Mick Sulley | 26 Jan 15:50 2015

Raspberry Pi rpi-update

I created a new 1-wire system on Raspberry Pi the other day and could 
not get it to work.  The cause was that /dev/i2c was missing. I fixed it 
with help from the Pi forum, for details see
http://www.raspberrypi.org/forums/viewtopic.php?f=28&t=97639&p=677652#p677652

One thing I learned from this is that it is not a good idea to run 
rpi-update unless you need to and know what you are doing, which I did 
not it seems :)

Be aware!!

Mick

------------------------------------------------------------------------------
Dive into the World of Parallel Programming. The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net/
Sven | 26 Jan 12:57 2015
Picon

Problems with DS2413

Hallo Mailinglist,

I have problems with a DS2413 device.
I do not have files like

|-- PIO.A
|-- PIO.ALL
|-- PIO.B
|-- PIO.BYTE

in my OWFS tree. I only see files like

driver
id
name
power
rw
subsystem
uevent

Was there a change in how to read/write to DS2413?

Other devices like DS18B20 work fine on the same bus.

I use 1-Wire bus connected to GPIO of a raspberry pi with OWFS.

Thanks a lot
Sven

------------------------------------------------------------------------------
(Continue reading)

Peter Hollenbeck | 23 Jan 03:11 2015
Picon

Basic Question

I have a LinkUSB and an MS-TV temperature and voltage sensor.

http://localhost:2121/ shows:

10.67C6697351FF
05.4AEC29CDBAAB
bus.0
and more

How do I know where to go to find temperature and voltage readings.

Ubuntu 12.04
owfs installed today with apt-get

Thanks for any input,
Peter

------------------------------------------------------------------------------
New Year. New Location. New Benefits. New Data Center in Ashburn, VA.
GigeNET is offering a free month of service with a new server in Ashburn.
Choose from 2 high performing configs, both with 100TB of bandwidth.
Higher redundancy.Lower latency.Increased capacity.Completely compliant.
http://p.sf.net/sfu/gigenet
_______________________________________________
Owfs-developers mailing list
Owfs-developers <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/owfs-developers
Howell, Larry (Contractor | 21 Jan 14:56 2015

OWFS random failure issue

Hi list,

 

I'm an embedded software engineer working on a project that uses DS2430A and DS2505 slave devices to store run-time and configuration data.  I'm building everything from source and have installed them under Linux 2.6.35 running on a Freescale i.MX53 (ARM Cortex-A8) CPU. Only 5 channels of the ds2482-800 bus master are used.  The application accesses uncached data.  The slave device are in attachments that can be connected to the embedded system via a 5 m cable.  The attachments contain brushless DC motors.  The embedded system is a DC power supply/motor controller.  The 1-wire generally functions well, but randomly a physically connected attachments' slave devices "disappear" from the owfs file system for some random period (a few ms to a few seconds).  Mounting owfs with --debug --error_level=9 shows no errors, but has provided confirmation that owfs is acting as expected on the data that fuse.ko/libfuse provides.  This suggests the issue is occurring with fuse or in kernel space.  Is it possible to pass parameters to fuse to enable fuse debug output?  Any suggestions on debugging this issue would be appreciated.

 

General details:

Bus Master:        ds2482-800         channels 0-4 connected, 5-7 not connected

Slave devices: ds2430a and ds2505

Fuse version: 2.9.2

OWFS version: 2.8p15

Typical OWFS mount command-line: "/usr/bin/owfs --i2c=/dev/i2c-0:0 -m mnt/1wire --allow_other"

 

Thanks in advance!

 

Larry Howell (Contractor)

Software Design Engineer

------------------------------------------------------------------------------
New Year. New Location. New Benefits. New Data Center in Ashburn, VA.
GigeNET is offering a free month of service with a new server in Ashburn.
Choose from 2 high performing configs, both with 100TB of bandwidth.
Higher redundancy.Lower latency.Increased capacity.Completely compliant.
http://p.sf.net/sfu/gigenet
_______________________________________________
Owfs-developers mailing list
Owfs-developers <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/owfs-developers
Paul Alfille | 6 Jan 14:59 2015
Picon

New Release 2.9p9

Minor improvement -- basically fixes the owfs problem for Markus Gaugusch's report:

"cat: temperature: Message too long"

The problem is that some tools (like cat) ask for an arbitrary (but large) read to gather all the data. It used to be 64K but now seems to be 128K which passes some internal limits.

