Weird Propeller-to-VGA problem (SOLVED)

Phil Pilgrim (PhiPi)

This one's got me stumped. I just got a combo NTSC/VGA/HDMI/DTV unit from Target. I'm trying to drive the VGA input with the Prop. I finally got the timing right (the display's a little fussy in that regard), and when the program first starts, I get a great picture: nicely saturated colors, dark darks, light lights, etc. But ever so gradually, the image on the screen begins to brighten until the colors are completely washed out.

Thinking I got a bum unit, I tried it with my laptop. No problem: same great display with no washout, no matter how long it runs. So I checked the voltage levels coming from the Prop, thinking the display might have termination issues. (The Propeller Demo Board has source impedance issues. With an incorreclty-terminated cable, the signal levels will be way too high.) But I got an unvarying 1V P-P with the cable plugged in, which is correct.

To restore the color to its former glory, all I have to do is interrupt the output, either by restarting the program or pulling the VGA plug and reinserting it. Then the gradual washout starts all over again.

Has anyone else experienced anything like this?

Thanks,
-Phil

Post Edited (Phil Pilgrim (PhiPi)) : 6/23/2009 4:14:29 AM GMT

CounterRotatingProps

That's really bizarre behavior, Phil.

What does a scope on the VGA lines show? --- this is obvious, but what does that prop board do on another monitor? (Scoping it too.)

(Recall the 'debug thread' ? Now you can say "I told you so!")

- H

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

Phil Pilgrim (PhiPi)

Scope shows 1V P-P. Prop board on another VGA monitor works fine: no washout. Signal levels are the same: still 1V P-P. Same cable used for all experiments, BTW.

-Phil

ericball

Shooting in the dark here.

Bias problems?· Try putting a capacitor in series with each of the signal lines to block DC.· Try grounding pins 6,7,8· (RGB return).

It's like you're gradually saturating a capacitor connected to the input.· Pretty soon, everything is white.


▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Composite NTSC sprite driver: http://forums.parallax.com/showthread.php?p=800114
NTSC & PAL templates: http://forums.parallax.com/showthread.php?p=803904

mikediv

Phil I bought a HDTV/VGA/S-video all the bells and whistles from Sams its a 26 inch and it did the very same thing you are talking about when I tried to use the proto board ,, for mine anyway I had to manually set the resolution in the monitor if I let it auto detect
it would work fine with PC or laptop but the proto board would work for a while as long as I was doing something if I let it sit it almost acted like it was going into screen saver but of course my prop did not have a screen saver the fix for me was went into TV menu and set the video to 60HZ and 640X480 and walaa it works fine I have not tried a higher res because I just didn't need it Oh the TV is a Westinghouse

Rayman

Is this a custom board? Sounds like a grounding problem to me...

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
My Prop Info&Apps: ·http://www.rayslogic.com/propeller/propeller.htm

Rayman

Actually, now that I think about it... I vaguely recall having that problem once... I think it was a pin or two on the VGA connector that I didn't solder right...

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
My Prop Info&Apps: ·http://www.rayslogic.com/propeller/propeller.htm

Phil Pilgrim (PhiPi)

Eric,

I think you're right about a cap getting charged, but I'm not sure where in the display circuit that might be. Here's a funny thing: I was displaying stuff that had only red and blue components on a black background. The background slowly became purple, then brightened from there. So I changed the display to green-only on black, expecting the background to start out dark green. But no: it starts at black, moves to purple, and then to magenta. The red are blue outputs are at a constant zero volts while this is happening, so the charge is not building up in the cable.

It does this on two different Demo Boards, too: an old model and a new model. I double-checked the five VGA grounds on the Demo Board, and they're all intact.

Now, looking at the VGA spec, I see the signal levels are supposed to be 0.7V P-P. the Demo Board outputs at least 1V P-P into a 75-ohm load at full brightness. So I tried displaying at reduced brightness, which produced a 0.7V P-P signal. I still get the gradual washout, though.

I checked the VGA spec, and it calls for DC-coupled signals, so I'm sure that's what my laptop (which works with the display) is providing.

