We have networks with ~60 2406's on them that we "poll" using the alarm method several times a second. We tend to get results in very quickly (faster than we can measure), and with the alarming we never miss any "real" data. (we have an issue with one of our hardware designs where the sense pin on a 2406 "floats" somehow and alternates open/close very, very rapidly. So when we read that, we'll frequently get bunches of alarms for the same event (open, open, open), but that's an electrical problem that needs to be addressed separately).
Now if only the DS2450's and 18B20's would alarm properly....We still have to poll those, as we can't make the alarming actually trigger using owfs. --Jim On Fri, Apr 24, 2009 at 4:05 PM, Michael Markstaller <[email protected]> wrote: > Paul, thanks for pointing me to this thread. It gave me the right directions > (latch, alarm) and I got it working today this way (which I knew about > before), I just wanted to make sure I haven't missed anything fundamental ;) > > Michael > > -----Original Message----- > From: Paul Alfille [mailto:[email protected]] > Sent: Friday, April 24, 2009 1:12 AM > To: [email protected] > Subject: Re: [Owfs-developers] Notification or polling switch/presence > > This is a frequent question. > > 1. Many chips have a latch that is changed when the triggered event occurs. > That way you can catch transient events, though perhaps not in real time. > > 2. Other chips set the alarm state, which can be polled by looking at /alarm > -- although not efficient in the individual case, it is a good way to scan a > large group of switches for a single event. > > 3. Many have found that a slight delay in polling (0.1 second) can reduce the > polling overhead considerably. > > 4. The presence pulse isn't captured It's better suited to microprocessors > that can poll the line. > > 5. you can always set up a small embedded system (openWRT router or the like) > to poll the line and read the reult over the network. > > For instance, look at this thread in the forums: > https://sourceforge.net/forum/message.php?msg_id=6994214 > > Paul Alfille > > > > On Thu, Apr 23, 2009 at 2:16 PM, Michael Markstaller <[email protected]> wrote: > > > I just wondered wether there's any better way to get i.e. a switch > status or presence of an i.e. DS1990R on the bus without doing "owdir > /uncached" as fast as possible. > This looks rather inefficient and ressouce-intense to me, is there > some smarter way to get something like "notifications" or similar for > time-critical things like an iButton getting connected (maybe only for a > short time) or the change of a input on a DS2406 ? > > The DS1990R for examle has some "presence pulse", any way to benefit > from this to trigger something when ist connected as fast as possible ? > > Michael > > ------------------------------------------------------------------------------ > Crystal Reports - New Free Runtime and 30 Day Trial > Check out the new simplified licensign option that enables unlimited > royalty-free distribution of the report engine for externally facing > server and web deployment. > http://p.sf.net/sfu/businessobjects > _______________________________________________ > Owfs-developers mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/owfs-developers > ------------------------------------------------------------------------------ Crystal Reports - New Free Runtime and 30 Day Trial Check out the new simplified licensign option that enables unlimited royalty-free distribution of the report engine for externally facing server and web deployment. http://p.sf.net/sfu/businessobjects _______________________________________________ Owfs-developers mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/owfs-developers
