Specifying PAV Pan Levelness and Warpage

Dr. David Anderson, PE, Professor Emeritus Saman Barzegari, PhD Student, Civil Engineering

Pennsylvania Transportation InstitutePenn State University

Presented to

2018 Petersen Asphalt Research ConferenceLaramie, WY

July 15-18, 2018

Edited for Presentation to

AASHTO Select Committee on MaterialsAugust 7, 2018

Variation in Film Thickness and Properties of PAV Residue

Long-time issue

Brought to head within ASTM and AASHTO Task Force

Dave Anderson

Karl Zipf, Delaware DOT

Maria Knake, AASHTO-resources

Task Force Charge

Means for measuring and specifying allowable warping in PAV pans

Specification for thickness of pans

Developed into complicated and protracted problem

Measurement of pan warpage via a simple and economical means

Design of current pan support systems and influence on film thickness

Levelness of pan supports and means of measurement

Pan Support Systems

First we need to have an understanding of how the pans are supported inside the PAV vessel

Three systems

Three-point Edge Support SHRP – Never used

Center support – flat plate early use and still in use

Continuous Edge support – ring at periphery most common system

Each system is affects variability in film thickness differently

Must be evaluated separately

Need to have a measure of film thickness when pans are on actual supports and not tilted on flat surface!

May not be practical for everyday use but necessary to evaluate systems

Slide -3-

Center Support – Early System

Early Prentex System (?)

Small tilt in support has little effecton pan orientation

Advantage in required levelness

Sweet spot (center) is support point

Advantage in measuring profile and placement repeatability

Binder has to be poured in exact centerto avoid upsetting center of balance – once off balance will stay tilted

Center is not open so - cannot measure support levelness

Need to systematically disassemble from top down

Disadvantage

Slide -4-

Ring Support System

Most recent and popular system

Narrow ledge around periphery of pan supports pan

Disadvantage in that plan placement is not repeatable

Disadvantage with build up of residue

Center is open so levelness of each support can be verified while rack is on level surface

Advantage

Can be retrofitted with flat plate as support

Advantage

Slide -5-

Factors That Can Affect Film Thickness

Pan Warpage

Pans warp in random manner

Convex (upward) at center or (concave downward) at center

Points around rim are not colinear

Levelness of pan supports

Levelness of pan supports must be checked with pan supports (rack) in place with a flat plate in the top shelf of the rack

Point(s) where pan is supported

Center support on flat surface

Very repeatable

Edge support on circular rim that is partially open

Not repeatable, no sweet spot at center of pan

Narrow ledge means pan poorly supported

Slide -6-

Is There a Support System Difference ?

Preliminary data obtained at Penn State indicates “YES”

Need to confirm this result with more robust experiment

Error bars are for replicate measurements

Slide -7-

Observations on Pan Warpage

30 pans from 8 laboratories were evaluated for warpage

Warpage may be in convex or concave direction

Outside periphery of pans may not be coplanar

Conclusion – pans warp in random and unpredictable pattern

When supported at center or at periphery pans assume preferred position

Position can be repeated by placing pan on flat plate or ring at periphery

Referred to a preferred position – typically at center of pan

Preferred position is repeatable by simply pushing downward with index finger

Tilt of pans in preferred position is different for edge support than for center support

Slide -8-

Sample Pan Profile

Slide -9-

Another Sample Pan Profile

Slide -10-

Another Pan Profile

Slide -11-

How Can We Estimate Pan Warpage?

Spinning pan

Only tells us that pan is supported at center

Finger push – push down and measure gap on opposite side

Depends on radius of curvature of pan under finger

Beam clamp - clamp and measure gap on opposite side

Depends on radius of curvature at clamping point

Direct profile of bottom of pan

Need to replicate support mechanism – flat or ring

Difficult to do – not OK for use in a test method

Profile of casting of pan

Flip casting over – most accurate method but involved

Difficult to do – not OK for use in a test method

Slide -12-

Methods for Measuring Warpage

Many go-no-go methods

Pressing on edge

Straight edge

Spinning

Etc.

In each case these techniques do not give true picture of warpage

Profiling

By directly profiling pan surface

By profiling or measuring cast of pan

Good for investigative work but not foe specification use

Flatness - Check for Upward Bow

Invert and check for gap at center of pan

-14-

< 0.05 mm?

Straight edge

Neither of these procedures account for non-coplanar warpage at circumference

Flatness - Check for Downward Bow

Press on one side of the pan

Opposite side should not raise by more than 0.2 mm

Rotate pan 90° and repeat

Note: The test method does not require this procedure and its tolerances.

< 0.05mm?

