BraveSirRobbin
Moderator
Is the Compose Firewall immune to noise generated from devices like this? If so, this may be a viable option until the other technologies mature a little (Insteon & UPB mainly).jlehnert said:
If you could start from scratch-Lighting
- Thread starter ALEDOUX13
- Start date
AutomatedOutlet
Senior Member
With the Compose firewall, you have 2 things that will help you here.
1) Not all circuits are on the firewall. Only circuits that you want to control.
2) If the device that is causing problems happens to be on the firewall, the firewall will also filter noise on the circuit. Granted that circuit will still have noise but it at least won't mess everything up.
The firewall will also allow you to check for noise and signal strength levels on a circuit by circuit basis.
As with X10, it the device causing problems is on a firewalled (not really a word I guess) circuit, the you will still need a plug-in filter.
At least the firewall reduces the issues caused and will also help you find the problem.
1) Not all circuits are on the firewall. Only circuits that you want to control.
2) If the device that is causing problems happens to be on the firewall, the firewall will also filter noise on the circuit. Granted that circuit will still have noise but it at least won't mess everything up.
The firewall will also allow you to check for noise and signal strength levels on a circuit by circuit basis.
As with X10, it the device causing problems is on a firewalled (not really a word I guess) circuit, the you will still need a plug-in filter.
At least the firewall reduces the issues caused and will also help you find the problem.
Beyond the things Martin mentioned, the firewall checks for "legitimate" signals and zaps anything that isn't right. That's why you cannot use Leviton Extended code devices in conjuction with a firewall, as the firewall filters out the extended codes. All the extra circuitry for this is why the firewall is more expensive than all the other coupler/repeaters.
Exactly the point I was (trying) to make earlier.
Martin
I haven't done any work linking multiple firewalls together, but you might be able to use it to solve your two building problem. IIRC they use cat-5 to connect the two firewalls.
Exactly the point I was (trying) to make earlier.
Martin
I haven't done any work linking multiple firewalls together, but you might be able to use it to solve your two building problem. IIRC they use cat-5 to connect the two firewalls.
BraveSirRobbin
Moderator
The only real downside I see with the Compose Firewall is you more or less have to get it installed with the house build (or heavily modify circuits near/in your breaker boxes.
Also, what happens when you want to add some lighting control on a circuit that isn't included in the firewall later on down the road?
I have a friend who will be building a house later this summer and I might need to decide between all of these options sooner than I want.
I did convince him to go with an Elk M1 Gold, so maybe UPB is the better choice since (as far as I know) the Elk already supports it.
Also, what happens when you want to add some lighting control on a circuit that isn't included in the firewall later on down the road?
I have a friend who will be building a house later this summer and I might need to decide between all of these options sooner than I want.
I did convince him to go with an Elk M1 Gold, so maybe UPB is the better choice since (as far as I know) the Elk already supports it.
If I were to start over, I'd use UPB. I just don't trust smarthome to get it right anymore. I'm even more afraid of how they could manage to mess up insteon.
I mean, if they can't even get a simple thing like congestion backoff working, then what hope do they have on something complex?
Pop quiz: what happens if you query three smarthome switchlinc-relay devices for status together?
Answer: you get to make a trip to the breaker box to power cycle them. because they're generating massive amounts of 120KHz noise and totally trashing any hope you have of using X10 again.
I had two switchlinc-relay devices next to each other for the kitchen lights. My wife has the knack of pressing the toggles in such a way that they're close enough to simultanious as to cause them to get into the retransmit loop and take out the X10 network. Using a grouped status_request is dead simple to trigger it - it takes more skill to trigger it from the toggle buttons. Grouped X10 status requests are silly commands to send, but they demonstrate the problem spectacularly and easily. The bug is far more pervasive than that though - I can reproduce it by simply power cycling the breaker (about 1 in 8 times on our house), or as often as once every 1-2 days when my wife is home.
Also, the 130KHz-ish insteon carrier is awfully close to the 120KHz X10 carrier and we all know how many things absorb the 120KHz signal. If Insteon uses an x10-style tuned circuit to transmit the carrier, then they'll have the same problem with multiple transmitters absorbing each other's signal. I sure hope they get the every-device-is-a-repeater algorithms sorted out and working!
I've checked and can reliably control my test UPB appliance modules when plugged into a neighbor's house, 6 doors away, other side of the street. I can't even reach half of the 120KHz devices in my own house reliably. I did resort to a passive phase coupler to improve the upb signal in the other phase. It was amusing that I got better UPB connectivity to the other phase at the neighbor's house than to the other phase in my own house. Anyway, I get "excellent" signal (100% signal strength, 0 noise) everywhere I've tested in my house on the same phase, even through a GFCI. I get "good" signal (40% signal strength, 0 noise) on the other phase. No dead spots, even when the modules are plugged into surge arresters etc.
