short rod correction.pdf

DISCUSSIONS AND CLOSURES

Dow

nloa

ded

from

asc

elib

rary

.org

by

Uni

vers

idad

De

Con

cept

ion

on 0

3/26

/13.

Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

Discussion of “Review of StandardPenetration Test Short Rod Corrections”by Chris R. Daniel, John A. Howie,R. Scott Jackson, and Brian WalkerApril 2005, Vol. 131, No. 4, pp. 489–497.

DOI: 10.1061/�ASCE�1090-0241�2005�131:4�489�

Luciano Décourt11Luciano Décourt Engineering Consultores Ltd., Av. Brig. Faria Lima,

01451-001 Sao Paulo, SP Brazil.

The authors are to be commended for bringing up the matter ofwhether or not to correct the NSPT values for rods shorter than10.0 m.

The standard penetration test �SPT� International ReferenceTest Procedure �IRTP� �Décourt et al. 1988� has been mentionedin this paper; and the discusser, as one of the authors of thatdocument, would like to offer additional information on thismatter.

First, the document mentioned in the paper was not actuallythe official SPT International Reference Test Procedure. Rather, itwas just a preview of the official document that the SwedishGeotechnical Society published one year later, during the 12thInternational Conference of Soil Mechanics and FoundationEngineering, held in Rio de Janeiro in 1989.

The former document presented an analysis of energy mea-surements that was based on the pioneer work by Schmertmannand Palacios �1979�; and, as mentioned by the authors, used theF2 method. Schmertmann, by the way, was also one of the authorsof the IRTP. Nevertheless, within the committee that elaboratedthe IRTP, no consensus existed about the adequacy of this method

Fig. 1. Variation of the torque ratio, T/NSPT, with depth

JOURNAL OF GEOTECHNICAL AND GEOE

J. Geotech. Geoenviron. Eng.

for evaluating the enthru energy of the SPT in any circumstances.Therefore, it was decided to omit that part of the report in thefinal and definitive version of the SPT International ReferenceTest Procedure.

In the same year, 1989, the discusser had the opportunity topresent a state-of-the-art report on the SPT �Décourt 1989�. Inthat paper, on analyzing the proposed correction factors for rodlengths of less than 10.0 m, the discusser stated, “This tendency,however, is against the practical experience of the author. Disre-garding sandy soils, for which the ambient pressure is fundamen-tal, and considering for example over-consolidated homogeneousclays, one should expect a decrease in the penetration resistancewith depth. But this almost never happens.”

Clearly, the engineers’ intuition contrasted with the theory ac-cepted at that time. Décourt and Quaresma Filho �1991� had pre-viously introduced torque measurements onto the SPT procedure,and that test became known as the SPT-T. Torque measurements�T� are ‘static,’ as opposed to NSPT measurements, which are dy-namic. Therefore, whether or not the NSPT is affected by thelength of the rods, the torque measurements are not affected.Also, the experience with the SPT-T demonstrated that for a givensoil, a relationship exists between T and NSPT. This knowledgeprovided the discusser with the first experimental evidence thatthe proposed corrections for rod lengths were wrong. The dis-cusser analyzed many cases of SPT-T carried out in uniform lay-ers of stiff clays �Décourt and Quaresma Filho 1994� and in sandsof constant relative density �Décourt 2002� and observed that theso called torque ratio �T /NSPT� was practically constant withdepth �Figs. 1 and 2�. If the NSPT values decrease with depth, thetorque ratio would have to increase with depth rather than beconstant or vary randomly, as happened in all the cases.

Matsumoto et al. �1992� made criterious measurements in twopoints of the rods and concluded that the energy transferred to the

Fig. 2. Variation of the torque ratio, T/NSPT, with depth

NVIRONMENTAL ENGINEERING © ASCE / DECEMBER 2006 / 1633

2006.132:1633-1634.

Dow

nloa

ded

from

asc

elib

rary

.org

by

Uni

vers

idad

De

Con

cept

ion

on 0

3/26

/13.

Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

sampler was independent of the rod lengths. These authors mea-sured not only the energy transferred to the rods but also thepenetration of the sampler into the ground during the time.

