenshot using my own Owon scope, and a GreenAntz transcoder (SCART version), looking at:

a) composite sync on the Y signal, as it is going into the TV (on top, in RED)

b) composite sync input via SCART to the GreenAntz (on bottom, in YELLOW)

This snapshot is taken during the vertical sync period. The actual vertical sync pulse is the long "valley", with 3 spikes (pulses) poking up, that you can see on both lines. All the other short "dips" down are actually horizontal sync pulses.

Looking at the Y signal at top (RED), you'll see that there are some horizontal sync pulses, or blanking lines, before and after the vertical sync pulse. Collectively this is the "vertical sync period", and it tells the TV when to go back to start and draw a new frame. You get approximately 59.94 vertical sync periods per second for NTSC (obviously it depends on vertical frequency).

Before and after the vertical sync period you can see some tall "mountains". They carry the "luma" information, which (roughly) equals luminosity. Each little mountain between the pulses represents a single horizontal line worth of screen data, about 64us (microseconds) worth.

You can get a good black and white picture with just this Y signal.

The bottom line (YELLOW) is measuring sync input, so all you see there are horizontal+vertical sync pulses (composite sync) with no luma/picture information.
 
Composite sync is negative for both H+V (usually), so the pulses are negative. That means that you'll see a long "plateau" at your sync voltage (you said 4.9v before?), and each pulse sees the voltage drop to essentially 0v or ground (If you had positive sync, it would be the other way around). If all you have is a plateau at 4.9v, then that isn't really sync. It must have the pulses, so that is what you look for.

If you look carefully back at the red (Y) line and compare to the yellow (sync input), you'll see the Y sync pulses going into the TV are a little bit more clearly defined, and the plateaus are squared-off better, than they are on the sync input. That is because the GreenAntz has cleaned it up a bit and adjusted the voltage levels to standard.

Something to note: The Y signal is identical between both component and svideo. The only difference is that with svideo, all the color is carried by single C (chroma) signal (thus the Y/C notation). With component the color data is further split up into Pb and Pr "components" and carried by separate cables (thus YPbPr, or YUV).

[Calamity]

Awesome post Zebidee 

[feltz]

wow thanks for so much detail.
I'm sure something is wrong, cause now when I re-test, voltages seem way off?!

This is from R200 series card with VGA -> RGB/S cable, and sync plugged directly into oscilloscope
I dont really have a way of testing the sync while plugged in, i assume id need a T connector or something?

[Zebidee]

wow thanks for so much detail.
I'm sure something is wrong, cause now when I re-test, voltages seem way off?!

On the probes there is a probe attenuation slider, with 1x and 10x positions. Slide it back to 10x (or maybe it is the other way around! Has been a while). There is a probe attenuation setting in the software as well. RetroBob covers this in his video (you linked it above).

While you are at it, take a little time to calibrate the square-wave of the scope. Bob also shows you how to do this in his video.

In the pic you posted we can see the H sync pulses clearly, but you need to also see the V sync. It will be easier if you "zoom out" by using a larger time scale (like 500us or 1ms), pause and centre on the V sync, then "zoom in" (to a shorter time scale like 50 or 20 or 10us). Drawback is detail is not as good this way. You can also use the trigger function on the scope to try an get it to lock onto the Vsync period, though this can be difficult as the scope can get confused between H & V sync.

Alternatively you can leave the scope at 100us and just randomly hit pause until you get lucky and it stops with the V sync period visible.

This is from R200 series card with VGA -> RGB/S cable, and sync plugged directly into oscilloscope
I dont really have a way of testing the sync while plugged in, i assume id need a T connector or something?

Yeah, it can be a bit tricky to get the probes securely in the right spot for the output to TV. Best to do your measurements "under load", while connected. In the industry they get around this by using BNC T connectors. You can then use an RCA to BNC adapters to connect your signal to the scope and then carry on to the TV. I know the Svideo doesn't use RCA connectors, but composite video does and you could still run the Y signal through to the scope and then to composite and still get a B&W picture.

Here is another video with RetroBob and Ste from HD Retrovision demonstrating how to use a Rigol scope. You'll see they use the BNC T-connectors. They also use some kind of custom SCART breakout board to get the signal out of SCART    You'd still have to hack a connection somehow.




When I took my measurements before, I was using my component-modded China TV. For that TV I had added some ceramic capacitors (104 I think) in series on the input, and 75ohm terminator resistors (to ground). Anyway, I left the termination resistor legs exposed and raised *slightly* at the signal end, leaving a small arch, which made it easy to hook the probe leads on. Obviously the back of the TV was off so I could get the probe leads in there.

