ISY on WiFi bridge?

[jtara92101]

While my Insteon network has gotten more reliable with installation of more devices, I'm considering ways to improve reliability further.

 

Right now my ISY is on a UPS and sitting on top of a rack cabinet in my office. The PLM is plugged into the same duplex outlet as the cabinet, but the computer equipment has a huuuuuuuge filter in front of the UPS.

 

Breaker box is just outside of the kitchen, and there is an outlet directly below. Out of bounds for aesthetics. 

 

There's a duplex behind the refrigerator, not much more wire than the one directly below the breaker box. I am willing to run a zip cord up through the cabinet above the refrigerator. No way I can reasonably fish stuff out of the back of that cabinet anyway, so much of the space will go unused.

 

Closest place with opposite (well, almost opposite - 120 degree...) phase is under the sink, where there is a duplex for the dishwasher and a duplex for the waste disposer. It is a split neutral. I will also be powering a UV sterilizer and auto-flush valve for reverse osmosis filter into one of those circuits.

 

(They don't have a tie bar on the breakers, but wasn't required by code in 1999 when building was renovated - it was only required if both circuits were on the same duplex, and these are not - it is two duplexes side by side in the same box... Current code does require the tie bar. The developer cheaped-out in EVERY POSSIBLE WAY, but were magnanimous about providing extra outlets under the sink! I am still dubious about 20 amp breakers and 20 amp outlets with split neutral and 120 degree rather than 180 degree phase difference.... I think the breakers and outlets should have been derated to 15A but still with #12 neutral. Or does #12 have enough capacity to handle the neutral for two 20A circuits at 120 degrees? Mostly an academic question, but those 20A outlets are an invitation to plug some 20A thingie into it - an attractive nuisance! In real life, I have NEVER SEEN any device with a 20A plug with the T on the neutral...)

 

Distance from the refrigerator outlet is 10 feet as the crow flies.

 

There is no reasonable way for me to get an Ethernet cable from my router to that location above the refrigerator, though.

 

So, I was thinking of using a WiFi bridge. I have an old router that I could set-up in bridge mode and place in the cabinet over the refrigerator. Then I could put the Wifi bridge, ISY, and PLM in the cabinet over the refrigerator, and plus a range extender into the (almost) opposite phase under the sink (the disposal circuit). Or, vice-versa, might actually be better as no holes to drill and have a dead corner in the under-sink cabinet that is hard to use and equipment could go back in there, and there are a couple of convenient boards to mount equipment on rather than just sitting on floor of the cabinet. But over the fridge gets the PLM closest to the breaker box.

 

This gets me injection on one phase very close to the breaker box, and RF bridge to other phase not much further away. (Or other way around).

 

I guess there isn't much functionality of the ISY without power to the home. It can't talk to any Insteon device, wired or wireless, as the PLM wouldn't be powered (even now, as of course it is not on UPS). It could talk to z-wave, and if Internet works, still connected to net. (Assuming Cox has no site-powered equipment needed. My old place, I did have connectivity during a power outage. I went on a walk during that big power outage, and utility boxes could be heard with their little generators humming away...) I can't think of any functionality that would be important to me, but others might.

 

Does anybody else run their ISY on a WiFi bridge? There will be no issue with the WiFi, as my router covers the apartment well. Any issues, gotchas, or other considerations?

[Teken]

WiFi is convenient - It should not be the primary method to interconnect a persons home automation if this is a serious effort. For those casual users sure WiFi is just fine but realize you just added one more thing to trouble shoot when things don't work as expected.

 

It doesn't matter what kind of router you have WiFi is notorious for dropping out . . . 

 

RE: 12 AWG wiring is indeed rated for 20 amperes and has no baring on anything else. Like you, I have rarely if ever seen a 20 amp plug in a consumer grade appliance. In my part of the world the 20 amp circuit is intended to supply sufficient power to those high current devices with out tripping a standard 15 amp circuit.  

 

They are most common in the kitchen to supply power to microwaves, counters, and islands.

[paulbates]

It doesn't matter what kind of router you have WiFi is notorious for dropping out . . . 

 

What you have to watch for is multiple routers from your neighbors close by you. The problem is much worse on 2.4ghz band as the channels overlap significantly. For instance, if your router is on channel 6 on the 2.4ghz band, its impacted potentially by other routers using channel 3 through channel 9. There are only 11 2.4ghz channels to choose from, depending on your router, so there are not really any good options.

