Explain This Regarding Dual Band Devices

[TJF1960]

Is it possible that the noise floor is high enough to prevent the plm from detecting the 0 crossing?

[IndyMike]

Lou,

 

I really am not trying to defend or implicate SH in any way. I'm simply trying to present my observations in the hope that they may help others.

 

We come from different backgrounds: I still have X10 devices in my home and noise is a real problem for these devices. That said, I use only three filters in the home (I've replaced the dimmer on the 8 CFL's with a relay - no longer an issue). I'm very curious what type device you have that is causing these problems. I've been searching for a reliable noisemaker for test purposes - you may have found one.

 

TJF1960,

 

Noise problems should occur in frequencies far above the 60Hz powerline. I would hope that the zero crossing detection would incorporated a low pass filter to eliminate these frequencies. It's been a long time since any schematics were made available for this area of the devices ('2005? - and those were sketchy). I have no clue what the new devices are using for the zero crossing.

[Brian H]

Maybe his. Three phase well pump. On a single phase to three phase converter.

viewtopic.php?f=28&t=7639

[IndyMike]

Interesting that the well is a 3 phase unit. While it could certainly cause transient noise on start-up, I would think that continuous noise while running would be fairly easy to diagnose. Shut down the breaker - observe communications. This of course assumes a point noise source (not an additive effect).

 

As a point of reference, I have a Grundfos variable speed well pump on a 30 Amp breaker. While this device has been identified as causing problems for UPB communications, I have never been able to confirm a problem with either X10 or Insteon.

[apostolakisl]

Maybe his. Three phase well pump. On a single phase to three phase converter.

viewtopic.php?f=28&t=7639

 

 

Yeahh, that is one of them. Probably not something you are going to pick up just to test out noise.

 

The other is either my kegerator or my small wine refrigerator. Both are on the same circuit. I think it is the wine refrigerator as it has a screwy electronic cooling system. The beer unit just has regular old small fridge compressor. I didn't even want the wine cooler. Mom and Dad gave it to us as a "building your house warming gift" and insisted we install it even though we put in a 2000 bottle walk in cellar.

 

I have found that style program is a very good test of noise.

 

If
       Status  'Alexis Room / Alexis BR-Overhead L' is Off
   And Control 'Alexis Room / Alexis BR-Overhead L' is switched Fast Off

Then
       Set 'Alexis Room / Alexis BR-Overhead L' 25%

Else
  - No Actions - (To add one, press 'Action')

 

It only works properly in a very clean environment. As I have added this to all of the bedrooms as a night light thing, I have confounded my family with its failure to work. I would like to get that fixed!!!

[Brian H]

Teken; Almost missed your question.

The tests with the scanner didn't show too much. Other than with all my Access Points disconnected so only the 2413S was sending RF. I could start something like a display link table for a device known to have many links. Then walk around the house listening for the data busts.

[apostolakisl]

Interesting that the well is a 3 phase unit. While it could certainly cause transient noise on start-up, I would think that continuous noise while running would be fairly easy to diagnose. Shut down the breaker - observe communications. This of course assumes a point noise source (not an additive effect).

 

As a point of reference, I have a Grundfos variable speed well pump on a 30 Amp breaker. While this device has been identified as causing problems for UPB communications, I have never been able to confirm a problem with either X10 or Insteon.

 

I don't think the motor is the problem, I think it is the electronics that do the 1 to 3 phase conversion. This is the latest fad in efficiency, turning 1 phase to 3 phase. 3 phase is really a hell of a lot better and I don't understand why it is only used in commercial buildings. Three phase is running down our street and probably most streets. My HVAC does the same thing, but I am pretty sure it does not cause trouble as I have never seen any change in behavior when it is running or not.

 

The well pump was quite obvious. This morning, the switches on the same subpanel as the well screwed up. The sprinkler was running. I shut down the sprinkler, all was fine. Turned it back on, screwy. Probably need to shut off the sprinkler for the season anyway. Maybe I should try to actually figure out the irrigation module.

[apostolakisl]

Maybe we should just have a conference call. Seems like all of us are posting at once here!

[IndyMike]

Lou,

 

I may be out of date, but the only methods I'm familiar with for converting single to 3 phase power are static and dynamic. Many years are a friend and I used the dynamic method on a CNC/Lathe setup - it requires the use of a 3 phase motor.

 

I've never tried the static method. I wouldn't call either method efficient.

