United StatesDepartment ofAgriculture

Forest Service

ForestProductsLaboratory

ResearchPaperFPL 387

December 1980

Development ofAn ImprovedHardboard-LumberPallet Design

Abstract

Low-grade wood residue from logg-ing and sawmilling operations couldbe used in the manufacture of hard-board for pallet construction, ifshown to be satisfactory for this pur-pose. Generally, this evaluation com-pared the performance of notchedstringer, partial 4-way entry oakpallets having 1-inch-thick medium-density hardboard decks with that ofsimilar red oak pallets underlaboratory and service-type condi-tions. The hardboard-lumber palletswithstood handling impact better;were more (or less) rigid in handling,depending on support conditions;racked much less because of corner-wise drop testing (but also failedsooner); and performed similarly,overall under indoor and outdoorhandling and storage conditions.Preliminary work indicated that stand-ard helically threaded pallet nails canbe used satisfactorily for hardboard-lumber joints for both exterior and in-terior exposure. This informationshould be pertinent to palletmanufacturers and users of pallets,especially to those employingmechanized handling systems.

Acknowledgement

The author wishes to acknowledgetechnical assistance provided in thiswork by FPL research staff membersJ. W. Koning, Jr., D. J. Fahey, G. C.Myers, and J. D. McNatt.

FPL 387 ADDENDUM

The information presented in this report accurately reflects the effects

of laboratory testing and outdoor handling and storage for hardboard-lumber

pallets and their all-lumber counterparts observed over a period of about

6 months. However, after an additional 9 months of outdoor storage, visual

observation indicated that the all-lumber control pallets deteriorated much

less with time than their hardboard-lumber counterparts. Generally, the

portions of the hardboard decks not impacted directly during handling and

shuttling were undamaged, though swollen progressively about l/8 inch. How-

ever, all of the portions of the panel near the contact points with the forks

exhibited increased shear, compression, and tension loading effects. In

general, it appears that prolonged outdoor exposure merely extended initial

damage. In view of the foregoing we now recommend that use of hardboard-

lumber pallets of the type evaluated be limited essentially to interior

applications.

United StatesDepartment ofAgriculture

Forest Service

ForestProductsLaboratory1

ResearchPaperFPL 387

Introduction

Development ofAn ImprovedHardboard-LumberPallet Design

BYROBERT K. STERN, Technologist

Logging and sawmilling wastecould be used for hardboard manufac-ture and subsequent pallet construc-tion, if the resulting pallets were com-petitive with existing lumber counter-parts. This would enhance the use ofwood residues. To investigate thispossibility Forest Products Labora-tory (FPL) personnel have conductedresearch on different aspects of theproblem. This report describes theresults of the latest work towardevaluating the use of hardboard com-ponents* for the construction of re-turnable, exterior-type pallets. Allcomparisons involving full-size palletsin this work were based on the perfor-mance of notched stringer, partialfour-way entry, flush type, nonreversi-ble 48- by 40-inch pallets, because oftheir widespread use in shipping to-day.

Previously, Superfesky and Lewishad defined the basic engineeringcharacteristics of three medium-density hardboards (10).3 Initial FPLwork with the use of hardboard aspallet component material byKurtenacker indicated that it might beused favorably for this purpose (4).Later, Stern reaffirmed that the per-

formance of block-type pallets withl-inch-thick medium-density hard-board decks compared favorably withsimilar pallets having nominal 1-inch-thick red oak lumber top decks (8). Aseparate study conducted at aboutthe same time evaluated the perform-ance of pallets of the same size (48by 40 in.) as in (8), but of the notched-stringer, partial four-way entry,reusable pallet style, and made withhardboard or insulation board decksof varied density with all other partsred oak. The performance of thesewere compared with that of all-lumbercounterparts. This work (9) indicatedthat notched stringer pallets withl-inch-thick hardboard top deckswould probably yield similar perform-ance as their red oak counterparts,except in resistance to diagonal im-pact distortion (produced by“free-fall” cornerwise drop testing).For this type of rough handling, thepallets with lumber decks racked onthe order of 10 times more than thetest pallets with hardboard decks, butthey also held together for manymore impacts.

One limitation in all of the previousFPL work involving use of hardboardas a pallet material was that urea for-maldehyde resin was used as abinder. This was necessary because

the panels were made for FPL byprivate companies according to theirmanufacturing procedure for commer-cial fiberboard. However, althoughfiberboard made with urea binder isadequate for indoor use and limitedoutdoor exposure, phenol-formaldehyde or phenol-resorcinolbinders were considered necessaryfor general outdoor exposure becauseof their superior water resistance.