I wish I could take some photos of how I attached the probes to the TV, but I've since dismantled the whole thing! I actually took that screenshot I posted earlier of the sync months ago. I'm currently re-organising my work areas to get back more space.

You'd have to take the back off and have a good look if you want to do similar to what I did. Try to find a spot (like a component leg) on the signal path before the next component after the termination resistor (probably a capacitor in series). If you can't find a good spot the probe leads to, you could try lightly soldering a resistor leg near to where your Y signal terminator is, so you have something to hook your probe to.

The good news is that you don't need to have the TV turned on to do this. In fact, you can leave it off while using the scope, it won't make a lot of difference, the signal termination is still there.

A less invasive way to do this, without opening the TV, would be to hack a svideo cable. Cut it in half and use a small piece of breadboard to reconnect the signal paths (Y, C, GND and shield ground) between the halves. Put in some loops, using old cutoff component legs, to make it easy to hook your probes onto.

You could also take the covers off your transcoder and see if you can find a good spot to access the Y signal from there. It would have to be either at the signal outputs or just before.

[Zebidee]

Awesome post Zebidee 

Thanks very much Calamity

The way sync and video signals work can be pretty confusing, I hope it makes understanding a bit easier for some.

[feltz]

This is super appreciated, and i'm sure will help other folks in the future testing this stuff. For now, i'm waiting on some parts (breadboard, extra s-video cable, bcn t splitter, etc) so i can make sure im testing these signals under load. I should have results in a few days

[feltz]

Ok I think i got it all somewhat working. Trying to get it all under load. Here's my setup:

VGA -> RGBS
RGBS BNC Sync cable into BNC T connector, one into Scope CH2, the other into RGB-2YC encoder
S-Video output from RGB-2YC cut and split into breadboard, and output to Scope CH1 and TV

It looks like its working? On CH2 i set the mark cursor to show voltage

[Zebidee]

Ok I think i got it all somewhat working. Trying to get it all under load. Here's my setup:

VGA -> RGBS
RGBS BNC Sync cable into BNC T connector, one into Scope CH2, the other into RGB-2YC encoder
S-Video output from RGB-2YC cut and split into breadboard, and output to Scope CH1 and TV

It looks like its working? On CH2 i set the mark cursor to show voltage

Good news is that you seem to be getting normal looking sync there. Input sync level a little on the low side, but you can clearly see the sync there on the output at roughly the right level.

Bad news is this doesn't explain your losing sync when moving windows issue.

I see no luma/picture content on your svideo output to TV. Can you try again with some image content, ideally while the wandering picture issue is happening.

I note that you run windows in an interlaced video mode, and that is where you were losing sync in original video. This may be important because interlaced modes can be tricky to make composite sync for. Interlaced modes have "odd" and "even" lines that are "interlaced" to make a whole image. The TV needs to know odd from even to display the image properly.

So, for every alternate line, the long vertical sync pulse starts and finishes with a a half-length period between horizontal pulses.

You can see this in the scope pic of the sync I posted above. The distance between the last Hsync pulse and the start of the long V sync pulse is shorter than normal, and the distance (period, because we're talking about time) after the V sync pulse to the first H pulse is also shorter.

These shorter periods between H pulses tell the TV to expect a odd or even line (I can't remember which is which, but it doesn't matter).

Now compare that with the scope pic you just posted, where those periods between H pulses before and after the sync look normal or full length.

Old CRT TVs tend to be pretty forgiving with regard to these long/short periods for odd/even interlaced sync, but it could be an issue in some cases   

[Zebidee]

I've attached an image of the sync highlighting (in blue/turquoise) what I meant above by those shorter periods between H sync pulses.

You can see how there is a shorter period between the last H pulse and the start of the V pulse. Note that the start of a V pulse is effectively the same as a Hpulse, so this equals a shorter H pulse period. And again, the period to the first inverted H pulse (the spikes you see during the H pulse period) is shorter than normal.

We see the same pattern of shorter H pulse periods repeated at the end of the V pulse.

[feltz]

I tried a bunch more and i think i finally got a snap shot that has those shorter pulses you mentioned.. But its not in all them, is that normal?
My original screen shot was with video on the screen at a 240p resolution. I dont understand why it wasnt showing video information.
Here's a couple more with 480i running on windows desktop

[feltz]

So i attempted to get some captures near the sync time while the issue is occurring on screen.
In these cases, i ran a game that is mostly dark (sinistar), and it runs in 480i mode. This game causes my TV to flicker in and out of sync during gameplay, so i paused it at a moment when it was out of sync.

One thing that might be an important factor, is it always seems to occur in darker games, where there is a lot of black on the screen, Sinistar is one, and R-Type is another.

