THE EFFECT OF SELECTED MECHANICAL

THRESHING PARAMETERS ON KERNEL DAMAGE

AND THRESHABILITY OF WHEAT

F. M. VasJunior Member CSAE

Department of Agricultural EngineeringUniversity of Alberta

Edmonton, Alberta

byH. P. Harrison

Member CSAE

Department of Agricultural EngineeringUniversity of Alberta

Edmonton, Alberta

INTRODUCTION

The Saskatchewan AgriculturalMachinery Administration in its testwork with combines, noted threshingdifficulties particularly with the 1963harvest. In one report from that year(17) the follownig was observed: "Inmany wheat crops, it was not possibleto adjust the cylinder of the Cock-shutt 431 so that the cylinder losswas acceptable, while keeping cracksand white caps at a reasonable level."In another report (MF 300) for thesame year: "Best adjustment thengave 2 percent crackage with a cylinder loss of about 1 kernel per head.Note that most of the cracks are lostover the shoe with the chaff." Withregard to the last comment, it isnecessary to point out that the testgroup assessed crackage on the basisof the quantity delivered to the hopper with the clean grain and did notconsider the "cracks" lost over theshoe.

Investigations into the effects ofmechanical threshing parameters onkernel damage and threshability ofsmall grains appear in the literatureas early as 1934, when Bainer andBorthwick (6) studied mechanical injury to seed beans. Since then, otherstudies (2, 3, 10, 14 and 15) havebeen carried out. These investigations indicate that cylinder speed isthe primary influencing parameterwhile concave clearance, although oflesser significance, is an importantfactor as well. Because these studieswere far removed from the climaticconditions of Western Canada, thepossible varietal differences and thedifficulties noted by Saskatchewan'stest group; a project with the following objectives was initiated.

1. Ascertain the importance of cylinder speed, concave clearanceand feed rate with respect tokernel damage and threshability.

2. Determine if an optimum levelof cylinder speed and concaveclearance exists for minimumgrain loss as suggested by Saskatchewan's test group.

3. Explore the possibility thatthreshing is an impact and/orfrictional process.

Figure I. The stationary laboratorythreshing unit used in this study.

EQUIPMENT ANDEXPERIMENTAL PROCEDURE

The stationary threshing unit usedappears in Figure 1 and 2. The cylinder was a standard commercial rasp-bar type being 24 inches wide and21 inches in diameter. It was usedwith the appropriate make of concave. The threshing unit was fedfrom a 50 foot canvas conveyor. Thecylinder was driven by an IHC 706diesel tractor. The conveyor andfeeder chain were driven by a 3 hpelectric motor, while the straw-walkers and the pan-under-walkerswere driven by a 1/4 hp electricmotor. The threshing unit lackedcleaning facilities such as sieves andfans, thus the total threshed sample(including grain, straw and chaff)was cleaned using a separate system.This system consisted of a series ofthree screens. The material firstpassed over a 3/8 by 7/8 inch expanded metal screen which removedstraw as well as straw with un-threshed heads from the sample. A14/64 inch round hole screen removed portions of the chaff and forfinal cleaning, a "Clipper" cleanerwas used. The system was partiallydictated by the necessity to avoidpneumatic separation until the laststep in order to minimize losses ofcracks and chips.

CYLINDER rFINGER GRATE STRAW7 WALKER

GRAIN ELEVATOR

FEEDCONVEYOR

COLLECTION BOX

Figure 2. The components of the laboratory threshing unit.

CANADIAN AGRICULTURAL ENGINEERING, VOL. 11, No. 2, NOVEMBER 1969 83

The threshing material usedthroughout the project was the hardred spring wheat variety, Park. Theaverage grain-to-straw ratio was 1.00/1.71 and the average grain moisturecontent was 10.4% (wet basis). Thewheat was taken from a field in theLeduc, Alberta area in sheaves andstored at the Ellerslie farm of theuniversity until used.