MaterialsHardboard

Previous research (8,9) indicatedthat medium-density hardboard ofl-inch thickness probably would berequired to produce similar perfor-mance under all-weather service con-ditions, as nominal l-inch red oakboards for some of the more impor-tant strength requirements in palletconstruction. Because no sources ofhardboard of this type were available

1 Maintained at Madison. Wis., in cooperationwith the University of Wisconsin.

2 The term “hardboard” is used in this reportto denote panel material made from wood fiberbound together with waterproof phenol-formaldehyde adhesive.

3 Italicized numbers in parentheses refer toliterature cited at end of report.

at the time, it was decided to makepanels 1 inch thick by bondingtogether two 1/2-inch sheets of hard-board. The composition of the chipsused to make the hardboard wasabout 85 percent oak. The remaining15 percent consisted of sweetgum,elm, and traces of other species. Thewet-formed mats were made with 3percent phenol-formaldehyde binderand 1 percent wax sizing under a con-tract with the North Little Rock (Ark.)Division of Superwood Corporation.Prior to expeditious shipment to FPL,the mats were prepressed to removeexcess free water. Shortly after theirdelivery at the Laboratory, the wet-formed fiber mats were pressed againto a 1/2-inch thickness in a hot pressheated to 375° F. Next, the mats wereheat treated for 2 hours in an ovenheld at 300° F. Later, the mats werebonded together with phenol-resorcinol adhesive into panels 1 inchthick, allowed to set under pressure,and trimmed for use as pallet stock.The ovendry density of the panelswas 38.6 pounds per cubic foot.

LumberRed oak pallet parts used in this

work were cut from No. 3A and No.3B common lumber obtained locally.From the time of arrival at FPL thelumber was stored underwater untilmachining just prior to palletassembly.

NailsThe two types of nails used in this

work were the following:

Description

2-1/2-inch tempered, hardened,SSDP, helically threaded,diamond pointed. (Nailheaddiameter-about 1/4 inb)

2-1/2-inch “Grip-Tite roofing,galvanized. (Nailhead dia-meter-about 3/8 in.b)

Figure 1.–Pallet designs evaluated in this work. (A) lumber control pallet withspaced top deckboards and (B) experimental pallet with hardboard top deck.

(M 146 547)

Preliminary Research

ObjectivesThe first goal of this portion of the

work was to determine the details ofa fair but accelerated laboratory ex-posure program for the hardboard-lumber pallets to represent exteriorservice conditions. Thus, screening ofvarious arbitrary exposure conditionswas necessary for determining the ac-

SupplierMIBANT valuea

range average(degrees)

American Nail Co-op, 16-21 18Earth City, MO.

Wire Products Co.,Hortonville, Wis.

44-58 52

a Based on tests of 25 nails of each type (note: For greater detail on the MIBANT test apparatusand procedure refer to (6).

b The manufactured nailheads were not uniformly round for the approximate sizes given.

celerated laboratory testing scheduleto be used for the bulk of the testingincluded in this study.

A second objective concernedselection of a suitable nail design forjoining the hardboard decks to lumberstringers. The results of this portionof the preliminary investigation wouldapply to both laboratory and field-type testing of loaded pallets con-ducted later. In previous FPL (ac-celerated) research involving loading,such as occurs in service, the headsof standard helically threaded palletnails tended to pull through hard-board, instead of failing at the higherloads associated with axialwithdrawal of the nails from the str-ingers (8,9). Therefore, before con-structing the test pallets, it was alsoconsidered necessary to estimate theinfluence of nailhead size upon the“pullthrough” tendency under axialloading for similar hardboard-lumberjoints. The average area of the roofingnailheads, eventually selected forcomparison with standard pallet nailsin this work, was about 120 ± 10 per-cent greater than that of the palletnails.

ResultsThis work indicated that ac-

celerated cyclic exposure testing wasexcessive and that the outdoor ser-

Figure 2.–Apparatus for handling im-pact testing.

(M 148 390)

vice tests were necessary. Also, useof large head conventional roofingnails did not seem preferable overconventional helically threaded,hardened steel pallet nails. Details ofthe work supporting these conclu-sions are given in the Appendix.

Principal ExperimentalEvaluation

This part of the work involved testsof pallets in two phases begun atabout the same time: (A) under ac-celerated laboratory testing condi-tions, and (B) involving handling andstorage representative of service con-ditions. The laboratory tests includedmeasurement of handling impactresistance, bending stiffness, anddiagonal rigidity. Previously untestedpallets were stressed repetitively todestruction for each type of test.Service-type handling and storagetests were conducted with loadedpallets on FPL property, in order toassure complete accountability andaccessibility of the pallets for inspec-tion during and after testing.

