I seem to get the most imaginative inspectors. I just had my electrical inspection and failed. The inspector had never seen ICF or at least my ICF. I used Rastra and routed out some of the electrical wiring into the form after it had been poured. This is per the manufacturer's install manual and exactly how I learned when I worked on prior projects (and became a certified installer). The NEC says that NM wire (e.g., Romex) can't be "embedded" in a masonry or concrete wall. Okay, it's not. The concrete was cured long before the Romex went in the wall. The NEC also says that a channel can't be cut into a mansonry wall and the Romex embedded in plaster to hold it in the channel. Okay, it's not. The channel is routed into the form, not the masonry. The plaster holds the wire back in the channel (to meet the code required distance from the finished surface), it is not embedded in plaster. Anybody have any experience with inspectors on this issue? I don't know quite where to look for info. I may have to use a "common sense" argument and those don't carry much weight with some inspectors. Mark

I would recommend contacting your form manufacturer for help. If they recommend this method they should have the technical expertise to argue your point for you. The ICF companies I have worked with have these resources available and are always glad to help.

How does the inspector say you need to do it? There is no reason you shouldn't be able to recess the NM in the composite ICF material.

Your box looks crooked.

Are the wires deep enough in the ICF? They look right near the surface. That won't pass code. If I recall correctly, there needs to be 1.5" of clearance from the face of wall to the wire. Otherwise, you risk drywall installers driving a nail or screw into the wire. Unless you cover the wires with steel protection plates.

Actually, the requirement is 1 1/4" from the face of the framing member, in this case the face of the block. With 1/2" drywall that puts it 1 3/4" from the face of the wall. Otherwise it must be protected by a 1/16" thick steel protective plate.

The Romex is back far enough, the picture is deceptive (although maybe the box is crooked). The ICF material is 2" thick. The notch is cut in the foam clear back to the concrete using an electric chainsaw as described in the manufacturer's install manual. It ruins the chain, of course, but the wire is actually a jam fit into the cut and is back 2". It's tapped into place using a shim, just like caulking a wood boat with oakum, if you're familiar with that process. Assuming Romex is 1/2" wide, that puts it back 1.5" prior to plaster (which is applied directly to this ICF). Plaster is 1/2 to 1" thick, so I've got more than enough distance from the finished surface, somewhere between 2-2.5". The inspector says conduit will meet code. The wire runs were cut freehand, so they are not perfectly straight and the bends don't match conduit bends. It would take a couple days of hogging out the forms, several hundred dollars worth of conduit, and several days of placement to accomplish whatever it is that is to be accomplished. I can't find anything in the Code Commentaries as to the problem caused by embedding Romex in concrete or plaster (not that that's what I'm doing, but I'm curious). I called the rep. He'd never heard of a problem and is calling the manufacturer. I'm putting together my info for the inspector's boss and will probably visit them with a sample of the material. I'm assuming that the "masonry" in an ICF is the material that was grouted into the forms, and not the forms themselves. The fact that nobody here has heard of this as a "problem" leaves me to believe that it isn't a problem. Maybe I'll look for an electrical website to post the question. Mark

It shouldn't be hundreds of dollars as PVC conduit is pretty cheap. The expensive part will be the wire. You can't reuse the Romex by running it through the conduit. You'll have to get single strand wire, and 3 singles strands, hot, neutral, ground, are a lot more expensive than romex.

Keep at it till you win. The inspector just needs to be educated. All you need at this point is persistence, patience, and lot of tactfulness!

Actually romex is allowed in conduit, as long as it is not in a wet/damp location (unless the local code officials have an amendment forbidding it). Problem is that you probably would have to go larger than 0.75" conduit due to fill percentages. Not something you want to have cut into an existing wall.

Yeah, you're right. What I had in mind was you can't strip the outer sheath off of romex and use the single wires in conduit because of the lack of identification. The romex outer sheath does of course have all necessary identification.

I think you're misinterpreting the term "masonry" as used with this cementitious product. I seem to recall that Rastra is autoclaved lightweight concrete?

Anyway, you cannot embed romex in concrete and pour. I don't think this is your problem because the concrete(wall) in this case has already been "poured" so you are only making a trench or raceway for the wire. However, assuming the wire cannot be then embedded in plaster or masonry patching products, I suggest you discuss with the inspector inserting a proper sized backer rod material typically used as a backer for caulking. This would keep the wire back from the face of wall, and provide a bond breaker material between the wire and the plaster finish.

