Service entrance head on roof

This is related to my previousl post, but not exactly the same. I found something called a Service Entrance Head (1/2 and 3/4 inch sizes) that are designed to attach to emt conduit and provide a weather tight entry for wires.

It looks to me like my best bet is to take the PV string wires and run them directly to a service entrance head go get them below the roof where I can connect them to the combiner/fuse box (all in EMT, of course).

Does that sound reasonable/code compliant? It will mean between 2 and 4 roof penetrations, but I know that I can do those with no issues.

Thanks for all the help, so far, I can't say how valuable reading through the various threads has been to me as I finalize the smaller issues on my design. I'm shooting for starting work this fall/winter when the room and attic reach reasonable temperatures.

Comments

  • dwh
    dwh Solar Expert Posts: 1,341 ✭✭✭
    Re: Service entrance head on roof

    Yes.

    Wiles covered this in one of his "Code Corner" columns for Home Power Magazine:

    http://docs.google.com/gview?a=v&q=cache:J2YHbRpdHuoJ:www.nmsu.edu/~tdi/pdf-resources/cc31.pdf+weatherhead+nec+solar&hl=en&gl=us


    "Exposed, single conductor cable is only allowed for
    module connections. At some point near the modules, the
    wiring method must be changed to one of the other
    methods that meet the requirements of the NEC. The
    exposed, single conductor cables could be routed to a
    weather head and into conduit and then into the building
    and to the PV disconnect switch."


    Also, keep in mind the strength of the pipe - for 1/2 or 3/4 EMT with a weatherhead on it I wouldn't want to see more than maybe a foot sticking up out of the roof unsupported.

    If I had to have more than a foot of 1/2 or 3/4 pipe sticking up out of a roof with a weatherhead on it, I would probably use ridgid (RMC) rather than EMT (just for the bit sticking through the roof - the rest of the run could be EMT).


    But, as I mentioned before, don't forget to consider de-rating. Wire is rated to carry a certain current in a certain temperature range. Wire in a conduit is a bit different. The wire has to be de-rated if the conduit reaches a certain fill level, or if the temperature reaches a certain point.

    I'm not sure how many wires of what size you are dealing with, but it doesn't take a whole lot to reach A) the fill level of 1/2 or 3/4 conduit. And on the roof it doesn't take much to reach B) the temp level.

    In the case of A - you have to increase the conduit size to have a fill level large enough to handle however many wires you have.

    In the of B - you have to increase the wire size to carry the same current at the higher temp, which means you will, again, be increasing conduit size.


    EDIT: Oh, and weatherheads aren't actually raintight - water can get into them. So - don't forget to leave "drip loops" on the wire.
  • Eric Ernst
    Eric Ernst Registered Users Posts: 16
    Re: Service entrance head on roof

    Thanks dhw,

    This gets curiouser and curiouser. I'm sitting here with NEC tables 310-16 and 17 and the de-rating factors. The panels I'm using are Sharp 235w with an Isc=8.6a that come with 12ga USE-2 cables. If go to the tables, I see that 12ga use-2 is good for 30a (ignoring the fuse issue). Then if I use the required 1.56 x 8.6a = 13.4a for my design current before correcting for roof temperature. At 80C the temperature correction factor is 0.41 so I have to use wire that can accommodate 13.4 / 0.41 = 32.7a! In other words, even though the panel is UL rated with 12 gage wire, I can't connect to it with 12 gage to go to the combiner box. Am I doing something wrong?

    Very good points about RMC and leaving drip loops!
  • sub3marathonman
    sub3marathonman Solar Expert Posts: 300 ✭✭✭
    Re: Service entrance head on roof

    I too ran into those same problems, both with the service head and also with the wiring from the panels to the combiner box.

    For the roof penetration I was extremely worried. Eventually I went with a 1 1/4" Rigid Aluminum Conduit, with the weatherhead. I got the flashing from Home Depot, and the electricians installed it up on the roof. UNFORTUNATELY, the flashing was for "1 1/4 inch to 1 1/2 inch conduit," so it didn't really seal around the conduit. So one suggestion that would probably help people is to go with the largest conduit for the flashing you're buying. I wondered, but it said 1 1/4" was OK. After putting it up I found out it wasn't. The options now are either get the two-piece flashing seal or removing the conduit and putting up 1 1/2" conduit. I still haven't decided.

    I too looked at the Solodeck flashing, and it looked very good. The only drawbacks were that I couldn't just get it locally, and also it is limited to the size of the roof penetration, so I think I would have needed two. So I went with just one roof penetration, but a bigger hole. I wonder though if the Solodeck wouldn't have worked out better after all.

    For the wire size, that is the big problem when it is run in conduit on the roof. The derating factors became such a dilemma that I just eliminated the conduit and put the money into longer MC cables instead, and then ran the cables along the rails under the panels and then down through the weatherhead. That eliminated the wire splices on the roof, which made the electricians happy too. They said in Arizona those splices would probably be OK, but in Florida there would eventually be problems.
  • dwh
    dwh Solar Expert Posts: 1,341 ✭✭✭
    Re: Service entrance head on roof
    I too ran into those same problems, both with the service head and also with the wiring from the panels to the combiner box.

