Boundary scanning on 8M Mini eval board is failing 50% of the time

[adevries]

I posted this question as a reply to another question, but since that question was already marked as answered, I'm not sure it will get seen.

Has anyone else noticed the current consumption of the eval board dropping or rising significantly during boundary scan testing? I'm trying to do some boundary scanning on my 8M Mini EVK and I've run into this problem. I will power up the eval board, and run a test that tells me if two pins I manually shorted are connected. The boundary scan will tell me they are connected (the correct answer) at which point the current reading on the 5V power supply will drop from 120mA to 55mA. I will then run the same test again, and this time it will say the two pins are not connected (incorrect answer). At the completion of that test, the current will go back up from 55mA to 120mA. If I run the test a third time, the results are the same as the first test, with the current dropping back to 55mA. This cycle repeats forever. I am able to consistently reproduce this on the 8M Mini EVK, so I'm curious if this is a known issue with the processor, or if anyone else has encountered this problem.

I have seen this same behavior on board I designed with the 8M Mini on it. In that case, the drop in current was due to VDD_ARM_0V9 and VDD_SOC_0V8 not drawing current anymore. The PMIC didn't report a fault, and the voltage levels seemed fine, but VDD_SOC_0V8 went from drawing 205mA to 10mA. VDD_ARM_0V9 went from drawing 80mA to 5mA. I'm betting the same thing is causing the current drop on the EVK, but I cannot figure out why this is happening. 

When I run the boundary scan on the eval board, I verified that the configuration bits are set according to the hardware development guide: SW1101 [1-10] = 1111110001

[Yuri]

Hello,

  I think, that some side effect just for the custom board takes place in the case.

Have a great day,

Yuri.

-------------------------------------------------------------------------------

Note:

- If this post answers your question, please click the "Mark Correct" button. Thank you!

- We are following threads for 7 weeks after the last post, later replies are ignored

Please open a new thread and refer to the closed one, if you have a related question at a later point in time.

[adevries]

Hi Yuri,

I don't really understand your response. The issue I'm describing occurs on both the i.MX 8M Mini evaluation board as well as the custom board I made. But I would be fine ignoring my custom board and just figuring out why this problem occurs on the evaluation board. Whenever I attempt to do boundary scanning on the eval board and determine if two pins are connected, the test is failing every other time. Based on this information and the information I provided above, do you know what could be causing this failure on the EVK? Was boundary scanning successfully tested on this board? And if so, what JTAG controller did they use to perform the tests?

Thanks.

[Yuri]

Hello,

  What does it mean: "two pins .. manually shorted"?

Is this correct to short the pins?  Outputs must not be shorted. 

Regards,

Yuri.

[adevries]

Hi Yuri,

I manually shorted TP16 and TP17 together when I doing the boundary scan tests. It was correct to do this because when you do boundary scanning the pins are configured as inputs by default, and then one of the lines is configured as an output. So one pin is an output, and it drives the second pin which is configured as an input.

However, even when I didn't have two pins shorted together the weird behavior continued. When I used the boundary scan tool to see if two unconnected pins were connected, the current consumption would rise as fall as I described in my original question.

Do you know if the boundary scanning feature was tested on the 8M Mini eval kits? And if so, what JTAG controller did they use when testing it?

Thanks

[adevries]

It almost seems like the boundary scanning is somehow causing the cores to be held in reset. Is it possible for that to happen?