system design help sought

Posted on
Fri Oct 01, 2010 9:33 am
baadpuppy offline
User avatar
Posts: 16
Joined: Oct 01, 2010
Location: Virginia

system design help sought

Hi. My name is Jim, and I've just started dabbling with Indigo and Insteon technology.

This is a long post, and I apologize for that. I'm trying to be complete with the information needed.

I suppose I should give you some background. I'm in the process of converting a bus into an RV for full time living and traveling. While I'm doing this conversion, I'm living on my parent's property, in the garage apartment.

My parents live in a 2 story house with a basement. There's an underground power line run from their house to the quarter kitchen (this is old farm land, and the quarter kitchen was the old farmstead summer kitchen). From there, there is an extension cord over to the chicken coop for keeping things from freezing in the winter. This is all on the "main" power meter.

The garage/apartment is a 2 car garage kit, concrete floor, with the right half of the garage having been built as an apartment instead of being a garage bay. There's a bathroom on the center line, with doors to both sides. I'm living in that apartment, and it's quite cozy.

To the right of the garage, there's an underground power line (2 legs) over to the pool shed. The pool pump, salt water chlorinator thingy, a small fridge, and some lights are run from that. There's another underground power line run from that building to the lean to that is attached to the side of the garden shed. More power outlets, and lights, and the charger for the mower. Also, some bug zappers.

To the left of the garage, there's an underground power line to the "old garage", which is a 12'x24' shed with a garage door entrance. So a garage door, some outlets, and some lights. There's another underground power line run over to the last shed, another 12'x24', which just has a couple of lights in it. We haven't finished wiring it, so no outlets yet.

If you picture the yard as a square, the house would be the top line of the square, and the quarter kitchen and chicken coop would be the right line of the square. The storage shed, old garage, new garage/apartment would be the bottom of the square, and the pool, pool shed, lean-to, and garden shed would be the left side of the square. If you then disconnect the top/right from the bottom/left, you get an idea of how the electrical is set up, and approximately how close they get.

For the data network, the heart lives in the basement. There is a crappy DSL line coming into a firewall. That then feeds 2 internal networks, those being the internal LAN, and the guest wireless with a captive portal. The internal LAN also has wireless, but it isn't wide open like the guest wireless. From the basement switch, there's gigabit fiber to the 2nd floor (dad's computer room), where there is dad's iMac. Mom's 13" MacBook is in the dining room on the 1st floor, and my niece's iMac is in another room on the 1st floor.

There is some buried conduit between the basement and the garage. Inside this conduit, there is a CAT5 wire which only carries the phone line, and a gigabit fiber link to the garage network switch. Past experience has shown that having any electrical connection (other than the phone) between the main house and the garage results in fried equipment when there are bad storms. So while fiber is more expensive, it is cheaper than yearly equipment replacements.

In the apartment, there is my iMac and my MacBook Pro. There are other network devices, but they're pretty much unimportant for this I think. (access points, time capsules, dish network dvrs, a linux server in the basement).

I also have a mac mini I'm planning to repurpose into my home automation server with Indigo.

What I have right now is Indigo running on my iMac in the apartment. I have the 2413U PowerLinc Modem attached. I have one ToggleLinc Relay to control the light in the bathroom (fluorescent fixture). I have on order 3 ToggleLinc Dimmers from machomestore.com. They should arrive next week. I plan to use them for the two main fixtures in the apartment, one of which is a 2 way switch setup.

I also have two old X10 appliance modules for the lamps beside the bed. Eventually, I want to replace those with LampLinc modules. I'm also using 3 of the old X10 motion sensors to detect motion inside my bus, in the garage, and in my bathroom. Additionally, there's an X10 mini controller beside the bed I'm using to trigger things within Indigo.

I'm currently running the 30 day trial of Indigo, but I'll be purchasing the Pro license Oct 13 (payday).

In the main house, and currently untouchable by Indigo in the apartment, is an appliance module for the living room lights. There's one of the old clock/X10 controller units (don't remember what it is called) that is used for manual control of those lamps. My plan is to eventually replace with a lamplinc.

As I said, I'd like to put Indigo on the Mac Mini. Ideally, that would go in the basement of the house (the biggest UPS is there), and would control all the Insteon devices in both sets of buildings. The X10 stuff can eventually go away. What I'm seeking is the best way to get the signals from a PLM on one leg in the basement of the house to interface with devices in both sets of buildings. I realize that X10 between buildings is a lost cause, and I'm not requiring that in the solution.

Short term, until I get everything in the apartment on Insteon and off X10, the Indigo server would stay in the apartment. Long term though, I want to integrate it all. I'd rather not have to run 2 separate servers if at all possible. Is this possible?

