Re: Dallas 1-Wire Problem/question

[Marc F Hult <MFHult@xxxxxxxxxxxxxxxxxxxxx>]

On 25 Feb 2006 23:48:29 -0800, mcbill20@xxxxxxxxx wrote in message
<1140940109.523822.137320@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>:

>Hello all. I have an IOM142A board from phanderson.com that I have been
>using for a couple of years now. I put temperature sensors in every
>room of the house, as well as the garage and basement. Overall it works
>pretty well except that I am getting a lot of read errors on certain
>temperature sensors.
>
>I have three wire runs. The first run is connected to the first input
>on the IOM142A and has six sensors. It is a single line and ends in the
>attic with a single temperature sensor on the very end of the line. The
>second and third runs consist of a single sensor at the end of each and
>are connected to the second output on the IOM142A. However, I was
>forced to use a star configuration on these last two.
>
>My problem is that the sensors on the end of each wire run have a lot
>of read errors. Occasionally some middle sensor will log a lot of
>errors, but generally they go weeks without a single error but the end
>sensors log hundres of errors in that time (my software issues a read
>every 15 minutes).
>
>All the wire is CAT5E. The sensors are soldered to a three inch piece
>of CAT5E which is crimped to a modular plug with gold contacts.  The
>total length of the run with six sensors is less than 100 feet. The two
>runs in the star configuration are approximately 15 feet and 35 feet
>respectively.
>
>It's strange to me that it's only the end sensors that have problems.
>Should I be using some type of terminating resistor at the end of the
>line? Any suggestions?
>
>Thanks.
>Bill


First thing I'd try would be a 4.7K pullup resistor at the end of the line
(4.7K from the data line to 5 volt supply). This is different from a
"terminating resistor" which would be installed from data to ground. If that
doesn't fix it, read on ...

This application note:  http://pdfserv.maxim-ic.com/en/an/app159.pdf
recognizes that great care is needed in programming to achieve very high
reliability with 1-wire.

But your problem seems more severe, and on the order of the severe problems
addressed in this app note that is on my web site and seems to have
disappeared off the Dallsemi/Maxim site..

http://www.econtrol.org/ds1820_Errors.pdf

This app note dated March 2000 addresses what Dallsemi called
"peculiarities" and "inaccuracies" and is especially relevant if you should
happen to have some older DS1820 sensors that were (in my blunt assessment)
defective. Unfortunately, it may not always be possible to distinguish good
sensors from the 'bad' ones by their markings. As I understand it, you don't
want DS1820's at all and you want DS18S20's with rev B5 or later.

What happens when you swap the misbehaving sensors with ones that are
seemingly reliable?

Also: you don't say if you are using these in parasitic mode. My advice is
that you don't.

Also, make sure that your supply voltage is adequate with no glitches at the
end of the line.

HTH ... Marc
Marc_F_Hult
www.ECONtrol.org