kohta ueno joseph w. lstiburek, ph.d., p.eng....

© buildingscience.com

Monitoring of Two Unvented Roofs

with Air-Permeable Insulation

in Climate Zone 2A

Kohta Ueno

Joseph W. Lstiburek, Ph.D., P.Eng.

December 5, 2016

Monitoring of Two Unvented Roofs with Air-Permeable Insulation in CZ 2A 2


Unvented Roofs:

Background



Unvented/Cathedralized Roofs

Insulation at roof deck, rather than attic floor

Brings HVAC ductwork/equipment into conditioned space

Can improve airtightness (ceiling plane vs. roof)

Wind driven rain, hurricane roof tear-off

Moisture risks with unvented roofs (vs. vented)

Condensation of interior moisture at roof sheathing

Code-compliant (IRC§R806.4) roofs—“air impermeable

insulation” (spray foam/SPF, exterior insulation)

Current research: air permeable insulation, CZ 2A

Lower cost, environmental impacts of SPF

Houston and Orlando test houses

Can moisture risk be managed?



Why Unvented + Fibrous Risky?

Different than walls?

Moisture risks at sheathing

Interior-sourced air leakage

Vapor contributing too?

Zero-perm exterior (“wrong

side perfect vapor barrier”)

Night sky radiation cooling

Stack effect in winter



2000’s Cathedralized Roofs-TX & FL

Houston climate (CZ 2A)

had moisture at ridge

Concentrated only at

ridge—rest of roof OK

Similar problems in

Jacksonville FL (CZ 2A)

No interior air/vapor

control (not practical)

Possible solution:

allowing release of

moisture at ridge?



Test Site &

Instrumentation Setup



Unvented Roof Test Houses

Houston: asphalt shingle, #30 paper

Unoccupied model, humidified in winter 3 of 3

Orlando: concrete tile, self-adhered membrane

Occupied by family of 4, 2 winters of data

Houston Orlando



Unvented Roof Test Houses-Insulation

Houston: adhered fiberglass, R-38/RSI 6.7

Orlando: netted/blown fiberglass R-38/RSI 6.7

Vapor open insulation: 120 perm-in. or

170 ng/(Pa·s·m)); no interior vapor control layer

Houston Orlando



Diffusion Vent (Houston)

200+ perms diffusion vent

Roofsheathing

2-in. holes at 2-3 ft. of hip

Tape for air & water barrier continuity

Ridge slot opening

Vapor-permeableridge membrane

Ridge vent

Asphalt shingles

30# felt underlayment



Diffusion Vent (Orlando-Tile)

Tape for air & water barrier continuity

Ridge metal & cap tile

Vapor-permeableridge membrane

Ridge slot opening

Roofsheathing

Concrete roof tiles

Self-adhered roof membrane



RIDGE

RIDGE

RID

GE

RID

GE

RID

GE

VALLEY

HIP

HIP

HIP

HIPHIP

HIP

HIPH

IPVALLEY

VALLEY

VA

LLEY

VALLEY

Sensor Key:

Relative humidity/temperature

Moisture content/temperature

Ridge “package”

UNVENTED

UNVENTED

ATTIC T/RH W (HIGH/LOW)

DV1 UV1

UV2 DV2

DV3

DV4

DV5

ATTIC T/RH N (HIGH/LOW)

ATTIC T/RH S (HIGH/LOW)

Roof Instrumentation (Houston)

Unvented

(membrane)

control vs.

diffusion vent

experimental

Ridge packages

Hip packages

“Downhill”

sensors

Interior T/RH



Roof Instrumentation

Temperature, relative humidity

(RH), wood moisture content (MC)

“Wafer” sensors

Sensor Key:



Moisture content wafer

Ridge wafer & T/RH

Sheathing Top MC/T

Upper MC/T

Ridge top MC/T

Upper MC/T

2'-0"



Air Leakage Testing

Guarded testing:

some attics tight,

others noticeably

leaky

Depressurization

w. infrared

Air leakage

occurs at roof-

wall connections,

not field of roof



Results: Houston

(Asphalt Shingle)



Houston Attic Conditions (Summer)

Temperature tracks interior main space T

Warmer summer, cooler winter (ΔT 5-7 F/3-4 C)

Temperature high-low stratification

Dewpoint shows diurnal adsorption/desorption

from sheathing, dewpoint stratification



Houston Unvented Peak Conditions

Strong seasonal swing in moisture levels

Wet in winter, dries out in summer

Interior humidification→major effect, high risk

Winter 1

Construction

MoistureWinter 2

Unoccupied

model

Winter 3

Humidified



Houston Diffusion Vent Peak Conditions

MCs only risky in 3rd humidified winter, drier than UV

Ridge RH shows minimal seasonal swing →

coupled to exterior conditions

Humidified winter ~40-65% RH interior attic



Houston UVR vs. DVR Comparison

Diffusion vent wetter in summer, but well within

safe range

Wintertime diffusion vent prevents accumulation



0

20

40

60

80

100

UVR1 UVR2 DVR1 DVR2 DVR3 DVH4 DVH5 DVR6

Re

lati

ve H

um

idit

y (%

)