It was demonstrated that the penetration of the sampler intothe ground was completed only after about 18 ms. The influenceof the first wave ended in a much shorter time, typically 4 to5 ms.

These authors proved that the energy of waves, other than thefirst one, contributed most significantly to the penetration of thesampler into the ground.

A quite different procedure was followed by Aoki and Cintra�2000�. On the basis of the Hamilton principle of conservation ofenergy, they proved that the energy transfer to the rods decreaseda little bit with the increase in their lengths.

The discusser had the opportunity to participate in the doctoralthesis of Cavalcante �2002�.

Using both force and velocity measurements, it was demon-strated that the energy transferred to the rods was independent oftheir length.

Odebrecht �2003�, following an original approach, also ob-served a slight decrease in the energy transfer with the increase inthe rod lengths. All this information is now brought by the dis-cusser to support the authors’ main conclusion that no correctionshould be applied to the NSPT values for rod lengths shorter than10.0 m.

It is time for the specialty to recognize that the recommendedcorrection factors for short rod lengths, even though they havebeen proposed by outstanding engineers, are not adequate; andtheir use should be immediately discontinued.

References

Aoki, N., and Cintra, J. C. A. �2000�. “The application of energy conser-vation Hamilton’s principle to the determination of energy efficiencyin SPT tests.” Application of stress-wave theory to piles—Quality as-

surance on land and offshore piling, S. Niyama and J. Beim, eds.,Balkema, Rotterdam, The Netherlands, 457–460.

Cavalcante, W. H. �2002�. “Theoretical and experimental investigationon the SPT.” Doctoral thesis, COPPE, Rio de Janeiro, Brazil�in Portuguese�.

Décourt, L. �1989�. “The standard penetration test: State-of-the-art re-port.” 12th ICSMFE, Vol. 4, Rio de Janeiro, Brazil, 2405–2416.

Décourt, L. �2002�. “SPT; SPT-T Brazilian practice: Advantages, limita-tions, and critics.” ABMS, technical publication �in Portuguese�.

Décourt, L., Muromachi, T., Nixon, I. K., Schmertmann, J. H., Thorburn,S., and Zolkov, E. �1988�. “Standard penetration test �SPT�: Interna-tional reference test procedure.” Proc., 1st Int. Symp. on Penetration

Testing, Balkema, Rotterdam, The Netherlands, 3–26.Décourt, L., and Quaresma Filho, A. R. �1991�. “The SPT-CF: An im-

proved SPT.” SEFE II, Vol. 1, São Paulo, Brazil, 106–110.Décourt, L., and Quaresma Filho, A. R. �1994�. “Practical applications of

the standard penetration test complemented by torque measurements,SPT-T: Present stage and future trends.” 8th ICSMFE, Vol. 1, NewDelhi, India, 143–146.

Matsumoto, T., Serigushi, H., Yoshida, H., and Kita, K. �1992�. “Signifi-cance of two point strain measurement in SPT.” Soils Found., 32�2�,67–82.

Odebrecht, E. �2003�. “Energy measurements in the SPT.” Doctoral the-sis, UFRGS, Porto Alegre, RS., Brazil �in Portuguese�.

Schmertmann, J. H., and Palacios, A. �1979�. “Energy dynamics of SPT.”
J. Geotech. Engrg. Div., 105�8�, 909–926.
1634 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGIN


Discussion of “Review of StandardPenetration Test Short Rod Corrections”by Chris R. Daniel, John A. Howie,R. Scott Jackson, and Brian WalkerApril 2005, Vol. 131, No. 4, pp. 489–497.DOI: 10.1061/�ASCE�1090-0241�2005�131:4�489�

Fernando A. B. Danziger, A.M.ASCE1;Bernadete R. Danziger2; and Erinaldo H. Cavalcante3

1Associate Professor, COPPE and Escola Politécnica, Federal Univ., ofRio de Janeiro, Rio de Janeiro, RJ, Brazil. E-mail: [email protected]

2Associate Professor, Rio de Janeiro State Univ., Rio de Janeiro, RJ,Brazil. E-mail: [email protected]