The wheat sheaves were selectedfrom storage at random and broughtinto the laboratory where they wereplaced on conditioning racks. Thesheaves were left on the racks forfive days before being used, in orderto allow them to reach a moistureequilibrium. One hundred pounds ofthe conditioned material was weighedand placed on the 50 foot conveyorwith the heads up and in line (8, 9).A combination of cylinder speed,concave clearance and feed rate wasselected. Cylinder speed was considered at 5 levels: 800, 900, 1000,1100 and 1200 rpm; cone&ve clearance was considered at three levels:1/4, 1/2 and 3/4 inch and feed ratewas considered at three levels: 100,200 and 300 lbs/min. This resultedin 45 different treatment combinations from which one was chosenusing a random numbers table. Theselected parameter values were seton the threshing unit and the materialthen run through. The catch wascleaned and weighed. The portion ofthe material containing unthreshedgrain was run through the unit againto give the unthreshed portion. Therethresh material was cleaned andweighed.

Samples were taken from the original free grain catch using a sampledivider. One hundred gram sampleswere used for grain moisture contentdeterminations using whole kerneloven-dry methods (11). A 100 gramsample was taken for visual damageanalysis (1, 4, 7).

ANALYSIS AND RESULTS

The data collected was analyzedon the basis of a randomized simplefactorial experiment with four replkcations of all treatment combinations.Definitions of the dependent and independent variables (parameters)considered in this study were:

YT — Threshability — ratio ofthe wheat kernels (byweight) removed fromthe ear during initial

84

treatment to the total expressed as a percent.

YD = Kernel Pamage —ratio ofthe wheat kernels (byweight), that show mechanical damage as determined by visual examination, including anybroken, cracked or chipped kernels; to the totalexpressed as a percent.

X- — Cylinder Speed — -speedof the cylinder measuredin rpm.

Xc — Concave Clearance —thedistance between the cylinder beater bar and thefront concave bar measured in inches.

XF — Feed Rate - the totalmaterial including grain,straw and chaff whichpassed through thethreshing uait measured.in lb/min.

XSXC, X$Sf and XqX? are the

two factor interactions.

Analysis of VarianceThe results (Table I) indicated

that cylinder speed and concaveclearance had a highly significanteffect on threshability. The interac

tion of cylinder speed and concaveclearance, which was also highly significant, indicated the differential response of threshability to cylinderspeed depending on the concaveclearance. Mean threshability in-

<ILI

2

V _l_

900 1000 1100 120(

CYLINDER SPEED - RPM

Effect of cylinder speed onthreshability.

_±_ _L

Figure 3.

<x

z<

th1/4 1/2 3/4

CONCAVE CLEARANCE - IN.

Figure 4. Effect of concave clearanceon threshability.

TABLE I. ANALYSIS OF VARIANCE (THRESHABILITY - - YT)

Source of Degree of Mean1Variation Freedom Squares F

S = Cylinder Speed 4 22.030 158.5**

C = Concave Clearance 2 11.570 83.2**

SC 8 1.270 9. 1**

F = Feed Rate 2 0.012 <1

SF 8 0.189 1.4

CF 4 0.085 <1

SCF 16 0.038 <1

Error 135 0.139

Total 179

* Significant at .05 level.** Significant at .01 level.1 Three decimal places were retained for calculation purposes

only.

CANADIAN AGRICULTURAL ENGINEERING, VOL. 11, No. 2, NOVEMBER 1969

creased with increasing cylinderspeed (Figure 3), but decreased withincreasing concave clearance (Figure4).

Variations in mechanical damagebetween cylinder speeds, concaveclearances and feed rates were statistically significant (Table II). Thesignificance of the two factor interaction terms indicated that the factors did not act independently andthat the influence of a factor ondamage depended on the level of theother factors. Mean mechanical damage increased with increasing cylinder speeds (Figure 5), but decreasedwith increasing concave clearance(Figure 6) and increasing feed rate(Figure 7).

