Gen1 UPB vs. Gen2 UPB

[JonW]

Ano - have you used a PCS signal bridge or just the SAI ones? I would be curious as to whether there is any performance difference. If you have not tried a PCS one, I may still have one I could send you to try. It would be interesting to see if there is a difference. It doesn't sound like the SPR is the ideal solution since it is a Gen II only device.

At the risk of beating a dead horse here, it really sounds like your environment has a noise and/or attenuation problem that is just killing your performance and maybe no bridge or SPR is going to help. Just like in the X10 days, we ran around unplugging things or turning things off to find the problem. I can't see you getting to the bottom of the issue unless you do this step (or have you?). Like I mentioned earlier, I think turning off the breakers and doing isolated circuit tests is your next step.

 

[IndyMike]

My repeater is enabled in UpStart. The reason that the strength measured on the opposite phase for GEN I with the repeater is that the return from the GEN I switch is only sent once (so repeater does not repeat it) and it passes through my phase coupler to UpStart. UpStart and the repeater are working fine with mixed GEN I and GEN II -- I understand that the responses FROM the GEN I switches will not be repeated on the other phase (because GEN I switches do not send multiple xmits) but they do receive repeated xmits from HomeSeer through thru the repeater.

Why do you think UpStart would not see GEN I switches on the other phase when the repeater is in use and UpStart knows about it? The GEN I switches still respond to UpStart, they are just not repeated through the repeater -- they are coupled (at low signal strength) across the phases.

tenholde

It looks like tenholde may have nailed it.

Here's a quote from the PCS "Tools" site:

2.1. When Can I Use An SPR?
Any home that uses 60Hz split phase 120/240VAC electrical power can take advantage of an SPR to improve the UPB™ communication reliability. The Split-Phase Repeater is a Generation II UPB™ device. It uses a mechanism that is built into the Generation II UPB™ protocol, called multi-packet messaging, to operate (refer to section 7. for an explanation of the multi-packet messaging mechanism). Generation II UPB™ devices, controllers, and setup/configuration tools that fully support multi-packet messaging can take full advantage of the SPR. The Split-Phase Repeater is designed to be plug-and-play – just install it and forget about it. There are no special setup/configuration steps necessary to begin using the SPR1.

2.2. When Can’t I Use An SPR?
The SPR is not recommended for use with Generation I UPB devices:
Unfortunately, old Generation I UPB™ devices did not fully support the multi-packet messaging mechanism. These “older” devices only allow multi-packet messaging to be used for one-way communication. Even with this limitation, Generation I UPB™ devices may still make use of the SPR for performing one-way command transfers. Many, if not most of, existing Generation I UPB™ systems do not rely on two-way communication for normal operation. These systems only require commands to be sent from one (or more) transmitting device(s) to one (or more) receiving devices. They do not expect the receiving devices to transmit anything back. In this situation, the SPR will undoubtedly improve the one-way communication reliability.

The document is located here: PCS SPR Description

The above description pretty much fits your problem statements. In retrospect, I'd say you overbought - a passive coupler would not have these issues.

 

[ano]

There are lots of things being discussed here. Passive bridges do support all modes, but the problem is they don't work for everyone. I've never tried the PCS bridge, but I have tried Simply Automated inverting and non-inverting bridges with not much luck with either. With three bridges in parallel I get 4 or 5 and things generally work, but its not ideal.

A repeater isn't overkill. It actually works very good with Gen1 switches from the originating device to other devices which is what I need 99% of the time. Signal at the switches has increased from 4 or 5 to 30 or 40+ BUT, as the documentation says, the downside is no return signal on the opposite phase. So what does that mean?

From my experience, and someone shout out if you see it differently, it appears that ELK and HAI don't make use of this return signal anyway. If I turn a switch on or off with my HAI panel, it assumes its on or off even if the switch is not there. CQC does seem to transit multiple times if a signal is not received, but that is all. If you use a repeater this really doesn't matter because the main signal is so good it always works, so the reply doesn't matter anyway, in my opinion. So the repeater is great.

