Well, I'm still workng on it. Digitemp does better than OWFS. It sees 20 of the 100 temperature sensors, and can read them.
OWFS sees 1 or 2. There is still more going on. Both are inconsistent, and some sensors are almost always seen (the DS2405)s and DS2438s. My guess is that there are timing issues, and I'm trying to find them. Paul -----Original Message----- From: [EMAIL PROTECTED] on behalf of Darryl VanDorp Sent: Mon 4/3/2006 12:07 AM To: owfs-developers@lists.sourceforge.net Subject: Re: [Owfs-developers] ds18s20 reading So your seeing problems with digitemp as well? I appreciate your efforts Paul, had I known about this issue I wouldn't have planned my whole "network" on a hub. -darryl On 4/2/06, Paul Alfille <[EMAIL PROTECTED]> wrote: > > Ok, I've spent the last week chaing this down. > > 1. We conform to the DS2409 datasheet. > 2. We use the same algorithm as digitemp (though digitemp only supports > one level of branching). > 3. We use the same byte sequence as digitemp. > > So what gives? > > I noticed a few things > 1. I wasn't turning off branches properly -- I'll fix > 2. digitemp doesn't really work either. > > Yup. I have a 100 sensors network that is fully discovered in the root > branch of either digitemp or owfs. Only a few sensors are seen in a main or > aux branch. > > Conclusion -- the 1-wire hubs are poorly suited for multiple sensors. > > Next fix attempt -- I'll attempt multiple "discoveries" when working on a > branch to perhaps get over the problem. Not efficient, but working is > important. > > Paul > > > On 3/30/06, Darryl <[EMAIL PROTECTED]> wrote: > > > > I just tried powering the ds18s20 and it made no difference still reads > > 85. Don't know if that narrows down any variables..... > > > > > -- http://randomthoughts.vandorp.ca
<<winmail.dat>>