3Associate Professor, Sergipe Federal Univ., Aracaju, SE, Brazil. E-mail:[email protected]

The electric device developed by the authors might indeed repre-sent a powerful tool to better analyze the contact history of thehammer and anvil during a standard penetration test �SPT�. Theauthors collected data from a laboratory study on SPT that re-vealed the occurrence of secondary impacts before the time atwhich the hammer and anvil are generally assumed to separate�2L /c�. The authors also emphasized that these impacts can bepredicted by a proper modeling of SPT. However, they did notprovide a clear explanation of the reasons for those impacts. Thediscussers have made some hypotheses about the causes of thoseimpacts; the first one is associated with an impedance ratio reduc-tion, since it seems to be related to the transition from the NWanvil rod to the AW rod. However, the distance from the top ofthe anvil rod to the transition is around 1.7 m, so the elapsed timebetween the first impact and the first disconnection should be0.7 ms, but the measured value was around 1.3 ms. Other possi-bilities might be related to the particular shape of the safety

Fig. 1. Force �F� versus time, theoretical and experimental; safetyhammer, AW rods �adapted from Schmertmann and Palacios 1979�

EERING © ASCE / DECEMBER 2006

2006.132:1633-1634.

Dow

nloa

ded

from

asc

elib

rary

.org

by

Uni

vers

idad

De

Con

cept

ion

on 0

3/26

/13.

Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

hammer-anvil-rod system used. In fact, when the compressionwave resulting from the first impact propagates downward in theanvil rod, a compression wave also propagates upward in theupper �solid� part of the hammer. When this wave reaches thehammer top �at the time Lh /c, where Lh=148 mm is the length ofthe upper part of the hammer�, it reflects as a tension wave that

Fig. 2. Force and velocity versus time; rod length of

Fig. 3. Force-versus-time curve; data from Fig. 2; elapsed time corres

Fig. 4. Energy and displacement versus time; rod length

combines with and cancels the upward compression wave, giving



the particles a total velocity equal to twice the previous value. Atthe time 2Lh /c, no strain exists in the hammer �Fairhurst 1961;see also Palacios 1977 and Schmertmann and Palacios 1979�. Ifthe hammer had a cylindrical shape, the hammer velocity at theinterface would abruptly change at 2Lh /c �Fairhurst 1961�. Thecontinuation of the process would generate a stepped-shape curve,

m; test performed in a sandy clay �Cavalcante 2002�

g to the period of separation between hammer and rods was removed

.39 m; test performed in a sandy clay �Cavalcante 2002�

16.39

pondin

of 16

as shown by Fairhurst �1961�, Palacios �1977�, and Schmertmann


2006.132:1633-1634.

Dow

nloa

ded

from

asc

elib

rary

.org

by

Uni

vers

idad

De

Con

cept

ion

on 0

3/26

/13.

Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

and Palacios �1979�. However, at the time 2Lh /c, the tensionwave starts to propagate also at the sleeve of the hammer. Thereflected compression wave, associated with the downward ten-sion wave at the sleeve, progresses upward in the sleeve, and issuperimposed on the waves traveling in the upper part of thehammer. How this superposition generates a disconnection is notclear, but if the authors could provide an explanation for suchunexpected behavior, it should be interesting.

Furthermore, although the authors have stated that “the forceand velocity measured prior to �2L� /c� decay exponentially onlyin a very rough sense,” this observation seems to be valid only forthe particular hammer-anvil-rod system that the authors used. Infact, the data obtained by Palacios �1977� and Schmertmann andPalacios �1979�, who also used a safety hammer, did show aforce-versus-time curve that could be well represented by astepped shape �or approximated by an exponential decay curve�,as shown in Fig. 1. The authors should provide an explanation ofthe reasons for the differences.