So the down side? As far as I can tell, the only program that really makes use of the return switch data is UPStart. UPStart actually does update the device status based on the return data. From what I have seen, UPStart runs in one of two modes, Uni-packet return mode or multi-packet return mode. From what I've seen, if you tell UPStart you are using a repeater, it ONLY watches for multi-packet return data. If you tell it NOT to use your repeater, then UPStart watches for direct data from the switch. This is why I was surprised by what tenholde said because mine doesn't seem to work that way, but I did notice I had a 6 month-old UPStart, so maybe things changed.

When I put UPStart in REPEATER mode, UPStart can't setup or "see" switches on the opposite phase, which makes sense to me because UPStart is only looking for repeated multi-packets or switches on the same phase which don't need repeating. When I turn OFF repeater mode, UPStart can see all switches, but the signal level is marginal. This seems to make sense, or at lease that is how it works with with my opposite phase GEN1 switches.

I do have some GEN2 switches on the opposite phase, and those ARE seen by UPStart in Repeater mode.

So again, the repeater is still very helpful in a GEN1 switch world, as long as you are aware of the UPStart problems, and I'm still not positive if HomeSeer or CQC really makes use of the repeater for return data. It MAY, just use the data returned like UPStart does when its NOT in repeater mode.

So this is how I understand it, and please correct me if you do understand it differently. I upgraded to the newest UPStart so maybe some things changed.

Also, I would be interested if tenholde could go into UPStart, activate it to use your repeater, then do both a Repeater Comm Test, and a Network Comm Test. (Under the Tools Tab.) For me, all Gen1 switches on the opposite phase fail both tests. On the Repeater test it actually says FAIL for each switch. On the Network Comm Test the return signal level is blank. So in both cases, no-go. Now when I restart UPStart, and tell it NOT to use the repeater, the Network Comm test passes ALL switches. (No repeater test, because no repeater was used.) Do you get the same or different results?

UPDATE: Updated to the newest UPStart, but it behaves exactly the same.

 

[pete_c]

I've never tried the PCS bridge, but I have tried Simply Automated inverting

I removed my SA passive bridge and replaced it with a PCS passive bridge and it made a difference in the opposite side UPB devices.

Other than using Upstart and the HAI PIM; what means of diagnostics would you use to see what HAI is doing with the UPB HAI PIM?

I am curious why a dual phase HAI PIM doesn't exist. What would make the engineering of this product difficult?

 

[tenholde]

There are lots of things being discussed here. Passive bridges do support all modes, but the problem is they don't work for everyone. I've never tried the PCS bridge, but I have tried Simply Automated inverting and non-inverting bridges with not much luck with either. With three bridges in parallel I get 4 or 5 and things generally work, but its not ideal.

A repeater isn't overkill. It actually works very good with Gen1 switches from the originating device to other devices which is what I need 99% of the time. Signal at the switches has increased from 4 or 5 to 30 or 40+ BUT, as the documentation says, the downside is no return signal on the opposite phase. So what does that mean?

From my experience, and someone shout out if you see it differently, it appears that ELK and HAI don't make use of this return signal anyway. If I turn a switch on or off with my HAI panel, it assumes its on or off even if the switch is not there. CQC does seem to transit multiple times if a signal is not received, but that is all. If you use a repeater this really doesn't matter because the main signal is so good it always works, so the reply doesn't matter anyway, in my opinion. So the repeater is great.

So the down side? As far as I can tell, the only program that really makes use of the return switch data is UPStart. UPStart actually does update the device status based on the return data. From what I have seen, UPStart runs in one of two modes, Uni-packet return mode or multi-packet return mode. From what I've seen, if you tell UPStart you are using a repeater, it ONLY watches for multi-packet return data. If you tell it NOT to use your repeater, then UPStart watches for direct data from the switch. This is why I was surprised by what tenholde said because mine doesn't seem to work that way, but I did notice I had a 6 month-old UPStart, so maybe things changed.