Multiple Regression

The general model considered forthe multiple regression had the following form:

Y - A0 + A,XS + A2X2S +

A3XC + A4XF + A5XSXC +

+ A6XSXF + A7XCXF 1

Y — dependent variable (threshability or mechanical damage)

A0 = constant

A1?A2 . . . A7 = multiple par

tial regression coefficients.

A computer program for stepwisemultiple regression (13) was used todetermine the regression equations.All of the independent variablesgiven in the general model weremade available to the program, butonly those which reduced the sum ofsquares by one percent or more wereretained in the final multiple regression equations. The results of theregression analysis for threshabilityand mechanical damage are given inTable III and IV respectively.

The analysis of variance for theregression in both cases, indicated ahighly significant reduction in sumof squares due to the regression. Themultiple correlation coefficients of0.91 and 0.98 indicated a good relationship between the dependentand the independent variables selected. The regression equations wereconsidered to be valid within therange of values used in their determination. A necessary limitation of

the equations would be that they are influencing threshability and kernelonly valid when crop variables are damage (3).similar to those used in this experiment, since moisture content, grain- Minimization of Grain Lossto-straw ratio, and varietal differ- In order to determine if some levelences are important crop parameters of the threshing parameters consid-

TABLE II. ANALYSIS OF VARIANCE (MECHANICAL DAMAGE - YD)

Source of

VariationDegree of

Freedom

Mean1Squares F

S = Cylinder Speed 4 515.950 1943.2**

C = Concave Clearance 2 76.870 290.0**

SC 8 2.827 10.7**

F = Feed Rate 2 16.969 64.0**

SF 8 1.692 6.4**

CF 4 0.668 2. 5*

SCF 16 0.312 1.2

Error 135 0.265

Total 179

TABLE III. REGRESSION ANALYSIS FOR THRESHABILITY

Analysis 'Of Variance For The RegressionSource of

Variation

Degree ofFreedom

Mean1

Squares F

Attributable to

Regression 4 29.393 204.1**

Deviation from

Regression 175 0.144

Total 179

TABLE IV. REGRESSION ANALYSIS FOR MECHANICAL DAMAGE

Analysis Of Variance For The Regression

Source Of

Variation

Degree ofFreedom

Mean1Squares F

Attributable to

Regression 3 737.083 1151*7**

Deviation from

Regression 176 0.640

Total 179

VOL. 11, No. 2, NOVEMBER 1969 85CANADIAN AGRICULTURAL ENGINEERING, VOL. 11, No

14

13

12

11

i 10

s

2«

3'1*

Z 4<UJ

* 3

2

1

H^ 900 1000 1100

CYLINDER SPEED - RPM

Figure 5. Effect of cylinder speed onmechanical damage.

ered would result in an optimumsituation of high threshability andlow kernel damage, a measure oftotal grain loss was defined as follows:

YL^ (100 — YT) +where YL — total grain loss (%),

100 - Y, unthreshed

grain (%) and

YD — damaged grain (%).

The regression equation obtainedfor threshability is:

YT — 87.9010000 + .021689XS -

0.000010X2S - 8.549019XC +0.006806XSXC 3

Multiple Correlation Coefficient= 0.91.

Cumulative Proportion of Sumof Squares Reduced =- 0.82.

Standard Error of Estimate —0.38%.

The regression equation obtainedfor damage is:

YD =* 36.200000 - 0.082244XS +

0.000054X| - 0.004576XSXC ....4

Multiple Correlation Coefficient— 0.98.

Cumulative Proportion of Sum ofSquares Reduced — 0.95.

86

<

z ,<Ioj 6UO

?S 5

z2< 4

1/4 1/2

CONCAVE CLEARANCE - IN.

Figure 6. Effect of concave clearanceon mechanical damage.

<y* 7 -z , r<

SisZQ

< A

100

FEED RATE

200

LBS/MIN

Figure 7. Effect of feed rate onmechanical damage.

Standard Error of Estimate =0.81%.