Specimen Preparation andMoisture Conditioning

The test specimens consisted ofsixty 48- by 40-inch nonreversibleflush-type notched stringer pallets.Top decks of one group of 30 palletswere made from the 1-inch-thick hard-board panels,4 and the remaining por-tions were made from red oak lumber.The 30 comparative control palletswere similar, except that top decks

Figure 3.–Bending stiffness test of hardboard-lumber pallet with load being ap-plied at the quarter points of the 36-inch span.

(M 147 301-2)

consisted of nominal i-inch, uniform-ly separated boards as shown infigure 1. The moisture content (MC) ofthe lumber at the time of pallet con-struction was above 30 percent–thefiber saturation point-in all in-stances during nailing. The fastenersused were standard helically threadedpallet nails of the type describedearlier in this report. The newly con-structed pallets were stored in aheated building at ambient at-mospheric conditions for a minimumof 30 days prior to testing.

Laboratory TestingHandling impact.–The primary

reason for conducting tests of thistype was to measure the resistanceof typical pallets to this very commonform of rough handling encounteredin service. Specifically, the perfor-mance of five experimentalhardboard-lumber pallets was com-pared to that of five all-lumber controlpallets. Prior to each impact, theforks were tilted 4 degrees forwardand positioned so that the uppersides contacted the pallet 8 inchesfrom the right angles of the forks. Im-pact speeds of the forklift truckagainst the pallets averaged 2.0 miles

per hour and ranged from 1.84 to 2.14.Testing of each pallet was repeateduntil it would have been unfit for fur-ther service unless repaired. Testingof hardboard-lumber pallets washalted if “forkbite” (i.e., indentation ofthe leading edge of the hardboard topdeck) progressed to a depth of2 inches or more. All of the tests wereconducted with a conventionallyequipped forklift truck, i.e., without at-tachments such as the FPL “impactpanel” (7). A sketch of the test con-figuration is shown in figure 2.

Bending stiffness.–Tests of palletrigidity were considered to benecessary because of the danger,cost, and inconvenience caused byfailure on this account. Pallet bendingstiffness becomes most importantwhen the loaded pallet is (a) beingtransported by forklift truck (espe-cially over rough terrain), or (b) isstored in a rack without support, ex-cept near its ends-as happens in“drive-in” or “drive-in, drive-through”racks. Therefore, as is specified byASTM D1185-73 (2), each of the

4 A detailed description of the hardboard andlumber composition and processing are givenearlier in this report.

3

Figure 4.–Pallet orientation showingdiagonals during diagonal rigiditytest.

(M 148 460)

pallets was loaded from zero tofailure in bending at the quarterpoints of the 36- and 44-inch spans asshown in figure 3. The loading wasapplied at 0.2 inch per minute by auniversal testing machine, and deflec-tion at the midpoints of each spanwas sensed by the movable elementsof two linear variable differentialtransformers (LVDT). The LVDT’s weremounted on metal yokes suspendedfrom the ends of the two outer string-ers, and the resulting voltages werefed into an x-y recorder. Thus, sets of5 pallets per group and four groupsmade a total of 20 pallets tested.

Diagonal rigidity.–Racking ofpallets out of rectangular shape dur-ing use can produce costly timedelays-especially if they are beingused with an automatic palletizer of acontinuous conveyor or in computer-controlled automated systems. Ob-viously, it would also be costly torepair or replace them, if they becamedamaged excessively. Therefore,these tests of 10 pallets (i.e., 5 eachof experimental hardboard-lumberconstruction, and a second group ofall-lumber pallets) were conducted asa part of the general evaluation ofhardboard for pallet construction. Theobjective of these tests was to com-pare the diagonal rigidity ofhardboard-lumber pallets with that ofthe oak pallet controls by cornerwise,free-fall drop testing.

Each pallet was dropped from 40 ±1/8 inches above a 2-inch-thick steelplate embedded in the concrete floor

Figure 5.–Representative condition of pallets after failure from handling impacttesting: (A) Lumber control and (B) hardboard-lumber pallet.

(M 147 272-2)(M 147 269-8)

supported by bedrock. Prior torelease, the pallet was suspendedfrom such a position on the corner sothat the diagonal across the top deckof the pallet was perpendicular to thedropping surface. Individual palletswere dropped 12 times or until thepallet damage reached a point that itwould have been unserviceable‘unless repaired. Racking of eachpallet was monitored by changes in

the original length of diagonalsmeasured approximately 1.4 inches infrom either corner. Changes in thelength of pairs of diagonals on theadjacent top and bottom sides of thepallets, as indicated in figure 4, wereaveraged for each drop and used as ameasure of pallet rigidity. The testprocedure and data report compliedwith that specified in ASTM D1185-73(2).

