Running Cable for Water Heater
I’m running 10/2 NM cable from my main panel to a Rheem hybrid water heater. The cable will run through the attic and then out the wall above and behind the water heater. I find mixed opinion about whether or not the exposed cable has to be protected with conduit. What I can’t find out, after a few hours of looking, is what the cable passes through to exit the wall surface and enter the room. The Rheem instructions, 2012 IRC, and dozens of articles are silent on this. Is it just something like an escutcheon if NM cable is okay going from the wall to the water heater, or if conduit is required, is there something the conduit attaches to at the wall where the NM cable inside the wall first enters it? A couple of low-res photos show something looking like a junction box on the wall, but if the NM cable is continuous and just enters conduit as it enters the room, is a junction box needed (with no junction between wires)?
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
The real answer to your question is at the whim of your local inspector.
Some locations the inspectors have decided the NM cable does not require additional protection other locations find that the cable must have protection if the room has finished walls IE drywall.
When you need protection I like to use a whip
https://www.lowes.com/pd/Southwire-6-ft-10-Gauge-3-Conductor-Non-metallic-Liquidtight-A-C-Whip/3616544
Walta
Thank you, Walter. If the NM does not need protection, what would it pass through as it exits the wall? Just some kind of escutcheon?
Around me, connection to the water heater needs to be with MC cable. You can run Romex to an octagonal junction box on the ceiling, cover plate with knockout, cable clamp, MC down to the water heater.
If you want to air seal, get an air tight (sometimes called vapor tight) box. These have a gasketed flange that will seal against the drywall.
Thank you, Akos. Would coming down from the ceiling still be the best choice if coming out through an interior wall is an option? If I already have plenty of NM cable, would running it into conduit be a good alternative to connecting it to MC cable?
Jeff,
I'm not sure what the connections would look like for NM in conduit (If it's even allowed. Typically only wire is allowed in conduit, not cable). Both cover-plates aren't set up for conduit. It's not worth reinventing the wheel for such a short distance. I'd just use a junction box and MC cable.
Thank you, Malcolm. If the cable doesn't need protection, what would you recommend for where it exits the wall? Are the junction boxes you usually see mounted inside the wall like a receptacle box with an outlet for conduit, or are they surface-mounted?
Interestingly, neither NM-B (i.e. Romex) or MC are acceptable when subject to physical damage, yet sometimes installers and inspectors feel better about MC than they do about NM-B in such situations.
You can sleeve NM-B in conduit. It's often done in this exact situation. The key thing that needs to be done is the NM-B needs to be clamped as it enters the conduit, so you need something like this: https://www.homedepot.com/p/Halex-1-2-in-Standard-Fitting-Electrical-Metallic-Tube-EMT-Combination-Coupling-91530/202328186
No fancy fitting is needed at the box end.
Note: The jacket on the NM-B should *not* be stripped inside the conduit. The whole cable assembly should run from the initial clamp to the box.
In case that link dies in the future, I've attached a picture for posterity.
Thank you, Patrick. Your comment about the code concurs with what I saw cited several times from the NEC in my research before posting my original question:: "334.15(B) Protection from Physical Damage. Cable shall be protected from physical damage where necessary by rigid metal conduit, intermediate metal conduit, electrical metallic tubing, Schedule 80 PVC conduit, Type RTRC marked with the suffix -XW, or other approved means." I saw all kinds of different practices described, probably due to local enforcement or customs, from plain NM cable to MC cable to conduit. That little clamp you linked to looks like the perfect solution if my local official wants me to use conduit.
I think the issue might be that you are not allowed to bridge open spaces with NMD. To use NMD, you have to run a stud down to the water heater and run the cable along it, to me that doesn't look "proper".
Thank you, Akos. I did see something about that. I imagine bridging at least a bit of open space is unavoidable, though, if only just the gap between the wall and the water heater.
