“SECRETARY OF LABOR,Complainant,v.GULF AND WESTERN, ENERGY PRODUCTSGROUP, BONNEY FORGE DIVISION,Respondent.INTERNATIONAL ASSOCIATION OFMACHINISTS, LOCAL LODGE 1830,Authorized EmployeeRepresentative.OSHRC Docket No. 79-4053_DECISION ___Before: FOULKE, Chairman; MONTOYA; and WISEMAN,Commissioners.BY THE COMMISSION:At issue in this case is one citation alleging that Gulf and WesternEnergy Products Group, Bonney Forge Division (\”Bonney Forge\”), exposedthirteen employees to excessive levels of noise during the manufactureof pipe fittings and elbows at a facility in Allentown, Pennsylvania.Item 1a of the citation alleges that six employees in the machine shopand three employees in the production grinding area were \”not provided\”with \”[p]rotection against the effects of\” noise exposures exceedingthose listed in Table G-16 of 29 C.F.R. ? 1910.95(a).[[1]] Theseemployees were not required to use, and at the time of the allegedviolation were not in fact using, personal protective equipment such asear plugs or ear muffs. Item 1b of the citation further alleges thatthese employees and four others in the forge shop (where personalprotective equipment was required and used) were \”subjected to soundlevels exceeding those listed in Table G-16\” because \”feasibleadministrative or engineering controls were not utilized\” as required by29 C.F.R. ? 1910.95(b) (1). [[2]]Bonney Forge contested the citation and the case came before formerAdministrative Law Judge Benjamin G. Usher for a hearing. During thehearing and in its post-hearing brief, Bonney Forge contended that thealleged violations should be vacated because the Secretary of Labor hadfailed to meet his burden of proving:(1) That all thirteen employees were exposed for impermissible durationsto noise at excessive levels;(2) That Bonney Forge knew or with reasonable diligence should haveknown of any excessive noise levels or durations of exposure;(3) That Bonney Forge did not provide hearing protection to employees inthe machine shop and production grinding area;(4) That the implementation of engineering controls capable of reducingthe noise to permissible levels would have been technologically feasible;(5) That those engineering controls would have been economically feasible;(6) That the abatement dates specified in the citation were reasonable.In his decision, Judge Usher addressed only the first matter: he decidedthat the Secretary did not prove exposure to excessive noise levels forimpermissible durations. On this basis he vacated the citation, notingthat employers have no duty to comply with either of the cited standardsunless employees are being exposed to daily noise doses greater thanthose permitted in Table G-16.The Secretary seeks reversal of the judge’s decision on the basis thatit is contrary to the preponderance of the evidence. We have closelyexamined the record and the judge’s findings, and we conclude that thepreponderance of the evidence, fairly weighed, does establish thealleged overexposures to noise. Accordingly, we reverse the judge’sdecision and remand the case for disposition of the remaining fivematters raised by Bonney Forge, as listed above.As we have indicated, Table G-16 establishes permissible noiseexposures, which are maximum daily doses delimited by decibel level(\”dBA\”) and time. Exposure to 90 dBA of noise is permitted for as longas 8 hours, but at higher noise levels the time limits progressivelyshorten. For example, at 95 dBA the daily time limit is 4 hours. Ifnoise levels fluctuate throughout a day, as they do in many workplaces,overexposure can be determined by adding up the several exposure timesin the manner specified by a formula, which is set out in the table’sfootnote (see note 1 _supra_).There are two instruments that the Secretary’s Occupational Safety andHealth Administration (\”OSHA\”) can use to measure an employee’s noiseexposure. One is a sound level meter and the other is a dosimeter.The sound level meter consists of a microphone to detect noise and ameter to point out the instantaneous noise level. To use a sound levelmeter to show that an employee is exposed to excessive noise forimpermissible durations, OSHA would have to measure the noise level inthe employee’s hearing zone over a substantial period of time and show(1) that a particular noise level listed in Table G-16 was exceeded forthe corresponding exposure time (e.g., that the employee was exposed to90 dBA for more than 8 hours), (2) that for a particular time periodlisted in Table G-16 the noise level exceeded the correspondingpermissible limit (e.g., that throughout an 8-hour period the employeewas consistently exposed to at least 90 dBA), or (3) that the cumulativeexposure times at various noise levels amounted to overexposure underthe formula.The dosimeter includes a microphone to detect noise; however, instead ofa meter pointing out the instantaneous noise levels, the dosimeter feedsthe noise data as it accumulates into additional circuitry whichautomatically applies the formula. Thus, the dosimeter can handlefluctuating noise levels accurately and efficiently since it eliminatesthe need to record the length of time an employee is exposed to eachmeasured noise level and the need to perform the calculations of theformula. A dosimeter reading of over 100% indicates that the noise is inexcess of permissible limits. _See Collier- Keyworth Co.,_ 13 BNA OSHC1208, 1210-11, 1986-87 CCH OSHD ? 27,867, pp. 36,509-10 (No. 80-2848,1987), _vacated per stipulation,_ No. 88- 2140 (1st Cir. March 22, 1989)(describing measurement of noise using sound level meters and dosimeters).In this case, an industrial hygienist from OSHA obtained dosimeterreadings that exceeded 100% for all thirteen employees. The industrialhygienist also took sound level meter readings to represent two noiseconditions to which the employees were exposed: the background noise attheir work stations, whenever their own machines or processes were notoperating, as well as the operating noise arising whenever their ownmachines or processes were operating along with other machines andprocesses. These sound level meter readings revealed that, during normaloperations, the noise levels reaching each employee’s ears as he workedat his work station generally exceeded 90 dBA–often by a considerableamount.From the dosimeter readings, the industrial hygienist computed eachemployee’s \”equivalent DBA.