Discussion of Results

Handling lmpact Tests

Figure 6.–Pallet failure patterns resulting from bending stiffness testing con-trols loaded (A) across 36- and(B) along 44-inch spans. Also, hardboard-lumberpallets loaded (C) across 36- and (D) along 44-inch spans.

(M 148 347-2)

Lumber control pallets.–As testingprogressed, the damage patternsobserved were quite similar to thoseencountered in (9). Specifically, all ofthe lumber control pallets failed by acombination of splitting and nailheadpullthrough of the top leadboard.Usually, splits were associated withunrelieved stress caused by the nails.The total number of impacts requiredfor pallet failure ranged from two tofour and averaged about three.Typical condition of a lumber controlpallet after testing is shown in figure5A.

Lumber pallets with hardboarddecks .-As expected, failure occurredin lumber pallets with hardboarddecks in every instance because of“forkbite”–i.e., indentation at thecontact points between the forklifttruck and the leading edge of thehardboard deck. When progressiveforkbite reached a cumulative total of2 inches in depth, further testing ofthat particular pallet was halted (fig.5B).

All of the nail joints remained in-tact in this test series.5 As shown bytable 1, lumber pallets with hardboarddecks withstood 2 to 13 impacts andaveraged about 6. Thus, for this typeof service-simulated testing the ex-perimental hardboard-lumber palletswere better than their all-lumbercounterparts.

Bending-stiffness tests.–Typicalfailure patterns obtained are il-lustrated in figure 6. The all-lumbercontrols in figures 6A and 6B repre-sent quarter-point loading across the36- and 44-inch span lengths, respec-tively. Excessive bending stress wasthe cause of failure of the bottomleadboard at its center, and thestringers failed mainly because oftension perpendicular to the grainalong with excessive shear stress atthe neutral axis. Figure 6C representshardboard deck failure at the quarterpoint in the 36-inch span, while figure6D illustrates failure caused by ten-sion and shear stress failure begin-ning at the notches. The left arrow in-dicates how a typical split has begunto open at the upper right portion ofthe left notch. Similarly, the presence

5 Note: Previous work reported in (8,9) had alsopredicted that a I-inch thickness of hardboardpanels was adequate to prevent inadequate jointstrength and the resulting nailhead pullthrough.

5

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

— — — — — Impacts — — — — —

Lumber (controls) 1 4, 2, 3, 2, 3 2-4 3

Table 1 .–Handling impact test results of loaded pallets

Handling impact dataPallet Pallet Individual

composition description pallets Range Nominalaverage

Lumber withhardboard decks 1 2, 4, 6, 4, 13 2-13 6

1 See “Materials” and “Specimen Preparation and Moisture Conditioning” of “Principal Experimen-tal Evaluation.”

In. — — — — — Lb — — — — — — — — — — — — — Lb/in. — — — — — — — —

Table 2.–Deta summary for bending stiffness test results

Maximum load Pallet stiffness

Pallet Unsupportedcomposition span Range Average Range Average Index for hardboard

pallets/controls- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

Lumber(controls)

136 11,750-13,620 12,750 7,184- 8,547 7,983

Lumber withhardboarddecks 136 5,950- 7,500 6,365 5,576- 6,608 6,040

Lumber(controls) 244 2,875- 4,700 3,990 10,101-10,989 10,480

0.753

1,380Lumber with

hardboarddecks 244 3,350- 9,400 6,420 11,976-17,143 14,465

1 Loaded across stringers2 Loaded along stringers.

Table 3.–Distortion resulting from diagonal rigidity testing

Number of cornerwise drops Average pallet distortionPallet Individual Group ∆ ∆ diagonals

composition pallets average A-B,E-F∆∆ diagonals

C-D,G-H- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

Pct In. PctIn.

Lumber (controls) 12, 12, 12, 12, 12 12 -1.36 -2.29 + 1.30 + 2.19

Hardboard deck, 6, 3,lumber parts otherwise

2, 2, 2(to failure)

-0.34 -0.57 + 0.11 + 0.18( to failure)

3

6

Figure 7.–Typical pallet condition after diagonal rigidity testing. (A) lumber con-trol and (B) hardboard-lumber pallet.