You generally need to use MC or equivalent to go from the wall or ceiling to the connection box on the water heater. If you are running the cable down the surface of the wall and not inside the wall, then I would run the cable in EMT on the surface of the wall using one of the clamps mentioned earlier, then to a 4” square box at the end (which gives more room to work compared to a single gang handy box). Run MC out the bottom of the 4” square box, and connect it to the water heater’s connection box using 90 degree clamp type MC connector if the connection box is on the top of the water heater. If your connection box is on the side, you’ll probably want a straight MC connector. The goal here is to minimize the chances of the MC making any sharp bends that could potentially split it open.
I would not run NM cable inside of greenfield to be “MC like”. Use proper MC, or THHN (T90 for our Canadian friends) in greenfield instead. Run a separate ground wire in there too, don’t trust the conduit itself to be the ground — it’s not reliable.
If you run the NM cable INSIDE the wall, I would do everything the same except that I’d use a mud ring on the four square box to bring a single gang opening to the surface of the finished wall. You can get a metal cover with a single knockout in the middle, then use that to connect your MC whip. This makes a clean installation in a finished space.
Bill
Greenfield (or, in NEC parlance, "Flexible Metal Conduit") has the same challenges as NM-B and MC if you have a particularly difficult inspector. It can't be used where subject to physical damage.
It all depends on the inspector (which is why I suggest asking them in advance if concerned). If an inspector won't approve NM-B run to a water heater because it's subject to physical damage, theoretically they shouldn't be allowing MC or FMC either.
Unless you are going to use MC elsewhere in your projects it is best to avoid buying and learning how to properly install MC. The premade whip costs less than the MC cutter.
Thank you, Bill, Patrick, and Walter. It's very helpful to have the details of how to implement whichever option the local inspector deems appropriate. I try to limit my questions to him to simply which rule applies when I can't determine it on my own, as in this case, where there seems to be so much variation. Even the NEC rule depends on how one judges the potential for physical damage. How to follow the rule, which is usually a much more complicated question, would be too much to ask him to explain. I expect this conversation will prove a most valuable resource for others like me who are getting ready to run cable to a water heater.
I find it curious that a dryer and range can be connected to a receptacle via a plastic-coated cord that's typically at a height that seems more susceptible to damage than an NM cable (admittedly, with thinner plastic) running from a wall to the top of a water heater, especially if behind the water heater. I know when I move my range, I have to be pretty careful that its sharp sheet metal isn't hitting its cord.
A water heater application isn’t going to be considered to be “subject to physical damage” in this case. We use greenfield in very large sizes (4” sometimes) to connect to transformers in industrial facilities and it’s not even a problem there. Places where damage could be expected would be outdoors where things might hit it, or walls around truck wells where you could reasonably expect physical contact. A water heater installation in a residential setting can’t be reasonably expected to be exposed to severe physical damage. If an inspector were to try to claim that, I would go up the chain and talk to someone else. If I were consulting on a larger development, I’d actually take an issue like that up to the state level to get the local inspector overruled, and in this case I’m actually sure that the local inspector WOULD be overruled by the state level inspector.
If you really read the code, the only conduit considered suitable in cases where the system will be exposed to severe physical damage is rigid conduit, RMC — the stuff that has threaded connectors and looks like galvanized water pipe. This stuff is almost never required on residential projects.
I work as a consultant on complex/unusual electrical systems. I can tell you that residential projects really don’t get into very many unusual code situations. If you’re really concerned, a quick call to your local building department will get you an answer, and they’re going to be fine with MC (or equivalent) unless you have a very unusual installation planned. This really isn’t something you need to lose sleep over :-)
BTW, you don’t need any special tools to work with MC cable. The easy way to work with it is to cut it with your lineman’s pliers, then bend it sharply where you want to remove the spiral armor. When the armor splits, cut it close with a diagonal cutter. Trim the armor tight to the wire with your diagonal cutter, taking care to remove any sharp points on the armor without damaging the wire. Slip on an anti-short bushing before putting on the connector and you’re good to go. If you’re careful with a speed lock connector, you can usually get by without the antishort bushing (speedlocks have a sort of integrated bushing) but it takes a little practice. I can’t even count how many of those I’ve installed this way. Note that it helps to wait to remove the plastic wrap over the conductors until after you install the connector since the plastic wrap helps to keep the wires tightly bundled which makes the connector easier to install.