Many years ago I refoamed the romex in a regular type ICF wall, and the inspector said don't do it anymore; they want the wire to typically have an air space as small as it sometimes is! BUT, again, here is where there is another difference. When insulators spray ceilings and walls with open or closed cell foam, the wires are embedded in the foam.

Discuss all the facts with the inspector tactfully again to determine exactly the issue he/she is concerned with!

Had a nice talk with the "boss" inspector yesterday. Rastra is a strange material (to them) and that's what got them going about "embedded" in masonry or concrete. Rastra is post-consumer polystyrene mixed with a "proprietary" cementitious binder. The proprietary binder seems to be cement with a high amount of acrylic binder, so Rastra's "patent" hasn't proven to keep copiers away and there are now 4 or 5 seemingly identical products. But the question is "is it cement or is it foam." The NEC isn't clear as the purpose of 334.12(b)(2), so that makes it a little difficult to determine compliance. Although the ICF foam block has a small amount of concrete binder, I stacked the 12" x 12" x 60" blocks myself and cut them by hand with an old cross-cut saw. Sounds more like foam than concrete. The grooves for the Romex were cut with an electric chainsaw. Try that on masonry. So we agreed that it wasn't embedded in masonry, concrete, adobe, etc. as prohibitted in NEC. Whether the wire was going to be "embedded" in plaster was another issue. It was only going to be touching plaster on one side of the Romex, but because the NEC doesn't provide the reason for prohibitting embedment in plaster, we were kind of at a loss to determine what was needed. I proposed using canned foam or backer rod to secure the wire in the groove so that it 1) could not even touch the plaster, and 2) would be held back from the finish the required distance. There are foams that are polystyrene safe because they don't have a blowing agent that effects plastic (that's what I used to glue the blocks in place prior to grouting). I got an okay on that, although he said he would prefer to see backer rod. I've still got foam in a foam gun, but I've also got a few rolls of backer rod. I'll see which works best. I'm still not certain on the setback required for the wire. If the code requires a 1.25 inch setback from the face of the framing member, and there's no framing member, how is the distance determined. I'm assuming that plain vanilla buiding construction would be a 2x4 with 1/2" drywall, giving a minimum distance from wire to finish of 1.75". A two-coat plaster is usually >1/2", something more like 3/4 to 1", depending on how straight the walls are. So I'm looking at a minimum finish-to-wire distance of 2" and probably closer to 2.5 inches. My code book is over at the job site. Anybody know if there is a code setback measured from the finished surface? Mark

Where installed in grooves, to be covered by wallboard,
siding, paneling, carpeting, or similar finish, wiring methods
shall be protected by 0.0625-inch-thick steel plate, sleeve, or
equivalent, a listed steel plate or by not less than 11/4-inch
free space for the full length of the groove in which the
cable or raceway is installed.

Quoted from IRC 2006, TABLE E3702.1
GENERAL INSTALLATION AND SUPPORT REQUIREMENTS FOR WIRING METHODS, which would be a quote or restatement from the NEC.

That's interesting because the code doesn't provide any credit for the thickness of the covering and considers 5/8" wallboard as "similar" to carpeting. So, I guess covering with 1" of plaster would be "similar" to covering with wallpaper and it is only the depth of the groove that's taken into consideration. It looks like everything will work out (until he finds something at the next inspection). Thanks. Mark

I suspect it has mostly to do with acceptable risk. Wall board, regardless of thickness, will likely be nailed or screwed with fasteners having a max penetration into the framing member less than 1 1/4". The 1/1/4" probably comes from what's reasonable to achieve when you're drilling holes for wires through a 2 x 4. Also, it's the only number that can be verified visually during the construction process. Distance from front surface is a concern based on what homeowners are likely to use for fasteners to hang pictures, etc. To put a value on that would be a crap shoot at best, so 1 1/4" is as good a number as any to avoid having nails and screws driven into the wire. I'm sure there are romex cables more than 1 3/4" away from the face of 1/2" drywall that have been penetrated by 2 1/2' screws used by overzealous homeowners! I'd be willing to bet that most picture hanging fasteners, and the like, seldom extend more than 1" behind the wall board. You will virtually never see wall material of any kind that is less than 1/2" thick.

It's all a game of what's likely to be used, what's a reasonable safety margin, and what's the risk if that safety margin is violated. Any fixed number in a code is a matter of compromise. The safety idealists push for the tightest, the builders push for the loosest. The side with the greatest clout at the moment is usually the winner.