I'm sorry this post has been so long. I appreciate you taking the time to read it and any advice you can give me towards achieving my goal.

Thank you,
jim

Posted on
Fri Oct 01, 2010 10:14 am
seanadams offline
Posts: 489
Joined: Mar 19, 2008
Location: Saratoga, CA

Re: system design help sought

Agree, you need to replace your X10 devices with Insteon.

I would say there is a good chance this will all just work, as long as your distances to each building are under about 200 feet. That's not a hard rule - I have gone further, but it seems to be about the range where you're pressing your luck.

The best topology for an Insteon network with lots of subpanels is a star, where the PLM is close to the center of the star. I.e. get it close to your main panel. Hopefully these outbuildings have a direct line to the main panel, as opposed to daisy chained?

Install hardwired signalinc phase bridges on your main panel AND any distant subpanels. If you have a problem with a remote device, try moving it to the other phase (or reversing phases to its subpanel).

Use the Insteon health tool.

If you are having problems with devices cooking in bad weather, you should check a couple things. First the phone lines: In your phone box, make sure the ground wire is connected to a proper earth ground rod AND the electrical panel's ground. If your phone service box (the part owned by the phone company) is the old style with screw-in carbon elements, see if you can get them to replace the box or at least replace the protection elements. Also, install an Elk-952 to protect your voice lines (I'm not sure if this unit will pass DSL signals, give it a try).

For electrical surge protection, install TVSS devices in every subpanel. They are available for most panels and they just fit in the space of a 2-pole breaker. They must go in the top slot, and you should cut its neutral wire as short as possible to reach the bus bar. Also check that your neutral is bonded to ground only in the main panel. Subpanels for separate buildings should NOT be bonded and must have separate ground and neutral wires feeding them. This has changed in newer revisions of the electric code - old installations with a remote ground rod instead of dedicated ground wire are allowed but you should make sure that it was at least done correctly in either case. I.e. if you have both ground and neutral conductors, make sure it's done the new way (not bonded), and if you only have neutral, make sure it IS bonded and you have a good ground rod. You are allowed to install additional ground rods in either case and this is highly advisable if you are having surge problems.

Posted on
Fri Oct 01, 2010 10:52 am
baadpuppy offline
User avatar
Posts: 16
Joined: Oct 01, 2010
Location: Virginia

Re: system design help sought

Thank you Sean.

I didn't know about the SignaLinc modules. I'll definitely incorporate that into the plan.

Electrically, the buildings are daisy chained. I'll wait to worry about them individually if problems crop up or the health script indicates problems.

The house main panel is on the back porch, and the apartment main panel is near the front door, leaving approximately 90-100 feet between the two.

So if I put the PLM in the house, as close to the apartment as possible, then put an access point in the apartment, that should effectively bridge insteon between the two main panels? And putting SignalLinc in each main panel will bridge the legs, keeping things as simple as possible? The PLM would transmit a signal on both the RF and the powerline band. The SignaLinc in the main panel would bridge it over to the other leg, the Access Point in the apartment would bridge the RF over to the powerline band, and the SignaLinc in the apartment main panel would bridge those two legs together. And of course, the reverse would work the same way. Do I understand this correctly?

I'm doing my best to buy only dual-band stuff if there's a choice. I know it costs a bit more, but I can't help but think it'll improve the performance of everything.

As for the lightning damage, long ethernet cables act like a long wire antenna. A sudden discharge nearby (like lightning hitting somewhere within a mile or so) induces a significant current and voltage in the long wire antenna, frying switch ports, switches, motherboards, etc on its way back to ground. Since removing all the long ethernet runs, we've had no lightning induced damage. Of course, nothing would help against a direct strike, but we've been fortunate in that regard.

Everything on the property is wired to code. Dad insists on that. :mrgreen:

Thanks again,
jim

Posted on
Fri Oct 01, 2010 11:11 am
seanadams offline
Posts: 489
Joined: Mar 19, 2008
Location: Saratoga, CA

Re: system design help sought

baadpuppy wrote:
Thank you Sean.

I didn't know about the SignaLinc modules. I'll definitely incorporate that into the plan.

Electrically, the buildings are daisy chained. I'll wait to worry about them individually if problems crop up or the health script indicates problems.


Make sure each building has at least one Insteon powerline device in it. They act as repeaters, so it will help the signals propagate farther down the chain.

The house main panel is on the back porch, and the apartment main panel is near the front door, leaving approximately 90-100 feet between the two.