Bill
Thank you, Bill. Details such as you've provided on how to cut MC are hard to find.
Bill --
Just curious, would it be legal to put a 30A plug and cord on a water heater and plug it into a 30A outlet? Like the way a dryer or stove is normally hooked up?
This is a big no.
It is perfectly fine to hard wire a plug in device, but not the other way around. This is related to Bill's previous comments on how devices get certified.
For non-permanent installations, this is usually fine. If you put the water heater on casters you might be able to get away with it.
I pretty much agree with Akos here. If you make the water heater "portable", then a cord is OK. You're not supposed to leave such devices connected for periods of longer than 30 days if I remember correctly (there is a time limit in the code, I just forget exactly what it is since I very rarely deal with this particular issue). I would not put a cord on a device that is supposed to be permanently installed.
Remember that dryers and ranges are plug-in appliances. They can be moved, they aren't permanently installed (I know that's a bit of a stretch on a range, but that's how those are classified too). Electric cooktops are intended to be permanently installed, and they have hardwired electrical connections -- not cord sets.
Bill
Bill
> I find it curious that a dryer and range can be connected to a receptacle via a plastic-coated cord that's typically at a height that seems more susceptible to damage than an NM cable (admittedly, with thinner plastic) running from a wall to the top of a water heater, especially if behind the water heater.
Curious, indeed. The simple reason for the seeming cognitive dissonance is that the NEC doesn't govern appliances themselves. That's more the domain of the NRTLs (nationally recognized test laboratories) such as UL, ETL, TUV SUD, etc. (https://www.osha.gov/nationally-recognized-testing-laboratory-program/current-list-of-nrtls).
Makes sense, Patrick; thank you.
The national electrical manufacturers association (NEMA) also is involved with “inside” of appliances. There are some areas of overlap though. For example, the reason you’re not permitted to install 15 or 20 amp receptacles on a circuit with a breaker or fuse larger than 20 amps (a rule in the electric code) is because that would create a situation where you could potentially have something with an 18 gauge cord connected. UL only permits 18 gauge for circuits of 20A or below, regardless of the current requirements of an appliance, because of the risk of 18 gauge wire failing in a spectacular way (fire!) prior to the protection device opening the circuit.
The general rule is the NEC covers the installation of things, NEMA covers the construction of things, and UL, etc, set testing standards for all the materials and devices (which makes them “listed for the purpose”, in code terminology for the NEC). All three essentially work together to try to make the system as a whole “safe”.
If you get really fancy though, we engineering folk can do things that are outside of all of the above, with the “under engineering supervision” exception, although usually we just design within the confines of the regular ol’ code rules just like everyone else — it’s simpler, and it usually ensures a safe design. We only get into the fancy engineering stuff when we have to, which usually only comes up on large industrial projects, and not even always then.
Bill
In our area, code requires that a disconnect box be installed near the hot water heater. Others have addressed the wiring, but I think you will also need a DC box at the source. When I installed my hybrid I had to upgrade my electrical to include DC box as originally we were directly wired to breaker panel to hot water heater.
Thank you, Eric. As I understand it, a disconnect is not required if the breaker is in the line of sight.
My inspector was fine with me running NMD cable inside waterproof, flexible PVC conduit from the ceiling to the water heater. Same material as in the "whip" pictured above, but hollow. It accepts threaded connectors on the end, so it connects seamlessly to the entry on the water heater. Might be worth asking your inspector about that configuration.