0

5

10

15

20

25

30

UVR1 UVR2 DVR1 DVR2 DVR3 DVH4 DVH5

Waf

er

Mo

istu

re C

on

ten

t (%

)

Houston Ridge RH and Wafer Box Plots

RHs all have peaks 95-100%; diffusion vent roofs

higher RHs in summer

Wafer sensor shows long-term accumulation

UVR1 & DVH5 wafer wettest outliers

Box = upper/lower quartile

Whiskers = max/min

Ridge RH Sensor Ridge Wafer Sensor



Results: Orlando

(Concrete Tile)



Orlando Attic Conditions (Summer)

Dewpoint diurnal cycling

Highest dewpoints consistently at “high” locations

Interior DP higher than exterior DP at peaks



Orlando Unvented Peak Conditions

Very high RHs and MCs first winter (construction

moisture); missing data after drywalling

Second winter—100% RH peaks, but sheathing

MCs not as concerning



Orlando Diffusion Vent Peak Conditions

Diffusion vent: lower MCs, RH peaks

“Smear” RH data at ridge vs. unvented



0

10

20

30

40

50

60

70

80

90

100

0

10

20

30

40

50

60

11/1/14 12/31/14 3/1/15 4/30/15 6/29/15 8/28/15 10/27/15 12/26/15 2/24/16 4/24/16

Rel

ativ

e H

um

idit

y (%

)

Mo

istu

re C

on

ten

t (%

)

UVR1 Peak Wafer DVR1 Peak WaferDVR1 RH UVR1 RH

Orlando UVR vs. DVR Comparison

Again, wafer and ridge RH patterns show safer

behavior with diffusion vent, less accumulation



0

20

40

60

80

100

UVR1 DVR1 DVR2 DVR3 DVH1 DVH2 RW1 RW2

Re

lati

ve H

um

idit

y (%

)

0

5

10

15

20

25

30

35

40

UVR1 DVR1 DVR2 DVR3 DVH1 DVH2 RW1 RW2

Waf

er

Mo

istu

re C

on

ten

t (%

)

Orlando Ridge RH and Wafer Box Plots

Unvented roof (UVR1) outlier, 100% RH; 35% MC

Diffusion vent roofs higher RHs in summer

Wafer sensor shows long-term accumulation

DVH2, RW1, RW2 wetter than other DVs

Box = upper/lower quartile

Whiskers = max/min

Ridge RH Sensor Ridge Wafer Sensor



Orlando Roof Shading Patterns

Higher RH areas in shade (DVH2, RW1, RW2)

DVH2

DVH1

RW1

RW2

DVR3

N



Decommissioning &

Disassembly



Unvented Roof Disassembly/DV Retrofit



Houston Hip Diffusion Vent

Monitored results showed low drying—closer to

unvented behavior vs. diffusion vent

Limited drying available at 2” hole “diffusion ports”

Covered up by asphalt shingles

Shingles cut back-retrofit



Orlando Roof Peak Conditions

Spotting on netting near ridge, ~4 ft./1.2 m down

Worst issues at unvented (not diffusion vent):

Spotting on truss vertical member/king post

Spotting on upper T/RH sensor jacket, not lower

Water stains on ceiling drywall under ridge

First winter construction moisture condensation



Orlando Attic Vertical Measurements

Handheld T/RH probe

Late afternoon June 2016

Temperature and dewpoint

differences w. height

-10

0

10

20

30

40

50

60

70

80

90

50 60 70 80 90

Hei

ght

(in

.)

Temperature/Dewpoint (F)

Temperature (F)

Dewpoint (F)

-10

0

10

20

30

40

50

60

70

80

90

50 60 70 80 90

Hei

ght

(in

.)


Temperature (F)

Dewpoint (F)

-10

0

10

20

30

40

50

60

70

80

90

50 60 70 80 90

Hei

ght

(in

.)


Temperature (F)

Dewpoint (F)

Upper Attic Upper Attic Lower Attic



Conclusions &

Recommendations



Houston/Orlando Results

Diffusion vent avoids wintertime ridge

accumulation problems (ridge peak RHs/MCs)

No failures at low interior RH, bigger difference at

higher RH (interior humidification)

Airtightness disappointing in some cases-no SPF

0

10

20

30

40

50

60

70

80

90

100

0

5

10

15

20

25

30

35

40

45

50

10/1/14 1/9/15 4/19/15 7/28/15 11/5/15 2/13/16

Re

lati

ve

Hu

mid

ity

(%)

Mo

istu

re C

on

ten

t (%

)

DVR1 Peak Wafer UVR1 Peak Wafer

DVR1 Peak RH UVR1 Peak RH

Unvented

Dif. Vent.

Dif. Vent.

