Hi,
Just wondering if anyone has used, or tried to use this device yet. I have a design I did for the ProAsic+ that I converted to use the PA3 as a test. Went fairly well. Only real issue I had was that when I set the option for Designer or move flip flops to the IO cells, it did not implement them correctly. Specifically the async resets where the wrong sense. Since The device I am targeting is not in production yet so I can only simulate the back annotated design. This is where I discovered the problem so I do not know if the problem is with the simulation model or with designer.
I had a couple of problems with the Plus and notice some changes to the PA3 data sheet; primarily concerned with the JTAG TRST pin. Under the pin description section, They "recommend" the following:
"TRST Boundary Scan Reset Pin
The TRST pin functions as an active low input to asynchronously initialize (or reset) the boundary scan circuitry. There is an internal weak pull-up resistor on the TRST pin. In the operating mode, a 100 ? external pulldown resistor should be placed between TRST and GND to ensure that the chip does not switch into a different mode."
I had a power-up problem on some of the Plus device and found out that if I ground the TRST pin , the device would start working. I report this to Actel and had them evaluate the parts that exhibited the problem. There recommendation: "ground the TRST pin". Sounds to me like there is a problem with the TAP port on both the Plus and PA3 devices and the
100 ? resistor is the "patch" to fix it. I am curious if anyone else has had problems with the TRST pin.The other problem I have had is a high programming failure rate while using the FlashPro programmer, mostly exit 11 errors. Actel was not able to helps us solve this problem. We did not press them on this since eventually we would be getting the devices programmed by our supplier and it would become a non-issue after that. The curious thing is that if a device successfully programmed the first time, it would more then likely always program successfully. I have reprogram a single device 50 to 60 times with no problem. I suspect a marginal problem with the device itself or a problem with the programming algorithm and not with my programming fixture.
It bothers me that Actel will not admit problems with their devices. Xilinx has no problem with admitting problems with devices and then publishing a work around to the problem until a permanent fix to the silicon is implemented. Why is Actel reluctant to do this. Maybe this problem with the TRST was already know to them, and if they had published an errata on this then maybe I would not have spent over a week debugging this problem.
Why do I use Actel if I am unhappy with there devices? the truth is it is the only reprogrammable FPGA that fit the application.
Would like to hear about any experiences that other people have had with the Actel Flash parts.
Thanks Dave Colson