Good to know, Trevor; thank you. A continuous cable from breaker to water heater would seem to have significant merit.
Jeff,
I'm not disagreeing, but wonder what you see as the advantages of a continuous cable? I've always found the circumstances in a h0use build where that occurs - baseboard heaters come to mind - disrupt both the air-sealing and insulation, and are difficult to finish at the wall surface.
It's just fewer potential points of failure, Malcolm. Even though wires are joined by wire nuts, etc. all over a house, and the likelihood of something coming loose is tiny, it's zero within a continuous cable. How does creating a junction between two cables improve the ability to air seal and insulate?
Jeff,
If you run the continuous cable through a junction box I guess there is no difference in air-sealing and insulation (which circles back to your original question). However during construction the cable is left hanging into the finished space, susceptible to damage and impedes the drywalling and painting. It just seems neater all round to add an easily replaceable cable after the fact.
Thank you, Malcolm. I see your point, but if I go with a continuous cable, I don't expect it to get in the way significantly before I poke it through a hole in the drywall, and not much else would be going on in that area other than painting. Replaceability is good, but if the exposed cable were ever damaged, I could add a junction box then, if needed.
You're not really gaining anything with a "continous cable", and you are required to bond the box to ground -- assuming a metal box -- so you'll have to splice into the cable there anyway. There are somewhere around 453 gazillion wire nuts in service out there, and very, very low failure rate. I tell my crews to "tighten'em until your fingers just start to hurt". If you put them on too loose, then you can have problems. A properly installed wire nut installed on non-twisted solid wire will twist the wires tightly together during installation, and the square wire spring inside bites into the wires to keep them tight together and to keep the plastic cap of the wirenut in place.
A junction box will also greatly improve the serviceability of the installation and I highly recommend using one. If you really want to avoid a splice there, put in a disconnect switch instead.
Bill
Bill,
Saves the fingers: https://toolmonger.com/2009/10/28/ideal-ratch-a-nut/
Malcolm, one of these days I'm going to have to visit Vancouver to see a project or two of yours to admire your artful designs, and buy you dinner for telling me about that tool. That thing is AWESOME! I did find a cheaper version that can go into a cordless driver too ("Ideal 30-902 Spin-Twist Wire Connector Socket"), which I just ordered myself.
I will have to tell my crews about these! I've never seen them before. When we wire up rows of racks in datacenters, my crews install hundreds of wirenuts since we typically have 2 or 4 twistlock receptacles about every 2 feet (the approximate width of a server rack). Fingers get pretty achy after installing that many wirenuts!
Bill
Bill,
A few years ago I was complaining about my fingers hurting when working with my electrician and he looked at me like I was a bit simple and showed me his screwdriver. I don't know if they are commonly used or not, but they are great.
That's now my go-to screwdriver.
The only (minor) flaw is that the bits are stored in the handle, and when you use the wire nut driver it tightens the cap to the bits quite tight. It can be quite a challenge to get the cap off.
Also, a screwdriver for electricians should have the two most commonly used tips -- #2 Philips and small flat head -- opposite each other on the same bit. But those are minor quibbles. It has a nice ratchet action, it's a great tool.
Thank you, Bill. I don't see why I would need a box at all with a continuous cable, so there would be no need for grounding or splicing, and given that I already have enough cable, using it--assuming I'm allowed to--avoids waste.
Jeff,
I'm sure what you are doing is fine, and I shouldn't belabour the point.
If possible leave some slack in the wall that can be pulled through if the cable gets damaged close to the surface so that you have some extra to make the splice in a box. Leaving a Courtesy Loop between the last staple and the box is good practice with all wires, as long as it won't disrupt the insulation behind (see figure #13 in the link). https://mytrainingbc.ca/youthexploreskills/activity/Electrical/pdf/WiringWallSection.pdf
Thank you, Malcolm. Leaving an extra bit of loop makes great sense, and I'm glad to add the reference you cited to my collection.