AnsweredAssumed Answered

DSP56F807: What is the reason for undefined voltage levels of GPIO-Pins with 30kohms pulldown at low temperatures

Question asked by Patrick Sebbes on Feb 24, 2016
Latest reply on Feb 25, 2016 by xiangjun.rong

I'm using the GPIOA5…GPIOA7-pins of the DSP56F807 as an input for detecting a high level (3.3V) or low level (DGND), being the output of a 74HC1G14 which is an inverting Schmitt Trigger. To protect the circuit there are 30kohms as a series resistance between the input of the DSP and the output of the Schmitt Trigger, nothing else. If the output of the Schmitt Trigger is high, everything is fine. There are 3.3V at the output and at the input of the DSP as well.
But if the output is 0V we have a very strange behavior at low temperature (below 0°C): There are still 1.6V at the input of the DSP. The internal Pullup is deactivated by the corresponding bit in the register (activating raises the level to around 2.2V which is plausible).
I do not understand where this level comes from. In the Appnotes I cannot see any reason for that (like figure 7-4 of the DSP56F800 user manual Rev. 8). For me it seems that the internal buildup of the input causes a residual voltage that is not defined in the data sheet (without pullup/pulldown the current should not exceed 1µA). I can see it at other floating ports where a voltage of 2.5V can be measured.
Now my questions: What is the reason for this behavior? Is it because of a resistor in the positive feedback of the Input? Is there a minimal load needed to get 0V at the input? How should the external impedance, i.e. the series resistance in my case, look like to guarantee me a defined low level at the input of the DSP under every operating condition? Is it possible to get a scheme of the internal buildup to get more understanding of it?

Thanks for your assistance.