Construction Water sealing a large corrugated roof - after-the-fact

[DaveC426913]

TL;DR Summary

Is there a practical way of sealing the leaks in a large corrugated patio roof?

Some dummy decided to
a] customize his patio roof by substituting a couple of clear lexan corrugated panels along the opaque aluminum corrugated panels, knowing they are not compatible.
b] do it himself, to save money.

So now the roof leaks where the joins and screws are. Not bad, but what use is a patio roof in the rain if you still get rained on?

I'd hoped I could spray some sort of stopleak on the underside, but I suspect that won't actually stop the leak for long. The water will pool up and unseal the leak. So it probably has to be applied from the top?

Trouble is: it's 16 feet wide by 12 feet deep. The cross braces, where the leaks occur, are at 4 feet and 8 feet - well beyond my ability to reach them from the front edge with a ladder.

Suggestions?

It's a lot like this, but huger:

 

[Averagesupernova]

I'd like to see a few pix. Does it leak on every screw?

 

[BillTre]

Did the assembler use the kind of screws or nails that come with a little rubber grommet to seal to the rook when used?

 

[phinds]

Use an extra long-handled paint roller with black tar roofing goop. Male your own handle-extender if necessary. Make sure the tar and preferably also the weather are warm when you do it. And yes, of course it has to be applied from the top.

 

[Baluncore]

Does it only leak when it is windy?

 

[sophiecentaur]

Does it only leak when it is windy?

This is very relevant. Wind can drive water uphill and filler could be applied to joints from underneath. Any filler has to be able to move / flex with any relative motion over joints or it will crack and you get your leak back. If does move during wind then would it be possible to reduce flexing / movement? Could you access from below and put in extra rafters or an overlapping strip of corrugated sheet?

 

[DaveE]

Aluminum butyl flashing tape might work from below, it's very sticky. Clean the surface first, of course. It's a kludge and should be on top.

I wonder if you could make a long handled tape roller to apply it on top from a ladder?

 

[DaveC426913]

Weird. I have not received any alerts for any responses to this thread. I have auto-subscribe on, and the thread is still set to "watch", so I don't know why.

I'd like to see a few pix. Does it leak on every screw?

I will post pix ASAP. No.

It leaks under only a very few screws in the aluminum panels (~12').

Most of the leaks are under the lexan panels (so that's a 4' run at the 4' mark and another 4' run at the 8' mark)

The primary reason for this is because the lexan panels use screw fasteners that are designed to go into wood, but the roof crossbeams are metal. This forced the dummy to pre-drill holes and then try to put the screws into the holes. Unfortunately, alignment is far from perfect and many screws went in at an angle, defeating the watertight seal.

Did the assembler use the kind of screws or nails that come with a little rubber grommet to seal to the rook when used?

Yes, but see above.

Use an extra long-handled paint roller with black tar roofing goop. Male your own handle-extender if necessary. Make sure the tar and preferably also the weather are warm when you do it. And yes, of course it has to be applied from the top.

:slaps forehead:
I was thinking about a paint brush on a long stick, but I had low confidence I would get good cover.

I don't know why I didn't think of a paint roller. That would work very well.

Does it only leak when it is windy?

No. I can say this with certainty because when I was marking the leaks there was zero wind and the rain was merely a drizzle.

Any filler has to be able to move / flex with any relative motion over joints or it will crack and you get your leak back.

This is my assumption, yes.


A paint roller will work very well.

My only "stretch goal" now (see what I did there?) is wishfully thinking that I might find a transparent sealant goo, because opaque goop is going to look hideous through the clear lexan panels - particularly when applied by a dummy with a long-handled paint roller from several feet way.

 

[DaveC426913]

I just realized, if worse comes to worst, I could also run a little eavestrough along the underside of the two crossbeams to run any remaining drips off to the sides.

Terribly kludgy, although almost invisible if I got the right colour. It would only need to be a trough that's, say, 2" wide x 1" deep. Even though it would be 16' long it would be hardly noticeable.

 

[DaveE]

opaque goop is going to look hideous through the clear lexan panels

That's one reason I like my butyl tape idea. At least the edges will be straight (sort of). Still, if you're really careful with the roller, goop could look OK.

