Does anyone have any familiarity with, or comments on the ZigBee protocol?
that’s a pretty open ended question. I assume it’s in relation to you “long range” question.
If you are building it yourself, zigbee is very complex. Mesh networking is pretty cool but you will have to provision each node. It’s great because of the overall redundancy and “auto” reorganization. not for the faint of heart nor beginners.
It’s a possible side trek from the long range issue. Just that I noticed how much cheaper the ZigBee offerings from Maxstream and Aerocom were than their regular offerings. And the range isn’t too bad either.
I’m not afraid of complexity so much as going to the trouble, and then finding there’s a show stopper somewhere in there where the ZigBee protocol just won’t do what I want.
I also don’t figure the docs that I’ve seen (the Maxstream part) that declare the units default to transparent serial ports, and later say that the default is Unicast mode, presumably there’s gotta be some command mode setup in there somewhere.
The other assumption that seems to have crept in is that the units are basically stationary. And the network discovery process seems to rely on that. Well it happens that if I use ZigBees some of my units will be stationary, and some will be on vehicles. So who knows what might happen 
Anyhow my question was deliberately general. I was just hoping someone might rave on happily about one and give me the confidence to actually try it out
For the application in your other posting, I think “Zigbee chipsâ€
part of the problem with zigbee modules is they typically don’t have long distance so you will need to put out a lot of them to cover a large area. I think it becomes a major headache to manage.
by the way, have you looked at the aerocomm stuff? they have some fairly long distance stuff. the mouser catalog pages are a good summary:
http://www.mouser.com/catalog/626/122.pdf
http://www.mouser.com/catalog/626/123.pdf
http://www.mouser.com/catalog/626/124.pdf
I don’t have any experience with their products but the spec’s seem like they might work for you.
I would love to start with a chip, but unfortunately I’m not up to paying the 100k or so it would cost to get type approval (and that’s probably just for this country 
)
The mouser people don’t seem to have stock of the Aerocomm ZigBee parts yet. Which leaves me with the Maxstream part I guess.
Its a shame noone has addressed the issue of ultimate long range properly. When they design these standards they tend to settle for a fixed bit rate and that’s it, but anyhow that’s a subject for my other thread.
Thanks for the links
Philba:
part of the problem with zigbee modules is they typically don’t have long distance so you will need to put out a lot of them to cover a large area. I think it becomes a major headache to manage.by the way, have you looked at the aerocomm stuff? they have some fairly long distance stuff. the mouser catalog pages are a good summary:
http://www.mouser.com/catalog/626/122.pdf
http://www.mouser.com/catalog/626/123.pdf
http://www.mouser.com/catalog/626/124.pdf
I don’t have any experience with their products but the spec’s seem like they might work for you.
I’m using several vendors’ modules in a project in my “day” job.
The range I’m getting is far more than I expected, considering how many walls (drywall) they have to penetrate. Each drywall-pair is about 4dB at 2.4GHz; masonry is far more (see NIST’s report on this).
The MaxStream and others’ 30 and 60mW radios do very well in range. The 1mW radios do pretty darned good too.
With ZigBee in place (an option as you know), I’ve found that self-forming and self-reforming multi-hop networks are now doggone near trivial.
Since '15.4 is 2MHz bandwidth not 20MHz like WiFi, and with 60mW '15.4 chips, comparable to the power out of 802.11g chips, the range is greater because the bandwidth is one tenth and the bit rate is 250Kbps. The modulation order is O-PSK which tolerates a poor signal to noise ratio far moreso than the 802.11 modulation modes.
Pretty good for $20@1ea (1mW) and $40 (60mW) ready-to-use radio modules. Especially nice that the MAC/PHY are standards based, non-proprietary and availabe from about 8 chip makers.
steve
Perhaps I’m guilty of not having read the manual fully, but I was left with the impression that in the Maxstream zigbee modules the mesh discovery and routing process isn’t automatic or transparent - you have to figure it out yourself. Or am I missing something?
I like their pricing. What intrigues me is they seem to be undercutting the pricing of their regular radio modems
steve, what ranges are you actually seeing. I’m getting interested in taking the plunge and the prices are looking better.
Be sure to distinguish IEEE 802.15.4 from ZigBee. They are different standards. The 15.4 standard defines the MAC and PHY, kind of like 802.11a/b/g. ZigBee is not an IEEE standard, nor is WiFi. Many applications need only 15.4 with fixed addresses or proprietary network layer routing. ZigBee-less.
