Penetration into roof via junction box?

[Xplode]

Is the hardware on the L brackets and rails , Stainless ? I hope so.

I can't tell, which is the uphill side, but the downhill side should have minimal goop, so any leakage water, can drain away and not be trapped under the L or the box

Yes to all of it. The fastener hardware is all stainless, I've also used anti-oxidation compound on all the fasteners, as well as the grounding connections, and I put a complete ring around the penetration underneath the box, but horseshoe shaped sealant around the outside edges of thebox... ie there isn't a bead across the downhill side of the box (it's the side under the black nylon glands)

The L Brackets are actually mounted using stainless hanger bolts with a rubber seal. It was what was suggested to me by the suppliers, although I have since found out that Kinetic has some other options. Didn't love the look of the seal and didn't trust it to survive a decade or two of -20C to +40C so I sealed the actual penetration with quadMAX and then coated the whole thing up to the shank of the stainless bolt again. I'm not normally an over-sealer kind of guy, but I really don't want to have this thing leak and make me look like an ass to the family.

IMG_20180806_125210134.jpgIMG_20180806_125216148.jpg

[Mike90250]

Is the hardware on the L brackets and rails , Stainless ? I hope so.

I can't tell, which is the uphill side, but the downhill side should have minimal goop, so any leakage water, can drain away and not be trapped under the L or the box

[Xplode]

Well I went for it this past few days... Got the system turned on today. Ended up not being able to turn the box diagonally due to the profile of the roof. but with the amount of QuadMAX i used, I feel pretty confident it'll stay watertight for years to come. if it doesn't, I have access to the underside of the penetration through the soffit material below, so I'll be able to see what's going on. The racking/panels are also only 2 feet from the peak of the roof, so minimal water should end up running at the box as well.
Due to some crappy service and poor stock choices from the local shop in BC here, I ended up with 300Watt Canadian Solar panels instead of the 280W I originally wanted, but the sales guy was kind enough to mark them down considerably for me ($250CDN/panel instead of $300CDN).

IMG_20180806_145645935.jpg
I probably could have centered the hole in the box a bit more, but I was thinking about how stiff the #10AWG PVU wire is, and wanted to have it be able to lay down in the box nicely. The two screws on the uphill side actually grab into the roof supports below, so the side that will see water flowing at it is cinched down tight to the roof. I sealed around the hole on the roof, and then again around the end of the box, plus the screws themselves were coated before install, and then around the screws where they are exposed as well.


IMG_20180808_093423154.jpg
After much reading and people not being keen on using wire-nuts for the splice (transitioned to teck90 cabling to get to the breaker down with the charge controller), I used some insulated terminal strip, but honestly I think wire nuts are probably more solid... I may go back and change it later. Panels are only capable of up to 9.8A max, so the #10awg should handle it with no heat at the joint. The cables enter the box from the downhill side, and I ended up picking up a 3/8NPT tap and just drill/tap the box instead of using the nuts. Should improve the sealing.

Normally I'd make the sealant look a little nicer, but in this case you can't see it (panel over top of it) and I just wanted lots of sealant to make sure it holds.

[foo1bar]

It's still a penetration of about the same size....

Sounds like you have to have a penetration of a ~1.5" conduit.
It's just how are you going to keep that penetration watertight.
Is it going to go up into a Jbox? And the jbox has a rubber seal compressed between the jbox and the roof, keeping that penetration waterproof? (along with some caulk that helps keep the jbox tight in place and helps with waterproofing?


Or is it going to be a 1.5" or 2" conduit going up to a weatherhead? or over to a jbox? or over to gland(s)?
And the weatherproofing for the roof penetration is standard boot type weatherproofing.
Then the weatherproofing for the inside of the conduit is handled by the weatherhead/jbox/gland(s).

I think either approach for weatherproofing the penetration is viable. I'd probably go with the one I think I'd be more comfortable with executing.

[Xplode]

That is a good call and I ALWAYS try to use anti-seize compounds on hardware. Electrical connections often too (although not little splices - more like lugs and stuff) I work in a lot of nasty industrial environments and that stuff has saved my ass more times than I can count. Thank you for the reminder - I'll put it on my packing list to make sure it goes with me for the install.

[Mike90250]

> I think if I rotate the JB 45degrees so it's angled downstream of the roof, it'll help it not "catch" water

That will help, and if your sealing goop is protected from the worst of the weather by a panel, thats good too. Maybe put some anti-size on that panel's mounting hardware so you can get it out easily.

[Xplode]

This sounds reasonable to me.