(M 147 357-2)(M 147 357-8)

of a knot and associated grain distor- The quantitative magnitude of thetion indicated by the arrow near the load and stiffness values involved areupper portion of the notch on the shown in table 2. In comparing theright has diverted grain separation performance of the two types oftemporarily, but then large scale grain pallets, it is obvious that the topseparation (horizontally in the string- deckboards of the lumber controlser) and pallet failure ensued. were functional when the pallets were

loaded as 36-inch spans-i.e., acrossthe stringers. This produced anaverage stiffness of 7,983 pounds perinch, i.e., about 32 percent higherthan the comparable value for thehardboard-lumber pallets. However,the average stiffness of hardboard-lumber pallets was about 28 percenthigher than the lumber control palletswhen the 44-inch span was involved.The cause of this reversal can bereadily understood when one con-siders the two types of pallet con-struction. For the 36- inch span, thecomposite stiffness of the boards ex-ceeded that of the hardboard deck.However, because the boards of thelumber pallet decks were separated,they did not function when loadedalong the 44-inch span, but the hard-board panel did. This inherent dif-ference between the two versions ofpallet construction might favor use ofall-lumber pallets, if they are to bemoved over rough terrain. Hardboard-lumber pallets would, however, havean advantage if they are moved overnormal pavement and are stored tem-porarily in racks without bridging be-tween the front and rear rails.

Diagonal rigidity tests.-As couldbe expected, damage and failure pat-terns for this type of testing repeatedthat experienced during earlier workinvolving notched stringer lumberpallets and experimental pallets madewith hardboard decks and lumberparts (9). Without exception, thelumber control pallets withstood theprescribed maximum of 12 drops from40 inches, but they also rackedseverely. In contrast, the lumberpallets with hardboard decks tendedto fail after an average of three drops,but they also racked much less up tothat stage. The magnitude of thecomparisons is shown in table 3. Therelatively good condition of thelumber leadboards after the required12 impacts is indicated by figure 7A.However, the same illustration alsoshows the moderately racked condi-tion of the lumber and “nailhead pull-through” that resulted from the12-drop series. The comparable condi-tion of a hardboard-lumber palletafter three impacts is shown in figure7B. As indicated, the top deck crush-ed much more (on one-fourth as manyimpacts), but also racked much lessthan the lumber pallets. This failuremechanism often included shearfailure at the notch, as shown infigure 7B.

7

Service-TypeHandling and Storage

TestingWhile the laboratory testing

schedule was an accelerated evalua-tion procedure based on logical testsof the pallets’ most importantstrength characteristics, a fair com-parison of the performance ofhardboard-lumber pallets requiredthat actual service environmental ex-

Started Completed

posure be conducted at about thesame time with a group of similarpallets. Therefore, to ensure 100 per-cent accountability of the palletsused-a factor essential to dataanalysis-the pallets were tested onFPL property only. Specifically, 20pallets (10 lumber controls and 10 ex-perimental hardboard-lumber pallets)were each loaded with a tare weightof 1,800 pounds, consisting of twosteel 55-gallon drums filled with sand,and given the following storage andhandling exposure:

Length ofexposure

Days

ShuttlingCycles

First Outdoor Storage and Handling

12-04-78 12-27-78 23* 3*

Indoor Storage

12-27-78 1-17-79 21 4 *

Second Outdoor Storage and Handling

The condition of the pallets wasmonitored periodically, except onweekends, during the regularlyscheduled 9 weeks of the programand after the additional 190-day staticoutdoor storage period that followed.Handling and shuttling were done inall instances with the Laboratory’s6,000-pound capacity LP gas operatedforklift truck shown in figure 8. Thespeed employed for handling wasabout 0.2 mile per hour and 0 to 10miles per hour for shuttling andtransfer.

First outdoor storage andhandling.– The 20 loaded pallets werestored outdoors for 23 days and shut-tled a total of three times during thisperiod. The first half of a single shut-tling cycle consisted of moving theloaded pallet about 600 feet byforklift, leveling the loaded pallet,placing it on the pavement, and thenwithdrawing the forks. The secondportion of the cycle involved reversalof the procedure. During transfer, themast, forks, and loaded pallets werealways tilted 4 degrees toward the

1-17-79 2-07-79 21 6rear of the forklift. Also, during lateraltransfer the forklift and its load

Additional Outdoor Storagereceived vertical jolts of variedmagnitude as the truck ran over

2-07-79 8-17-79 190 —cracks, expansion joints in the con-crete, and other irregularities.

At this time it was decided that the

*Some variation in the length of exposure and amount of handling occurredarbitrarily adopted handling condi-

unavoidably because of inclement weather and necessary forklift truck servic-tions were excessively severe and

ing during the testing.nonrepresentative of service handling.Accordingly, the dummy loads andsupports were repositioned on thepallets to allow subsequent handlingfrom the opposite, previously un-damaged ends and subsequent shut-tling for each trip was reduced to adistance of 100 feet.