Improved Measurement of Warpage

Slide -16-

“T” Square to locate positions at 90o

increments'

Clamped Beam For Concavity

Slide -17-

Jig for Verifying Pan Levelness

Slide -18-

Why the New Jig?

Easier to manipulate and make gap measurements than finger or squeeze clamp method

Clamps are awkward to use

Two modes may be checked

Coplanarity pf edges

Warpage

Both important for both flat support or for circular support racks

Third check is to place straight edge on bottom of pan and measure maximum gap

All measurements above should be checked at 15°increments

Allowable gap for any measurement will be small, <≈ 0.2 mm

Slide -19-

Comparison of Clamp and Finger Method

Table with finger, beam, center

Slide -20-

Gap Size (mm)Point 1 2 3 4

Pan1Bar and Clamp 0.28 0.28 0.18 0.25Finger Pressure 0.30 0.18 0.53 0.28

Pan2Bar and Clamp 0 0 0.06 0Finger Pressure 0.18 0.33 0.08 0.33

Pan3Bar and Clamp 0.48 0.39 1.51 1.69Finger Pressure 0.63 0.69 1.31 1.69

Pan4Bar and Clamp 0.3 0.3 0.23 0.23Finger Pressure 1.94 1.51 0.71 0.69

Pan5Bar and Clamp 0.04 0 0.63 0.63Finger Pressure 0.53 0.32 1.06 1.31

Pan6Bar and Clamp 0.92 0.88 0.96 0.94Finger Pressure 0.88 0.88 0.63 0.71

Gap

Gap

Gap Gap

Profile Gage (Prototype)

Moving gage laterally and rotating upper glass plate allows profile at any point on pan

May make casting of pan and profile casting by turning casting upside down

May simulate each of three support systems

Center, ring, and 3-point

Slide -21-

Why Is Profiling Needed?

Three possible warping modes:

1. Pan bottom may be concave

2. Pan bottom may be convex

3. Circumference may not be co-planar

None of the previous methods address warpage

Support method may also affect film thickness

With edge support angular location of pan will affect thickness if circumference is not coplanar

With shelf support pan assumes repeatable position as affected by gravity and possible warpage in shelf

Profiling should be done with support consistent with equipment in laboratory where pans are used

Each support system can be simulated

But the Pan Moves When the Measurement is Attempted!

Problem with support systems 1 and 2

Problem solved recognizing that pans are supported at unique point

Procedure for ring support or center support

Place small dot of fast curing gel-type epoxy under pan at periphery of simulated support

Allow epoxy to adhere to pan

Epoxy serves as a pedestal for placing pan in profiler

Slide -23-

Making a Casting

1. Level the support using a machinist’s level or equivalent

2. Line the inside of the vertical lip with modeling clay— Needed because some lips tilt to the inside

3. Place the pan on the support

4. Pour the desired mount of self-leveling casting resin onto the pan and allow to harden – typically 24 hr.

5. Remove the cast from the pan

6. Profile with jig or measure with micrometer

Slide -24-

What About PAV Rack Levelness?

Obscure requirements given in initial version of test method

Measured dimensions of rack

Unrealistic and never enforced

AASHTO and ASTM quiet on this issue

Issue has been discussed periodically

Varying rack design complicates measurement

Levelness of oven not reliable

Warping of vessel can affect rack levelness

Probably less critical than pan levelness

Should measure pans as they sit on their support

Easily done with edge support – small electronic level

How level?

Impossible with shelf support

Slide -25-

Levels – Choices

Inadequate – bubble gages commonly found in lab

<$15

Preferred – Machinist gage

Starrett 98-4

Has markings

≈ $90

Electronic Gage

Wixly Model WR365

Good experience with discontinues 320 Type 1

Avoid 320 Type 2 and Imported knockoffs

Slide -26-

Making Measurement

Place rack in vessel and heat to test temperature

Allows for vessel warpage

Index position of vessel

Place flat plate on top shelf

Measure levelness of top pan

Adjust PAV device cabinet to level pan

Forget about leveling screws!

Check rack position to determine if lateral movement affects levelness

Experience indicates YES

Solution?