So if I put the PLM in the house, as close to the apartment as possible, then put an access point in the apartment, that should effectively bridge insteon between the two main panels? And putting SignalLinc in each main panel will bridge the legs, keeping things as simple as possible? The PLM would transmit a signal on both the RF and the powerline band. The SignaLinc in the main panel would bridge it over to the other leg, the Access Point in the apartment would bridge the RF over to the powerline band, and the SignaLinc in the apartment main panel would bridge those two legs together. And of course, the reverse would work the same way. Do I understand this correctly?

I'm doing my best to buy only dual-band stuff if there's a choice. I know it costs a bit more, but I can't help but think it'll improve the performance of everything.


Search for my previous posts on signal quality and dual band issues. Dual band is better than not, but for best results you should optimize the powerline first, and consider that your primary medium - especially between buildings. We haven't talked about filtering noisy devices but that is another thing you'll need to look for when you start playing with the health tool.

As for the lightning damage, long ethernet cables act like a long wire antenna. A sudden discharge nearby (like lightning hitting somewhere within a mile or so) induces a significant current and voltage in the long wire antenna, frying switch ports, switches, motherboards, etc on its way back to ground. Since removing all the long ethernet runs, we've had no lightning induced damage. Of course, nothing would help against a direct strike, but we've been fortunate in that regard.


That is true, and fiber certainly solves the problem for those particular devices. However ethernet is already transformer-isolated and provides a few KV of common-mode rejection by itself, and if it was in a buried conduit it's not going to be acting as much of an antenna. I would say if you were seeing these issues before installing the fiber, the ethernet run between buildings was not necessarily the root of the problem. You may still be susceptible to high potentials between buildings and this could be a threat to your new Insteon gear. That is why I suggested the TVSS and ground rod measures.

Everything on the property is wired to code. Dad insists on that. :mrgreen:


OK, but it wouldn't hurt to confirm _which_ version of the code you're using on this bonding issue. Ground/neutral bonding is one of the more abstruse areas of the electrical code and there are many professionals that don't fully understand it. There are many things you can do to improve performance above and beyond the minimal code requirements.

Posted on
Fri Oct 01, 2010 11:40 am
baadpuppy offline
User avatar
Posts: 16
Joined: Oct 01, 2010
Location: Virginia

Re: system design help sought

seanadams wrote:
Make sure each building has at least one Insteon powerline device in it. They act as repeaters, so it will help the signals propagate farther down the chain.


I'll make sure to do that.

Search for my previous posts on signal quality and dual band issues. Dual band is better than not, but for best results you should optimize the powerline first, and consider that your primary medium - especially between buildings. We haven't talked about filtering noisy devices but that is another thing you'll need to look for when you start playing with the health tool.


I'll look for those posts and start reading them ASAP. I've already learned a lot in the past week reading the posts here. And I'm already running into some noise issues, probably because we have almost all CFL lighting. I'm switching the apartment back to incandescent for good dimming results. I don't look forward to the discussion with dad for the house though.

That is true, and fiber certainly solves the problem for those particular devices. However ethernet is already transformer-isolated and provides a few KV of common-mode rejection by itself, and if it was in a buried conduit it's not going to be acting as much of an antenna. I would say if you were seeing these issues before installing the fiber, the ethernet run between buildings was not necessarily the root of the problem. You may still be susceptible to high potentials between buildings and this could be a threat to your new Insteon gear. That is why I suggested the TVSS and ground rod measures.


The conduit under the ground isn't shielded, it's just black pvc pipe (irrigation pipe). Over the years, the vehicles and weather has worn the parking area down so that there's probably only 6" of dirt over the conduit in that area. Also, when we originally relocated things from the upstairs room to the basement, we did a quick and dirty 100' ethernet cable out the window, down the wall, and in the basement door until we could pull wire for that. It was a costly quick and dirty. Now that I think about it more, we also had an ethernet running out of the garage and over to an old RV for a while during that time. This was back in the fall of 2004, so my memory is a bit sketchy. Most of the equipment damage was on the two exterior runs. But when we fixed it, we switched all those runs to fiber. None of the original network gear is still in use.

I have no doubt that there are voltage differentials between the two main panels. They're 90ish foot apart, and fed from different power meters. The power from the power company goes from their transformer on the pole by the road through the air (2 separate sets of lines I think, but I'll have to check on that) to a pole in the back yard that they installed. From there, one set of lines runs through the air to the house, and down the side of the house to the meter. If I remember correctly, that is then grounded to a heavy ground rod less than 6 feet from that meter. The other power lines run down the pole and underground over to the apartment, then up to its meter. The wires then feed straight through the wall into the back of the apartment main panel. If I remember correctly, the heavy ground rod for that meter and panel is directly beneath that meter.

