MCT8316A: Frozen in MOTOR_IDLE state

Hi Marcel, 

Could you provide a capture of the sleep pin, AVDD, and VM when entering and exiting sleep mode?

Regards,

Joshua

Hello Joshua, 

I've done the measurements. It seems the MCT does not go into sleep -> just hangs?

Powering up:

The AVDD comes up right away, which would suggest the chip is not in sleep. It should be after 200ms, that is the time the WAKE pin should be kept low to enter sleep. 

Waking up:

The wake pin comes up, but nothing really changes in the MCT. 

To go (back) to sleep:

The AVDD pin stays high, also after 200ms or longer. 

Meanwhile, the MCT does not respond to anything anymore on the I2C bus. I't wont ACK any byte I am sending. I'll play with the drvoff function and wake pin to hopefully get some life back into it.

While looking into it using a logic analyzer, things do look ok

Thanks for your response!

Marcel

Hi Marcel,

Do these issues occur with multiple MCT8316As?

Are you using our EVM, If you are not would you mind sharing your schematic and layout?

Regards,

Joshua

Hello Joshua, 

I've happened to see this behaviour before on another prototype. I am not using the EVM. Here is the layout. 

The values for the decoupling and other capacitors are all from the reference design. 

I rather not share the schematics for this project (in public).

Note that this all started since I've set the DEV_MODE to sleep mode. Before, all worked fine. 

Thanks

Marcel

Hi Marcel,

I will need some time to look into this. I will aim to get back to you by next Wednesday

Regards,

Joshua

Hi Marcel,

Sorry for the delay in response. I sent you a private message on E2E, could you send the schematic though that private chat and we can continue this discussion from there?

Regards,

Joshua

Hello Joshua,

I've sent the schematics. Did you receive them? 

Thanks!

marcel

I have tried to tie the wake pin to the 3.3V psu to have the chip enabled right at boot, but it makes no different. The MCT still does not ACK I2C messages. 

It is not possible it's address has changed, right? 

The fact that it is still the same after a week of powerdown worries me a bit. 

Hopefully we can get some life back into it!

Thanks,
Marcel

Hi Marcel,

I did receive the schematics and have had a chance to look over it and nothing looks incorrect or that it could be causing an issue.

The address of the MCT8316A should not change from 0x00. 

Just to check, are you implementing a 100us delay between every byte transferred over the I2C bus?

If you replace the MCT8316A are these issues still present?

Was the AVDD voltage in the captures you provided measured at the pin?

Regards,

Joshua

Hello Joshua,

I'll do a swap of the MCT today and report my results. The AVDD measurement was done in the pin indeed.

Thanks,

Marcel

After replacing the MCT, I found it does *more*.

It tries to start the motor, but the default configuration is not capable. Therefor it tries to start the motor, but the motor only starts beeping. 

On the I2C bus however, still only NAKS are received. I've connected a scope at the MCT side on the I2C pins to make sure the voltage levels are correct.

They are. 

There is a bit of crosstalk, but that shouldn't be enough to have the whole chip go dead...

Note that it used to work before on this hardware. Isn't it strange a new MCT also does not ack?

Thanks,

Marcel

Hi Marcel,

I will need to look into this. I will aim to get back to you by next Wednesday

Regards,

Joshua

Hi Marcel,

Apologies for the late reply.

Can you check that the voltages of the AVDD, DVDD and BUCK are all near their expected values before you send any I2C commands?

AVDD should be around 3.3V, DVDD should be around 1.5V, and the Buck voltage should be above 3.3 with the voltage depending on BUCK_SEL.

If you have access to our EVM I would recommend attempting to use its MCU to communicate with the MCT8316A.

Regards,

Joshua