A strange problem of TMS320F2812 caused by an eCAN bug

Aguilera

A strange problem of TMS320F2812 caused by an eCAN bug [Copy link]

Recently, when I was adjusting F2812 to implement the CANOpen protocol, I found a strange problem. When reading the CAN message data registers MDL and MDH of F2812, the reading was 0, as shown below. The local variables DataL and DataH have a probability of having a value of 0. In desperation, I added a read-back judgment statement, which greatly reduced the probability, but did not fundamentally solve the problem, and I could not explain the problem.

After consulting with TI's technical staff, we finally found the explanation and solution of the problem from the errata sheet, as shown below:

This means that when the eCAN module accesses the eCAN register, if the CPU reads the eCAN register at the same time, the reading will fail. Our phenomenon is described in Case 1.

In the errata table, solutions and reference codes are also given

For Case 1, an additional judgment statement is required to determine whether the read value is all zero. If it is all zero, then the second step of reading is performed. At the same time, TI provides a reference code solution to add a secondary readback function as follows:

Attach the code link:
http://www.ti.com/general/docs/lit/getliterature.tsp?literatureNumber=sprc180&fileType=zip

sanhuasr

Thanks for sharing and learning

dongfanghanke

What specific model are you using? I sent you a private message, but I don’t know if I can send it to you.