We now trim the read request before processing.

Paul Alfille
------------------------------------------------------------------------------
Dive into the World of Parallel Programming! The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net
_______________________________________________
Owfs-developers mailing list
Owfs-developers <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/owfs-developers
Jan Kandziora | 5 Jan 02:14 2015
Picon
Picon

Re: DS2482-800 long wire problem

Am 04.01.2015 um 17:33 schrieb Jan Sennesael:
>  <at> Jan: in the document linked by Paul
> (http://www.maximintegrated.com/en/app-notes/index.mvp/id/148) they start by
> saying "The scope of this document is limited to 1-Wire networks that use
> Category 5, twisted-pair copper wire and have 5V bus power supplied by the
> master" 
>
I think this is the only place where they mention Cat5e cables.

Anywhere else they talk about limiting bus capacitance, and using a
twisted pair for 1W/GND is sure the worst thing you can do when your
goal is limiting capacitance.

> So i guessed that this was the preferred cable... (in addition, hobbyboards
> also talk about this type of cable and connectors)
>
Well, it's a poor choice. You don't gain any noise immunity by using a
twisted pair for 1W/GND, because the pair is driven unbalanced. But you
gain a lot of additional capacity, which is very bad when your bus is
open-drain instead of push-pull. Like onewire is. So don't do that.

> Normally when I crimp the rj45 connector, normaaly it is secured with the
> piece of plastic.
>
The plastic notch of a standard RJ45 plug is designed to cut into the
soft, flat mantle of a 8-wire phone cable. Nothing else will fit, even
6- or 4-wire cable mantles will not connect tightly. There even exist
variants of the RJ11/RJ12 and RJ45 connectors which have a smaller cable
opening, just for the purpose of having tight fit for 6- or 4-wire
cables. So this is crucial for the function of the connector.

For cables with round mantles, the mantle has to be fixed another way,
e.g. by melting it into additional plastic.

> But it's a great tip if i run into some errors.
> I am getting some improvement by using 7/8 as gnd (i can succesfully connect
> a sensor at the end of the long wire, but not yet the actual sensors (they
> have their own cat5e cable of 8m). So still trying ;-)
>
Another tip: Use as many ground wires as possible. Long cables produce a
ground-lift (especially when far devices are bus-powered), and keeping
the ground level on the far end as seen by the host below 0.7V is also
crucial for the bus function.

Kind regards

	Jan

------------------------------------------------------------------------------
Dive into the World of Parallel Programming! The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net
Jan Sennesael | 4 Jan 17:33 2015
Picon

Re: DS2482-800 long wire problem

 <at> Jan: in the document linked by Paul
(http://www.maximintegrated.com/en/app-notes/index.mvp/id/148) they start by
saying "The scope of this document is limited to 1-Wire networks that use
Category 5, twisted-pair copper wire and have 5V bus power supplied by the
master" 
So i guessed that this was the preferred cable... (in addition, hobbyboards
also talk about this type of cable and connectors)
Normally when I crimp the rj45 connector, normaaly it is secured with the
piece of plastic. But it's a great tip if i run into some errors.
I am getting some improvement by using 7/8 as gnd (i can succesfully connect
a sensor at the end of the long wire, but not yet the actual sensors (they
have their own cat5e cable of 8m). So still trying ;-)

Thanks 
Jan

------------------------------------------------------------------------------
Dive into the World of Parallel Programming! The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net
Stefano Miccoli | 4 Jan 14:07 2015

pyownet: announcing release 0.8.0

Happy 2015 everybody!

I have released version 0.8.0 of pyownet:

Highlights:


pyownet can be considered a replacement for ownet (which can be found in module/ownet/python in the owfs source tree). The main reason for writing this code was the observation that the ‘official' ownet has an incomplete set of features, it is not maintained anymore, and that there is no support for Python3.