Thus, all unthreshed grain and damaged kernels were considered to contribute to grain loss in this definition.Small kernel fragments for example,are likely to be lost due to pneumaticseparation (5) which is direct loss.Cracked kernels may result in dockage when wheat is sold for millingpurposes and can be considered as aloss too. Damage may also result inpoor germination if the grain is usedfor seed. Thus, any damage is undesirable and can be considered as aloss. Unthreshed heads which are retained for rethreshing often result indamaged kernels and can also be considered as a loss.

The regression equation for totalloss (YL) in terms of the independent

variables becomes:

YL — 48.299000 - 0.103933XS +

0.000064X2S + 8.549019XC -0.011382XSXC 5

To determine the minimum of YL,

methods using partial derivatives(16) were used. The resulting solution indicated that minimum grainloss would occur at a cylinder speedof 751 rpm and .69 inch concaveclearance. In this investigation therange of cylinder speeds was 800 to1200 rpm. Since extrapolation of theregression equation beyond the values

considered in the experiment is invalid, the solution is in question.However, the procedure is useful inthat it indicates future experimentsshould consider lower cylinder speedsif minimization of grain loss is to beconsidered.

DISCUSSION

Threshability and Grain Damage

High cylinder speeds and smallconcave clearances resulted in increased threshability and increasedmechanical damage (Figures 3, 4, 5and 6). According to Arnold (3), anear of grain is subject to large impulsive forces at high cylinder speeds.Decreasing concave clearance mayhave:

1. Increased the chance of an earof grain being struck by the cylinder beater bar and,

2. Increased the chance of multipleimpacts to the ear before itpassed from the threshing zone.

It is convenient to describe the aboveas an "impact" process or model.Thus, the impact model may accountfor the increased threshability andkernel damage with increased cylinder speed and decreased concaveclearance.

An incresae in feed rate resultedin a decrease in mechanical damage(Figure 7). This might be due to a"cushioning" (5) effect at the higherfeed rates. That is, the crop streambetween the cylinder and concavemay have been denser at the higherfeed rates, thus providing a cushionfor the kernels. TTie cushioning effectwould fit the impact model. On theother hand, changes in the feed ratehad no significant effect on threshability (Table I). The cushioningeffect which may have decreaseddamage at the high feed rates, shouldalso have decreased threshability.Since this was not supported by theresults, a "frictionar model may beindicated. That is, increased feedrate may have increased the cropstream density which in turn increased the frictional forces betweenparticles in the crop stream. Thus,the cushioning effect of the impactmodel may have tended to decreasethreshability but was offset by theincreased rubbing action of the frictional model. The net result could belittle or no change in threshability.

CANADIAN AGRICULTURAL ENGINEERING, VOL. 11, No. 2, NOVEMBER 1969

The rubbing effect of the frictionalmodel would not likely result in damaged kernels (12). In other words,the effects of cylinder speed, concaveclearance and feed rate on threshability and kernel damage may bedescribed on the basis of an "impact-friction' model.

Minimization of Grain Loss

As noted previously, the minimumtotal loss (YL) would occur at cylin

der speeds below 800 rpm (4396 ft/min). General recommendations (18)on settings of cylinder speed andconcave clearance for combines indicate a cylinder speed of approximately 5300 ft/min and a concaveclearance of % inch for the front and1/8 inch for the rear. These recommendations were for threshing wheatunder very dry conditions and werebased on a summary of recommendations found in operator's manualspublished by combine manufacturers.In terms of this research, 5300 ft/minwould indicate a cylinder speed of980 rpm for the laboratory threshingunit. According to the findings of thisresearch, the cylinder speed for anoptimum situation of high threshability and low kernel damage wouldbe below 800 rpm cylinder speed.This would imply that the generalrecommendations, as found in operator's manuals for setting cylinderspeed, are high.

It is believed that farmers adjusttheir combines according to the operator s manual or by considering thethreshability without due regard tomechanical kernel damage. If this isthe case, they are probably usingcylinder speeds which are unnecessarily high and the result could beexcessive damaged kernels along withhigh threshability.