An extensive field research program was undertaken by thediscussers �Cavalcante 2002; Cavalcante et al. 2004�. A total of1,393 blows have been recorded �not considering the blows re-corded in the 150 mm initial setting�. Different kinds of soil andsoil resistance have been found in 12 boreholes where the mea-sured data have been obtained. Most of the instrumented testshave been performed in depths limited to around 16 m. The SPTequipment consisted of a pinweight hammer and 3.23 kg/m rods,as in Brazilian practice. Although no device similar to the au-thors’ device has been used to verify the separation between ham-mer and anvil, the position of the instrumentation �about 0.5 mbelow the anvil top� and the shape of force and velocity-versus-time curves allow the discussers to assume that no separation hasoccurred before 2L /c in all blows recorded.

Moreover, the force-versus-time curve in almost all casescould be associated with the exponential decay type �as reported

Fig. 5. Efficiency versus length of rods �Cavalcante 2002;Cavalcante et al. 2004�

by Fairhurst 1961, on the basis of previous work by Timoshenko



and Goodier �1951�, who derived the expression to the rigid ham-mer case; see also Yokel 1982� until the impedance change whichcorresponds to the top of the sampler �2L� /c, where L��distancefrom the instrumented section to the sampler top�, as illustrated inFig. 2 for a blow related to a 16.39 m rod length. A second impactcan be observed from the figure. If the part of the curve corre-sponding to the elapsed time where anvil and rod are separatedbetween the first and second impact is removed, the exponentialdecay type of curve is even easier to observe �Fig. 3�.

Furthermore, the authors have mentioned that the assumptionof Schmertmann and Palacios �1979� that any energy transferredduring post 2L /c impacts would not contribute to samplerpenetration has never been properly assessed in real soils. Thediscussers did obtain such data for different kinds of soils�Cavalcante 2002; Cavalcante et al. 2004� and found that thisassumption is definitely not true. The discussers do agree with theauthors that the additional energy is delayed, rather than lost, anddoes contribute to sampler penetration, as shown in Fig. 4.

In all cases where a tension-reflected wave has occurred�which happened in most of the tests performed�, at least a secondimpact has occurred in a single blow, each one contributing to theenergy delivered to the rods.

The authors also speculated about the validity of their findingsin real soils, since their research was performed in the laboratory.The discussers’ experience is that those speculations are generallycorrect, and their data support them. In fact, the discussers con-cluded �Cavalcante 2002� that the timing of the secondary im-pacts is affected by the soil stiffness and strength, related by thediscussers to the N� value, where N� is defined as the ratiobetween 300 mm penetration divided by the sampler final pen-etration �in mm� in that instrumented blow �Schmertmann andPalacios 1979�. Thus, the length of time between successive im-pacts decreases with the increase of the N� value for a given rodlength.

Finally, the number of impacts after 2L /c, the elapsed timebetween successive impacts, and the amount of energy transferredin each impact vary with the N� value. As another example, for anN� value as low as 1.9 and a very short rod �2.39 m�, sevenimpacts have been measured. The first impact has contributedonly 41% to the maximum energy delivered to the rods. However,the maximum energy �or energy efficiency� from the hammer tothe rods does not show any trend of variation with depth, asillustrated in Fig. 5. Data in the figure represent the average of allblows in each SPT. Therefore, the discussers’ experience, whichis based on extensive field-testing research, supports quite wellthe point raised by the authors that no short rod correction isneeded at all.

References

Cavalcante, E. H. �2002�. “Theoretical-experimental investigation ofSPT.” D.Sc. thesis, COPPE, Federal Univ. of Rio de Janeiro, Rio deJaneiro, Brazil.

Cavalcante, E. H., Danziger, F. A. B., and Danziger, B. R. �2004�. “Es-timating the SPT penetration resistance from rod penetration based oninstrumentation.” Proc., Geotechnical and Geophysical Site Charac-terization, Millpress, The Netherlands, Vol. 1, 293–298.

Fairhurst, C. �1961�. “Wave mechanics of percussive drilling.” Mine andQuarry Engineering, 27�3�, 122–130.

Palacios, A. �1977�. “Theory and measurements of energy transfer duringstandard penetration test sampling.” Ph.D. thesis, Univ. of Florida,
Gainesville, Fla.

2006.132:1633-1634.

Dow

nloa

ded

from

asc

elib

rary

.org

by

Uni

vers

idad

De

Con

cept

ion

on 0

3/26

/13.

Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

Schmertmann, J. H., and Palacios, A. �1979�. “Energy dynamics of SPT.”J. Geotech. Engrg. Div., 105�8�, 909–926.

Timoshenko, S., and Goodier, J. N. �1951�. “Theory of elasticity.” 2ndEd., McGraw-Hill.

Yokel, F. Y. �1982�. “Energy transfer in standard penetration test.” J.Geotech. Engrg. Div., 108�9�, 1197–1202.

Closure to “Review of StandardPenetration Test Short Rod Corrections”by Chris R. Daniel, John A. Howie,R. Scott Jackson, and Brian WalkerApril 2005, Vol. 131, No. 4, pp. 489–497.

DOI: 10.1061/�ASCE�1090-0241�2005�131:4�489�

Chris R. Daniel1; John A. Howie2; R. Scott Jackson3; andBrian Walker4

1Doctoral Candidate, Dept. of Civil Engineering, Univ. of BritishColumbia, 6250 Applied Science La., Vancouver BC, Canada V6T1Z4. E-mail: [email protected]

2Associate Professor, Dept. of Civil Engineering, Univ. of BritishColumbia, 6250 Applied Science La., Vancouver BC, Canada V6T1Z4. E-mail: [email protected]

3Research Technician, Dept. of Civil Engineering, Univ. of BritishColumbia, 6250 Applied Science La., Vancouver BC, Canada V6T1Z4. E-mail: jscottj@@civil.ubc.ca

4Project Engineer, ADI Limited, 7071 Bayers Rd., Suite 119, Halifax NS,Canada B3L 2C2. E-mail: [email protected]

The writers thank the discussers for their interest in the paper andfor supporting the suggestion that short rod correction factorsshould not be applied. The intent of the study was to demonstratethat the logic underlying short rod corrections is flawed and thatthe total transferred energy is insensitive to rod-length variations.The field data presented in both discussions appear to support thisassertion. The writers agree with the discussers that continued useof short rod corrections is undesirable but note that the modifica-tion of an accepted standard practice, regardless of its merits,must be preceded by a careful assessment of consequences. Thisclosure addresses this issue and provides additional informationon early secondary impacts and stress wave shapes, as requestedin the discussion by Danziger et al.

Discontinued Use of Short Rod Corrections

Standard penetration test �SPT� �N� values are corrected for varia-tions of overburden pressure, energy, borehole diameter, rodlength, and sampler details to yield �N1�60 values �Youd et al.2001�. This long list of corrections creates uncertainty and evenconfusion for practitioners. Discontinuing the use of unnecessaryshort rod corrections would help reduce this uncertainty. Shortrod corrections have been in use for more than 20 years, however,and empirical SPT correlations developed during that time arelikely to be based in part on short rod corrected data. Many prac-titioners believe that discontinuing the use of short rod correctionswould be unconservative for this reason. In fact, one consequenceof the invalidity of short rod corrections is that correlations thatare based on short rod corrected data are inherently unconserva-
tive. Some of these correlations therefore need to be reassessed


without short rod corrections to avoid continued unconservativeinterpretation of SPT blow counts.

Consider the extreme case of a correlation that is based en-tirely on calibration-chamber data. These data are typically col-lected by using a single test setup, including a short rod string. Ifshort rod corrections are applied, all blow counts are decreased bythe same amount, introducing a bias error in the data. For ex-ample, Skempton �1986� proposed the following relationship be-tween �N1�60 and relative density �Dr� for “Reid-Bedford model”sand:

�N1�60 � 36 · Dr2 �1�

The relationship is based on calibration-chamber data collectedby using rods of 2.4 m �8 ft� in length, to which Skempton �1986�applied a rod-length correction of 0.65. Noting that no reductionof energy is likely to have occurred because of rod length, theshort rod correction can be removed to reveal the true relationship

�N1�60 � 55 · Dr2 �2�

Next, consider the interpretation of an �N1�60 value of 20 re-corded in a field deposit of the same sand using rods that aregreater than 10 m long. The interpreted relative density would be75% if the short rod corrected relationship was used �Eq. �1��,compared with 60% if the true relationship was used �Eq. �2��.Most engineers would recognize that the difference of 15% iswithin the level of accuracy of the relationship, but this exampleillustrates that seemingly conservative short rod corrections canlead to unconservative correlations.