However, it never hurts to check on things. I'll suggest the TVSS to dad. Ultimately, it is his property, and I can only do so much to change it.

OK, but it wouldn't hurt to confirm _which_ version of the code you're using on this bonding issue. Ground/neutral bonding is one of the more abstruse areas of the electrical code and there are many professionals that don't fully understand it. There are many things you can do to improve performance above and beyond the minimal code requirements.


Actually, I don't even have to consult my memory on this one. I know for a fact that the only bond between neutral and ground in any panel or box on the property is done inside each of the main panels. When dad brought the electrical in the house up to snuff (GFCI outlets, proper grounding on all circuits, etc) he decoupled neutral and ground everywhere but the main panel in the house, and I helped. When he did the wiring for the garage/apartment, he just wired it that way from the beginning. I assisted in that wiring far more than I wanted to... :?

Next time I'm in each panel and box etc, I'll also make sure all the connections are tight. I know over time, sometimes they work loose and introduce significant noise in the lines.

Posted on
Fri Oct 01, 2010 12:26 pm
seanadams offline
Posts: 489
Joined: Mar 19, 2008
Location: Saratoga, CA

Re: system design help sought

baadpuppy wrote:
The conduit under the ground isn't shielded, it's just black pvc pipe (irrigation pipe). Over the years, the vehicles and weather has worn the parking area down so that there's probably only 6" of dirt over the conduit in that area.


So it's NOT all to code then. ;) I haven't seen black PVC pipe - I think you mean ABS. That is not listed for electrical use, and the minimum cover requirement for this application is 18 inches (300.5). I don't think this has anything to do with your high-pot issues but raises concerns as to what other problems you have.

I have no doubt that there are voltage differentials between the two main panels. They're 90ish foot apart, and fed from different power meters. The power from the power company goes from their transformer on the pole by the road through the air (2 separate sets of lines I think, but I'll have to check on that) to a pole in the back yard that they installed. From there, one set of lines runs through the air to the house, and down the side of the house to the meter. If I remember correctly, that is then grounded to a heavy ground rod less than 6 feet from that meter. The other power lines run down the pole and underground over to the apartment, then up to its meter. The wires then feed straight through the wall into the back of the apartment main panel. If I remember correctly, the heavy ground rod for that meter and panel is directly beneath that meter.


Whoah! Now we're getting to the meat of it. In the absence of a low-impedance ground path (rods aren't) between buildings, this arrangement certainly allows for high potentials between buildings, and any low voltage wiring between them is likely to become the discharge path.

If you're going to have any wire at all between buildings, you need to make sure that is addressed. If there is an existing metallic cold water pipe between buildings that might be sufficient per 250.52(a)(1).

Either that, or ensure you have NO alternate ground paths whatsoever between buildings - that means no gas lines, no phone, no ethernet, no cable/satellite coax. Fiber with nonmetallic sheath is OK.

Just to be clear, what I am suggesting is that your equipment was being damaged not necessarily by energy coupled onto the interconnecting wires, but because they acted as your lowest impedance ground path between structures.

Actually, I don't even have to consult my memory on this one. I know for a fact that the only bond between neutral and ground in any panel or box on the property is done inside each of the main panels. When dad brought the electrical in the house up to snuff (GFCI outlets, proper grounding on all circuits, etc) he decoupled neutral and ground everywhere but the main panel in the house, and I helped. When he did the wiring for the garage/apartment, he just wired it that way from the beginning. I assisted in that wiring far more than I wanted to... :?


OK, so you DO have 4-wire feeders? And those structures have local ground rods, right?

Posted on
Fri Oct 01, 2010 1:29 pm
baadpuppy offline
User avatar
Posts: 16
Joined: Oct 01, 2010
Location: Virginia

Re: system design help sought

seanadams wrote:
baadpuppy wrote:
The conduit under the ground isn't shielded, it's just black pvc pipe (irrigation pipe). Over the years, the vehicles and weather has worn the parking area down so that there's probably only 6" of dirt over the conduit in that area.


So it's NOT all to code then. ;) I haven't seen black PVC pipe - I think you mean ABS. That is not listed for electrical use, and the minimum cover requirement for this application is 18 inches (300.5). I don't think this has anything to do with your high-pot issues but raises concerns as to what other problems you have.


You're right, I meant ABS.

When the garage was built 7 or 8 years ago, and the water and sewer lines were installed, I wasn't here. If I remember what dad said correctly, he put the ABS pipe in at the last minute so he could later pull direct burial phone cord through it without having to dig new trenches and miss the newly buried pipes. He probably should have planned better and put in some 2" or 3" direct burial electrical conduit. That's one of the downfalls of a last minute change. The installation crew didn't give him much time to get his "conduit" and get it in the trenches before filling them in. Later, we decided to pull cat5 instead. After that, we pulled the fiber and new cat5.