What did you use to hold the box in the wall? Are inspectors satisfied by simply having foam hold them in place?

Is that question directed to me? I'll answer anyway. The single gang boxes are Carlon blue plastic with the mounting tab on one side. We screwed the tab through the foam to the ICF web. Just had to be careful not to screw down too much and sink the tab into the foam. These boxes are just large enough for 3 romexes and one device. You can't run 4 romexes into them and be code compliant. Double and triple gang boxes were anchored with can foam all around the box. They were held in place while the foam set up with a stick over the box and screwed to adjacent webs. The multi gang boxes were Carlon blue with nailing ears. I cut the nailing ears off. Strictly speaking that was probably a non-approved modification to a UL listed box (a code violation) but the inspector either didn't notice, or didn't care. And I sure as h*** didn't point it out to him!!

For interior and exterior wall mounted lights I used plastic round boxes that had wells in the bottom for metal screws attached to a mounting plate. I tossed the mounting plate and used Tapcons to fasten the box to the concrete. I put a PVC medallion around the exterior boxes for the fixtures to set against, and fit the siding around the medallion. The box depth matched the thickness of the foam plus the medallion, or sheetrock. In a non-metallic box you can't have metal screws that the wires can touch. I think the outside boxes were fiber reinforced plastic and the inside regular blue plastic. Their depths are different.

My boxes in the ICF are held in place with only foam. Most of the single boxes are 18 cubic inch, but I did use a few 20 ci where I needed more wires (or double-gang boxes). Because my ICF produces a "screen grid" of concrete, there are places where the foam form is deeper than 2" and I could use a 20ci box, which I think is about 2.25" deep. In one spot I actually used my air hammer a little bit because I really wanted the box in that exact spot. I also used plastic boxes on my steel stud framing, and they are connected to the steel studs with metal screws. As was pointed out, this is probably a code violation (metal screws extending into the box) that the inspector didn't see or ignored. However, several of the boxes are grounded to the framing, so the whole steel framing is grounded. Because wood stud framing is what the MegaStores are geared up for, a plastic box with nails is 20 cents and a box for steel (if available) or some other configuration is $1.50 or more. I snipped off the nails and tabs on the cheap boxes, foamed the hole in the ICF, stuck the box and wires in, and added more foam around the edges. I had to change a few and they are not easy to remove. About the same as removing them when nailed to a stud. The only issue I saw was that it was possible for the foam to expand through the Romex opening into the box. An 18 ci box could become a 16 ci box. I went around and pulled out any foam the expanded into the box. I passed my electrical with the State after foaming the channels over the Romex, as I had planned. I then called the County for my framing, DWV, mechanical, potable, electrical (smoke detectors), and insulation inspection all in one shot. Passed them all. Merry Christmas to me. Now, it's on to plaster/stucco. Mark

I was an apprentice electrician many years ago (more than 30) and had little experience with plastic boxes. I learned the trade while attending a vocational school, which included studying the code. A lot has changed since them. I had no idea that you would not be able to put a metal screw inside a plastic box! I guess it's time to pick up the latest copy of the NEC. Thanks for the input.

Seeing that panel cover made me think of something else. Do most people surface mount their service panels on a piece of plywood on top of the foam, or do they mount it directly to the concrete? I see trade-offs in doing it either way. My ideas are that: Surface mount - 1) keeps the walls better insulated 2) allows easy access to knockouts to add additional romex 3) looks worse (with all the wires hanging out) if panel is in a location where you see it all the time (although you could still use conduit) Flush mount on concrete 1) looks much cleaner 2) more prep involved to run spare conduits into the attic for future expansion 3) loss of insulation since the inside foam will be gone (might not be a concern if located in garage.) I'd like to hear others opinions on this.

That's basically what I did, except I put it into a chase. I surface mounted two 2 x 4s on the foam from floor to ceiling and ran all the wires in the chase to the attic space or to the crawl space. I do have this 18" wide by 3 1/2" deep projection on the garage wall, but it looks a lot better than wires hanging out! Some wires leave the chase by the side in the foam.

A standard panel box is deeper than 2 1/2" of foam plus 1/2" of drywall and will stick out about 1". Fastening it to the concrete in a foam cut out may not look that pretty either.

Before I poured the garage slab or the first part of the house wall I installed four 2 1/2 or 3" conduit bends from the garage to the crawl space for romex and PEX channels.