TV TFT display problems
MagIO2
Posts: 2,243
On another thread I reported that my first propeller-board ever died: http://forums.parallax.com/showthread.php/149533-Quickstart-Rev-A-not-found-on-PC
Now the same thing happened with my second QS. In both cases it happened when I attached it to this 4.3" TV TFT display:
http://www.pollin.de/shop/downloads/D120915D.PDF
I really have no glue how this can happen because it has it's own external power supply and is only connected to P12-14 of the propeller via resistors and of course to ground.
Another problem I have with the display is, that it has some problems to display blue. It looks like waves of different brightnesses are overlaid over the blue areas. If I change the frequency fpal in the TV driver the pattern changes, but it never dissappears. Without color the display works quite fine.
So, my questions:
Is there a way to decouple the display? Would an opto-coupler work? Any idea which one?
Any ideas about the problem with blue color?
Any idea why the display can destroy the FTDI?
Now the same thing happened with my second QS. In both cases it happened when I attached it to this 4.3" TV TFT display:
http://www.pollin.de/shop/downloads/D120915D.PDF
I really have no glue how this can happen because it has it's own external power supply and is only connected to P12-14 of the propeller via resistors and of course to ground.
Another problem I have with the display is, that it has some problems to display blue. It looks like waves of different brightnesses are overlaid over the blue areas. If I change the frequency fpal in the TV driver the pattern changes, but it never dissappears. Without color the display works quite fine.
So, my questions:
Is there a way to decouple the display? Would an opto-coupler work? Any idea which one?
Any ideas about the problem with blue color?
Any idea why the display can destroy the FTDI?
Comments
How did you power the Quickstart board?
If you wanted to add isolation, maybe you could capacitively couple the video signal. That would leave only the grounds with a direct connection.
Maybe just use an 0.1uF ceramic capacitor in series on the video line...
This may require also some pulldown resistors on either side of the capacitor, not sure...
So, I measured AC voltage between USB shield of the Netbook which was attached to the QS and the power supply of the display and it was ~40V. So my guess is that there was a weak connection in the wiring (video connector to pins plugged into breadboard and a wire from breadboard to QS header).
I think in future I'm gonna put more efford in good ground connections.
Here is a 4.3 option for you with Touch from Newhaven. Rayman has a controller board for it, and this is a very nice display.
And to be honest, for more quality I'd prefere this display: http://www.mikrocontroller-praxis.de/data_article/ITS-ITDB02-4.3/DS_IM120419007_ITDB02_4.3.pdf
For 36 Euros it also includes touch, an SD card slot and has it's own display memory. My expectation is that the controller should also work in 8-bit or 9 bit mode in this module. Will see ...
This time I connected my scope to the breadboard setup and only connected ground of scope to the board. When switching on the scope the QS stops working. Now being prepared I removed the power supply from my Netbook and everything worked fine again.
So it looks like ground sometimes has a different potential ....
Another display option is this :
http://www.digikey.com/product-detail/en/VM800B43A-BK/768-1194-ND/4332002
Includes a smarter display engine, and a bezel (and of course a higher price)
The transmitter power supply for the sonar system I worked with in the Navy had a floating ground... about +75V above 'earth' ground. If you used the wrong o-scope when troubleshooting or doing planned maintenance, you would get a nasty shock when you tried to change any of the settings on the scope. Only made that mistake a couple of times!
(The ZIP includes an MP4-file)
The color fringing in the blue you showed us is due to the Propeller not producing a stable enough color reference. We've had this problem with PAL the whole time. NTSC can tolerate very significant variances from the standard. If you are anywhere even close, it mostly works. PAL has very tight tolerances and it needs to be spot on to avoid artifacts.
Can you switch your display to NTSC? NTSC is a downgrade in terms of resolution, but it's going to be a lot more stable for color.
Edit: After reviewing your movie, NTSC is going to bring another problem and that is artifacting on your higher resolution display elements. Standard composite video NTSC would not look as good as what you've got going on right now.
Does your display offer S-video input? If so, you could use NTSC on it and get that kind of look without the artifacts. Your connection to the display and the driver would both require a tweak.
Overall, the best PAL display is Chip's TV driver. We've attempted lots of solutions and they all seemed to come with some fringing in PAL. Tweaking your driver to your display is the best option. When it's close, it's close. Looks like that is the one you are using too.
You might consider a color palette. Look for the ones that display the fewest artifacts. Maybe avoid color area fills too, outlines only. The fringing will still be there, but at a lower impact.
Nice display BTW!
The display can also do NTSC, but I would like to stay with PAL, as you find more devices that support PAL here in germany. So I'll go the route to find the stablest settings/colors and give the option to disable color.
I also like the display and have to buy some more just for the case I need more in future ;o)
Maybe I try the 5" one. If I remember right that one also supports S-video.
He was trying to help you diagnose the display problem.
Switch to NTSC ans see if the problem disappears,
If so, then you know.
If not then it's most likely bad connections.
The S-video may help you with PAL too. Not sure it's worth the trouble though. PAL resolution is good enough for your display over composite. You might try the setup for S-video, but feed both signals into composite hack I mention below on PAL to see how that performs. I've not done it.
One thought may be to modify your video interface some to allow for less overall color saturation. Those fringes will still be there, but much less pronounced with a lower overall color saturation value. The easy way to do this is with an S-Video type configuration. Just insert a resistor on the color signal to dial back it's saturation level. I've used this as a hack to improve displays on composite as well. Just connect both the color signal and the luma signal to your composite input. Color gets a bit more washed out when doing this, but it does improve the display some. Not best practice, but potentially useful to you.
I suppose the other option is to get a PAL converter chip and feed it RGB. This increases your cost and part count, but it will look really good.
Well, sort of. That color fringing is a known problem on Propeller PAL video output. Really, I was trying to communicate some options more than anything else. However, if there were similar artifacts on an NTSC display, then shielding on the video signal cables would be warranted for sure.
I wrote a little test-program to scroll through all possible fg and bg color values and there is rarely a stable color.
Thanks a lot for your help!