    For the roof penetration I was extremely worried. Eventually I went with a 1 1/4" Rigid Aluminum Conduit, with the weatherhead. I got the flashing from Home Depot, and the electricians installed it up on the roof. UNFORTUNATELY, the flashing was for "1 1/4 inch to 1 1/2 inch conduit," so it didn't really seal around the conduit. So one suggestion that would probably help people is to go with the largest conduit for the flashing you're buying. I wondered, but it said 1 1/4" was OK. After putting it up I found out it wasn't. The options now are either get the two-piece flashing seal or removing the conduit and putting up 1 1/2" conduit. I still haven't decided.

    Strange. A metal flashing like that would come with a hole pre-cut for the smaller pipe, and would have to be opened up a bit to fit the larger size. Some flashings even have lines embossed around them to indicate where to cut for each size pipe. Or was this a "rubber-ring" type flashing?

    Either way, I blame the electrician - he damned well should have noticed the poor fit when he was installing it...and he should have said something before it was too late.


    The way I do it is what I like to think of as "craftsmanlike" and some people think of as "custom".

    I like to buy a flashing that's a bit too small, and then use snips to open up the hole. One reason is that roof pitch varies, thus so does the plane that the opening lies in. By manually trimming the flashing I can get a perfect tight fit around that particular pipe with that particular pitch on that particular roof - every time.

    That's of course for the all metal flashings, I never use the rubber-ring type.
  • niel
    niel Solar Expert Posts: 10,300 ✭✭✭✭
    Re: Service entrance head on roof

    it's also possible it was bought with the enlarged hole already there as people in stores do play, but the electrician should have noticed this.
  • dwh
    dwh Solar Expert Posts: 1,341 ✭✭✭
    Re: Service entrance head on roof
    Eric Ernst wrote: »
    Thanks dhw,

    This gets curiouser and curiouser. I'm sitting here with NEC tables 310-16 and 17 and the de-rating factors. The panels I'm using are Sharp 235w with an Isc=8.6a that come with 12ga USE-2 cables. If go to the tables, I see that 12ga use-2 is good for 30a (ignoring the fuse issue). Then if I use the required 1.56 x 8.6a = 13.4a for my design current before correcting for roof temperature. At 80C the temperature correction factor is 0.41 so I have to use wire that can accommodate 13.4 / 0.41 = 32.7a! In other words, even though the panel is UL rated with 12 gage wire, I can't connect to it with 12 gage to go to the combiner box. Am I doing something wrong?

    Ya well...welcome to the club! :D

    This is why I mentioned the derating thing. It's a PITA, and it sounded to me like maybe you hadn't properly considered it.


    I note a few points (in no particular order or relevance):

    * That 30a rating for the #12 USE-2 is from table 310-16 (enclosed)...310-17 (free air) shows 40a.

    * You are using the full 80C to derive the .41 derating factor - but according to the above mentioned Wiles article:

    "After determining the temperature ratings of the module terminals (75°C or 90°C — marked on the back of UL-listed modules or from the manufacturer), the appropriate temperature derating factor can be selected from Table 310-16 or 310-17 in the NEC.

    Cable with 90°C insulation must use a 75°C conductor rating for the ampacity calculations, if the terminals are rated for only 75°C. The temperature derating must then be applied to this ampacity. Most installations should use a temperature of 65°C to derate the conductors. In hot locations, with no ventilation provided for the back of the modules ( e.g. mounted directly on a roof), a 75°C temperature should be used...

    ...The modules are mounted on a rack on a brown shingled roof, but for esthetic reasons, the spacing is only two inches....

    ...If the modules were spaced six or more inches from the roof, the maximum operating temperature would drop to about 65°C on hot, sunny days. In this case, the derating factor is..."


    Soooo...

    What are the terminals of your panels rated at (the Sharp spec sheet doesn't tell me)?

    http://solar.sharpusa.com/files/sol_dow_NU-U235F1.pdf

    How far off the roof will the panels be?

    Will there be ambient temps > 40C?


    You might come up with a way to run a bunch of #12s down from the roof in conduit...but, personally - I think you might end up having to combine on the roof (or just under the roof - in the attic) and run larger wire down from there.
  • Eric Ernst
    Eric Ernst Registered Users Posts: 16
    Re: Service entrance head on roof

    Wow, I feel like I'm back in kindergarden - and I thought I'd done my homework.

    You are right. I punched the wrong table. I have both tables on the same page, and just "spaced" it.

    I never heard of the "Cable with 90°C insulation must use a 75°C conductor rating for the ampacity calculations, if the terminals are rated for only 75°C." but it makes sense. The Sharp spec sheet I have, under "Mechanical Characteristics" has "operating temperature (cell)" of -40 to 90 C. Does that count for the entire panel including the terminals?

    I was using the 80C de-rating factor because the table I have only has 70 and 80 degrees and I wanted to be conservative.

    Now, it seems like: if I use the possibility (worst case?) that the panel only has 75C terminals then 310-17 allows (under 75C) 35 amps for 12 ga. Even if I use the 80C de-rating factor, I have to deal with 32.7 amps which puts me under the 35 (barely).

    Gad, I think I might be starting to "get it."

    This is fun
    hmmm, that probably means I'm sick :p

    E