TMS320F28031: ADC abnormal behavior when connecting AIO to ground.

Young,
This does sound strange, since these pins are shared with both ADC, Comparator, and AIO; I'd like to eliminate some variables, can customer try/confirm the same behavior happens when the AIO functions are disabled, as well as make sure comparators are off?

i.e. the only time the ADC is used is to sample B1; but still flip the switches to see if the problem happens again.

I'm wondering if somehow the customer has enabled the output of the AIO and is driving against ground and we get too much current sink at 3 IOs to ground. I think the easiest way though is to eliminate some variables first.

Will see what customer finds.

Best,
Matthew

Hi Matt,

Thanks for your reply. I have done more testing as follows.

1. turn off DIP1~DIP8, which means all the ADC pins are pulled high with 3.3K resistors.  A voltage probe is used to monitor the voltage of ADCINB0. ADC_SOC example code is used to test customer's board. Comparator's CLK is disabled.

2 After powering on, ADCINB0's voltage is 3.3V. When PC run to the following code(Band Gap enabled), ADCINB0's voltage is 3.08V.

3. After powering on ADC reference, ADCINB0's voltage is 2.46V.

4. After powering ADC module, ADCINB0's voltage is 2.23V.

5. Whether enabling or disabling AIO, the testing results are same.

6. If removing all the pull-up resisitors except DIP8's and AIO pins are floating, ADCINB0's voltage changes to 1.83V. And ADC sampling result is fixed to 4095.

7. Other doubtful points: TEST2 pin is connected ground directly, VREFHI is connected to ground directly.

Thanks.

BR,

Young

Young,
I'm checking on TEST2, as you said datasheet requires this pin be left unconnected at all times. I believe from above customer is using the 64 pin package where VREFHI is a shared pin with ADCINA0 so if this is tied to VSSA it is OK as long as external reference is not used.

Does this happen on multiple devices, or just verified on one device, from the above I'm considering that the device may have been damaged.

Best,
Matthew

Hi Matt,
Yesterday customer tested 50 pcs PCBA, and all the boards had the same issue. And issue is "If more than 3 AIO are connected to ground, ADC result registers are 0x0FFF, and can trigger ADC interrupt normally. "
I have gotten one board, and will start to test in TI lab.
Thanks.
Young

Young,

Is there a single ground on the board or is it split between analog and digital ground?

-Tommy

Hi Tommy,

There is a single ground on the board.

Thanks.

BR,

Young

Hi Tommy/Matt,

Today I tested the code using TI F28035 ControlCard, and all the ADC pins were connected to the ground directly except ADCINB0. And ADCINB0 was pulled high with a 2K Ohm resistor. The example code can work well on ControlCard. 

Then I removed all the DIP switches of customer's board. After that all the ADC pins would be floating. And all the ADC sampling results are 0x0FFF, even if some of ADC pins were connected to GND.

At last, I checked the Boot ROM, and found both of customer's F28031 and Controlcard's F28035's OTP checksum were 0xFC00

Would you kindly help to check the root cause?

Thanks.

BR,

Young

Tommy,

Please find the customer's F28031's register contents as follows

Regards,

Young

Tommy,

Please find the TI's F28035 controlcard's register contents as follows

Regards,

Young