I have no doubt that there are voltage differentials between the two main panels. They're 90ish foot apart, and fed from different power meters. The power from the power company goes from their transformer on the pole by the road through the air (2 separate sets of lines I think, but I'll have to check on that) to a pole in the back yard that they installed. From there, one set of lines runs through the air to the house, and down the side of the house to the meter. If I remember correctly, that is then grounded to a heavy ground rod less than 6 feet from that meter. The other power lines run down the pole and underground over to the apartment, then up to its meter. The wires then feed straight through the wall into the back of the apartment main panel. If I remember correctly, the heavy ground rod for that meter and panel is directly beneath that meter.


Whoah! Now we're getting to the meat of it. In the absence of a low-impedance ground path (rods aren't) between buildings, this arrangement certainly allows for high potentials between buildings, and any low voltage wiring between them is likely to become the discharge path.

If you're going to have any wire at all between buildings, you need to make sure that is addressed. If there is an existing metallic cold water pipe between buildings that might be sufficient per 250.52(a)(1).

Either that, or ensure you have NO alternate ground paths whatsoever between buildings - that means no gas lines, no phone, no ethernet, no cable/satellite coax. Fiber with nonmetallic sheath is OK.


At this point, the cat5 between the house and garage carries the phone signal, but it isn't actually plugged into anything in the apartment. No equipment is attached to it. However, it would probably be good to disconnect that link entirely and remove it from the ABS pipe. We have some DECT phone system in use, and that works fine for voice.

The water and sewer pipes are plastic (schedule 40 pvc I believe -- white plastic).

Remember to treat the collection of buildings attached to the garage power system as a separate service from the house and the one building attached to it.

Just to be clear, what I am suggesting is that your equipment was being damaged not necessarily by energy coupled onto the interconnecting wires, but because they acted as your lowest impedance ground path between structures.

Actually, I don't even have to consult my memory on this one. I know for a fact that the only bond between neutral and ground in any panel or box on the property is done inside each of the main panels. When dad brought the electrical in the house up to snuff (GFCI outlets, proper grounding on all circuits, etc) he decoupled neutral and ground everywhere but the main panel in the house, and I helped. When he did the wiring for the garage/apartment, he just wired it that way from the beginning. I assisted in that wiring far more than I wanted to... :?


OK, so you DO have 4-wire feeders? And those structures have local ground rods, right?


From garage to pool shed is 2 10-2 with ground direct burial lines. That goes into a breaker box. Grounds are tied to Grounds. Neutrals are tied together. The two hot legs are fed down either side of the breaker box. I believe the box holds 4 or 6 breakers.

That box then feeds 2 10-2 with ground direct burial lines to the garden shed/lean-to. I forget if there's a distribution box there or not. I'll have to check. No low voltage lines associated with those buildings.

From garage to old garage is 2 10-2 with ground direct burial lines. I'll have to check in the building tonight for the specifics of its hookup internally. That then feeds on to the last shed on that side with 2 10-2 with ground direct burial lines. No low voltage wiring for that set of buildings either.

We had a lot of 10-2 with ground direct burial line (brand new on the spool, left over from a previous project at another location) ready to be used. It was cheaper than buying new wire.

All the grounds for the external buildings (except the quarter kitchen) tie back to the main panel in the apartment/garage. Neutral/ground bonding is done in that panel.

The quarter kitchen has 1 12-2 with ground direct burial lines run to it. That was there when we moved in. Replacing that run would be problematic due to crossing the driveway. The ground for that building is in the main panel in the house.

There are GFCIs scatted through all that, but I can't remember exactly where and which circuits. Basically, we installed them on every outlet circuit.

Are you saying we need to add local ground rods for each small shed? How would that account for the fact that ground potential changes over such distances? Wouldn't that put a possible voltage differential on those ground links between the buildings?

Ugh, you're going to make me scare up a copy of the NEC and try to decipher it again aren't you? :lol:

I'm still reading through all your older posts. There's a lot there.

thanks,
jim

Posted on
Fri Oct 01, 2010 2:06 pm
seanadams offline
Posts: 489
Joined: Mar 19, 2008
Location: Saratoga, CA

Re: system design help sought

baadpuppy wrote:
You're right, I meant ABS.[..] Later, we decided to pull cat5 instead. After that, we pulled the fiber and new cat5.


Oh, I thought you were talking about the feeders. For low voltage, that requirement does not apply. I would still recommend grey PVC for future purposes.