Indoor storage.– Following the first23 days of outdoor storage with in-terim shuttling, the modified, loadedpallets were moved to an indoorstorage area. For this stage in theoverall program the 20 loaded palletswere divided into two equivalentgroups and stored in racks, or directlyon the concrete floor nearby. Theloaded pallets, in storage, in theracks were supported only by theirends. (Note: this is also typical in ser-vice in “drive-in” or “drive-in, drivethrough” rack storage systems.) Dur-ing indoor storage the loaded palletswere handled and shuttled four timesalong the concrete floor for a

Figure 8.–The FPL forklift truck in operation handling a loaded experimentaldistance comparable to that used dur-

hardboard-lumber pallet.ing the first and last parts of this por-tion of the work.

(M 147 215-8)

8

Followup outdoor storage.–As athird step, the loaded pallets werethen transferred to the originalstorage location and given similartreatment to that described for the“first outdoor storage and handling”period, except that the loaded palletswere stored for 21 days and handledand shuttled six times to a distanceof about 100 feet.

Following this last scheduledstorage and handling period theloaded pallets were left in static out-door storage for an additional 190days. They were then examined andfurther testing was ended.

ResultsThe thickness of hardboard decks

of hardboard-lumber pallets swelled amaximum of about 2 percent duringthe entire exposure. Much of this oc-curred during the first 2 weeks. Also,there was no problem evident fromnailhead pullthrough.

Another observation concerned theserviceability of hardboard-lumberpallets when compared with their all-lumber counterparts after months ofexposure and handling. In general,the two groups of pallets performedsimilarly, but there were differences.A brief summary of deleterious ef-fects of service on both types ofpallets in the rack storage portion ofthe indoor tests is given in the follow-ing:

I.

II.

III.

IV.

Phase

Initial 23-day outdoorstorage and handlingperiod.

Indoor storage andhandling for 21 days.

Second outdoorstorage and handlingfor 21 days.

Storage for anadditional 190 days.

Although this comparative studywas ended after a total of 8-1/2months-practically all of which in-volved outdoor exposure, roughlyequivalent pallet resistance to handl-ing and atmospheric exposure wasobserved for both types of pallets.The fact that the hardboard panelsdid not soften markedly with extend-ed exposure to moisture probably waslinked directly to the high waterresistance of the phenol-formaldehyde binder.

As might be expected, damagesuch as is shown in figure 10C didnot occur in any of the loaded palletsthat were stored directly on the floorduring indoor testing. The major dif-ference between this type storageand storage in racks was that themiddle portion of the bottoms of theloaded pallets was supported in floorstorage, but not in rack mounting.

It became quite clear as the workprogressed that the transfer of loadedpallets between buildings, i.e., long-distance shuttling, produced much ofthe damage to loaded pallet ends.This probably was associated withtilting of the forks, and pavement ir-regularities.

In general, allowable cross-grain (5)in stringers and deckboards did notaffect the results. Similarly, theaverage unloaded pallet weights of65.6 and 83.2 pounds for all-oak andhardboard-oak pallets, respectively (adifference of about 27 pct), did notappreciably affect the results of thiswork.

Principal Changes

Shear failures of the top deck leading edge ofsome of the lumber controls and tension failuresin the hardboard-lumber pallets (fig. 9). Also,checks and splits tended to enlarge in all lumbercomponents.

Leading edges of some leadboards tended tolift and split (figs. 10A, 10B). A product of thisstorage mode: split, beginning at the notch,caused by bending stress of pallet in storage(fig. 10C).

A minimal progression of damage describedunder Phase I.

Practically no change in the density of palletcomponents or in damage produced prior tothe beginning of this period.

Figure 9.–Contrasting conditions of pallets during the initial 23-day storage andhandling exposure: (A) lumber pallet (control) in good condition, (B) damagefrom vertical shear, (C) relative/y unaffected hardboard deck, and (0) verticalshear effect upon its counterpart.

(M 147 216-6)

10

Conclusions

Based on this and the referencedwork in (8,9), the following conclu-sions can be made about the use ofhardboard of the type evaluated forstandard pallet construction:

1. In many types of comparativeevaluations, hardboard-oak pallets ofthe size and type evaluated will per-form about as well or better than redoak returnable pallets of similar sizeand style-especially when used withmechanical handling systems.

2. This use represents a goodoutlet for limbs and sawmill waste,such as short lengths and miscut ortwisted lumber. As the price of all-lumber pallets continues to increasethe use of hardboard constructionbecomes more attractive economical-ly.

3. Hardboard-oak pallets of thissize and style rack less, but failsooner, than similar red oak lumberpallets under cornerwise impacttesting. Therefore, they might bemore suitable for automatic pallethandling operation, but less suitableif considerable cornerwise droppingis expected.