Unvented

Summer WinterWinter



Questions?Kohta Ueno

kohta [at] buildingscience [dot] com

Project Partners/Sponsors: Cosella Dörken, David Weekley

Homes, DuPont Building Innovations, Johns Manville, and

Owens Corning



Document Resources Building Science Digest 149: Unvented Roof Assemblies for All Climates

http://buildingscience.com/documents/digests/bsd-149-unvented-roof-assemblies-for-all-climates

Building Science Insight 043: Don't Be Dense—Cellulose and Dense-Pack Insulation

http://buildingscience.com/documents/insights/bsi-043-dont-be-dense

Building Science Insight 088: Venting Vapor

http://buildingscience.com/documents/insights/bsi-088-venting-vapor

Building America Report 1511: Field Testing of an Unvented Roof with Fibrous Insulation, Tiles,

and Vapor Diffusion Venting

http://buildingscience.com/documents/building-america-reports/ba-1511-field-testing-unvented-

roof-fibrous-insulation-tiles-and

Building America Report 1409: Field Testing Unvented Roofs with Asphalt Shingles in Cold and

Hot-Humid Climates

http://buildingscience.com/documents/building-america-reports/ba-1409-field-testing-unvented-

roofs-asphalt-shingles-cold-and

Building America Report 1001: Moisture-Safe Unvented Wood Roof Systems

http://buildingscience.com/documents/bareports/ba-1001-moisture-safe-unvented-wood-roof-

systems/view

Building America Report 1308: Moisture Control for Dense-Packed Roof Assemblies in Cold

Climates: Final Measure Guideline

http://buildingscience.com/documents/bareports/ba-1308-moisture-control-dense-packed-roof-

assemblies-cold-climates/view



Reserve Slides



Ventilated Attics—Best Choice

Roof sheathing dries to

ventilated attic-moisture safe

Interior moisture (air leaks)

ventilated away in winter

Air sealing at ceiling critical for

best performance (e.g., spray

foam air barrier, detail with

sealant)



Then Why Unvented Roofs?

Living space built into roof

Vented cathedral assemblies—

often poor performance

Complicated rooflines, hip

geometries—how to vent?

Unworkable air barrier at

ceiling line

Blown-in rain (coastal)

Hurricane tear-off

HVAC in vented attic



Ducts in unconditioned attic = huge energy losses

Industry reluctant to move ducts out of attic

Ice dam issues due to duct losses

Solution: bring ducts into conditioned space

Unvented/conditioned attic—keeps ductwork in

conditioned space, duct leak issues eliminated

Unvented Roofs & HVAC Placement



Fibrous Insulation Unvented Roofs

Dense pack insulation of unvented roofs common

in cold-climate retrofits

Moisture risks (see BSI-043 “Don't Be Dense—

Cellulose and Dense-Pack Insulation”)—2 in 10 failure?

Violates I-codes (see IRC§R806.4)

“Ridge rot”—localized problems (SIPS same problem)



Spray Foam/Exterior Insulation Roofs

2006 IRC: R806.4 Unvented attic assemblies

Minimum R-value of “air impermeable insulation”

Not ratio of R-values… don’t get me started…

Nail base needed with rigid foam on roof deck



Why Fibrous Fill Unvented Roofs?

Unvented roofs without spray/board foams could

reduce costs and increase market penetration…

IF moisture damage risks are addressed

Retrofit opportunities (existing uninsulated living

space at roof line, without removing finishes)



Why Unvented + Fibrous Risky?

Risk reduced by:

Airtightness of ceiling

Dense insulations-less

airflow

Solar drive

But white roofs, shading

Lower interior RH (winter)

Why many of them work?

Lower permeance interior

Assumes good airtightness—

vapor retarder not bypassed

Moisture accumulation:

what gets in vs. gets out



0

10

20

30

40

50

60

70

80

90

100

0

5

10

15

20

25

30

35

40

45

50

2/10/16 2/11/16 2/12/16 2/13/16 2/14/16 2/15/16

Re

lati

ve

Hu

mid

ity

(%)

Mo

istu

re C

on

ten

t (%

)



Houston UVR vs. DVR Comparison

0

10

20

30

40

50

60

70

80

90

100

0

5

10

15

20

25

30

35

40

45

50

8/10/14 8/11/14 8/12/14 8/13/14 8/14/14 8/15/14

Re

lati

ve

Hu

mid

ity

(%)

Mo

istu

re C

on

ten

t (%

)



Summer (August) Detail-5 Days

Winter (February) Detail-5 Days



Results: Inward Vapor

Drive Experiment

(Houston)



Sensor Key:



Moisture content block

Inward Drive Experiment Setup

Sealed volume (2’/0.6 m square footprint), vapor

impermeable interior, airtight seal

T/RH and wafer sensors measure accumulation at

interior side (captures inward vapor drive)



Inward Drive Experiment Results

RH and wafer sensor on interior interface has

seasonal variations, but not hitting 100% RH

Pattern correlates to 24-hour average of

sheathing temperature

Moisture adsorbing/desorbing from sheathing

kohta ueno joseph w. lstiburek, ph.d., p.eng....

Documents