 

A 5 ghz band router and AP would be preferable for this purpose, and do your best to find the most unused channel in your building. There is an android app called Wifi Analyzer that is perfect for investigating this around your place, provided your phone supports 5ghz band,

 

Having said all of this, one day someone else sets up a router on your router's channel nearby you, and you could start having problems.

 

Paul

[stusviews]

1. As Teken indicated, 12AWG is rated for and the standard for a 20 amp circuit.

 

2. The electric supply in most homes is split, single-phase. The current in each opposite leg is 180º out-of phase with the other leg. Because it's 60Hz AC, the direction of the current flow changes 60 times each second. Only three-phase supplies have a 120º phase difference.

 

3. It's common to use a 3-wire cable to power two circuits, one circuit on each of the opposite legs and one neutral wire. The neutral wire carries only the difference of the load of both circuits which is never greater than any one of the circuits. In that case od supplying two separate circuits the breakers are not required to be (and shouldn't be) yoked. A yoked pair of circuit breakers is required when the circuits supply  a 220/240VAC device. That requirement precedes 1999 by decades.

 

4. "As the crow flies" is a most inaccurate way to determine wire length. That depends on the structure and design of the wiring. A more distant outlet may have less copper between that outlet and the panel that one that's physically closer to the panel.

 

5. I don't know what a "split neutral" is.

[G W]

 

 

They are most common in the kitchen to supply power to microwaves, counters, and islands.

And garage. And anywhere I live

 

Short and to the point.

Best regards,

Gary Funk

[jtara92101]

My building has 3-phase power. Each apartment is wired to 2 of the 3 phases. My phases are not "opposite". The two phases in each apartment are 120 degrees out of phase, rather than 180. With phases at 180 degrees, the neutral current will never exceed 20 amps. At 120 degrees, it will for part of the cycle.

 

A split neutral is what you said in (3). Maybe I am using a slightly incorrect term. I have seen "shared neutral" and "split circuit". I probably synthesized "split neutral".

 

As I understand it, until a fairly recent code revision, a yoked pair was needed IF the two circuits fed the two outlets of a duplex. (I think the wording is actually "same strap" or some-such). I'd assume this is for safety reasons. So that you don't turn off one breaker (or have it trip) and go to change the outlet and zap yourself on the other circuit. Or you have a faulty device, not really prudent to assume that if there is a faulty upper outlet in a duplex that the bottom one is OK. The latest code requires the bar it if the circuits feed outlets in the SAME BOX.

 

FWIW, all of my circuits and outlets are 20A and all of my wiring (save for big appliances) is #12. The one forward-thinking thing they did in the 1999 renovation. Before that, it was whatever they used in reinforced concrete buildings in 1927. (So, at least not knob-and-tube I would hope! My grandfather's house had that!) Hopefully, ALL of the "new wiring" was indeed "new wiring". (Unlike the "new plumbing" for which the HOA took a $6.5M settlement from the developer when it apparently turned-out was not all new and flooded.)

 

My "as the crow flies" comment was regarding RF. That is, the PLM and range extender to bridge the phases would be plenty close to each other.

 

I am confident that 5gHz WiFi will make it the 30 feet and two walls from the router. I have previously surveyed, and just now checked in that cabinet with CloudCheck WiFi SweetSpots on iPhone and get a minimum of 200mbps consistently. Should be enough for an ISY!

 

I do have WiFi Scanner on my Mac Mini and Macbook and so can survey WiFi. I've used it with a cantenna to see how many signals I could see. In my previous place, I could see at least 200 pointing it around. Probably even more here. In an outage, I can choose from many coffee houses, restaurants, banks, brokerages, and probably even the new federal courthouse they are building to be able to bring in the narcos to the rooftop helipad. (I mean, hey, they all have guest WIFI). That is IF I want to catch a virus! So, yes, it is a pretty crowded RF environment, but even so it is a concrete building and 5gHz is pretty quiet.

 

Point noted that WiFi shouldn't be a primary link.

