Original question:
We've had a couple motor drivers stop functioning with the state as reported in the thread title. It gives no other debugging information (no other bits are set in either of the fault registers). From probing it appears that logic and other supplies are ok. I imagine that we've done something to damage the IC, I'm wondering if anyone is aware of the scenario that might help us further debug the cause that would lead to this state.
Hi Lee,
I'm trying to get a better under standing of your problem. Can you clarify what you mean by address?
address: 2,
is this referring to control register 0x02?
and what do you mean by reg_out and reg_in? is this what you are writing/reading via SPI? When writing or reading a register there should be 10 bits.
Regards,
Yara
Hi Lee,
Thank you for clarifying your process!
A few things to mention
1.
I'm still a little confused and correct me if I am wrong but is your code printing out the SPI register bits in Hexadecimal or Decimal? Either way, some of these wouldn't translate to 10 bits
address: 2, reg_out: 1000, reg_in: 0
address: 3, reg_out: 1bff, reg_in: 3ff
address: 4, reg_out: 237f, reg_in: 37f
address: 5, reg_out: 2804, reg_in: 4
address: 6, reg_out: 3240, reg_in: 240
for example 1bff would be 0001101111111111 in binary, 16 bits, but it looks like you're only reading out 10 bits which would make sense because the device only knows how to read/write 10 bits. Make sure you are writing to the SPI registers using 10 bits binary, you can have your code take in hex or decimal as long as you are converting it to binary before you write it to the register. Use the following example of control register 2 as an example.
If I follow the settings I've selected and highlighted in Table 15 I would have to write the 10 bits you see in the blue squares to the register, 01010000000 in binary or 0x280 in hex (again you have to write in binary to the device).
2.
To address how to debug using SPI fault register is simple if you were to read out and print the bits from the register in binary. Look at the following table of fault register 1. If I read Fault Register 1 (0x00) and what came back was 0100100000 in binary (0x120 in hex). Looking at the table I can see bit 9 and 6 are set to 1, meaning in this scenario I get the following faults "Indicates VDS monitor overcurrent fault condition" and "Indicates overtemperature shutdown"
remember there are TWO fault registers, reading from both will be able to tell you a lot about what is going on.
Ultimately I have a feeling you're running into a SPI communication error. Trying to write more than 10 bits to your SPI registers will have you configuring certain registers unpredictably.
Let me know if any of this information helps
Regards,
Yara
This is sort of a different direction than my original question. I'm still wondering what causes the driver to get stuck in this fault only state.
But on the driver strength that is an appreciated point and good faq/app note. By PWM I meant the SH signal. That generally looks fine. Looking at SL and GL do look a bit troublesome. There is significant undershoot on GL/SL when the high side fet turns on, until I bring the high side gate drive strength down to 80 mA. I'd rather have a little bit less switching losses though. Are there any solutions to this undershoot problem? Do you think negative voltages on GL would cause this failure of the driver IC?
Yellow is GL at 1 A high side gate strength.
Now at 80 mA.