Remember to treat the collection of buildings attached to the garage power system as a separate service from the house and the one building attached to it.


Right, that's the source of the concern. If they were ultimately fed from the same main panel they would share the same ground<->neutral bonding point. Since they're not, you have to worry about ground/neutral loops and unintended fault paths. Besides fault conditions, without a good deliberate tie between the systems you can even end up with low voltage data cables carrying a portion of your neutral current!


From garage to pool shed is 2 10-2 with ground direct burial lines. That goes into a breaker box. Grounds are tied to Grounds. Neutrals are tied together. The two hot legs are fed down either side of the breaker box. I believe the box holds 4 or 6 breakers.

That box then feeds 2 10-2 with ground direct burial lines to the garden shed/lean-to. I forget if there's a distribution box there or not. I'll have to check. No low voltage lines associated with those buildings.

From garage to old garage is 2 10-2 with ground direct burial lines. I'll have to check in the building tonight for the specifics of its hookup internally. That then feeds on to the last shed on that side with 2 10-2 with ground direct burial lines. No low voltage wiring for that set of buildings either.

We had a lot of 10-2 with ground direct burial line (brand new on the spool, left over from a previous project at another location) ready to be used. It was cheaper than buying new wire.

All the grounds for the external buildings (except the quarter kitchen) tie back to the main panel in the apartment/garage. Neutral/ground bonding is done in that panel.

The quarter kitchen has 1 12-2 with ground direct burial lines run to it. That was there when we moved in. Replacing that run would be problematic due to crossing the driveway. The ground for that building is in the main panel in the house.


It sounds like your ground fault paths may be OK, although you're not allowed to double up small feeders like that. It's only allowed for 1/0 AWG and larger. The issue is if one of those hots becomes open, you'll melt the other one.

There are GFCIs scatted through all that, but I can't remember exactly where and which circuits. Basically, we installed them on every outlet circuit.


GFCIs are a problem for Insteon because they act as signal attenuators. But as far as I can tell they are only a problem if the signal is trying to go _through_ the GFCI.

Are you saying we need to add local ground rods for each small shed? How would that account for the fact that ground potential changes over such distances? Wouldn't that put a possible voltage differential on those ground links between the buildings?


Grounding electrodes for separate structures are an absolute requirement per 250.32(A), whether you have 3 or 4 wire feeders. Earth ground potential does NOT change from place to place, that's the whole point of it. Ground rods are not a discharge path for high currents, that's what your feeder's ground is for in the 4-wire case, or the neutral in the 3-wire case. Otherwise during a fault condition the ground reference would be "lifted" and this is extremely dangerous since the breaker is unlikely to even trip. A ground rod doesn't prevent that, it's purpose is to keep everything at the same potential during non-fault conditions (eg due to a high impedance "leak" or EMI). They also provide a faster discharge path for rapid high-pot surges - even though the impedance to earth is a few ohms, this is overcome by the high voltage, and the shorter distance provides an immediate discharge instead of sending it all up the feeder (or your data cables). Ground rods will not interfere with data communications, they can only make it better.

Ugh, you're going to make me scare up a copy of the NEC and try to decipher it again aren't you? :lol:


It's good stuff. They could easily fill ten times as many pages just explaining the rationale for everything - that is really helpful to understand.
Last edited by seanadams on Fri Oct 01, 2010 4:21 pm, edited 1 time in total.

Posted on
Fri Oct 01, 2010 2:12 pm
seanadams offline
Posts: 489
Joined: Mar 19, 2008
Location: Saratoga, CA

Re: system design help sought

BTW, driving ground rods is easy and quite fun, with the right tool. If you don't have a roto hammer you can rent one, then use one of these: http://www.amazon.com/Bosch-HS2172-Grou ... 799&sr=8-6

You can even drill through a concrete slab and then run the ground rod down through that hole.

Posted on
Sat Oct 02, 2010 6:44 pm
baadpuppy offline
User avatar
Posts: 16
Joined: Oct 01, 2010
Location: Virginia

Re: system design help sought

seanadams, thanks for all your help and clarification.

Talked with Dad today. We'll be driving some ground rods for each building soon.

Would it be worthwhile to figure a way to tie the two different service panel grounds together? Would that make it safer? Would it be compliant?

I'll get 2 SignaLincs ordered asap (one for house, one for apartment), and access points for bridging the house and apartment.

thanks,
jim

Posted on
Sat Oct 02, 2010 9:02 pm
baadpuppy offline
User avatar
Posts: 16
Joined: Oct 01, 2010
Location: Virginia

Re: system design help sought

Oh, one more question... On the SignaLincs, it indicates I should install new breakers just for the SignaLinc. Is that required, or can it go on one of the 220 breaker pairs already installed?

