TMS320F280049C: GPIOs pulled up when start up or reset

Hello Pawel,

By default not only do GPIOs have pull-ups disabled, but they are inputs. They should not be outputting anything. Can you please tell me if you have anything connected to these pins with any sort of pull-ups on them? The device itself should not be exhibiting this behavior for GPIOs. Is there any code loaded to Flash that might be causing this? Does this pulse happen at even a small delay after reset? Please show a scope waveform demonstrating this for clarification.

There is a transistor circuit connected to the pin:

With this I have the pulse presented below:

(yellow - 3.3V supply, magenta - GPIO16)

When i disconnect the transistor circuit (by removing R162) It looks like this:

There are different voltage levels depending on whether there is transistor circuit or not so it seems that there is an internal pull-up that creates a voltage divider with R162 and R164. The code and transistor circuit are the same for few others GPIOs but I have this problem only for the three i mentioned.

I have noticed that the problem also occurs during turning off power supply:

Pawel Witaszek said:

When i disconnect the transistor circuit (by removing R162) It looks like this:

Please try disconnecting from the pin entirely, my point was to isolate whether the problem is the device or the connected circuit. If you leave the GPIO floating (nothing is connected at all) and you still see the issue, let me know. Otherwise, this is something caused by what is connected to the GPIO itself.

Pawel Witaszek said:

There are different voltage levels depending on whether there is transistor circuit or not so it seems that there is an internal pull-up that creates a voltage divider with R162 and R164.

If you can replicate this without anything connected, let me know. I'm not sure what hardware on the F28004x would somehow override the default pull-up disable configurations to instead enable it AND make the GPIO an output instead of a default input for the power-on reset. There is nothing particularly unique about these pins that would cause a problem, but if you're able to replicate this problem without anything physically connected to the pin then let me know and I can bring it up with the design team.

At the second waveform, after removing R162 there is nothing connected to the pin. GPIO16 is completely isolated.

Okay, I will contact the design and validation team to see if they know what may be causing this issue. I don't know if this is something they'll be able to replicate, but I'll let you know what I find out by early next week.

One other follow-up that someone reminded me of, was there a possibility that any of the pins could be damaged? This can happen due to over-current/voltage. Also, have you already tried testing on the device when nothing but power is connected to it (i.e. remove all other connections it may have)? Does this happen on all devices you have or just this one?

I've tried to test it with only power connected, situation is the same. I checked it on each of 5 prototypes we have and they behave identically.

Hello Pawel,

Omer Amir said:

was there a possibility that any of the pins could be damaged? This can happen due to over-current/voltage.

Can you confirm that there was no over-current/voltage on the pins? Also, would it be possible to show the physical setup of your device? I want to understand how the part is being tested and what, if anything, is in physical contact with any of the pins.