Roof valley ice melt from below. Thoughts and help with circulator and plate heat exchanger sizing

nancyl

Greetings.

We're contemplating a method to warm the underside of a somewhat flat roof valley using 1/2" pex, aluminum radiant heat transfer plates attached to the underside of 1/2"OSB and then covered in 1" closed cell spray foam. Something like you see in this picture.




The goal is to provide enough warming so on warmer winter days the heat from this will assist in warming the ice from below in the attic should a layer form, thus preventing an ice dam from getting a foothold. There are 7 roof valleys to address. Each loop of 1/2" pex ends up being about 300 feet in length between the ends that connect to the manifold.

The valleys will never all be warming at the same time, just one at a time as the situation call for it. We do not anticipate running it to melt all the ice or snow, just warm from below to 35F or so to get things melting and then turn it off.

We'll be adding this demand to our radiant heat system already in place heating our house. That system was oversized to accommodate this addition.

Each valley area of spray foamed pex ends up being about 150 sqft.

We expect to use a glycol mixture but don't know what percent yet.

We're trying to establish what size circulator pump and braised plate heat exchanger we might need to perform what I'm describing.

Would love to hear your thoughts and comments on the idea but are really mostly interested in the pump and heat exchanger question rather than alternate solutions like electric solutions or heat tape on the roof, etc.

Thanks all for your help.

Jamie Hall

Be absolutely certain that your melting is most effective at the edge of the roof. Otherwise you will make the situation worse.

pecmsg

Jamie Hall said:

Be absolutely certain that your melting is most effective at the edge of the roof. Otherwise you will make the situation worse.

Exactly 
start with the lower roof. 

Slowly warming and melting. 

Zman

Ice dams generally are formed because the roof surface over the heated space is being warmed due to poor insulation/air seal details. The ice dams form at the eves when the water flowing down the roof is no longer over the unheated space.

The vast majority of ice dams are a result of poor design and insulation details You are far better off designing the building correctly and avoiding ice dams to begin with.

Do not spray foam directly on tubes unless you get the tubing manufacturer's blessing. The heat generated by insulation curing can damage the tubing or O2 barrier.

Zman

To your original question, if you intend to heat the roof, be sure to extend the tubes to the edge of the roof so you don't add to the ice dam problem. I would think you would want to estimate your loads like a snowmelt system, >100 BTU/Ft.

hot_rod

 Can you just melt a narrow flow path. That seem to be how the electric cable systems work.


Similar to snowmelt you design to how much, how quickly.

Melting  as the snow falls or kicking it on to handle an accumulation.

Id agree, a 100 btu/ sq ft or higher

Could you put metal roofing in the valleys, keep snow sliding off before it melts and dams?  That is common in ski towns

I have always considered tube under metal roofing for melting and or solar thermal harvest

Mad Dog_2

Interesting idea.  When I was an Apprentice I worked for a great shop called GF Sheridan.  Our Shop was in sub basement of Rockefeller Center and we serviced all the famous buildings: Radio City Music Hall, The Restaurants all in and around The Rock Rainbow 🌈 Room, Hurleys Saloon, the ice skating rink and Saint Patricks Cathedral.   Saint Pat's had bad icing up in the copper roof gutters.  It was a very radical pitched roof and there was only about 12" that you could walk and work on.  Harnesses were useless without anything to tie off to. I was "The Kid" so I creeped out and scooted in the gutters, ran pipe out and then  soldered all the joints.  To supply the gutter mains,, we cut some tees off the steam risers in the catwalks high above the church floor. It was all wood frame catwalks and handmade wooden ladders. We ran out copper mains (black pipe would rust and react with the copper gutters) in to the gutter trenches and pitched them for parallel flow.   The condensate was wasted in to the gutters (great for the boiler, heh?)
and we peaned over the ends so the condensate would just piss out without losing much steam. It was simple and crude but it worked.  No more icy gutters for St Pat's. The Cardinal personally Thanked us.  Cardinal O Conner was my favorite Cardinal. Mad 🐕 Dog 

GroundUp

If you were to just properly insulate the underside of the sheathing, ice dams would never be present. No need for any sort of melting devices if the building is properly constructed.