My suspicion rests on the incorrect source impedance produced by the Demo Board, but I can't put my finger on a mechanism. The only issue that should cause is cable reflections.

-Phil

Phil Pilgrim (PhiPi)

Rayman,

Nope, it's a Parallax Demo board — two different ones, in fact, that exhibit identical behavior. And the grounds are intact.

-Phil

Phil Pilgrim (PhiPi)

One more experiment: output a completely black screen. 'Same result: gradual background intensity buildup. All I have to do to make it go away is disconnect, then reconnect the cable or restart the program.

-Phil

Phil Pilgrim (PhiPi)

Another experiment: switch to a display with a magenta background instead of black, with a few small black areas, along with red and blue. In this case, the black gradually changes to green. There's also a green stripe along the left margin and along the bottom. I take it that this is displayable area that the Prop is not outputting to.

-Phil

localroger

Phil, you might want to try installing a real resistor across each of the r, g, and b lines to ground. The Prop DAC resistor scheme assumes 75 ohms impedance forming the ground leg of the voltage divider; it sounds to me like the monitor has AC coupled that impedance, which would cause it to slowly become ineffective as the coupling capcitor charges, in turn causing the voltage from the dividers to rise and wash out to saturation. You might just need something like a 1K to keep the monitor input from holding a charge or a real 75 or 100 ohm resistor to take its place entirely.

Phil Pilgrim (PhiPi)

I've now got 150R resistors from each of the RGB pins to ground. This corrects the output impedance to 75 ohms. It didn't fix the problem, though.

-Phil

Peter Jakacki

The termination resistance being AC coupled should be immaterial as we are not dealing with a DC signal signal anyway. However it's worth a try but I feel that the problem may have more to do with power supply "grounds". Rather than inserting a cap into each signal just insert a 47uf cap into the ground line to the monitor, the + should ideally go to the monitor (should work) or just put two 100uf back to back.

Of course the easiest way to check is to run your Prop off a battery if possible.

*Peter*

jrjr

Phil,

I've seen sync pulse widths cause washout problems as described.

Sync pulse is slightly into the active area.

Perhaps shortening slightly or timing between H&V.

jr

Tubular

Phil you mention interrupting the plug restores the correct display. How about just interrupting the R G and B and leaving the Syncs run through - what happens then? Does it still restore? Or is it the interruption of the H and V sync causing the display to go back to a reset condition?

Perhaps just short pins 1 2 and 3 temporarily to ground on the back of the proto board. The magenta thing is interesting, wonder if the green channel has some special immunity

tubular

kwinn

Could also be a difference in the ground potential between the display and the prop. try connecting a ground wire between them.

Phil Pilgrim (PhiPi)

jrjr, you might be onto something. I've noticed that if I short the VSYNC output to ground, the problem goes away immediately and stays gone. The picture remains on the screen and starts to roll very slowly upward. When i release the VSYNC pin, the sync locks immediately, and the washout process begins anew.

-Phil

Phil Pilgrim (PhiPi)

The length of the vertical sync pulse was indeed the problem. I changed the vs driver setting from two lines to three lines, and I get a great picture with no creeping washout. But, geez, what a weird interaction!

Thanks to all for your time and assistance!

-Phil

Cluso99

Nice find and excellent detective work. Congratulations to all involved

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Links to other interesting threads:

· Home of the MultiBladeProps: TriBladeProp, RamBlade, TwinBlade,·SixBlade, website
· Single Board Computer:·3 Propeller ICs·and a·TriBladeProp board (ZiCog Z80 Emulator)
· Prop Tools under Development or Completed (Index)
· Emulators: Micros eg Altair, and Terminals eg VT100 (Index) ZiCog (Z80), MoCog (6809)
· Search the Propeller forums (via Google)
My cruising website is: ·www.bluemagic.biz·· MultiBladeProp is: www.bluemagic.biz/cluso.htm

Toby Seckshund

The monitor was probably black level clamping·in the last vertical period. If it wasn' there then it would "slowly" wander off to the wrong level in stead of locking onto the right one. Now you have been given a clue where the problem lies, tweat until best display happens, without upsetting any other displays though

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Style and grace : Nil point