GPIO Current Drive

[lance338]

My application uses the nRF51822 with a GPIO configured as output and another GPIO is configured as input. Connected to the GPIO_OUT is a resistive load (Rload) which then connected to GPIO_IN. When GPIO_OUT drives a '1' and if the load has low resistance, we would expecte a '1' at GPIO_IN.

GPIO_OUT ------------------ Rload -----------------------GPIO_IN 

Per spec, the VDD is 3.3V.

With GPIO_PIN_CNF_DRIVE is set to S0S1, the spec say this would drive 0.5mA. The max Rload measdured to be around 6K ohm.

When GPIO_PIN_CFG_DRIVE is set to H0H1, we get the same result.

I would expect the high-drive to allow for more resistance on the load. 

Is this correct? Why would the S0S1 and H0H1 have the same result?

[awneil]

The way you use the word "load" is suspicious.

The connection you've shown would not normally be described as a "load":

lance338 said:

GPIO_OUT ------------------ Rload -----------------------GPIO_IN 

That is simply a resistor connecting the OUT to the IN.

As Kenneth said, the IN has a very high impedance - so virtually no current wil flow through the "load"

But does that diagram fully represent your situation?

Why are you calling this thing a "load" ?

Are the OUT and IN pins on the same nRF51822 chip?

Can you post a full schematic of your setup?

Or some good, clear, in-focus photographs.

[lance338]

Thanks for your response. I used the word load loosely. Let's call them external resistor whose resistance varies depending on the amount  of chemical substances it interacted with. The circuit is exactly as shown. Both gpios or on the same nrf51822.  For our purpose, if there is no chemical substances, the external resistor is 1M ohm. With chemical present, the resistance is 20K ohm. 

[awneil]

So are you actually trying to measure the resistance, then?

Or just detect whether the chemical is present ("low" resistance) or not ("high" resistance)?

[lance338]

Just detect whether chemical present/not present.

[awneil]

So you want to detect whether the resistance is "high" (1M) or "low" (20k)

The arrangement you showed will not do that.

What you need is a potential divider, and user either the ADC or Comparator to look at the voltage and, thus, decide whether the resistance is "high" or "low".

[lance338]

Are you suggesting to use the GPIO as analog with the LPCOMP logic? I came to that conclusion as well. If that is the case, can you give/point to more details on how to set it up?

[awneil]

Yes.

I would read the section on the comparator in the Product Specification, and then look for an example in the SDK examples.

Note that the nRF51 chips are not recommended for new BLE designs.

[lance338]

Thanks. We will migrate to nRF52 eventually.

BTW, you said we need to use a potential divider. Can you please elaborate?

[awneil]

You need a hardware engineer on your team!

https://en.wikipedia.org/wiki/Voltage_divider

[lance338]

I know what a voltage divider is, just don't understand how that is required for the analog comparator (which admittedly I just learned about today)

[awneil]

You really do need a hardware engineer on your team!

The analogue inputs see a voltage. So you need some way to translate the variable resistance of your sensor into a voltage for the analogue input ...

This is basic electronics - not specific to Nordic. 

https://en.wikipedia.org/wiki/Voltage_divider#Sensor_measurement