02-09-2011 01:57 PM
I hooked up the kit right out of the box and got the "Detected Video Resolution 0x0" message. I followed the troubleshooting guide and checked pin PG_C2M and saw 2.5 V, which was the only thing the guide said to check.
1. If I toggle the test parttern from the main menu, I get the colorbars but then I can't changed to any other test pattern.
2. Resetting the FPGA and re-starting the app still leads to a 0x0 message
3. The board is hooked up to a Dell 10" monitor
11-28-2011 09:14 PM
The FPGA logic expects the image sensor to generate embedded BT.656 codes,
which includes codes to generate VSYNC.
If there are no embedded BT.656 codes, there will not be any VSYNC.
Without a VSYNC, the FPGA logic will fail to determine how many lines of video for each frame.
This is the most probable cause of your symptom: 1280x0
Can you describe your setup ?
- which reference design are you using ?
- which connector is the OV9715 connected to ?
- CAM1 (ie. CON300)
- CAM2 (ie. CON301)
- which camera is the software configuring ? does it match the hardware setup ?
11-28-2011 10:45 PM - edited 11-29-2011 01:19 AM
i am running the modified reference on ml605.
i've changed some of the settings :
fmc_iic_sw v2.01a - needed to use this else it would be stuck in while loop
in the main() here are some changes made
FMC_ID_SLOT1 to FMC_ID_SLOT2
other than that was the changes from 's6' to 'v6', all other fimware codes were copied from the 13_1 version
camera connected to CON300
hope i didnt miss anything else
looking at the fmc_imageov_camera_in using chipscope, i do see the CCIR656 pattern,
3ff 000 000 200 d0 d1 d2 .... 3ff 000 000 274, so i believe the ccir656 format is valid from the camera,
11-29-2011 05:39 AM
Chee Kin Tang,
Thank you for checking the embedded CCIR656 codes.
This confirms that the image sensor has been configured by your FPGA design.
The next probable cause may be that one of the VSYNC/VBLANK signals is disconnected somewhere between the fmc_imageov_camera_in and ivk_video_det pcores.
Can you send me your MHS file ? This is something I can check quickly.
11-29-2011 07:36 AM
well, for the problem, i've solved it, i was missing the hsync and vsync on the gamma ports. they weren't available for connection in the GUI in xps. had to check in the mhs
a new problem is the hdmi, nothing appearing on it, time for more debugging tomorrow.
11-29-2011 12:06 PM
Chee Kin Tang,
Good job !
For the DVI output, make sure that the "enable" port is active high.
The FMC IPMI software library uses the sg_i2c_controller pcore's GPIO as follows:
-  => fmc1_enable
-  => fmc2_enable
If you changed the FMC module location, please check the code for the
fmc_ipmi_disable( &fmc_ipmi_iic, FMC_ID_SLOT1 );
fmc_ipmi_disable( &fmc_ipmi_iic, FMC_ID_SLOT2 );
11-29-2011 06:35 PM - edited 11-30-2011 01:00 AM
thanks for the hint, there is output now on the monitor.
i had to use FMC_ID_SLOT2 for detect, and FMC_ID_SLOT1 for the enable/disable section.
colors are good for the camera, but i have further question,
the image output is rounded/fisheye output on LCD - i think its due to the lense issue that it has fish eye outputs, will have to source for some other lense, any recommendation ?
and is flickers/shifts, is this a frame issue ? it is shaking violently.
further difference in my system is the system bus is running at 100mhz instead of 50mhz. i dont think this is the problem, correct me if i am wrong.
edit. i am using the video clock labeled VIDEO_CLK_S6 (G3 on FMC) in the schematics, and this could be the problem when i am trying to use it on the ML605. will try correct this and give it another go to see if it solves the problem.
nope, still have the problem. i had to use
NET "fmc_imageov_video_clk_pin" CLOCK_DEDICATED_ROUTE = FALSE; to be able to compile without error.