I've got them ordered. Should ship monday, and be here before the following monday.

Thanks,
jim

Posted on
Sat Oct 02, 2010 11:42 pm
seanadams offline
Posts: 489
Joined: Mar 19, 2008
Location: Saratoga, CA

Re: system design help sought

You probably got the impression so far that I'm a total code nazi but that's not why I read it at all. I'm more interested in understanding why it is what it is, and usually there's a pretty good reason. One case where I'll admit to bending the rules is with signalinc installation. My preference is to remove them from their bulky gang box enclosures - the only functioning part of it is a little 1" square circuit board with an inductor and capacitor on it (for 130KHz resonance) . Let's just say it's a high profit margin item - if I had needed more than three I'd make them myself. Anyway, then I solder new leads and encase the PCB in two layers of heat shrink tube, so I have a small unit that can fit _inside_ the panel. I'll usually dedicate breakers to them only because I have some spare ones, but I wouldn't hesitate to double up on a pair of small breakers since these draw no current.

On tying the two panel grounds together: what you have with those two meters is presumably separate 3 wire feeders on the utility's side... I feel this is the same situation that you encounter if you had 3-wire feeders going to multiple structures on your own side of the meter. You are supposed to avoid low-voltage wiring between the structures in that situation. But I don't think the code prohibits you from tying the grounds together and I can't think of any way this is functionally different from tying them to a shared cold water main like I was suggesting (a cold water pipe is explicitly allowed as a grounding electrode and would tie the structures together anyway). If you do run a wire between them the best choice is probably #4 stranded in pvc conduit. It's normally sold in black but you can (and must) wrap green tape around the ends to identify it as a ground. I am not sure if this will really make a difference for you so it's hard to say if the cost is justified. Again this is pressing the boundaries of my code knowledge - there may be a specific way you're supposed to handle this in light of the separate meters but I'm not aware of anything to that effect.

Posted on
Sun Oct 03, 2010 12:45 am
baadpuppy offline
User avatar
Posts: 16
Joined: Oct 01, 2010
Location: Virginia

Re: system design help sought

Thanks sean,

No, I don't think you're being a code nazi about this. I had some incorrect beliefs about what was and was not allowed by code. You challenged my assumptions, and I looked up what you were talking about, and came to the conclusion that I was wrong.

One important thing I re-learned was that everything in the NEC is there because it is the *minimum* that is considered *safe*. Safety is very important.

I had read before where you had removed the actual circuit out and heat-shrinked and taped it. I think we'll end up doing the same thing. Lots of heat shrink tubing around here.

You're right that both meters have the 3 wire feed (Line 1, Line 2, Neutral) from the power company.

I think we'll remove the cat 5 from the ABS pipe for now, and consider it a bad idea to connect any low voltage wiring between the two services. If we decide we must have a low voltage interconnect between the services, then we can re-address the situation, making sure to keep safety first.

As for the runs of 2 10-2 w/ground wires to the buildings, I think that's fine. See, it is really 2 120V circuits, 30A capable (turns out it is 20A breakers feeding it). Each one contains a hot, a neutral, and a ground. We just happen to tie the neutrals to each other, and the grounds to each other, at the shed end. The two hots happen to be on opposite legs, but could just as easily be on the same leg. There's no 240V equipment in any of the sheds nor any intention to ever install any. The only reason for putting one per leg is for trying to keep load balanced.

The end result is basically that we have a double-sized neutral and ground. A 4 wire feed would have been cheaper in terms of copper use. Instead, we have a 6 wire feed.

In the garage/apartment, the county required GFCI on all circuits (except lights I think) since the floor was concrete. The apartment side has a wooden floor built upon the concrete, as does the bathroom. Everything was signed off on by the inspectors.

I've learned that Virginia tends to lag way behind in adopting new NEC revisions. Near as I can tell, they're still using the 2005 version (NFPA 70-05). USBC 2006 went into effect in 2008. USBC 2009 is in process, and will likely go into effect in 2011. I expect it will reference NFPA 70-08, or the NEC 2008 edition. Our county tends to be rather... lax to put it kindly. Backwards might be more accurate though.

Also, this property is zoned agricultural. As such, many of the stricter requirements are exempted. A building permit isn't strictly required for example. That still wouldn't be an excuse to wire things in a dangerous fashion though.

Anyway, I think my initial design questions have been answered. Also, my potentially dangerous low voltage interconnection between the two services is being removed. Additional safety ground rods will be installed and bonded to the grounds in each of the sheds. And we'll live with the GFCI circuits as they are.

