I just got this new SmartFusion Evaluation board from Actel. On it is a cool chip that a combination FPGA, analog processor, and Cortex-M3 microcontroller. The board has a built-in JTAG programmer that uses the Actel toolchain to download new designs into the chip, which is pretty nice. However, you can’t debug the software on the microcontroller with it.
So I decided to see if I could get OpenOCD working with my Olimex JTAG-Tiny USB JTAG tool, since there was a 20-pin ARM JTAG header on the board as well. On the first go, it didn’t work, so I decided to look at the signals on my ‘scope. Turns out that the TRST and TCK lines were wiggling, but only at about 1V when they should have been at around 3.3V. I downloaded the schematics from the Actel site, and this is what I found:
Turns out that those lines had very strong pull-downs on them. Coupled with the 330 ohm, series output resistors, these pull-downs created a voltage divider that was pulling the signal down to around the 1-2V level.
So I fired up my soldering iron, and replaced the resistors with some 4.3k ohm resistors that I had near my desk. And after hooking everything back up, I was able to halt the processor and look around in memory.
Hopefully I’ll be able to get the eNVM flashing over JTAG working in OpenOCD soon.
TL;DR: Remove R14 & R15 and replace with >5k ohm resistors.