4. Use of phenol-formaldehydebinder in hardboard panel manufac-ture can achieve similar all-weatherperformance capability as red oakdeckboards of pallets of the samesize and style, but this practice mightalso increase pallet cost above thatfor lumber pallets.

5. Nailhead pullthrough probablywill not be a serious problem in theuse of medium-density hardboard forpallet construction if phenol-formaldehyde is used as a binder.

Figure 10.–Damage to pallets being stored in a rack during indoor storage: (A) ageneral view, (B) split endboard (probably caused by transfer over an irregularpavement), and (C) tension perpendicular to the grain failure in stringer duringrack storage.

(M 147 241-11)(M 147 241-8)(M 147 241-3)

11

Literature Cited

1. American Society for Testing and Materials.1976. Standard methods of evaluating the properties of wood-base fiberand particle panel materials. ASTM D1037-72, Part 22. Philadelphia, Pa.

2. American Society for Testing and Materials.1973. Standard methods of testing pallets. ASTM D1185-73, Part 20.Philadelphia, Pa.

3. Kurtenacker, R. S.1965. Performance of container fasteners subjected to static and dynamicwithdrawal. USDA For. Serv. Res. Pap. FPL 29. For. Prod. Lab., Madison,Wis. June.

4. Kurtenacker, R. S.1976. Wood-base panel products for pallet decks. USDA For. Serv. Res. Pap.FPL 273. For. Prod. Lab., Madison, Wis.

5. National Wooden Pallet and Container Association.1962. Specifications and grades for hardwood warehouse, permanent, orreturnable pallets.

6. Stern, E. G.1971. The MIBANT quality-control tool for nails. VPI and State Univ. Bull.No. 100.

7. Stern, R. K.1975. Increasing resistance of wood pallets to handling impacts. USDA ForServ. Res. Pap. FPL 258. For. Prod. Lab., Madison, Wis.

8. Stern, R. K.1979. Performance of medium-density hardboard in pallets. USDA For. Serv.Res. Pap. FPL 335. For. Prod. Lab., Madison, Wis.

9. Stern, R. K.1979. Performance of pallets with hardboard decks of varied density. USDAFor. Serv. Res. Pap. FPL 340. For. Prod. Lab., Madison, Wis.

10. Superfesky, M. J., and W. C. Lewis.1974. Basic properties of three medium-density hardboards. USDA For.Serv. Res. Pap. FPL 238. For. Prod. Lab., Madison, Wis.

12

U. S. GOVERNMENT PRINTING OFFICE: 1980-750-027129

3 . 0 - 1 2 - 8 0

APPENDIX-PRELIMINARY

RESEARCH DETAILS

TestingExposure followed by tensile

testing of joints. –Ten nail jointswere constructed by nailing togethersingle 4- by 2- by 1-5/8-inch pieces ofhardboard to 6- by 3- by 1-5/8-inchpieces of red oak. A single standard2-1/2 inch, helically threaded, hard-ened steel pallet nail was used tofasten each piece of hardboard andlumber together to make thesimulated joint. A similar set of 10specimens was also made entirelyfrom red oak lumber, except that3/4-inch (nominal 1 in.) red oak wasused instead of 1-inch hardboard.

Following construction, the twosets were each given one exposurecycle of the six specified in the “Ac-celerated Aging Cycle” of (1).Specifically, this involved the follow-ing steps in the order shown:

1. Immersion in water at 120° ±3° F for 1 hour.

2. Spraying with steam and watervapor at 200° ± 5° F for 3 hours.

3. Storage at 10° ± 5° F for 20hours.

4. Heating at 210° ± 3° F in dryair for 3 hours.

5. Spraying again with steam andwater vapor at 200° ± 5° F for 3hours.

6. Heating in dry air at 210° ± 3° Ffor 18 hours.

Following accelerated aging ex-posure, the simulated pallet jointswere conditioned for at least 48 hoursin an atmosphere of 68° ± 6° F anda relative humidity (RH) of 65 ± 1 per-cent. At this time the two sets ofspecimens were given axial tensiontests at a loading rate of 17.5 inchesper second. The testing equipmentconsisted of a servo-controlled elec-trohydraulic testing system. Thedeflection transducer output wasdigitized by a high speed digitalstorage oscilloscope and transferredsubsequently to an x-y plotter.

Results of aging and subsequenttensile tests .–As accelerated agingtesting progressed, a large degree ofswelling occurred in the hardboard.This swelling suggested that use of amilder exposure schedule might bemore realistic and advisable to pre-

Figure 11.–Two failure mechanisms produced by nail withdrawal testing: (A) nailwithdrawal and (B) nailhead “pullthrough.”