In all your screen shots, your video signals peak much higher than mine, is that maybe the culprit that my output voltage is too low?
Attaching an animated gif to show what that looks like.

Here's the output from the scope at that time:

[Zebidee]

The mystery deepens!

That sync (svideo output) looks fairly normal to me, a little fuzzy/messy/noisy around the tops/bottoms/edges but mostly normal. However, it definitely seems to be a sync issue. At least for your interlaced 480i. Which brings us back to your TV. Maybe it is just very fussy? This would explain why your buddys using the same transcoder don't have issues.

The short-period H sync pulses are only for interlaced modes come every odd line, so you'll get them ~50% of the time.

Have you tried your Sony with an alternative svideo source like a DVD player? Or something else you can plug into your transcoder?

You can measure the sync voltage level present by placing those horizontal voltage bars at the black level and the bottom of the pulses. You could also use the vertical bars to measure the width (time) of those H pulses, should be roughly 4-5us.

Applying Occam's razor - now that we have eliminated the most obvious, it is a good time to explore the less obvious.

You can tweak the sync out from PC by editing the monitor.ini (sync pulse characteristics). I've never had much joy myself from using the NTSC preset myself, but that might be because the TVs I have all do both PAL and NTSC ranges already. Maybe you could try using a different preset like the generic 15khz (like abispac suggested above), just to see if that makes any difference. Seems unlikely but who knows, maybe the NTSC preset isn't quite right for some reason. Make a backup or just copy it over anew from your original CRT_emudriver files later.

You could also play around with the presets. For example, you could start with the generic_15 preset and replace the horizontal and vertical frequency parameters to match NTSC (italics), then perhaps the horizontal sync parameters for front porch/sync pulse width/back porch (bold), to see if it makes a difference.

crt_range0 15625-15750, 49.50-65.00, 2.000, 4.700, 8.000, 0.064, 0.192, 1.024, 0, 0, 192, 288, 448, 576

There is a stickied thread for monitor presets.

[feltz]

Yeah it seems to be leaning more and more towards the TV itself, which is good and bad news. Its such a beautiful screen when its working...
I did try the generic_15 preset, and the problem still occurs ..

One thing I did notice during all this troubleshooting, is after leaving it on for a long period of time (2+ hours) the problem seems to lessen. Also, the problem doesnt occur at all on Composite video input out of the same encoder.

I suppose at this point I will see what I can do to find another TV to replace it. Thanks for all the help Zebidee its been worth the work just to learn, and you've gone above and beyond what i could ask for. Thanks!

[Zebidee]

One thing I did notice during all this troubleshooting, is after leaving it on for a long period of time (2+ hours) the problem seems to lessen. Also, the problem doesnt occur at all on Composite video input out of the same encoder.

This is often a sign of tired electrolytic caps.

You could try replacing the electrolytic caps. I don't normally recommend "capkitting", but if you are only ever going to do one or two TVs then you may as well just go through and replace all the old electro caps.

Here's a trick for you that will cost nothing - try running the sync into the composite input! Take the sync directly from your source rather than the transcoder. As the raw sync voltage is too high for TV, put a resistor, say 470 to 750 in series on the sync line and then into the composite AV jack.

[feltz]

Yeah i considered recapping, ive done it before on other electronics successfully, but im deathly afraid of CRT monitor and the process of discharging it =\

i'll try the sync suggestion, although im not sure how the inputs would be handled with both inputs being fed into. The TV auto-detects the input source, and switches from Composite to S-Video automatically.

[Zebidee]

Yeah i considered recapping, ive done it before on other electronics successfully, but im deathly afraid of CRT monitor and the process of discharging it =\

All you need to discharge a CRT is a screwdriver and a cable with alligator clips on each end. This is my set:



Discharge to chassis ground or nearest large metallic thing. Last time I did it to my bike (pedal-power!).

Being scared is healthy. It stops being scary after a few times. Most 80's+ TVs will let the charge out slowly anyway, so you probably won't even get a pop if it hasn't been on in the last couple of days. I always feel disappointed when I don't get a pop. My tips are to hold the screwdriver in the hole for a few seconds so the charge fully dissipates, and keep your other hand in your pocket.

Anyway, there are many YT videos.

i'll try the sync suggestion, although im not sure how the inputs would be handled with both inputs being fed into. The TV auto-detects the input source, and switches from Composite to S-Video automatically.

Maybe you'll need to select manually with the buttons or remote.

I'm guessing it may work as it seems to for many RGB mods I've read about, even when they don't have the composite input selected. Fairly reasonable guess because once the sync is there, I expect it will feed into the same jungle chip processing matrix anyway.

[psakhis]

I always feel disappointed when I don't get a pop.

I'm still laughing