by Guest » Sun Feb 20, 2005 9:12 am
Hi Tim,
Finally back here in the states...
I powered up a version 1.2 AVRcam when I got back, and re-verified that all the commands worked, and they did. I was able to dump frames as well as track objects.
I then uninstalled the AVRcamVIEW software, and installed the latest one that was on the CD that shipped to you, since this was the only difference between my setup and yours. This was also successful in performing both the frame dumps as well as object tracking.
Couple of more questions for you:
1) You said that you do receive the ACKs after the DF/ET command, right? If this is the case, the AVRcam has entered into ether dump mode or tracking mode (depending on which command is sent). In these modes, the system is dependent on a few critical signals generated by the OV6620. Namely, the VSYNC, HREF, and PCLK signals are used to sample pixels, determine when a line of pixels is complete, and several other tasks. If any of these lines were not connected for some reason, the system would appear to hang after the commands were sent. I just checked the continuity on my board, and everything appears connected (which makes sense since I was able to get the proper response from the system). Try ohmming out the connections of these lines between the camera module and the mega8 to make sure that they are connected properly (see the schematic for the exact connections/pins). Also, verify that they are not tied down high or low permanently (accidental short to ground or something). If you have access to an oscilloscope, check to ensure that these lines are changing states when doing a frame dump.
2) What kind of power source are you using to power the AVRcam? I have seen very bizarre behavior with the system when I was accidentally powering it with less than 5V going into the 5V regulator (I was feeding it around 4.4V). Can you verify that, when powered up, your board is getting at least 5.5V at the Vin terminal?
Oh yeah...the tiny12 is used to turn on the external clock output of the OV6620 so that the mega8 has a clock source. It sends a couple of I2C commands to the OV6620, and then doesn't perform any further useful activity. I posted this in another thread as well, and will be getting the source code for this up on the site soon.
Let me know how these tests turn out...I'll keep thinking about what the culprit could be.