I have just come across an issue with programming a dspic30f5011 in the MPLab IDE (v8.83) using the ICD3. I have programmed these chips previously and have been successful. Nothing new has been added to the project or the circuit. I have attempted to do 10 of this board and have had a 70% fall out rate for the following error...  
The following memory regions failed to program correctly:
Program Memory
Address: 00000000 Expected Value: 00040100 Received Value: 00ffffff
Programming failed
 
If I try to read the chip it completes fine but all the locations are 0xFFFFFF. Its like it will not program. As I said, I believe nothing has changed from the last time I have programmed this chips (famous last words), and why would three out of the ten program fine?  I have looked at the configurations bits and they seem fine. I have tried to program another board (DSPIC30f4013) with the same tools and it works fine. Has anyone else seen this issue or have any suggestions on what to try next.

Show your real and complete schematics.

OK, I cant show the entire schematic, but I can show the connections to the microchip. I realize this is a very vague question. I know there are probably a few different things that could be going on. It is probably something I am doing, but I just dont understand how some can work and some dont. Thanks for your help.

Can't see almost anything. Repost a readable picture.

Agree with MBedder. The picture is too small and blurry to read.
Has this ICD3 been connected to a PC running MPLABX?
If yes, it will have updated the firmware to a nwere version incompatible with MPLAB8. You would need to force MPLAB8 to "update" (really downgrade) the firmware back to a version it is compatible with.

Gentlemen,
Again thank you for looking at my question, and excuse my tardiness in response to your request. I had been put on a higher temperature fire for the last day and a half. I believe I have fixed my issue, maybe someone can educate me specifically as to what the real reason is. As I stated, I was getting the error that looked as if I was not writing to the chip. I performed a read of the chip and saw that the first location did read all F's as stated. Then I looked further down in the code and found op-codes at different addresses. So that looked like some things were programming. So I started looking at the configuration bits. I found one section that I had not defined each possible option. I had done this because I thought that the IC came with all those locations defaulted to 0xFFFF. Once I set my config section __FBS to include  WR_PROTECT_BOOT_OFF, I was able to get a complete program and verify. I guess it was a function of that dang old "never assume" thing. Stay Safe