 

I've googled and don't see any "new" methods for conversion.

 

Are you sure this is a three phase motor? Is it possibly a variable speed DC motor (H bridge)? This would be similar to my Grundfos well.

 

From your last post - I for one would be happy to participate in a conference call (not sure if you were serious). If such a thing were to come about, I would suggest a published agenda/questions for discussion.

 

IM

[apostolakisl]

No, it's for sure a three phase.

 

I have the bf50 model.

 

http://www.pumpsandwells.com/miva/Balan ... 01-5hp.pdf

 

page 5 summarizes what you need to know. I have the 5hp pump.

 

My heat pumps use similar technology.

[TJF1960]

I don't think the motor is the problem, I think it is the electronics that do the 1 to 3 phase conversion.

Perhaps you can isolate and filter just the electronics.

[apostolakisl]

I don't think the motor is the problem, I think it is the electronics that do the 1 to 3 phase conversion.

Perhaps you can isolate and filter just the electronics.

 

Yeah, my intention is to put a filter in the pump controller box, not the breaker panel. I actually at some point might want to put some Insteon stuff down at the pump house.

[TJF1960]

When posting my last comment I was thinking you might be able to get away with a smaller filter after looking at the controller terminal strips which don't look like they handle 10 gauge wire. I was thinking there must be another module which handled the heavy amperage and switching to the motor. After looking thru the manual though I am not so sure now. I didn't see a reference to anything but the controller board.

[IndyMike]

Lou,

 

My apologies - I was stuck back in the dark ages with the 3-phase converter terminology.

 

You likely have a 3 phase motor that is being fed by a DC - AC inverter (i.e. PWM). In short, it's very similar to my well pump. If you're interested, there's a microchip app note here: http://ww1.microchip.com/downloads/en/appnotes/00843a.pdf

 

You can try filtering the low voltage side, but that is probably not where the damage is being done. The high voltage drivers (FETs, IGBT's, etc) that are actually driving the motor are most likely the culprits. These are the devices that are generating transients and serious current.

 

I did have problems with my controller earlier this year. It uses tank pressure as feedback to vary the pump speed. I found that my 1 gallon tank had become "waterlogged" and was causing the pump to "porpoise" (wild speed variations while trying to control pressure - under damped system). The speed variation was driving transients through the electrical system.

 

Check the air charge in your tank to make sure you have sufficient damping in your system.

[apostolakisl]

When posting my last comment I was thinking you might be able to get away with a smaller filter after looking at the controller terminal strips which don't look like they handle 10 gauge wire. I was thinking there must be another module which handled the heavy amperage and switching to the motor. After looking thru the manual though I am not so sure now. I didn't see a reference to anything but the controller board.

 

That guy has 6g wire going down the hole 650 feet. It cost me a pretty penny. I forget what it is from the breaker box to the controller. I think 8g. It is a much shorter distance.

 

I started to read your article Indymike. Sorry to say, it is too late at night to read that kind of stuff. But I do know that it is 240v 1 phase into the controller, and 3 phase 240 volts out of the controller. What happens in the guts, I don't know. It is variable speed and I don't know how that is done, perhaps it lowers the voltage for less than full speed. Maybe your article splains it.

 

I am familiar with the sea mammal to which refer. At certain gpm's it does tend to do that, but not terribly so. I screwed around with getting the tank pressure exactly as spec'd, and it still does it a little. For the most part, I have it cranked when running the sprinkler so that isn't so much an issue.

[Winter Dragoness]

I'm coming in two years late, but so far, I haven't found anyone else talking about this subject, so I thought I'd add my data points to the discussion.

 

I'm using MisterHouse through a latest-model Smartlinc rather than ISY so I don't have the scene scanning tool, but I was sometimes seeing corrupt responses from a 2486SWH6 keypadlinc when I scanned its link table. MisterHouse does some inefficient link scanning by individually requesting every link table entry, so it creates a lot of traffic and sends a lot of extended messages back and forth, resulting in a pretty good test. The keypadlinc is a discontinued single-band relay keypad, but it's not all that old since Mobilincstore still has them in stock. The corrupt responses were not surprising since the keypad is on the opposite power phase from the Smartlinc PLM and I had no physical or RF bridge between the phases. Usually there was no corruption but I could not track down a cause for the clean vs dirty signal.

