Hsync instabilities
AndrewP
Posts: 4
On the Propeller P8X32A-Q44 being used to display VGA data at 1024x768 I am having a recurring and very random problem that is extremely difficult to reproduce consistently.
Periodically hsync pulses will be sent at uneven intervals and even split into two pulses for every one.· This is causing the image to distort, image blanking, and sometimes full resets.
Unfortunately I can not reliably reproduce this problem.· It will sometimes occur immediately after posting, sometimes immediately upon receiving the first few bytes of serial data, sometimes spontaneously while displaying a screen for a long for a long period of time, and sometimes it will go for hours without a single hiccup.·
I can produce similar symptoms by touching the XI side of the crystal.· To rule out a cheap crystal, I tried with the crystal that came on the development kit.· I've tried adding 47pf caps to ground at the crystal lines as well with no effect.· I do not think it is a problem with the crystal, but the fact that I can trigger similar problems to those I am experiencing is hard to dismiss.
My schematic is virtually identical to the development kit, just simplified.· I have omitted all DTR circuitry as it is not needed for my application (the device does not transmit), as well as any circuitry that does not deal with displaying serial data through VGA.
I did not write the software being loaded nor do I have the source code, but I can pass along ideas if this is likely a software issue.· As I can not reproduce any of the problems using the demo spin files included with the Propeller Tool, that is where I am currently leaning.
All·assistance with·this issue is greatly appreciated.
Periodically hsync pulses will be sent at uneven intervals and even split into two pulses for every one.· This is causing the image to distort, image blanking, and sometimes full resets.
Unfortunately I can not reliably reproduce this problem.· It will sometimes occur immediately after posting, sometimes immediately upon receiving the first few bytes of serial data, sometimes spontaneously while displaying a screen for a long for a long period of time, and sometimes it will go for hours without a single hiccup.·
I can produce similar symptoms by touching the XI side of the crystal.· To rule out a cheap crystal, I tried with the crystal that came on the development kit.· I've tried adding 47pf caps to ground at the crystal lines as well with no effect.· I do not think it is a problem with the crystal, but the fact that I can trigger similar problems to those I am experiencing is hard to dismiss.
My schematic is virtually identical to the development kit, just simplified.· I have omitted all DTR circuitry as it is not needed for my application (the device does not transmit), as well as any circuitry that does not deal with displaying serial data through VGA.
I did not write the software being loaded nor do I have the source code, but I can pass along ideas if this is likely a software issue.· As I can not reproduce any of the problems using the demo spin files included with the Propeller Tool, that is where I am currently leaning.
All·assistance with·this issue is greatly appreciated.
Comments
The fact that you can reproduce this by touching the crystal unfortunately tells us nothing. That can interrupt the system clock and the whole thing stops for as long as the interruption lasts. You can have a loose connection involving the crystal socket which would cause the same thing to occur, but other loose connections on your board can do the same thing.