Mad Dog_2

Hmmm..kinda makes sense. I guess your a builder/roofer? Mad Dog

hot_rod

If I were to do a roof melt, use the aluminum transfer plates, keep  1/2@ looks at 200- 225 feet in length

the size of the HX and pump depends on how many loops you install

35- 40% glycol mix ought to do it

its going to take a good size boiler if you figure 100 btu/ sq foot of surface

nancyl

Thank you for the comments. I'll address some of them

This is an ICF building with a standing seam metal roof with a two foot eve. The pex and plates will be installed from just into the eve (<20" from the edge) and run loops up towards the peak in only the first 8 feet or so and 4 feet or so out from the center of the valley.

The attic is insulated with 24" of blown in fiberglass on top of 8" of fiberglass batts. The roof is well ventilated and a darker color which encourages rapid melting when the sun hits it.

The cause of the ice dams is a dryer vent which exits on one side of the valley 7' up from the edge and a large cooktop vent that exits on the other side of the valley. The roof angle is low enough that these cause some melting during their periodic use and resulting ice formation when they are off.

The goal is not snow melting, but ice warming enough to get melting started. Once melting starts on a sunny day the roof clears pretty quickly.

So... any help on the plate heater size? Is oversizing the heater a really bad idea or?

pecmsg

If the problem is the drier vent Move it!

hot_rod

The HX size depends on how much area or how much tubing you plan on installing. You sort of need that number to size both the HX and the required boiler capacity that you will need.
But it seems like an expensive fix for a couple exhaust fan vents?

Derheatmeister

Up here in Summit Country Colorado we use 130 BTU/sqft/Hr for our snowmelt designs..
-30 F is typical for the Glycol concentration on Snowmelts and Solar systems to avoid potential freezeups.
-30 F is very viscus which makes it harder on pumps (Must be oversized).
300 foot loops on 1/2" is WAY to long..I would keep it between 180 and 200 feet !
Advice is not to Spray foam directly on to tubes..

Seems like you have a cold roof which hopefully has a proper eave and ridge ventilation system installed..
Are you retaining snow with guards ?

With the proper engineering a roofmelt can be done but personally and as other have previously mentioned in this post i would address the source of the problem first before wasting large amounts of precious,steadily increasing energy resources and the funds on a roof melt..

nancyl

Derheatmeister said:

Seems like you have a cold roof which hopefully has a proper eave and ridge ventilation system installed..
Are you retaining snow with guards ?

Yes, cold roof with eve and ridge ventilation.
Yes snow guards. We've had 45" of very wet snow so far this year

PC7060

Why not extend the height of the problem exhaust vents. Seem much simpler and less error prone. 

Ciupaga_PL1

I think it's a bad idea to do it like that, if a pipe burst in the winter you're out of luck.
Is it leaking because of the ice dam? If not just leave it alone, you'd be surprised how many ice dams there are almost on every roof in the winter, just can't see them unless you go up on the roof. 
East and north sides will have more ice, south and north less.
If it ain't broken don't fix it.
If done correctly and the best way possible, standing seam double lock is the best roofing system you can put on a house. 

Ciupaga_PL1

It might not be the vents faults completely also. Water flows down standing seam much quicker than any other roof, your valley is possibly in a shady part, when the water gets there starts to freeze up again.
My theory is that if you put the heating the underside the ice will just form a bridge over the valley and just be melting whatever it can closest to the roof middle of the valley, Unless it's a very wide area  in the valley you're heating. 
Standing seam might have a high heat ice and water shield as underlayment (problem for you) , and a breather vent on the entire roof between the ice and water shield and metal roof. Breather vent is there so you don't get heat transferred from the metal to the attic in the summer instead it rises along the roof deck and vents through roof vents. If breather is installed it will probably condensate not melt much, cause more damage than good. 
Let us know how it turn out. I'm  curious because I haven't seen this system in 20 years of roofing, might work. 

nancyl

Will report back.  Thank you all for the thoughts and suggestions.

pecmsg

Think about it

What makes your roof and chimney so unique that only You have this issue?