When I put UPStart in REPEATER mode, UPStart can't setup or "see" switches on the opposite phase, which makes sense to me because UPStart is only looking for repeated multi-packets or switches on the same phase which don't need repeating. When I turn OFF repeater mode, UPStart can see all switches, but the signal level is marginal. This seems to make sense, or at lease that is how it works with with my opposite phase GEN1 switches.

I do have some GEN2 switches on the opposite phase, and those ARE seen by UPStart in Repeater mode.

So again, the repeater is still very helpful in a GEN1 switch world, as long as you are aware of the UPStart problems, and I'm still not positive if HomeSeer or CQC really makes use of the repeater for return data. It MAY, just use the data returned like UPStart does when its NOT in repeater mode.

So this is how I understand it, and please correct me if you do understand it differently. I upgraded to the newest UPStart so maybe some things changed.

Also, I would be interested if tenholde could go into UPStart, activate it to use your repeater, then do both a Repeater Comm Test, and a Network Comm Test. (Under the Tools Tab.) For me, all Gen1 switches on the opposite phase fail both tests. On the Repeater test it actually says FAIL for each switch. On the Network Comm Test the return signal level is blank. So in both cases, no-go. Now when I restart UPStart, and tell it NOT to use the repeater, the Network Comm test passes ALL switches. (No repeater test, because no repeater was used.) Do you get the same or different results?

UPDATE: Updated to the newest UPStart, but it behaves exactly the same.

I will try UpStart later today to verify.

HomeSeer DOES use responses FROM switches that it has sent a command to to verify that it worked only when sending commands to switches, not when sending links. HomeSeer is also able to take advantage of the repeater when it Xmits if you have it setup to repeat its transmissions.

I also assume, but have not been able to verify, that when a GEN I sends a link command to another switch on the opposite leg and the sending switch has been setup to send multiple packets, that the repeater will repeat these link commands.

 

[ano]

I will try UpStart later today to verify.

HomeSeer DOES use responses FROM switches that it has sent a command to to verify that it worked only when sending commands to switches, not when sending links. HomeSeer is also able to take advantage of the repeater when it Xmits if you have it setup to repeat its transmissions.

I also assume, but have not been able to verify, that when a GEN I sends a link command to another switch on the opposite leg and the sending switch has been setup to send multiple packets, that the repeater will repeat these link commands.

I checked CQC, and it claims to use also use the reply. There is no reply from a link, so what CQC does is wait a few seconds, then poll each switch in the link and set their state. I have had discussions a ways back suggesting that CQC should instead just assume the link worked and set states accordingly, because 1) UPB is very reliable, and because 2) polling afterwards takes a long time for large links and puts lots of traffic on the powerline.

I will do some more testing this weekend to see what the repeater does and doesn't do. If I turn off my bridge, I have ZERO signal across phases, so the perfect repeater test environment.

pete_c

I've always wondered why a dual-phase PIM doesn't exist also. The big problem would be installation, it would really need to be plugged into a 240V socket which might be a distance from their computer.

With your recommendation I might buy the PCS passive bridge to see if it does any better than the Simply Automated versions. It can't hurt to try it.

> Other than using Upstart and the HAI PIM; what means of diagnostics would you use to see what HAI is doing with the UPB HAI PIM?

I know for a fact that the HAI panel doesn't use the acknowledge from switches. You can see that by just turning on or off a switch you have removed. The panel will THINK its on or off. I bet ELK is no different. Other than that, its probably just trial and error.

When I shut-off my bridge, I'm going to run the following tests with the repeater, all with GEN1 switches....
1) Is the repeater forwarding the original PIM to switch commands across phases? (We think YES)
2) Is the repeater forwarding the switch reply across phases? (We think NO)
3) When a switch on one phase sends a link to a switch on the other phase, does the repeater carry that traffic?
4) When you manually change a switch, and it sends the status back to the PIM, does the repeater carry that traffic?

 

[pete_c]

I've always wondered why a dual-phase PIM doesn't exist also. The big problem would be installation, it would really need to be plugged into a 240V socket which might be a distance from their computer.

