Multiple R4's per Gateway?
vining
Posts: 4,368
I need to add 4 R4s to a job which already has 4 R4s which run on 2 gateways set to channels 15/25. Each gateway handles 2 R4s. Since I can't add more gateways (house isn't large enough to repeat channels) I'm stuck adding 2 more R4 to each gateway for a total of 4 R4s per gateway. I still don't really like R4s since they require a lot of consideration when designing/coding feedback and when running multiple R4s on a single gateway that only compounds the problems.
Since I've never tried more than 2 R4s per gateway I'm curious to learn what others have experienced. Am I just being paranoid? I already significantly control the feedback I send them so is anyone running 4 or more R4s per gateway with good results?
Since I've never tried more than 2 R4s per gateway I'm curious to learn what others have experienced. Am I just being paranoid? I already significantly control the feedback I send them so is anyone running 4 or more R4s per gateway with good results?
0
Comments
It will depend on your code and your environment, but 4 is our maximum recommendation. Our system is pretty demanding (lots of traffic) so we do a lot of work to pace the outgoing commands, but if it holds up for us, it will hold up for most.
FYI we found that the gateway can't efficiently send commands to the R4's while it gets a continuous flow of commands coming from the NetLinx, even if paced slower ala the AMX Australia queue module. We found the best results came if you give a break for a second or more every 40 or so commands. This lets them get out, and we actually found overall throughput went UP with this scheme... the second or so gap does NOT slow the appearance at the panel as much as it does when the gateway can't get a break. And it does wonders for keeping the buffer under the magic 300 that causes it to drop offline.
Check the buffer and connection logs in the gateway web browser often when tuning the queue or delays, they tell the whole story.
I would add a third gateway to keep the R4-per-gateway count minimized.
From day 1 I didn't like these being in the same spectrum as wi-fi especially since we had enough issues with that spectrum already and if I recall Zigbee comes in several flavors (spectrums) but the 2.4ghz had the highest data transfer rate.
Here's a link that I've posted several times since the R4's first came out that I've used for reference, I'll verify if it's still valid but AFIK the PRO release only affected traffic routing not bandwidth.
The old link is no longer valid but this link is an updated version of that and nows shows 15, 20, 25, 26 as being valid non interfering channels. Is 25 still greyed out in the gateways as an option? If not that would give me the 4 channels I want to keep everything 2 on 1.
http://www.mobiusconsulting.com/papers/ZigBeeandWiFiInterference.pdf
I found the old link which I printed out long ago (hanging up behing my monitor) and it shows the same non interferring channels as this new link. I swear I thought we were limited to channels 15 & 25/26, maybe I was brainwashed by the forum or maybe my brain is toast.
I also found another link:
http://www.zigbee.org/imwp/idms/popups/pop_download.asp?contentID=13184
which show channels 15. 16, 21 & 22 as being non interferring but they're using wi-fi channels 1, 7 & 13 so maybe they're non US. I thought Scheider Electric was based in America though??
http://en.wikipedia.org/wiki/Schneider_Electric
I was satisfied with 1-3 second delays on page changes, as they are data driven with lots of data. Now they are showing up in half the time.
Don't know how I missed this before, maybe I just didn't realize the impact of avoiding WIFI.
EDIT: More testing, found range extended too, but not by much (5-10 feet more, over the 50' I was getting before).
It's probably a good idea to always use one of the 4 non overlapping zigbee channels even when only one or two access points are installed just in case someone decides to add another AP down the road or simply change the RF channels and not knowing there's a potential of interfering with the zigbee. They probably wouldn't even know there is zigbee on the job and if they did they probably wouldn't know what that means.