 

The ISY does nothing very important. It makes it easy to set-up Insteon. It lets me write programs that will light a KPL lamp when lights are on in other parts of the apartment, and press the same button to turn them off. It lets me write home/away programs. (Which do not unlock any locks!) Some day it will set Nest to Off when a window has been open for n minutes. (Once I paint the windows and stick the Insteon sensors on them.)

[larryllix]

I used a Cisco router I had problems with (weak Tx power) and loaded the open source firmware into it. I use it in my shop building in bridge mode for a Webcam and my WC8 board for windspeed and temps to talk to my home router into my ISY. It was fairly reliable using 2.4Ghz but since converting it to 5GHz it has never failed.

 

I do have ISY programs to power cycle it, on failure to receive heartbeats, from the WC8 board, though.

 

 

 

With 120 degree phases your neutral current can never exceed the phase current, in either phase conductor.

 

At 20 amperes in either phase, the return current is the same = 20A, Same as in 180 degree phases.

 

With 20 amperes in each phase, the resulting neutral current is the vectorial sum of the two currents at their respective phase angles 120 degrees apart.

 

- 20A x cosine(120 deg) + 20A x cosine(0 deg) = 10 amperes at an angle exactly split between the phases (60 deg)

 

With 180 degree phases (single split into two phases) it becomes

- 20A x cosine(180 deg) + 20A c cosine(0 deg) = 0 amperes.

 

Different power factors on the two loads will vary results somewhat.

 

Sorry for the technical stuff. I spent most of my life analysing electrical systems.

[stusviews]

- 20A x cosine(120 deg) + 20A x cosine(0 deg) = 10 amperes at an angle exactly split between the phases (60 deg)

 

- 20A x cosine(180 deg) + 20A c cosine(0 deg) = 0 amperes.

 

cos(0º) = 1

cos(120º) = -0.5

cos(180º) = -1

 

-20A x cos(120º) + 20A x cos(0º) = 30A

 

-20A x cos(180º) + 20A x cos(0º) = 40A

[larryllix]

cos(0º) = 1

cos(120º) = -0.5

cos(180º) = -1

 

-20A x cos(120º) + 20A x cos(0º) = 30A

 

-20A x cos(180º) + 20A x cos(0º) = 40A

You have used a double negative for the first current in each calc.

We can't say it's negative and 120/180 degrees out of phase. Two different styles of representing this.

 

Either use

-20A x cos(180-120)

  or

+20A x cos(120)

 

It is more obvious for the 180 deg case. If you draw a vector at 180 degrees and then make it negative it results in the original angle 0. Thus the 40A result with two 20A in phase vectors/ phasors adding arithmetically.

 

We know split phase phase/leg currents don't add up to create neutral currents this way but rather subtract to become 0, in a balanced system.

 

I thought I had forgotten how to do cosines in my head.

[stusviews]

Using your corrected formulas gives:

 

-20A x cos(180-120)  = -10A

and

+20A x cos(120) = -10A

 

I'm not sure what negative amps represents in an AC circuit

[stusviews]

For a split, single-phase electric supply the neutral current is the absolute vale of the difference of the current in each opposite leg:

| I_L1 - I_L2 |

 

For a 3-phase electric supply, the neutral current is the positive square root of the sum of the squares of the current in each phase minus the sum of the products of the currents in each pair of phases:

   ___________________________________________

√ I²_L1 + I²_L2 + I²_L3 - I_L1I_L2 - I_L1I_L3 - I_L2I_L3

[larryllix]

On balanced systems that may work.

 

The neutral is the reversed phasor of the sum of all phase current phasors.

or

The phasor sum of all non-neutral currents, reversed.

or

Kirchoff's law states they all have to add to zero so the neutral is what is missing.

 

 

For lay people the simple arithmetic works better and in most cases is good enough.

 

I admit I have never seen your last formula, before. Interesting!

[stusviews]

The formula I provided is for an unbalanced 3-phase electric supply. For a balanced 3-phase electric supply the formula is more complex, but is not needed. The neutral current in a balanced electric supply, be it split, single-phase or 3-phase is zero (0).

 

In fact, the unbalanced formula evaluates to zero if the current in each phase is equal (i.e., balanced).

 

The simple formula you provided is incorrect. See post #8 and post #10.

[larryllix]

Your formula does not consider phase angles of the load currents and does not work for unbalanced angles.