The effect of short rod corrections on empirical correlationsthat are based on field data is expected to be less significant.Some of the data will have been recorded by using rods greaterthan 10 m in length, and relatively few data points will have beencollected by using rod lengths of 4 m or less, to which the largestshort rod corrections are applied. If the data are not grouped ac-cording to rod length or some indicator of rod length, such asoverburden pressure or test depth, the effect of short rod correc-tions should be relatively random. In that case, the short rodcorrections serve only to increase the uncertainty of already un-certain relationships.

Overburden correction factors developed by using calibrationchamber data are fortunately immune to the effect of short rodcorrections. The corrections are based on ratios of blow countsrecorded at different confining pressures. The short rod correc-tions appear in both the numerator and denominator of the ratioand thus cancel, having no net effect on the resulting overburdencorrection factor.

Secondary Impacts Prior to †2L /c‡

The propagation of stress waves through SPT hammer-and-rodsystems is highly complex because of the partial reflection ofstress waves that occurs at every change in cross-sectional area.This leads to exponential growth of the total number of stresswaves propagating within the system. The writers are thereforehesitant to ascribe early separations to any one feature of thehammer or rods. FVCALC can be used to demonstrate this com-plexity by investigating the effect of individual rod componentson separation histories.

A series of additional FVCALC simulations was completed forthe case of the safety hammer and the 6.49 m string of drill rods
described in the paper. For the first simulation, the rod string was

2006.132:1633-1634.



Dow

nloa

ded

from

asc

elib

rary

.org

by

Uni

vers

idad

De

Con

cept

ion

on 0

3/26

/13.

Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

modeled as a single NW rod of 6.49 m length with no changes incross-sectional area �Fig. 1, solid lines�. Simulations were thenrun with each of the following additions: �1� the NW-AW cross-over; �2� a string of four 1.0 m long AW rods and a samplerbelow the location of the NW-AW crossover; and �3� the anvil.The simulated NW-AW crossover and anvil are indicated by thedashed lines in Fig. 1. The reader is referred to Fig. 1 of the paperfor clarification of the hammer and rod geometries.

The axial force between the hammer and the top of the rodstring �FA� is monitored during FVCALC simulations. Hammer-anvil separation is simulated whenever the passing of a stresswave would cause the force to decrease below zero. The �FA�histories arising from each of the four previously described simu-lations are shown in the top four plots of Fig. 2. For a uniformrod, the hammer and anvil remain in contact until time �2L /c�,when the stress wave reflected from the bottom of the rod arrivesat the hammer-anvil interface. The next three plots illustrate howeach of the three added components affects the �FA� history ob-served during the uniform rod case. Early hammer-anvil separa-tions, indicated by periods of zero �FA�, are predicted in all cases.The fifth plot shows the �FA� history when all three componentsof the rod string are included in the simulation. This �FA� historycorresponds to Fig. 2 in the paper. The combined effect of thethree individual components leads to a rather complex contactforce history, including many more early separations than werecaused by any one component.

The paper noted that FVCALC simulates “perfect” stress waveformation and propagation, including perfectly sharp transitionsfrom loaded to unloaded states, whereas true stress waves exhibitfinite rise times. The bottom plot in Fig. 2 demonstrates the effectof imposing a user-defined rise time. This modification reducesthe number of secondary impacts to the observed number of two.This �FA� history corresponds to Fig. 3 in the paper.

These simulations illustrate that a comprehensive study wouldbe required to determine all the factors contributing to prematurehammer-anvil separations. It is anticipated that the wide variety oftest equipment in use will make it very difficult to predict allsituations for which these impacts occur. The writers reiterate,however, that their occurrence probably has little or no effect onthe transferred energy or measured blow count.