Thanks again,
jim

Posted on
Sun Oct 03, 2010 11:35 am
seanadams offline
Posts: 489
Joined: Mar 19, 2008
Location: Saratoga, CA

Re: system design help sought

Check back and let us know how it goes when you deploy more devices. I have been surprised how well Insteon works even going long distances through multiple panels. With your separate meters, as long as they are fed from the same transformer, there's a fair chance the powerline data will get up one feed and down the other just fine.

Posted on
Fri Oct 08, 2010 9:50 am
baadpuppy offline
User avatar
Posts: 16
Joined: Oct 01, 2010
Location: Virginia

Re: system design help sought

All of the things I've ordered have arrived.

At this point, I have 4 insteon devices installed. There is a ToggleLinc V2 relay for the bathroom vanity, a ToggleLinc V2 dimmer for the bedroom overhead, and 2 ToggleLinc V2 dimmers for the living room overhead installed. That involved some minor rewiring, as the power source for that light comes from the box the light is in, rather than from one of the switches.

I also have a TM751, and 3 X10 EagleEye motion sensors. I wish Indigo would automatically know that these have 2 X10 addresses, the motion detector, and the light detector. I also wish that I could specify I want a reverse condition for the light detector. When it sends an "ON", that means it is dark, yet it looks "lit" in the interface. All the motion detectors are on house code M.

Additionally, I have an X10 mini controller, configured for house code B. I have 2 X10 appliance modules I'm currently using for the two bedside lamps. These are on house code A (A15 and A16 actually).

In the pool shed, there is another mini controller, house code G. In the lean to, there are 2 bug zappers plugged into an appliance module with code G3. I know the mini controller in the pool shed reaches the lean to module. That's the one we use to control it.

In the house, there are a few of those clock/controller X10 devices, all on house code G. Dad used house code G for all the old stuff. There is an appliance module on G1 to control the two living room lamps. That has been in place for years.

I still have 1 ToggleLinc V2 dimmer to install for the office desk overhead lights, and 2 ToggleLinc V2 relays to install for the bathroom central light.

My 2 SignaLincs arrived yesterday, as well as 2 access points. I've installed 1 of the SignaLinc devices in the hot water heater disconnect box. This is the one for the apartment.

Before installing the SignaLinc in the apartment, all my X10 stuff worked internal to the apartment. Signals from transmitting X10 devices got to Indigo, and signals from Indigo to X10 devices got to the devices. When I first set it up, I had to move the PLM from one leg to the other to get it all working smoothly. I am unable to control any X10 device not within the apartment itself (that includes the bug zappers), nor anything in the house (not surprising, it never worked before either).

Now that I'm up to 4 insteon devices, those devices seem to be working as well as they did upon first installation. They're quite fast. Group links and scenes, etc are neat.

Since installing the SignaLinc, I've noticed that X10 devices have gotten laggy. I lose more commands or have more collisions or something. I'm not sure exactly what it is. For example, a motion detector that has been reliable the last week or so now only intermittently gets heard by Indigo. The *only* thing that has changed is the addition of the SignaLinc.

We're holding off on installing a SignaLinc in the house, and holding off on the access points, until we're sure they'll be needed. Dad thinks that both sets of service should be able to see each other since they are attached to the transformer at the same place. On the other hand, the new smart meters the power company put in less than a year ago might be filtering such signals, no?

Every outlet circuit in the garage is on a GFCI. This was required by the building inspectors due to the concrete floor. From what I've read, GFCI acts as an attenuator.

The PLM is plugged into a GFCI outlet, and all other X10 devices in the apartment and pool shed and lean to are plugged into GFCI outlets. I'm guessing that the GFCIs are significantly attenuating the X10 signals, making them weak. I think bridging the legs causes the signal to split at the bridge, losing more energy, taking it from a bit above marginal to right at marginal for the interior modules. And for exterior modules, there's just no hope for it. I suspect that my insteon signals are making it all the way to the house. However, I have only hardwired type modules, so can't test that theory easily.

I think tonight I'll temporarily wire an electrical plug to one of my not-yet-installed ToggleLinc V2 devices, capping the red wire for safety, and use that to walk around with a laptop and test how far the insteon signals actually go. The LED on the switch will indicate whether it is working. I'm not going to sweat the X10 stuff, as we plan to phase that stuff out as fast as we can. Once the apartment is insteon only, and once I get insteon bridging the gap between the two different services, I'll move my server into the house. Then all the X10 stuff will only be affecting the one structure.

jim

Page 1 of 1

Who is online

Users browsing this forum: No registered users and 8 guests

cron