I guess you could try clear silicone sealant at the seams from the bottom. I'm really not a big fan of silicon RTVs with any kind of environmental exposure, but I bet it would work for a couple of years. Then you could scrape the easy bits out and reapply.

 

[BillTre]

I guess you could try clear silicone sealant at the seams from the bottom. I'm really not a big fan of silicon RTVs with any kind of environmental exposure, but I bet it would work for a couple of years. Then you could scrape the easy bits out and reapply.

A goop product might work. (there are a variety of products related to shoegoo). It is much tougher to abrasion etc. than silicon sealer, but more difficult to apply well.

 

[sophiecentaur]

I just realized, if worse comes to worst, I could also run a little eavestrough along the underside of the two crossbeams to run any remaining drips off to the sides.

We're gonna need a bigger bucket.


If it's possible to get close to the join from underneath then (after thorough cleaning with a squirter) why not try a mastic gun to force the goo into the screw hole up from behind? But water can be sneaky stuff.

I have used some magic stuff called "Captain Tolly's Creeping Crack Cure". I used it to seal small leaks on a sailing cruiser and also on weeping joints in a conservatory. It works best from above as it drips through through gravity but it could be poured from a teaspoon on a stick, to get to the site of the problem. It's colourless, btw.

 

[Dullard]

I have successfully used the combination of a step-ladder, an extension ladder, and the house roof to 'bridge' that kind of porch roof. Have a trustworthy person brace the step-ladder.

 

[DaveC426913]

If it's possible to get close to the join from underneath

It's more than possible; it's trivial.
Don't even need a step ladder.

By the way, those two pieces of reflective tape dead centre show the worst leak.

And you can see how the wood screws with their rubber grommets, screwed askew, into the metal crossbar, do a terrible job of being watertight:

 

[DaveC426913]

I have successfully used the combination of a step-ladder, an extension ladder, and the house roof to 'bridge' that kind of porch roof. Have a trustworthy person brace the step-ladder.

I have lived six decades without a sprained or broken bone by systematically eschewing these kinds of solutions. And that was before internet videos came along with their Life Lessons.

(But more practically, the house roof is a second story - 10 feet higher up than the patio roof).

 

[Averagesupernova]

So is that some type of form underneath the clear sheets on top of the beam that goes across? Are they a substantial form? It's possible they are there to take the stress so the screw can be run in on top of the ridge. Seriously reduces leaks. Still an issue to get a screw made for wood to go through steel. And of course it's after the fact. Short self drilling screws made to go into steel are available with the rubber washer. The ones I've used were not very long and were made to install on the flat part of steel sheeting that looks similar to what you have off to the side of the clear plastic. Longer screws that self drill into steel are likely available.
-
Also, polycarbonate clear sheets are available that match (from what I can tell in your picture) the metal sheeting you show. I don't know how they perform laying flat with snow load and all.

 

[DaveE]

By the way, those two pieces of reflective tape dead centre show the worst leak.

Oh, now I get it. I thought it was the long seams. Duh.
I hereby completely rescind everything I said about tape. That would be nearly impossible to apply.

How about a dollop of caulking applied to each screw head. Maybe with a caulking gun fashioned with an extra, extra long nozzle (after the cleaning brush with the extra, extra long handle)? I don't think it would look bad to anyone but you, since it's above the beam. Even you wouldn't care after a year or so.

 

[DaveC426913]

So is that some type of form underneath the clear sheets on top of the beam that goes across?

Just that 16' long L-bracket.

Are they a substantial form? It's possible they are there to take the stress so the screw can be run in on top of the ridge.

I don't follow. Remember, this is custom.

The panels have wood screws, not compatible with going into a metal bracket. I had to pre-drill holes, and they were not well-aligned because the lexan panels are 12 feet long and there's a lot of torquing and flexing to be done.

Yes, self-tapping metal screws would have been better if I had done that from the start, but here we are.

Seriously reduces leaks. Still an issue to get a screw made for wood to go through steel. And of course it's after the fact.

Right.

Also, polycarbonate clear sheets are available that match (from what I can tell in your picture) the metal sheeting you show.

I do not think so.