CONCLUSIONS

Conclusions that can be drawnwithin the limitations of crop varietyand moisture content are as follows:

1. The mechanical parameterscausing significant variation inthreshability are cylinder speedand concave clearance.

2. The mechanical parameterscausing significant variations inpercentage mechanical damageare cylinder speed, concaveclearance and feed rate.

3. Threshability can be expressedin terms of the mechanical parameters by a regression equation of the form:

YT~A0 + A,XS+ A2X| +

A3XC + A4XsXc

where YT is percentage thresh

ability, A0 is a constant and

A], A2, A3, A4 are partial re

gression coefficients. X$ and X^

are the mechanical parametersof cylinder speed and concaveclearance respectively.

4. Mechanical damage can be expressed in terms of the mechanical parameters by a regressionequation of the form:

YD - B0 + B,XS + BjXf +

3**S C

where YD is percentage mechani

cal damage, B0 is a constant and

Bj, B2, B3 are partial regression

coefficients. X$ and Xc are the

mechanical parameters of cylinder speed and concave clearancerespectively.

5. A level of the variables (parameters) considered in this projectthat would result in an optimumsituation of high threshabilityand low kernel damage was notobtained.

6. In practical situations, combineoperators should consider kerneldamage as well as threshabilityin adjusting their combines.

REFERENCES

1. Agness, J. B. 1968. MeasuringMechanical Damage to Corn Kernels. Paper No. 68-620 presentedat A.S.A.E. Winter Meeting atChicago, Illinois.

2. Arnold, R. E. 1959. The Effect ofHarvest Damage on the Germination of Barley. J. Agric. Engng.Res., 4: 24-29.

3. Arnold, R. E. 1964. Experimentswith Rasp Bar Threshing Drums(Some Factors Affecting Performance). J. Agric. Engng. Res., 9:99-131.

CANADIAN AGRICULTURAL ENGINEERING, VOL. 11, No. 2, NOVEMBER 1969

4. Arnold, R. E. and M. P. Jones.1963. A Survey of Grain DamageIncurred and Drum Settings UsedDuring the Combine Harvestingof Capelle Desprez Wheat andProctor Barley. J. Agric. Engng.Res, 8: 178-184.

5. Bainer, R, R. A. Kepner, and E.L. Barger. 1960. Principles ofFarm Machinery. Wiley and Sons,New York, 381-410.

6. Bainer, R. and H. A. Borthwick.1934. Thresher and Other Mechanical Injury to Seed Beans ofthe Lima Type. California Agric.Expt. Station Bull. 580.

7. Buchele, W. F. and H. Waelti.1969. Factors Affecting Corn Kernel Damage in Combine Cylinders. Trans. A.S.A.E, 12: 55-59.

8. Cooper, G. F. and A. E. Neal.1968. Performance Testing ofCombines in the Lab. Agric.Engng, 49: 397-399.

9. Cuskas, L. 1964. Examination ofthe Flow of the Crop with Combine Harvester with Special Respect to the Possible Increase ofPerformance. Translation No. 169,Scientific Information Department, N.I.A.E, Wrest Park, Sil-soe, Bedford.

10. DeLong, H. H, and A. J.Schwantes, 1942. Mechanical Injury in Threshing Barley. Agric.Engng, 23: 99-101.

11. Feinstein, L, C. Golumbia, andJ. Hart. 1959. Oven Methods forPrecise Measurement of MoistureContent of Seeds. Marketing Research Report No. 304, U.S.D.A.Agric. Marketing Service.

12. Hall, C. W. and G. C. Zoerb. 1960.Some Mechanical and Rheologi-cal Properties of Grain. J. Agric.Engng. Res, 5: 387-398.

13. IBM publication H20-0205-3,SYSTEM/360, Scientific Subroutine Package, Version 111, Programmers Manual.

14. King, B. L. and A. W. RiddoUs.1960. Damage to Wheat Seedand Pea Seed in Threshing. J.Agric. Engng. Res, 5: 387-398.

continued on page 91

87

tactor, provide another current pathand the contactor stays closed eventhought the first microswitch has reopened under grain pressure.