Check at each loading

Slide -27-

Calculation of Film Thickness for Tilt

Slide -28-

0

1

2

3

4

5

6

7

8

0.5° 1° 1.5° 2°

Fil

m T

hic

kn

ess

(mm

)

Tilt Angle ( )

50 gr Binder

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

0.5° 1° 1.5° 2°

Fil

m T

hic

kn

ess

(mm

)

Tilt Angle ( )

25 gr Binder

0

0.5

1

1.5

2

2.5

3

0.5° 1° 1.5° 2°

Fil

m T

hic

kn

ess

(mm

)

Tilt Angle ( )

12.5 gr Binder

-0.5

0

0.5

1

1.5

2

0.5° 1° 1.5° 2°

Fil

m T

hic

kn

ess

(mm

)

Tilt Angle ( )

6.25 gr Binder

Rack Position

Position of the rack

Experience shows that moving the rack can affect the levelness significantly

Slight movement of rack gave change of 0.3 degrees

Equivalent to 1.3 mm at edge of pan

Conclusion

Need procedure for indexing rack position inside PAV vessel

Slide -29-

How Can We Estimate Effect of Levelness and Warpage?

Profile pan or casting

Use binder-specific equation for effect of film thickness on properties of PAV residue

Divide profile into segments and weight effect of each segment by profiled thickness

Slide -30-

Calculation of Methodology

Area Disance(mm)

1 0.00 532.22 0.00 58.61 58.61

2 22.50 532.22 22.43 54.15 58.61

3 45.00 532.22 41.44 41.44 58.61

4 67.50 532.22 54.15 22.43 58.61

5 90.00 532.22 58.61 0.00 58.61

6 112.50 532.22 54.15 -22.43 58.61

7 135.00 532.22 41.44 -41.44 58.61

8 157.50 532.22 22.43 -54.15 58.61

9 180.00 532.22 0.00 -58.61 58.61

10 202.50 532.22 -22.43 -54.15 58.61

11 225.00 532.22 -41.44 -41.44 58.61

12 247.50 532.22 -54.15 -22.43 58.61

13 270.00 532.22 -58.61 0.00 58.61

14 292.50 532.22 -54.14 22.43 58.61

15 315.00 532.22 -41.44 41.44 58.61

16 337.50 532.22 -22.43 54.15 58.61

17 0.00 638.66 0.00 35.29 35.29

18 45.00 638.66 24.96 24.96 35.29

19 90.00 638.66 35.29 0.00 35.29

20 135.00 638.66 24.96 -24.96 35.29

21 135.00 638.66 0.00 -35.29 35.29

22 225.00 638.66 -24.96 -24.96 35.29

23 270.00 638.66 -35.29 0.00 35.29

24 315.00 638.66 -24.96 24.96 35.29

25 0.00 425.78 0.00 13.98 13.98

26 90.00 425.78 13.98 0.00 13.98

27 180.00 425.78 0.00 -13.98 13.98

28 270.00 425.78 -13.98 0.00 13.98

Mid

dle

An

nu

lus

Inn

er

Cir

cle

Point IDAngle

(degrees)

Ou

ter

An

nu

lus

) ) ) )

Slide -31-

On-Going Inter-Laboratory Study

Profile six pans with different degrees and types of warping to 5 participating laboratories

RTFOT material supplied to each laboratory to minimize variability

Instructions for leveling PAV supports

Pan warpage measured in each laboratory using using finger and clamped beam

PAV runs in each laboratory with duplicate measurements

Conduct PAV aging with two support types

Edge support

Flat plate

Simulate flat plate by inserting metal plate in rack to support pan

Slide -32-

ANOVA – Pan and Support as Factors

Slide -33-

Conclusion: Support System is more important than warpage although it should be noted that the pan support is confounded with PAV run

Preliminary Conclusions

Effect of warpage less than originally assumed

Accept pans based on measurements with clamping jig 45o

increments around pan

Preferred method will depend on analysis of cooperative testing

Reject if clearance for any position is > 0.1mm (?)

Solution poor support from circular racks

Retrofit racks by adding flat steel plate to each shelf or circular referencing ring between pan and rack

Direct measure of levelness of top shelf using metal plate insert for circular support system

Above in process of being verified with on-going inter-laboratory experiment

Slide -34-

Preliminary Specification Requirements

Pans must be flat – end of story!

Use simple jig to check for edges being coplanar and not bowed upward or downward

Add flat plate, add indexing ring, redesign, or otherwise ensure proper seating on circular pan supports

Verify levelness of upper shelf of pan support

Top shelf must accommodate Starrett 98-4 level

Verify with level

Tighten other pan dimensions (in addition to flatness)

Increase thickness of pans

Specify outside diameter tolerances

Specify maximum radius at outside - bottom with go-no-go gage

Slide -35-

Acknowledgements

Saman Barzegari for his volunteered time

Nittany Engineers for Financial Support

Penn State for use of facilities

Maria Knake and Karl Zipf, ASTM and ETG Task Forces

Andrew Hanz (Mathy), Karl Zipf (DelDOT), Amir Golalipour (FHWA) and James Mahoney (Caplab) for testing

And to many others …………

Slide -36-