I have never used or seen a BDM Lightening board from P&E so please check with P&E any information I provide.
I have had a similar problem with a 5235 board and a P&E Rev f parallel port pod. Actually a number of them. I traced the problem down to the write to the clock control register that increased the clock speed. I was using the BDM software on Sourceforge and having written most of that software I knew the code to be ok. The software has been working for many years.
I spent many hours looking into the problem and could find no specific reason. I contacted P&E who provided me with the standard answer of run the test code. Their test code passes. This is to be expected as the test software does not increase the clock speed and the BDM interface works at the slower speed. I received an engineering answer and it is listed in the ticket I reference below. I will not post all those details here. In short they suggested I try the Cyclone product. I have to admit I ended up being disappointed with the outcome so ported the TBLCF support to the BDM project. This took less time than the time I spent on this issue.
The P&E support ticket reference is:
http://www.pemicro.com/SRS/view_reply_SR.cfm?issue_id=8330For the record I checked the BDM timing with a digital scope and logic analyser and it showed no issues. I did not probe around the BDM pod and did not look at specific voltage levels. I also tried all the clocking options in the pod. Nothing provided a specific solution. We had a number of the 5235 boards and pods around the world and we could not see a pattern. We even shipped board from a working system to a failing system and it did not solve the problem or provide any extra information. My only guess is a power supply and/or PC issue maybe even something related to grounds. This is all not important to me any more as the Axiom TBLCF pod I have works very well.
Regards
Chris
