1-Wire

[Fohnbit]

Hello,

I´m not sure about the maxium 1-wire sensors.

1. How many GPIOs I can use for 1-wire? Right now I enabled Pin 9.12 for 1-wire. Can I add a second or third pin to made several own lines?

2. How many 1-Wire I can use on one pin? I read on a website about up to 10 ... but this is not very much.

Thank you!

Hannes

[Chris Green]

Fohnbit <ged...@gmail.com> wrote:
> [-- multipart/alternative, encoding 7bit, 61 lines --]
>
> [-- text/plain, encoding quoted-printable, charset: UTF-8, 24 lines --]

There is a dedicated 1-wire mailing list where you will probably find
more knowledgeable specialist help.

It's on sourceforge:-
owfs-de...@lists.sourceforge.net

Don't worry about the 'developers' bit, they're a helpful crowd and
will do their best to answer any reasonable sort of 'user' questions.

--
Chris Green
·

[TJF]

Hi Hannes!

Am Donnerstag, 21. März 2019 17:39:06 UTC+1 schrieb Fohnbit:

1. How many GPIOs I can use for 1-wire? Right now I enabled Pin 9.12 for 1-wire. Can I add a second or third pin to made several own lines?

2. How many 1-Wire I can use on one pin? I read on a website about up to 10 ... but this is not very much.

1. I don't know.
2. Last time I checked the W1 kernel driver in 2015 it could handle only one sensor per pin (Dallas temp sensor).

For my aim this wasn't sufficient. That's why I made my own driver using a PRU to handle the tree-state pin with accurate timing. The driver is called libpruw1 and provides all features to handle multiple sensors at one pin. It also provides a bus monitoring feature to log the bus state during operation. That's helpful to debug wiring issues.

The W1 bus is using an 8 byte identifier, so the number of sensors at one pin is (nearly) unlimited. In case of Dallas temp sensors the sensor needs 750 mSec to sample, and the bus needs approx. 10 mSec to read one sample -> for 10 sensors you'll need 750 + 10x10 = 850 mSec for a cycle.

Regards