 

Neutral currents can only be calculated using vector addition, usually by cartesian cordinates (sine and cosine), some use the term imaginary numbers for it. Phase angles need to be considerd for unbalanced systems and straight arithmetic can't calculate it accurately.

 

Most systems can use simplified calculations due to mostly balanced loads (magnitude and phase angle) and aproximations being good enough.

[stusviews]

It's not my formula, it's the standard formula for an unbalanced 3-phase load. As with many formulae, including the complex formula for balanced loads, it can be simplified. A formula that works for an unbalanced load should give correct results for any possibility including a balanced. The formula I posted does that.The formulas you posted lead to incorrect results, as I've demonstrated.

[Teken]

wye dus 1+1 nott eqwells 4? 

[stusviews]

Here's the formula for a balanced 3-phase where θ is the angle in radians between phases, an inverted vector sum based on Kirchhoff's Circuit Law.

 

i = sin(θ) + sin( θ - 2π/3 ) + sin( θ + 2π/3)

  = sin(θ) - 2sin(θ) cos( 2π/3 )

  = sin(θ) - sin(θ)

  = 0

 

Wow, that was a chore to type.

 

Disclaimer: I was a class A Journeyman electrician, then contractor for 15 years and taught mathematics for nearly two and a half decades. Just for fun, I have a totally unrelated doctorate from UCLA.

 

Paper weight lifter. With effort.

[jtara92101]

Oh, oh, this is dredging up bad memories that make my head spin! And now I realize how backwards I got the neutral issue!

 

Thank you Mr. Kirchoff! You are the historical figure most responsible for my decision to go into Computer Science in college after suffering through your equations in high school. So sad you couldn't have made them more interesting, so that I would have retained them! (I took an Electronics curriculum at Cass Technical High School in Detroit.) Sorry, though, for the jokes we made about your name!

 

I was trying to reason it out "intuitively", even though somewhere in the back of my brain I knew better. And somewhere I knew the neutral netted-out, pretty sure we had to wire-up 3-phase circuits in lab and measure current to prove the magic of the neutral. I suppose today if they do teach this stuff in some high school, they are using computer simulations or some cute little 1V AC circuits. We were using 440V. We also used real chemicals in chem class that we could and did do Scary Things with.

 

But now I realize I can make a great "intuitive" analogy with the oddball (but quite efficient) HVAC system my building has. Each unit has a water-source heat pump. The water source is "neutral temperature" (in human terms) water circulated throughout the building. It is maintained in a range by a chiller and boiler, but even without either operating, the units can "work against" each other through the water source, and as typical in a high-rise, as the day progresses opposite sides of the building will have some opposing heat gain/loss. The circulating water is almost a literal analogy of an electrical neutral.

 

(Few who live here understand what they have. Ask anyone, and they will tell you that heating and cooling is "included". Well, it is, and it isn't... It's just that the individual units are so darn efficient - between compressors designed for 208V at 120degrees - which I gather from some reading is pretty ideal - the efficiency of using a water source vs air or ground, and the "help" from the boiler and chiller, you could easily look at your electrical bill and think that HVAC is "included".)

 

Neutral current will never exceed the current of one phase, and only when the current in the opposite phase is zero.

----

 

OK, back to WiFi bridging et al... 

 

Regardless of where I inject the signal into the line with the PLM, suppose it wouldn't hurt to put range extenders on each phase very close to the breaker box. I could put one behind the fridge and one under the sink, or both under the sink. I wonder if they will have trouble if they are right next to each other? (Don't know if they would fit on two duplexes in same box?) I need to run RO UV and auto-flush under the sink, and so I can mount a small power strip in under-sink cabinet away from the double-duplexes, and plug the UV, auto-flush, and one of the REs into the strip. 

 

From previous posts, I gather that my one working Access Point won't repeat I2 or I2CS messages, right?

 

Which gets me thinking - do I2 devices repeat I2CS messages? Is the checksum just additional  opaque payload that an I2 device will still happily repeat?

[stusviews]

Insteon Labs has indicated that Range extenders on each phase of a 3-phase electric supply adequately bridges the signal. Devices can repeat signals at their capability and below only. I1 devices repeat only i1 signals, I2 devices repeat both i2 and i2 signals. And i2CS devices can repeat i1, i2 and i2CS signals.