That being said, I'm going to tinker with insteon out of curiosity (I have a devkit). But I'm going to keep on gradually replace my troublesome smarthome switchlinc-relay devices with UPB devices as they annoy me. Fortunately, I've got rid of the two worst offenders (the kitchen). The reason I'm curious about insteon is because of the supposedly dramatically higher data rate. But I'm not sure if I'm willing to trade data rate for firmware bugs or unreliability especially when UPB has so far had a perfect reliability record for me. (every single command acked on first transmission attempt so far, according to the logs)
In case you didn't pick it up, I'm still peeved at smarthome for not having a clue how to reproduce this or what to do about it.. that's really great after spending something like $5000 (or more) on their stuff. I seem to get the blow-off response because the easiest way to provoke it is a useless (but valid) X10 command. I sorely regret buying things from them now. That reminds me, I have got to RMA some dead switchlinc-relay switches.
Z-wave is out for me, I already have too much 2.4GHz noise in the house and nearby. I don't need the aquarium controls to go offline because the microwave is turned on or something silly like that.
-Peter
I mean, if they can't even get a simple thing like congestion backoff working, then what hope do they have on something complex?
Pop quiz: what happens if you query three smarthome switchlinc-relay devices for status together?
Answer: you get to make a trip to the breaker box to power cycle them. because they're generating massive amounts of 120KHz noise and totally trashing any hope you have of using X10 again.
I had two switchlinc-relay devices next to each other for the kitchen lights. My wife has the knack of pressing the toggles in such a way that they're close enough to simultanious as to cause them to get into the retransmit loop and take out the X10 network. Using a grouped status_request is dead simple to trigger it - it takes more skill to trigger it from the toggle buttons. Grouped X10 status requests are silly commands to send, but they demonstrate the problem spectacularly and easily. The bug is far more pervasive than that though - I can reproduce it by simply power cycling the breaker (about 1 in 8 times on our house), or as often as once every 1-2 days when my wife is home.
Also, the 130KHz-ish insteon carrier is awfully close to the 120KHz X10 carrier and we all know how many things absorb the 120KHz signal. If Insteon uses an x10-style tuned circuit to transmit the carrier, then they'll have the same problem with multiple transmitters absorbing each other's signal. I sure hope they get the every-device-is-a-repeater algorithms sorted out and working!
I've checked and can reliably control my test UPB appliance modules when plugged into a neighbor's house, 6 doors away, other side of the street. I can't even reach half of the 120KHz devices in my own house reliably. I did resort to a passive phase coupler to improve the upb signal in the other phase. It was amusing that I got better UPB connectivity to the other phase at the neighbor's house than to the other phase in my own house. Anyway, I get "excellent" signal (100% signal strength, 0 noise) everywhere I've tested in my house on the same phase, even through a GFCI. I get "good" signal (40% signal strength, 0 noise) on the other phase. No dead spots, even when the modules are plugged into surge arresters etc.
That being said, I'm going to tinker with insteon out of curiosity (I have a devkit). But I'm going to keep on gradually replace my troublesome smarthome switchlinc-relay devices with UPB devices as they annoy me. Fortunately, I've got rid of the two worst offenders (the kitchen). The reason I'm curious about insteon is because of the supposedly dramatically higher data rate. But I'm not sure if I'm willing to trade data rate for firmware bugs or unreliability especially when UPB has so far had a perfect reliability record for me. (every single command acked on first transmission attempt so far, according to the logs)
In case you didn't pick it up, I'm still peeved at smarthome for not having a clue how to reproduce this or what to do about it.. that's really great after spending something like $5000 (or more) on their stuff. I seem to get the blow-off response because the easiest way to provoke it is a useless (but valid) X10 command. I sorely regret buying things from them now. That reminds me, I have got to RMA some dead switchlinc-relay switches.
Z-wave is out for me, I already have too much 2.4GHz noise in the house and nearby. I don't need the aquarium controls to go offline because the microwave is turned on or something silly like that.
-Peter
Regarding WAF.. so far, the WAF here has been reasonably good for UPB. The WAF for the smarthome switches started out as 'excellent' and degraded to 'get the damn things out!'. When the remote light controls in the bedroom stop working because the kitchen switchlinc-relay switches have trashed the network, that turns things sour pretty quick. The keypad can't send messages to the other dimmers in the same room, which means getting back out of bed. That REALLY makes WAF plummet quickly.jlehnert said:
The only thing that's stopped an 'excellent' WAF rating for the UPB switches has been the latency at the rocker switches. It seems that they pause for a moment to differentiate between a single and double tap. I haven't found a way to turn this off yet. There is an awful lot of programming control on these switches, I should re-read it and see if there is a threshold in one of the registers somewhere. Anyway, we can live with this easily. It is only a very minor annoyance.
One addendum.. Now that I've read through the Insteon SDK docs, their design looks fascinating!
The powerlinc that comes with it is actually a controller. You can download your own program to the powerlinc.