Atypical Stress Wave Data

The discussion by Danziger et al. takes issue with our observationthat force and velocity histories recorded during SPT follow apattern of exponential decay only in a very rough sense. Thewriters believe that the differences between the data presented inthe paper and those presented by Danziger et al. can be attributedto differences of scale and instrumentation details.

The data presented by the writers span a very short period oftime compared with the data presented in the Danziger et al.discussion. Deviations from idealized exponential decay curvesbecome more apparent when data are viewed in greater detail. Inaddition, the writers’ experience indicates that instrumentation de-tails can have a significant effect on the shape of the measureddata. For example, it is apparent that the Schmertmann and Pala-cios �1979� data presented in the Danziger et al. discussion havebeen heavily filtered, producing a relatively smooth curve. Thisfiltering occurred because a piezoelectric load cell was used tomeasure force, because of the intentional and unintentional filter-ing that is applied by all signal processing equipment and becausethe data were digitized by hand from photographs of an oscillo-

Fig. 1. Block model of safety hammer and rods �solid lines showmodel used for initial uniform rod simulation; dashed lines arecomponents added during subsequent simulations�

scope screen. Although this level of filtering would be considered


2006.132:1633-1634.

Dow

nloa

ded

from

asc

elib

rary

.org

by

Uni

vers

idad

De

Con

cept

ion

on 0

3/26

/13.

Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

extreme by current standards, modern proprietary systems con-tinue to employ mechanical filters such as soft accelerometermounts to protect instruments and low-pass antialiasing filters tolimit the amount of data required for each hammer blow.

The writers have made a concerted effort to develop equip-ment capable of recording the highest-quality data �Howie et al.2003�. The system that was used is capable of recording thehigher-frequency stress wave activity that is predicted during

Fig. 2. Histories of �FA� from FVCALC simulations of several rod crods�

simulations but is often absent from stress wave data because of



filtering. These higher-frequency components generally do nothave a significant effect on the calculated energy but are of con-siderable interest during a fundamental study such as ours.

Conclusion

Field and laboratory evidence backed by numerical modeling in-

rations �periods of zero �FA� occur when hammer has separated from
onfigu
dicate that the logic underlying short rod corrections is flawed.


2006.132:1633-1634.

Dow

nloa

ded

from

asc

elib

rary

.org

by

Uni

vers

idad

De

Con

cept

ion

on 0

3/26

/13.

Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

The evidence indicates that significant proportions of the appliedkinetic energy at impact may be transferred to a short rod stringduring secondary impacts. The effect of this delayed energy trans-fer on sampler penetration is unknown but is not considered tojustify the short rod corrections currently in use.

The writers recommend that the use of short rod correctionfactors be discontinued. SPT correlations derived by using shortrod data may need to be reevaluated, since unconservative errorsmay be introduced when they are applied to data collected byusing rod lengths greater than those represented in the database.This is of particular importance for correlations that are based oncalibration chamber data.

Erratum

The following correction should be made to the original paper:The third sentence of the second paragraph on page 496

should reads as follows:



“It is of interest to compare these factors to the ratio of theenergy transferred during the initial impact cluster to the maxi-mum transferred energy.”

References

Howie, J. A., Daniel, C. R., Jackson, R. S., and Walker, B. �2003�. “Com-parison of energy measurement methods in the standard penetrationtest.” Rep. Prepared for the U.S. Bureau of Reclamation, GeotechnicalResearch Group, Dept. of Civil Engineering, Univ. of British Colum-bia, Vancouver, Canada.

Schmertmann, J. H., and Palacios, A. �1979�. “Energy dynamics of SPT.”J. Geotech. Engrg. Div., 105�8�, 909–926.

Skempton, A. W. �1986�. “Standard penetration test procedures and theeffects in sands of overburden pressure, relative density, particle size,aging and overconsolidation.” Geotechnique, 36�3�, 425–447.

Youd, T. L., et al. �2001�. “Liquefaction resistance of soils: Summaryreport from the 1996 NCEER and 1998 NCEER/NSF workshops onevaluation of liquefaction resistance of soils.” J. Geotech. Geoenvi-
ron. Eng., 127�10�, 817–833.

2006.132:1633-1634.

short rod correction.pdf

Documents