Loading on Beaglebone Black Boot Pins?

[Greg]

I'm trouble shooting a problem on a simple cape for a BeagleBone Black.

I have a bipolar switch for an LED, very similar to that used for the "User LEDS".

So I have added an approximately 10Kohm load to one of the "boot pins", which

in header lingo is P8_44.

P8_44 is also one of the "boot" pins.  I'm interesting in using the PRU connection mode of this pin.

The System Reference Manuals says the boot pins must not be driven prior to coming out of reset (SYS_RESET goes high).

So maybe my understanding of the term "driven" in this context is incorrect, and it could also be interpreted as "loading",

as in excessive resistive loading?

I tried an experiment on both BBB and BBG with a 10kohm resistor to ground on P8_44.  Both refused to boot up and run.

So is it a requirement to keep resistive loads off this pin during the boot process?

Regards,

Greg

[Chris Morgan]

Yes.

You should avoid applying any voltage to any pin before SYS_RESET goes
high, or you'll likely permanently break something. In your case it
looks like you are changing the boot pin configuration (maybe it
defaults to an internal weak pullup?) and preventing your system from
booting. If that isn't what you intended you'll want to avoid loading
the pin until after SYS_RESET goes high.

What are you trying to accomplish with your cape and that io pin?
There are approaches that may let you both use the pin and avoiid
changing the boot strapping but it would really help to see the
circuit you are trying to use with that pin.

Chris

[Harvey White]

On Mon, 18 Apr 2016 17:12:37 -0700 (PDT), you wrote:

>I'm trouble shooting a problem on a simple cape for a BeagleBone Black.
>
>I have a bipolar switch for an LED, very similar to that used for the "User
>LEDS".
>So I have added an approximately 10Kohm load to one of the "boot pins",
>which
>in header lingo is P8_44.
>
>P8_44 is also one of the "boot" pins. I'm interesting in using the PRU
>connection mode of this pin.
>

>The System Reference Manuals says the boot pins must not be *driven* prior

>to coming out of reset (SYS_RESET goes high).
>
>So maybe my understanding of the term "driven" in this context is
>incorrect, and it could also be interpreted as "loading",
>as in excessive resistive loading?

Driven in this case (and lots of others) means a signal connected to
that pin.

A resistor going to VCC or ground does indeed "drive" that signal.

BOOT pins must completely float until the boot process is done, *then*
you can do something with them.

A good idea is to avoid boot pins
Another good idea is to use tri-state drivers on those pins, and turn
them on ONLY when the boot process is done.

There are a number of threads discussing this.

However, nobody has specifically said that resistors effectively drive
boot pins. They do.

Harvey

[Greg]

Thanks!  It never would have occurred to me that a resistive load in this context is a "driver".  Lesson learned!

Greg

[Greg]

I was attempting to emulate the PRU cape, which uses P8_44 as a PRU output.

However, the PRU cape uses a FET switch, gate connected to P8_44.  It does not load the pin during boot.

I used a BJT with resistor in series with the base, which was enough to disable the board!

I've changed the series base resistor to 47Kohm and now it works.  However, I suspect it might be a bit marginal.

So I need to get some FETs!  Too bad, I've got a lot of 2N2222s, they will have to be put to work elsewhere.

Greg