I don't know how they perform laying flat with snow load and all.

They did survive more than a foot of very wet packing snow - but just barely. Impressive, really.

Frankly, the roof itself was going to fail before the lexan panels did. I almost didn't notice that the whole roof was bowing in the middle by quite a few inches. Fortunately, the patio comes with temporary bracers specifically for winter use.

 

[Averagesupernova]

I do not think so.

I do. Unless that sheeting is different than I think it is. I'll snap a pic on a building I have and post it.

 

[DaveC426913]

I do. Unless that sheeting is different than I think it is. I'll snap a pic on a building I have and post it.

Alas, even if so, it is moot.

1. I'm not about to replace them. Not because I don't l want to but because I'm past the point of return. Once the roof is built, I can't get at the joints from the side on a ladder. The clear panels are in the middle - I have no way to get at them to remove them or replace them without removing panels just for access.

2. Even if I were of a mind to replace them, new panels will not line up, since the the roof has already been based on the width of the existing panels. I'd have to pull off half the roof and start again for the new edge.

Live and learn I guess.

 

[sophiecentaur]

@DaveC426913 from what you write I think it must be down to the screws. For an ideal seal, the threads should be smaller diameter to the holes and the shank be a good fit. As there application is so common, I'd bet that just the right thing is available with wood or machine / self tapping threads. rubber washers top and bottom and a tape in between the join would also help. The varying load means the seal would need a lot of help, imo, to allow for all that flexing.

What you do depends on how you feel about re-doing jobs. But it's often much worse in prospect than when you actually get to doing it. If you can cope with the idea of removing panels then look at what suppliers offer.

Having just moved into a house with a 'legacy' I share your pain.

 

[Averagesupernova]

Wood screw vs metal. Both are made for ribbed sheeting.



Polycarbonate that matches the steel.




Better pic of the type of ribs I am used to
9 inch on center between ribs. Each sheet covers 36 inches. Care needs to be taken when installing to get the rain groove under the adjacent sheet instead of over. Can't see the rain groove after installation.

 

[DaveC426913]

I should set expectations here. I don't think a re-do is in order. It is technically difficult and I do not think it will fix the leaks because the panels are now full of holes.

So I'm concentrating on remediating leakage without disassembly.

9 inch on center between ribs.

I think mine are 7 inches on-centre, but it's been 6 months.

 

[Averagesupernova]

Good luck.

 

[sophiecentaur]

So I'm concentrating on remediating leakage without disassembly.

I can understand that.

The interface between the rafter and each 'trough' with the screw through it is presumably accessible with a penetrating sealant. Captain Tolley could well work and I'd give it a go. It would flow down the slope and 'find' any splits / leaks. It may only be available in UK but the local Marine market may have similar. Failing that, clear silicone sealant could be pumped in from top and bottom. The overlap may be too long for the to work but maybe a small local hole in the lower sheet could allow a nozzle in close to the problem. (Strength could be compromised?)

This assumes that the drips come from the screw holes and not via the uphill end of the lower sheet. If it comes in that way it presumably would be blown uphill by wind. Perhaps some sealant along that uphill line could stop that.

 

[DaveC426913]

The interface between the rafter and each 'trough' with the screw through it

No troughs should have screws in them. Only the crests. Design feature.

This assumes that the drips come from the screw holes and not via the uphill end of the lower sheet. If it comes in that way it presumably would be blown uphill by wind. Perhaps some sealant along that uphill line could stop that.

True, the aluminum sheets are only four feet, so they overlap. And I did run a bead of caulking under the overlap. But the worst leaks are from the lexan, but they are uninterrupted 12 feet end-to-end, so no overlaps.

Regardless, it leaks plenty on a windless, drizzly day.

So, upshot: majority of leaks are coming from lexan screws, set in crests at janky angles, defeating rubber grommet seals.

 

[Averagesupernova]

No troughs should have screws in them. Only the crests. Design feature.

This would have cleared up a question I had yesterday. Oh well...

 

[sophiecentaur]

This would have cleared up a question I had yesterday. Oh well...

With hindsight, it's obvious that securing the peaks would keep pressure on the trough contacts. (Another school day for me)