It would be an issue. I made the assumption that all folks plug their PIM's right off their fuse panels. I did this initially with my X10 CM11 PIM in the late 1990's. In the new home I did the same thing with the Inteon PLC/PLM and UPB PIMs. I also made another assumption relating to said product because it doesn't bother me to work inside of my fuse panel.

With your recommendation I might buy the PCS passive bridge to see if it does any better than the Simply Automated versions. It can't hurt to try it.

The thread is buried here somewhere when I had problems (noise etc) and switched over to the PCS bridge base on the "evolution" of the passive bridge and the issues that everyone had with opposite phase UPB signal strengths.

 

[JonW]

From my experience, and someone shout out if you see it differently, it appears that ELK and HAI don't make use of this return signal anyway. If I turn a switch on or off with my HAI panel, it assumes its on or off even if the switch is not there. CQC does seem to transit multiple times if a signal is not received, but that is all. If you use a repeater this really doesn't matter because the main signal is so good it always works, so the reply doesn't matter anyway, in my opinion.

HAI actually polls the switches to get their status (when used in HLC mode).

 

[pete_c]

That is actually how I found out that I had been using a defective HAI PIM (converting my HAI OPII setup from initial UPB setup to HLC mode).

I initially assumed that I wouldn't see status because I never did early on (on the OPII) but did with HS / floating PIM.

This discussion has become very informative relating to UPB in general.

 

[ano]

That is actually how I found out that I had been using a defective HAI PIM (converting my HAI OPII setup from initial UPB setup to HCL mode).

I initially assumed that I wouldn't see status because I never did early on.

I don't use HLC mode, but it sounds like what CQC does, just poll all the switches in a link. It can take a LONG time if you have many switches.

 

[Digger]

From my experience, and someone shout out if you see it differently, it appears that ELK and HAI don't make use of this return signal anyway. If I turn a switch on or off with my HAI panel, it assumes its on or off even if the switch is not there. CQC does seem to transit multiple times if a signal is not received, but that is all. If you use a repeater this really doesn't matter because the main signal is so good it always works, so the reply doesn't matter anyway, in my opinion.

HAI actually polls the switches to get their status (when used in HLC mode).

How often does it poll? If you try and turn on a switch during the polling does it affect the reliability or is the HAI panel the main controller and it coordinates this seemlessly like the ISY does for Insteon?

 

[JonW]

How often does it poll? If you try and turn on a switch during the polling does it affect the reliability or is the HAI panel the main controller and it coordinates this seemlessly like the ISY does for Insteon?

It polls each switch of a room a few seconds after receiving a transmission either from a switch in that room or a keypad in that room. It does not continously poll all switches. It only polls when it sees activity and even then, it only polls the other switches in the room of which the transmission came from. I think this is the biggest reason for "programming on the 8's". I don't recall any problems with reliability during the poll process.

I have many lights set to use "last level" when they are turned on. HAI will immediately switch the status to "ON" when it sees the initial transmission, then it will quickly update it to 80% or whatever level as soon as it polls the exact status.

 

[Digger]

How often does it poll? If you try and turn on a switch during the polling does it affect the reliability or is the HAI panel the main controller and it coordinates this seemlessly like the ISY does for Insteon?

It polls each switch of a room a few seconds after receiving a transmission either from a switch in that room or a keypad in that room. It does not continously poll all switches. It only polls when it sees activity and even then, it only polls the other switches in the room of which the transmission came from. I think this is the biggest reason for "programming on the 8's". I don't recall any problems with reliability during the poll process.

I have many lights set to use "last level" when they are turned on. HAI will immediately switch the status to "ON" when it sees the initial transmission, then it will quickly update it to 80% or whatever level as soon as it polls the exact status.

If it only polls a group of eight (max) then I doubt there is a reliabilty issue. I guess you can keep it fairly accurate. I beleive the ISY for UPB will be a little more robust but I dont have one yet to see.