4. When the grain level rises to acertain height, grain will press againsta second paddle forcing open anothermicroswitch located in the AND circuit of the contactor.

5. a) The contactor releases andthe elevator motor stops. When thegrain level drops and pressure is released from the top paddle the micro-switch recloses. However, the OR partof the circuit is now open so that thecontactor remains open.

2. This is only necessary to preventthe elevator from running uselessly ifthe grain supply fails. Another pressure operated microswitch, heldclosed by the presence of grain, islocated a* the outlet of the main supply bin. The switch is connected intothe AND circuit of the contactor. Ifthe grain supply fails this switchopens and the contactor opens cuttingoff the elevator motor.

The unit can be shut down partway through a cycle by depressingthe stop button which is located inthe AND circuit. The contactor opensand the circuit remains off until thegrain level drops below the bottomgrain sensor. The addition of a hand-off-automatic switch instead of thestart-stop buttons adds more sophistication to the circuit. This allows theunit to be manually (hand) operated,locked off (off) or run automatically.If desired, the three microswitchescan be connected to operate indicatorlights to show at a remote locationthe condition of the system.

This one method of automating thisoperation is arrived at by the substitution of control elements forhuman action.

There is another solution. Thissecond solution uses a long tube ofreasonable diameter (3 to 4 inches).At the bottom of the tube a pressuresensitive switch is located which isconnected into the AND circuit ofthe contactor. Grain pressure holdsthis switch open. The tube is locatedso that when the bin is almost emptygrain runs out of the tube releasingthe pressure from the switch whichthen closes. The contactor closes andthe elevator motor starts. When thebin is almost full grain overflowsinto the tube containing the pressure

CANADIAN AGRICULTURAL ENGINEERING, VOL. 11, No. 2, NOVEMBER 1969

switch. Grain pressure opens theswitch and the contactor opens. Another pressure switch located at themain grain supply outlet, connectedinto the contactor AND circuit willcut the unit off if the grain supplyfails.

This second system does what isrequired, that is to automatically refill the bin. It is not, however, asversatile as the first nor as troublefree but, using fewer components, ischeaper. These two solutions showtwo approaches to the same problem.

REFERENCES

1. Canadian General Electric, Standard Motor Control Wiring Diagram, Canadian General Electric,Industrial Apparatus Department,Peterborough, Ontario.

2. General Electric, TransistorizedStatic Control, Bulletin GPC-B53D, General Electric, Blooming-ton, Illinois.

3. Maley, Gerald A. and Earle, John,The Logic Design of TransistorDigital Computers, Prentice-HallInc., 1963.

4. McCluskey, E. J. Introduction tothe Theory of Switching Circuits,McGraw-Hill, 1965.

5. Whitesitt, J. Eldon, Boolean Algebra and Its Applications, Addison-Wesley, 1961.

THE UNEXPECTED . . .

continued from page 41

ment of certain objectives in ourchosen profession geared to the betterment of the agricultural industry.Results by themselves are of little use,however, unless made public. Thephilosophers have not included intheir rules the necessity of publishingthe results of tests, or of experimentaland development projects. Here liesa most critical test of new theoriesand predictions.

THE EFFECT OF SELECTED . . .

continued from page 87

15. Kolganov, K. G. 1958. MechanicalDamage to Grain During Threshing. Translation. J. Agric. Engng.Res, 3: 178-184.

16. Protter, M. H. and C. B. Morrey.1965. Modern Mathematical Analysis. Addison-Wesley PublishingCompany, Inc., Reading, Massachusetts, 169-175.

17. Reed, W. B. and E. O. Nyborg.Cockshutt SP431 and Massey-Ferguson 300. Test Reports, Saskatchewan Agricultural Machinery Administration.

18 1965.The Grain Grower. InformationService, United Grain Growers,Winnipeg, Manitoba, 742-746.

91