If there is an owserver listening at port 4304 of server.example.com, the use of pyownet is as simple as:
>>> from pyownet import protocol >>> owproxy = protocol.proxy(host="server.example.com", port=4304) >>> owproxy.dir() [u'/10.A7F1D92A82C8/', u'/05.D8FE434D9855/', u'/26.8CE2B3471711/'] >>> owproxy.present('/10.A7F1D92A82C8/temperature') True >>> owproxy.read('/10.A7F1D92A82C8/temperature') ' 6.68422'
As pyownet is registered on PyPI, the recommended installation is ‘pip install pyownet’.
Bye

Stefano
------------------------------------------------------------------------------
Dive into the World of Parallel Programming! The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net
_______________________________________________
Owfs-developers mailing list
Owfs-developers <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/owfs-developers
Ekkehard Pofahl | 4 Jan 13:47 2015
Picon

Raspi : Debian get-apt install vs generic install

Hello,
after typing "apt-get install owfs" and editing owfs.config OWFS(2.8p15) worked on my Raspi. Real Magic !

I do have some issues with the programs reading out the OWFS server. So my plan was to install the freshest copy of OWFS to make sure, I did not haunt an issue, which is long resolved with the latest OWFS SW. 

I retrieved the sources of the latest OWFS package (2.9p8) into /usr/src/. Configure, make and make install seemed to work successful into /opt/owfs/. BUT did not replace the old SW (2.8p15) automatically in the system.

"apt-get remove owfs" did not remove OWFS completely. It did something, but I do not understand, what. Nasty text about owfs etc. shows up after reboot. I did some digging into init.d and other stuff : simply to much information for me, and no success for my little Raspi. Most probably I did not try hard enough.

A Google search for the best recipe failed for me. Do you know more than Google and have a recommandation ? Worst case : fresh Linux copy, don't use apt-get, ;-) .

Best regards
Ekki



--
--
Ekkehard Pofahl
Finkenweg 12, 51515 Kürten
mailto:ekkehard <at> pofahl.de
tel:+49-151-54860909
------------------------------------------------------------------------------
Dive into the World of Parallel Programming! The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net
_______________________________________________
Owfs-developers mailing list
Owfs-developers <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/owfs-developers
Jan Kandziora | 3 Jan 11:40 2015
Picon
Picon

Re: DS2482-800 long wire problem

Am 03.01.2015 um 09:49 schrieb Jan Sennesael:
>
>  <at> Jan: now i'm a bit surprised, since maxim themselves talk about using
> cat5(e).
>
Where? If yes, they forgot about all they've written in various
application notes before.

In my own experience, plain flat telephone wire is the best.

> Of course the cable is only used for 1-wire. I use a modified set
> of wiring based on hobbyboards. They use a pair for gnd/1-wire_data. My
> pinout is: 1and2 for +5v and gnd , 3and6 for return of 1-wire_data /
> 1-wire_gnd , 4and5 for the dépard, 7and8 not connected
> But i'll give it a try and use the pairs 7/8 as gnd for this (long) cable!
> 
If you use RJ45 connectors with round mantles, you have the additional
problem the mantle cannot be crimped into the connector plastic
correctly and the contact blades move together with the lose mantle when
plugged into a socket, resulting in lose contacts.

I'd stumbled across this in an installation someone else did with CAT5e
cables. After 5 minutes I had localised the problem. Too bad it was an
8h train ride away. ^.^;

Kind regards

	Jan

------------------------------------------------------------------------------
Dive into the World of Parallel Programming! The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net
Jan Sennesael | 3 Jan 09:49 2015
Picon

Re: DS2482-800 long wire problem

 <at> Paul: Yes i had read that website some time ago. It made me change (then)
from (mixed) star topology to my current (linear) bus. to accomodate this i
followed the suggestions of the list members and used the utp-wires 4and5
for the dépard and wires 3and6 for the return.
I did not try another channel since i only have one utp cat5e running from
the master to the attic (and the new sensors cannot get to the master
without passing through the attic); so splitting them up is not so easy

 <at> Jan: now i'm a bit surprised, since maxim themselves talk about using
cat5(e). Of course the cable is only used for 1-wire. I use a modified set
of wiring based on hobbyboards. They use a pair for gnd/1-wire_data. My
pinout is: 1and2 for +5v and gnd , 3and6 for return of 1-wire_data /
1-wire_gnd , 4and5 for the dépard, 7and8 not connected
But i'll give it a try and use the pairs 7/8 as gnd for this (long) cable!

------------------------------------------------------------------------------
Dive into the World of Parallel Programming! The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net

Gmane