(M 147 149-11)(M 148 120)

vent premature joint failure. Thissuspicion was reinforced when (un-typically) 7 of 10 all-lumberspecimens failed during the tensiletest by nailhead pullthrough.

Figure 11 shows the two failuremechanisms resulting from thesetests. Part A illustrates the result of aconventional nail withdrawal se-quence wherein the nailhead essen-tially remained in its original positionand the shank pulled away from thestringer section. In contrast, nailheadpullthrough obtained with the sametype of joint is illustrated in Part B.For this pattern the withdrawal forceprobably reached a maximum whenthe fiber directly below the nailhead(encircled) failed during the initialstage of failure. As the nailhead waspulled through the hardboard, thelower portion pulled free and thenfailed in bending.

Quantitatively, an average max-imum force of 514 pounds was re-quired for the hardboard-lumber jointsand 577 pounds for the all-lumberjoints failure by nailhead pullthrough.The remaining three all-lumber jointsrequired an average maximum forceof 683 pounds, and the average forthe entire set of 10 all-lumber jointswas 618 pounds. The results obtained

from this set and other combinationstried during this preliminary testingphase are given in table 4.

In view of the foregoing perfor-mance and because all-lumber palletsseldom fail in use by nailheadpullthrough, it was decided to con-tinue searching for a more practicalpreliminary exposure procedure.

Twenty-four-hour soaking followedby tensile testing.–In the second por-tion of the preliminary research, 20nail joints were given nail withdrawaltests immediately after a 24-hoursoaking in water. This rigorous en-vironment was selected because ofthe likelihood that the truesignificance of the use of hardboardin construction of outdoor-typepallets ultimately rests with itsresistance to moisture. The principaldifferences between this and the firstset of tests were that (a) a 24-hourwater immersion replaced the one cy-cle of the D1037 accelerated-agingprogram, (b) the large head galvanizedroofing nails described earlier wereincluded, and (c) the replication wasreduced from 10 to 5 in the fourgroups tested. A listing of the prin-cipal combinations of variables, aswell as the results obtained for thisseries, are given in table 4.

Table 4.–Tensile test details for simulated deckboard stringer joints following accelerated aging

Test specimens1Nail Type of failure Maximum force2

Composition Nail type joints/ Nailhead Shankgroup pullthrough withdrawal Range Average

----------------------------------------------------------------------------------------------------------------------------------------------

— — — — Lb — — — —

AFTER 1 ASTM D1037 CYCLE (1)

All oak Helically threaded 10 7 3 500-770 618

Hardboard-oak Helically threaded 10 10 0 400-695 514

AFTER 24-HOUR WATER IMMERSION

All oak Helically threaded 5 0 5 295-672 478

All oak Roofing 5 0 5 495-925 647

Hardboard-oak Helically threaded 5 0 5 435-507 461

Hardboard-oak Roofing 5 0 5 434-597 509

NONE

All oak Helically threaded 5 0 5 452-667 576

All oak Roofing 5 0 5 405-775 585

Hardboard-oak Helically threaded 5 3 2 443-621 486

Hardboard-oak Roofing 5 0 5 394-585 490

1 Each held by a single nail joint.2 Based on the approximate maximum force exerted by individual nails during the tension tests.

As indicated, all joints in thisseries failed by withdrawal of theshanks from the oak stringer sec-tions, and the averages of the peakwithdrawal loads for each kind of nailwere similar. The one exception in-volved joints of oak deckboards andstringer sections nailed together withlarge head roofing nails. Quantitative-ly, the average withdrawal force forthis combination was 647 pounds andranged from 495 to 925 pounds. Incontrast, the mean peak withdrawalforces for hardboard-oak joints were461 and 509 for the conventionalpallet nails and roofing nails, respec-tively. The maximum withdrawal

forces for the conventional helicallythreaded pallet nails ranged from 435to 507 pounds, while the correspond-ing range for roofing nails was 434 to597 pounds. However, it appeared tobe inadvisable to use large head roof-ing nails instead of standard helicallythreaded pallet nails because of thesharp loss of initial withdrawalresistance likely in service. Thisphenomenon is clearly shown byKurtenacker in (3), figure 7.

Nail joints of green memberswithout aging.–This, the last portionof the preliminary testing, was con-ducted without accelerated aging.

Results of tests of green memberjoints.– Considered collectively, themaximum tensile load sustained bythe all-oak controls averaged about19 percent higher than comparablesets of hardboard-lumber joints.However, the average for the max-imum force required to fail hardboard-lumber joints was similar, whether ornot the joints were previously im-mersed for 24 hours prior to tensilewithdrawal testing.

The failure patterns, averages, andranges of tensile withdrawal forcesustained by each combination aregiven in table 4.