GPIO pullups and direction

RE: GPIO pullups and direction

JohnBrown — Sat, 06 Jun 2015 13:03:56 GMT

No problem. Thanks for confirming things.

John

RE: GPIO pullups and direction

Susheel Nuguru — Thu, 04 Jun 2015 12:39:14 GMT

I have this confirmed and to my ignorance, I was testing it all wrong. I did two mistakes, I choose a GPIO pin that is connected to an onboard LED and my logic analyzer trigger was not sampling fast enough .Sorry for that John. The pullup configuration stays when you change the Pin direction.

RE: GPIO pullups and direction

Susheel Nuguru — Wed, 03 Jun 2015 19:15:48 GMT

This is very confusing, Logic analyzer should have caught that. I will ask the architect now to confirm on this and will update this thread tomorrow.

RE: GPIO pullups and direction

JohnBrown — Wed, 03 Jun 2015 14:08:08 GMT

Well we'll have to agree to disagree on that, then. I haven't connected a logic analyser, but I have used a low current LED with a 220R resistor in series, and I am getting a dim glow, consistent with a pullup being enabled.

Using exactly this code(LED is on P0.17)

   NRF_GPIO->DIRCLR = (1UL << 17);		// Set as ip
    NRF_GPIO->PIN_CNF[17] |= GPIO_PIN_CNF_PULL_Pullup << GPIO_PIN_CNF_PULL_Pos; // Set pullup

    SEGGER_RTT_printf(0, "PIN_CNF = %08x\n\r", NRF_GPIO->PIN_CNF[17]);
    nrf_start_timer_ms(2000);	// Set timeout at 2 seconds

    	while (NRF_TIMER1->EVENTS_COMPARE[0] == 0)
    	  {
    	  }


    NRF_GPIO->DIRSET = (1UL << 17);		// Set as op
    SEGGER_RTT_printf(0, "PIN_CNF = %08x\n\r", NRF_GPIO->PIN_CNF[17]);
    nrf_start_timer_ms(2000);	// Set timeout at 2 seconds

    	while (NRF_TIMER1->EVENTS_COMPARE[0] == 0)
    	  {
    	  }


    NRF_GPIO->DIRCLR = (1UL << 17);		// Set as ip
    SEGGER_RTT_printf(0, "PIN_CNF = %08x\n\r", NRF_GPIO->PIN_CNF[17]);

    while(1);

John

RE: GPIO pullups and direction

Susheel Nuguru — Wed, 03 Jun 2015 13:43:26 GMT

my bad, i just typed in the code in this answer, but when i tested i used DIR register. And to elaborate, the PIN_CNF still reads that the pullup configuration stays but I don't see that really effecting the pin state on logic analyzer.

RE: GPIO pullups and direction

JohnBrown — Wed, 03 Jun 2015 12:12:25 GMT

Sorry, the code tages got screwed up there! But as you can see, if you use DIRSET and DIRCLR correctly, the pullup does seem to persist.

John

RE: GPIO pullups and direction

JohnBrown — Wed, 03 Jun 2015 12:10:11 GMT

NRF_GPIO->DIRCLR = (1UL << BUTTON_NORTH); // Set as ip NRF_GPIO->PIN_CNF[BUTTON_NORTH] |= GPIO_PIN_CNF_PULL_Pullup << GPIO_PIN_CNF_PULL_Pos; // Set pullup

SEGGER_RTT_printf(0, "PIN_CNF = %08x\n\r", NRF_GPIO->PIN_CNF[BUTTON_NORTH]);
NRF_GPIO->DIRSET = (1UL << BUTTON_NORTH);		// Set as op
SEGGER_RTT_printf(0, "PIN_CNF = %08x\n\r", NRF_GPIO->PIN_CNF[BUTTON_NORTH]);
NRF_GPIO->DIRCLR = (1UL << BUTTON_NORTH);		// Set as ip
SEGGER_RTT_printf(0, "PIN_CNF = %08x\n\r", NRF_GPIO->PIN_CNF[BUTTON_NORTH]);

while(1);

0> PIN_CNF = 0000000E 0>
PIN_CNF = 0000000F 0>
PIN_CNF = 0000000E 0>

RE: GPIO pullups and direction

JohnBrown — Wed, 03 Jun 2015 12:06:42 GMT

I think you are using the wrong register. It should be DIR, rather than DIRSET. I just wrote the following snippet of code:

NRF_GPIO->DIR &= ~(1UL << BUTTON_NORTH);		// Set as ip
NRF_GPIO->PIN_CNF[BUTTON_NORTH] |= GPIO_PIN_CNF_PULL_Pullup << GPIO_PIN_CNF_PULL_Pos; // Set pullup

SEGGER_RTT_printf(0, "PIN_CNF = %08x\n\r", NRF_GPIO->PIN_CNF[BUTTON_NORTH]);
NRF_GPIO->DIR |= (1UL << BUTTON_NORTH);		// Set as op
SEGGER_RTT_printf(0, "PIN_CNF = %08x\n\r", NRF_GPIO->PIN_CNF[BUTTON_NORTH]);
NRF_GPIO->DIR &= ~(1UL << BUTTON_NORTH);		// Set as ip
SEGGER_RTT_printf(0, "PIN_CNF = %08x\n\r", NRF_GPIO->PIN_CNF[BUTTON_NORTH]);

and the output on the Segger RTT shows the PIN_CNF rgister to be: 0x0000000e 0x0000000f 0x0000000e

So I believe the pullup does persist if you use the DIR register to chage direction.

I will re-write this to use the DIRSET and DIRCLR registers...

while(1);

RE: GPIO pullups and direction

Susheel Nuguru — Wed, 03 Jun 2015 11:28:28 GMT

I think this came as a side effect of being able to set direction either from DIR register or in PIN_CNF register. i will bring this to notice of right people.

RE: GPIO pullups and direction

JohnBrown — Wed, 03 Jun 2015 11:16:43 GMT

Thanks for the answer, but I have to say that sounds like a crazy system to me! Some things about the nRF51822 are great, others, like this and the compromised PWM capability, are not so great.

RE: GPIO pullups and direction

Susheel Nuguru — Wed, 03 Jun 2015 09:54:12 GMT

I have tried the both ,

NRF_GPIO->DIRSET |= (1UL << LED_1);
NRF_GPIO->PIN_CNF[LED_1] |=  GPIO_PIN_CNF_PULL_Pullup << GPIO_PIN_CNF_PULL_Pos;	

NRF_GPIO->DIRSET &= ~(1UL << LED_1);

NRF_GPIO->DIRSET &= ~(1UL << LED_1);
NRF_GPIO->PIN_CNF[LED_1] |=  GPIO_PIN_CNF_PULL_Pullup << GPIO_PIN_CNF_PULL_Pos;	

NRF_GPIO->DIRSET |= (1UL << LED_1);
while(1);

and the pullup wasn't there. So i am concluding that it will be reset to no_pull everytime we change pin direction