I'd want it fastened tightly to the roof so there was minimal movement. Probably locknuts both above and below getting it tight as possible so that the jbox would not move much on that roof.


Another option is put the jbox on the rail for the solar module. And have a boot around the conduit.

Or don't do a jbox - just have a boot with the conduit going to a gland.

This isn't a bad idea either, except i have two things to bring down through the roof. a Teck90 cable (about 3/4-1" across) and a #6awg RW90 Ground wire. Not sure how to seal that nicely, unless I install a pipe and weatherhead type arrangement, but then I don't understand why that's a better solution than gluing down the jbox anyway... It's still a penetration of about the same size....

I think if I rotate the JB 45degrees so it's angled downstream of the roof, it'll help it not "catch" water, and it's under the panels so no direct driving rain or snow. Now I'm wondering if drilling a small 1/8" hole in the lower corner of the box will help ensure any water that does ingress won't fill up the box and end up in the cabin. I can use watertight connections on the cables entering the attic space so there is no way for water to leak through... All my cables entering the JB will be on the lower side again to prevent water ingress...

[foo1bar]

If I were to have to mount a J-box on the roof I would use a liberal amount of Butyl rubber sealant under the box followed by a clean bead around the perimeter with some quality urethane sealant. 3M 5200 comes to mind for the perimeter. Butyl rubber is interesting stuff. It never cures. It will stay soft and sticky indefinately. While nothing is forever this combo would remain sealed for a very long time.

This sounds reasonable to me.

I'd want it fastened tightly to the roof so there was minimal movement. Probably locknuts both above and below getting it tight as possible so that the jbox would not move much on that roof.


Another option is put the jbox on the rail for the solar module. And have a boot around the conduit.

Or don't do a jbox - just have a boot with the conduit going to a gland.

[Xplode]

Soladeck, whatever. Referring to the materials I'd use if I had to do this.

That is actually a cool looking product that would work very good on a shingle roof, which one of my cabins is, so I may investigate that further for that install.

But flashing and such doesn't work on a metal roof... Sigh... I really don't want it to leak, but I used some Geocel 4500 on a metal roof penetration at a factory a couple months ago and it seems to have worked awesome for bringing up power cabling to a motor on the roof. Bonds to metal and PVC, cures underwater and stays flexible... But in that case I built a ABS Pipe stack and sealed an EPDM style rubber boot around it... don't need a 3" pipe x 18" high this time tho... aiming for a little more subtle.

https://www.geocelusa.com/product/geocel-4500-roof-bonding-sealant/

The solar supply shop sells and recommends LePage Quad MAX for penetrations

http://www.lepage.ca/en/lepage-products/sealant-construction-sealant/lepage-quad/lepage-quad-max.html

Looks like these products should work well enough for sealing the box down...

[littleharbor]

Soladeck, whatever. Referring to the materials I'd use if I had to do this.

[Mike90250]

a j-box is not a usual device to seal a hole in a roof.

[littleharbor]

If I were to have to mount a J-box on the roof I would use a liberal amount of Butyl rubber sealant under the box followed by a clean bead around the perimeter with some quality urethane sealant. 3M 5200 comes to mind for the perimeter. Butyl rubber is interesting stuff. It never cures. It will stay soft and sticky indefinately. While nothing is forever this combo would remain sealed for a very long time.

[Xplode]

I think gluing a J-box to the roof is a poor idea. Something like a plumbing roof vent, real flashing, a weather head for the cables and then splice from the UV rated cable, to standard wiring , at a j-box in the attic, is what I'd do.

Thanks for the feedback Mike... I hesitate to make a penetration in the roof under a jb, but is a plumbing roof vent any better? (the roof is metal, not shingles) I don't have an actual attic I can access once the ceiling is up inside (it's currently under construction).. But I could transition to the teck90 cabling (which is also UV rated) on the roof and then go down through a vent if that really does seem like the better idea...

Why do you prefer having to seal the vent, vs sealing the jb?

[Mike90250]

I think gluing a J-box to the roof is a poor idea. Something like a plumbing roof vent, real flashing, a weather head for the cables and then splice from the UV rated cable, to standard wiring , at a j-box in the attic, is what I'd do.

[Xplode]

I don't do a lot of solar work obviously, but i can't see anything in the CEC about 2 strings or less not needing OCPDs. Everything I read seems to indicate there needs to be protective devices, they need to be rated at 100% of the capacity of the panels (or next standard higher value), and that everything needs to be installed in metal raceways. There are rules about requiring DC Arc Fault protection as well.


But again, the question I posed originally was more of a "what do you think of my method of penetrating the roof" and not so much a code argument.