Re meshing - any vendor who claims to have been ZigBee 1.0 certified (see their website) must implement self-forming/self-reforming meshing as an option.
In 1.0, there is one “coordinator”. All network nodes are called either Routers or EndPoints. An endpoint is a node that cannot relay traffic, whereas routers can.
Every node except the coordinator finds a parent by choosing among the neighbors that it can “hear”. ZigBee defines how this is done, using broadcast frames and hop counts and so on. It’s complex. ZigBee 1.0 uses a variant of the widely used (Internet) protocol called AODV. You can read that RFC. Unlike wired networks where every candidate route from A to B (via intermediate routers) has about the same error rate, wireless does not. So AODV with ZigBee must do route selection not based merely on fewest number of hops, but rather this plus avoiding unreliable candidate links (RF conditions poor).
As to range, like any wireless, it “depends” - on
Transmitter power - modules for sale range from 1mWatt to 30 and 60 and even 100mW. Each doubling of power is 3dB. 1mW = 0dBmW (or dBm as it’s abbreviated). So 2mW = +3dBm; 100mW = 20dBm. Each 6dB of more power adds a lot of cost to the radio. Compare prices.
Receiver sensitivity. All the brands are essentially the same. As is the case with WiFi too, due to the 802 standards. ZigBee is so far used only with IEEE 802.15.4 underneath ZigBee as the network layer.
Antenna gain, pattern and where each node is within the othe node’s antenna pattern.
RF Path obstructions: Walls of some material type (drywall is about 3dB of loss at 2.4GHz), trees, human bodies, metal, etc.
RF Signal Impairments: interference, reflections, multipath, etc. A very busy WiFi network on/near the chosen freq. for '15.4/ZigBee will of course reduce throughput. But 15.4 is just 2MHz wide whereas WiFi is 20MHz, so 15.4 sees WiFi on nearlyl the same freq as an increase in the noise floor.
15.4 has about a dozen non-overlapping 2MHz channels whereas 802.11 has just 3 non-overlapping 20mHz channels.
Also, since 15.4 uses just 2MHz and 250Kbps, it requires a much lower signal to noise than does WiFi. So it works well with received signals in the -'90s whereas WiFi struggles at lower rates at -80 to -90.
==============
Now all this said, marketing people like to advertise Line Of Sight range in the absence of all impairments and with just 2dBi gain antennas, or less, the range for 1mW is perhaps 300 ft and 100mW can be perhaps a kM.
It Depends.
Of couse, on 2.4GHz, we have nots of noise and interference so WiFi is a Wireless LAN and 15.4 is a wireless PAN, neither are wireless WANs.
I am very impressed with the indoor range I’m seeing penetrating many walls in an office building for 100+ ft range through say 10 walls.
===============
There now. You asked what time is it and I told you how to make a watch!
Russel:
Perhaps I’m guilty of not having read the manual fully, but I was left with the impression that in the Maxstream zigbee modules the mesh discovery and routing process isn’t automatic or transparent - you have to figure it out yourself. Or am I missing something?I like their pricing. What intrigues me is they seem to be undercutting the pricing of their regular radio modems
Their mainstream modems are 900MHz which have orders of magnitude greater range due to the lower freq and higher power. These are often point to point links rather than a LAN/PAN
some others to look at for hobby work (low quantities avail from distributors - and maybe SparkFun)
MaxStream
SiliconLabs = SiLabs
Helicomm + SiLabs
Panasonic (at Digikey or Mouser as I recall)
OKI (same like Panasonic)
Jennic (Digikey)
Ember (no know 1ea sources)
Non-ZigBee
MoteIV
Dust, Inc.
These are module-makers who use the 802.15.4 chips made by
Freescale (ex-Motorola Semi)
TI/Chipcon
sole source
Jennic
Ember
Ok, the Maxstream modules claim to be zigbee, and you’re saying that part of zigbee is the routing protocol, for mesh forming etc.
But after having had another look at the manual for the Maxstream part it’s still unclear whether they implement the routing part of zigbee.
That the gist of what I was asking. I have a tricky question about antennas too but I’ll put that in another thread.
At MaxStream, ZigBee 1.0 is in beta test status.
yes, they do comply with the meshing AODV capabilities of ZigBee 1.0.
