This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TCA6424A: Floating inputs after reset

Part Number: TCA6424A
Other Parts Discussed in Thread: TCA6424

During the time between device reset and device programming, the TCA6424A ports are configured as inputs. In some applications, the output load could be a resistor and a series diode (as in an optocoupler LED). The leakage current of this combination could be such that the Vil of the TCA6424 may or may not be met thus causing the potential for high leakage on the port drivers (Vcc to GND) or perhaps oscillation.

The question is - is there any potential for damage to the device during the time between power on and when the ports are subsequently programmed as outputs? The time may be as long as, say 30sec for example but we'd also like to know if the state were held indefinitely, would any damage occur.

Figure 29 of the April 2014 revision indicates that resistors are required if the input pins are floating. If this is the case, then because of the default configuration, wouldn't resistors (or equivalent loads) always be required since there is always a time when the ports are configured as input after power-up?

Please advise!

  • Hey Steve,

    If I am understanding correctly, please correct me if this is wrong, you are worried the device could be damaged because the datasheet recommends the ports be pulled either high or low during start up (ports stay inputs for a while with no reference voltage). The concern is due to the ports needing to be outputs which connect to LEDs which customer does not initially want to place pull up resistors on. With diodes, some leakage current is expected and you believe the pins (Pxx) may even see oscillation.

    ^My comment here is if you are using diodes then the leakage current will slowly charge up a parasitic capacitor between the diode and our port pin. This means the voltage will slowly build up to Vcc of what the diode is connected to (I am assuming the cathode is connected to our port pin and not the anode). [This also does not take in consideration the leakage current from the FETs on the output of our device]. Probably won't see oscillation during this unless the diode is being turned on and off without the use of the port pin. Take this opinion with some salt as I am not an expert on diodes.

    Also, ViL does not matter in this kind of situation because we just want to set them to outputs and don't care about input value necessarily.

    "The question is - is there any potential for damage to the device during the time between power on and when the ports are subsequently programmed as outputs?"

    I would probably need to see a diagram/schematic to 100% confirm but if my understand of what you presented is correct then we should not see damage. This is assuming you are not drawing excessive current when the device is an output and when it is still being set up (an input) the device does not see voltages outside of it's absolute max. Our device is 5V tolerant so even if Vcc is GND as long as the device does not go over 5V i suspect we should be fine.

    "Figure 29 of the April 2014 revision indicates that resistors are required if the input pins are floating. If this is the case, then because of the default configuration, wouldn't resistors (or equivalent loads) always be required since there is always a time when the ports are configured as input after power-up?"

    -I was not responsible for writing this datasheet, however I suspect the writer stated this because floating inputs are tied to a gate (the input to the D Flip Flop in figure 19) and this may turn a FET on or off and will likely violate a parameter in the datasheet if left floating (supply current will likely change as it turns on and off) as our testing of the device was likely done with the pins pull up to Vcc.

    I do not think leaving the pins floating will damage the device but will likely draw more supply current and fall outside the datasheet specs.

    Thanks,
    -Bobby
  • The customer decided to use the pull-down resistors to be safe