Even more interesting is that the docs talk about downloading your own programs into other Insteon devices too... How very interesting. I thought I read a mention of downloading SALad code onto a switchlinc, but I couldn't find it with a quick skim-read to double check.
They also describe the flood based repeater network that they use for insteon messages...
The powerlinc that comes with it is actually a controller. You can download your own program to the powerlinc.
Even more interesting is that the docs talk about downloading your own programs into other Insteon devices too... How very interesting. I thought I read a mention of downloading SALad code onto a switchlinc, but I couldn't find it with a quick skim-read to double check.
They also describe the flood based repeater network that they use for insteon messages...
PeterW - You might have good other reasons to reject Z-Wave, but don't worry about the operating band. It works at 900MHz - not in 2.4GHz.
Check this: http://www.totalvoicecontrol.com/ProtocolRoundUp.html
Check this: http://www.totalvoicecontrol.com/ProtocolRoundUp.html
Ahh, ok.
FWIW, Insteon-RF is 904MHz. Insteon-PLC is centered at 131.65KHz, but it is a FSK system. They should be able to get a decent transfer rate from that.
Is Z-wave a fully open spec? ie: can I go write my own software to use it on Unix for fun? (I can do that without any hassles at all for UPB, Insteon and X10). Some windows DLL or API or worse - NDA'ed docs is a deal killer for me. So far I haven't found much about Z-wave except for people grumbling about the lack of docs.
-Peter
FWIW, Insteon-RF is 904MHz. Insteon-PLC is centered at 131.65KHz, but it is a FSK system. They should be able to get a decent transfer rate from that.
Is Z-wave a fully open spec? ie: can I go write my own software to use it on Unix for fun? (I can do that without any hassles at all for UPB, Insteon and X10). Some windows DLL or API or worse - NDA'ed docs is a deal killer for me. So far I haven't found much about Z-wave except for people grumbling about the lack of docs.
-Peter
rocco
Active Member
Peter, it is not an open spec, but any one can play. All you need to do is buy the development kit and sign the NDAs (sorry), and have fun. All that fun will cost you about $10,000.PeterW said:
The price is the reason I went with ZigBee. I got the development kit for free, and ZigBee seems to be much more capable then ZWave (from what I can tell).
Am I the only one who'd prefer a custom solution? If working on a new build (like I was), it was really nice running CAT5 to each switch and each light back to a central location to eventually put dimmers in there with digital control. At the moment I haven't got that far, I have homemade QPROX light switches built behind blank plates into the CAT5 which just trip mains relays at the other end for simple on/off control. However my full control system is only a few months away I hope 
BraveSirRobbin
Moderator
Are you saying you are providing custom relay's that switch the AC to the lights and these switches are controlled via the Cat5 wiring? Not sure of your setup, but interested in the details.
yes and no.
at the moment I've built a simple QPROX touch sensor switch for each room. it's a simple circuit, using 2 pairs of the CAT5. one for +5/GND, the other for two logic signals. The chip I use gives me two channels, with two electrodes placed behind the wall plate. Touching the left side of the wall plate would control channel 1, the right, channel 2. The logic line (one for each channel) stays high the whole time the button is touched for.
now, temporarily, until I build my full serial controlled dimmers (about 80% done on the design and prototypes, 90% done on the PCB layouts and I have all the components I think I'll need), I have just some mains relays I got secondhand almost free controlling the lights.
Some PICAXE18X's are setup with simple code to basically flipflop 8 channels each. I have 2 setup, and 15 mains relays for the main lights in my house. So pushing (almost only the left one at the moment, as its pretty much one light per room) the faceplate, will just turn on/off the light.
Eventually I will have serial interfaces, dimming, etc setup.
The lights are all wired back to the 'control centre' using 1.5mm T&E, which then go to the relays for now.
Really smart. Can't wait to have the final project finished.
at the moment I've built a simple QPROX touch sensor switch for each room. it's a simple circuit, using 2 pairs of the CAT5. one for +5/GND, the other for two logic signals. The chip I use gives me two channels, with two electrodes placed behind the wall plate. Touching the left side of the wall plate would control channel 1, the right, channel 2. The logic line (one for each channel) stays high the whole time the button is touched for.
now, temporarily, until I build my full serial controlled dimmers (about 80% done on the design and prototypes, 90% done on the PCB layouts and I have all the components I think I'll need), I have just some mains relays I got secondhand almost free controlling the lights.
Some PICAXE18X's are setup with simple code to basically flipflop 8 channels each. I have 2 setup, and 15 mains relays for the main lights in my house. So pushing (almost only the left one at the moment, as its pretty much one light per room) the faceplate, will just turn on/off the light.
Eventually I will have serial interfaces, dimming, etc setup.
The lights are all wired back to the 'control centre' using 1.5mm T&E, which then go to the relays for now.
Really smart. Can't wait to have the final project finished.
Similar threads