Again, you may not need meshing for a small system where the location of nodes is fixed. A handheld button-box could be arranged to always be near the coordinator, RF-wise, so as to avoid the complexities of ZigBee and just use MAC addresses for all messaging. Kind of like running 802.11 WiFi in the ad-hoc mode.
I’ve been getting into ZigBee pretty heavily at work with the MaxStream modules, but I’m definitely not an expert thus far. Here’s what I’ve experienced - These devices claim to be ZigBee compliant in certain modes, but the most useful features have proprietary written all over them. When you only use MaxStream modules on your ZigBee network, they’re pretty powerful. We’ve tested the standard XBee module (1mW?) through sheet metal walls and a few cars with range of about 150-200 feet. Pretty impressive. We also have several XBee Pro modules that claim up to 1 mile line of sight @ 100mW (we haven’t tested range on these).
Overall impression: Happy with them. Very simple to use in transparent broadcast mode. Didn’t have to do a thing, the default settings on the modules brought several data streams seamlessly into one computer. When you want to get addressing information (to parse different sources with similar information), you just assign each node an address and a destination address and put it into API mode and it gives you framing information (source, length, checksum) along with the packet.
I’d definitely recommend a XBIB-U USB interface board. It’ll let you configure each module’s settings (source and destination addresses among others).
Good Luck,
Brad
what’s proprietary? I haven’t found that yet. The ZigBee plugfests I’ve seen don’t seem faked.
Examples of non-ZigBee network layer code are found in the product rooted in the UC Berkely code - such as Crossbow, MoteIV and Dust.
Reality is that there are so far just 2-3 implementaions of ZigBee - Figure8 Wireless, now owned by TI/Chipcon - is in Chipcon- and FreeScale-based designs, for the most part. And these are the majority of the products on the market now. This will change, since FreeScale must change ZigBee stacks because of the TI acquistion of Chipcon.
The host interface (USB or serial) is an API that of course isn’t in the ZigBee standard for the network layer on top of 802.15.4’s MAC/PHY.
The host interface commands, binary or AT-commands (Hayes-like) are part of that host interface.
MaxStream’s DN command in the AT command set is really neat and unique, but it uses the MAC/PHY/NWK standards. This allows a DNS-like way to address nodes by their given name rather than their MAC addresses. If you’re using ZigBee (it is of course an option), you can’t be sure of network addressess of nodes because of meshing. The MAC address of course is invariant but awkward to use.
The XBeePro’s 60mW makes a big difference in range, though it costs more than the XBee.
I spent some time looking at the xbee and xbeepro. Nice looking stuff. when I get back from vacation, I think I’ll dig more into it.
I noticed that the xbee datasheet says the pro is 100 mW yet I’ve seen other places where it says 60. In other places they talk about 19 mW for european limitations. pretty confusing. I’ll assume 60 but I hate it when companies cant get their story straight
also, maxstream shows digikey and future carrying them but, for the life of me, I can’t find them on futures site and digikey is charging over list price for them (36 vs 32 for the Pro, for example). where did you buy yours?
stevech:
At MaxStream, ZigBee 1.0 is in beta test status.yes, they do comply with the meshing AODV capabilities of ZigBee 1.0.
Again, you may not need meshing for a small system where the location of nodes is fixed. A handheld button-box could be arranged to always be near the coordinator, RF-wise, so as to avoid the complexities of ZigBee and just use MAC addresses for all messaging. Kind of like running 802.11 WiFi in the ad-hoc mode.
I couldnt understand that. What is a handheld button-box?
Philba:
The XBeePro are 60mW vs. 1mW in the XBee. The regulations for 2.4GHz vary somewhat by country, but seemingly all permit WiFi 802.11 which in the US can be up to 4Watts radiated with a certain narrowbeam antenna, and far less power as you get towards an omni antenna. Unlike other bands like 433/868MHz I don’t believe any countries have severe duty-cycle restrictions at 2.4GHz. Those restrictions are to enable coexistence of wireless telemetry like consumer thermometers.
The XBee-PRO-PKG is an XBeePro module in a little box with a USB or RS232 connector for plug and play.
MaxStream was just acquired by Digi. I hope this is a good thing. All too often, it isn’t.
As you may know, there are about 8 vendors of 802.15.4 modules (chipsets on thumb-sized PCBs). MaxStream is quite inexpensive, as is Panasonic and OKI, less so Silicon Labs.
I’m using two vendors’ modules and working with the beta test conditions of ZigBee which is the optional network layer for 802.15.4