 

I installed a latest model Micro On/Off module 2443-222 to RF bridge one power phase to two latest model SwitchLinc Relays 2477S on the other power phase. I did not expect that to solve the problem because the Insteon white paper says that a dual-band device will take a powerline signal and repeat it to RF, then another RF receiver will repeat that back to the powerline on the other phase, so I expected that on a noisy line, the original powerline signal from the keypadlinc would also pass between phases and reach the Smartlinc after it had gotten corrupted. The Smartlinc would then see one corrupt message and one good message passed over RF. Since the keypadlinc is too old to be sending I2CS with the checksum, I expected Misterhouse to get one corrupt and one good message and have no way to tell which was good, perhaps assuming they were both good.

 

In practice, every link scan I've done of that keypadlinc in the last three days since RF bridging has come back flawlessly, and I've done dozens of scans throughout each day. I've also tried adding noise to the line, turning on tons of CFL bulbs, plugging in every charging device I can find, turning on a bathroom fan, etc. I never get a corrupted link scan.

 

My only theory is that the newer firmware devices will wait a bit to see if they get an RF or powerline signal that appears to be a repeated signal, and they'll choose to act on or repeat only the strongest of the signals.

 

Another remote possibility is that I simply haven't had the same line noise that caused the original corruption I saw during any of my link scans. That seems unlikely, but I can't think of any way to test things further other than to deactivate the Micro On/Off module to unbridge the phases, but it's behind a cabinet and on the same circuit breaker as the Smartlinc and other networking equipment, so it's a big pain to test that. To unbridge from the other side I would have to physically disconnect a Switchlinc that is on the same circuit as the Keypadlinc.

[Xathros]

cdragon-

 

Why not just airgap the dualband switchlincs on the opposite phase for a test? Should have the same affect of breaking the phase bridging.

 

And to add some of my experiences to your report, I have noticed increasing occurrences of repeated responses to Insteon events as I have increased the number of dual band devices. I have had to add program delays in a few cases to filter out the dupes.

 

-Xathros

[Winter Dragoness]

Why not just airgap the dualband switchlincs on the opposite phase for a test? Should have the same affect of breaking the phase bridging.

 

And to add some of my experiences to your report, I have noticed increasing occurrences of repeated responses to Insteon events as I have increased the number of dual band devices. I have had to add program delays in a few cases to filter out the dupes.

 

Ah, airgap. The feature I always forget. I did that and had one experiment where MH timed out because it never got a response for five retries to a particular query, but other than that, no corruption so far.

 

I wish Smarthome wasn't switching to all dual-band devices because it just causes a lot of power waste, redundant wireless signals all over the place, and apparently duplicate messages in your case. That's why I hunted down the single-band Keypadlincs, well, that and they're smaller and cost less. I had to really stuff a keypadlinc into a single-gang box and I don't think the dual-band model would have fit. I do wish they used i2cs instead of i2 protocol, though. I wonder if there is any new command to disable the wireless on individual devices?

[LeeG]

cdragon

 

It is somewhat ironic that the smaller size of the powerline only devices was originally advertised as a positive reason for using Insteon devices.

 

There is no command to turn Off the RF, at least nothing that has been mentioned on any of the forums. I do not have a Developer subscription so it is possible but I think it would have "slipped out" by now if there was one.

 

The newest Dual Band KeypadLincs are I2CS.

 

The concern raised earlier about duplicate messages normally does not happen. The PLM is pretty good at not passing the same message to the application from different paths. Not always though as there are traced and documented cases where a Hops Left=0 message, which apparently took a long time (relatively speaking) to arrive at the PLM, is passed to the application. Considering the total volume of powerline traffic it is not a common occurrence but it does happen. I just switched to the latest PLM with v.9B firmware to see if the duplicate Hops Left=0 messages were eliminated at the latest firmware. They are not.

[Winter Dragoness]

3 or 4 days ago I added a MisterHouse script to do a link scan every 5 minutes. I never saw a corrupt link scan with RF bridging in place. I ran it for a day with no RF bridging and I actually only got one corrupt scan. So I guess corruption isn't a common issue at my house, but the RF bridging does help.

 

I also learned that even 24 byte extended messages in the I2 protocol include one CRC byte to detect corruption. So that CRC byte failed to detect the corrupted messages I did see, which is probably why I2CS adds a second CRC byte. 1-byte CRC is recommended to cover a message of up to 64 bytes, so it's somewhat surprising that corruption still gets through. I haven't researched the math to figure out the actual probability of missing a corruption on a 24 byte message, however.