\” That is, the hygienist computed theconstant noise level that would have produced the dosimeter reading. Forexample, a dosimeter would give a reading of 400% if worn for 6 1\/2hours in noise measuring a constant 102 dBA (which, as the followingtable indicates, is the equivalent dBA for employee Wilt, who worked inthe production grinding area, where the actual noise measured by soundlevel meter fluctuated between 80-100 dBA of background noise and103-107 dBA of operating noise).The entire results of the hygienist’s measurements and calculations areshown in the following table. [[3\/]] As it shows, the equivalent dBAfor each employee is within the range of his sound level meter readings:Shop \tEmployee \tDosim’r \tHoursWorn \tEquiv.dBA\tBkgd.dBA \tOper’gdBAMach.S. \tDiaz \t299% \t6.8 \t99 \t85-92 \t90-103 1\/2Mach.S. \tFelguiras \t247% \t6.8 \t98 \t90-94 \t90-104Mach.S. \tJoler \t293% \t6.7 \t99 \t88-94 \t91-102Mach.S. \tBachman \t375% \t6.6 \t101 \tunknown \t90-106Mach.S. \tCheck \t181% \t6.7 \t96 \t86-97 \t86-97Mach.S. \tKuder \t173% \t6.5 \t95 \t83 \t91-96 1\/2Forge S. \tReph \t410% \t1.95 \t110 \t93-105 \t110-112Forge S. \tMcCormick \t386% \t5.65 \t102 \t90-105 \t102-109Forge S. \tMcFarland \t749% \t5.6 \t107 \t99-105 \t108-115Forge S. \tRodriguez \t702% \t5.6 \t107 \t98-105 \t99-113Prod.G.A.. \tWilt \t400% \t6.5 \t102 \t80-100 \t103-107Prod.G.A. \tO’Donnell \t415% \t6.4 \t102 \t80-100 \t105-110Prod.G.A \tTorok \t539% \t6.5 \t104 \t97-100 \t102-108As can be seen from the table, the industrial hygienist’s sound levelmeter readings are consistent with the dosimeter readings.IIIn vacating the citation, Judge Usher concluded: \”I am not persuaded bythe testimony that Respondent’s employees were subjected to sound levelsin excess of those prescribed in [Table G-16]\” inasmuch as \”[t]he recordaffords much room for doubt.\” The judge’s key finding as to the sourceof doubt is stated as:The means and methods employed by [the industrial hygienist] in thecollection and recording of the sound level data at the various workstations were imprecise, confused, fraught with inconsistencies andconsequently were rebuttably unreliable.Leading up to this finding is a discussion in which Judge Usheridentified nine reasons for discrediting the industrial hygienist’s results:(1) The industrial hygienist’s failure to take readings throughout afull 8-hour shift for each of the thirteen employees, _i.e.,_ thefailure to make \”full-shift\” measurements;(2) His failure to insure that the noise collected by the dosimeters was\”emitted by the operations performed by the employees at their workstations and that noise alone\”;(3) His uncertainty about the times during which employee exposure toexcessive noise was detected throughout the day;(4) His overall lack of experience and training in noise measurement andcontrol;(5) His concession that three out of the six dosimeters used to measurenoise exposure in the machine shop were not within tolerable limits uponcalibration before and after use;(6) His failure to observe that all of the dosimeter microphonesremained in their proper position on the thirteen employees throughoutthe sampling period;(7) The possibility that his dosimeters took readings of noise at levelsbelow 90 dBA.(8) His failure to adjust his readings to compensate for impact noise; and,(9) His failure to apply an error tolerance factor greater than 2 dBA tocounteract the possible effects of high frequency noise on the measuredoverexposures.A. _Commission Precedent Upholding Use of Dosimeters_The first three of the judge’s rationales are inconsistent withCommission precedents, which have correctly held such reasoning to beunsound. With respect to the first rationale, the Commission has heldthat full-shift measurements are not always necessary in order toestablish overexposure to noise. _Sun Shipbuilding & Drydock Co.,_ 2 BNAOSHC 1181, 1182, 1974-75 CCH OSHD ? 18,537, p. 22,518 (No. 268, 1974).The soundness of the Commission’s precedent is evident. If the Secretarycan show that an employee is exposed to more noise in a fraction of aworkshift than is permitted for a full shift, there is clearly no needfor measurements to continue throughout the remainder of the shift.[[4]]The second rationale is untenable because the Commission has held,properly, that the standard regulates the total noise to which anemployee is exposed, not simply the noise from the employee’s ownmachine or process. _See_ _Collier-Keyworth, _13 BNA OSHC at 1223,1986-87 CCH OSHD at p. 36,522.[[5]] OSHA need not separate out the noiseto which an employee is subjected by surrounding operations orprocesses. The third rationale is unsound because the missing data isentirely superfluous. Dosimeters automatically take into account howlong the employees are exposed to particular noise levels. _SeeCollier-Keyworth, _13 BNA OSHC at 1210-11, 1986-87 CCH OSHD at pp.36,509-10. There is therefore generally no need for an industrialhygienist to keep a separate record of the time periods that employeeswere exposed to the particular noise levels.B. _The Industrial Hygienist’s Sampling Practices_The next three rationales (Nos. 4 through 6) are frivolous, for there isno evidence of any problem from them. As to rationale No. 4, althoughthe industrial hygienist was young, slightly trained, and lacking inoverall experience, it does not follow that his readings in this caseare inaccurate. Michael P. McSherry had been an OSHA industrialhygienist for almost two years. He had a degree in biology and had takenOSHA’s industrial hygiene courses, one of which covered use of soundlevel meters and dosimeters. Moreover, he testified that he restudiedthe dosimeter manufacturer’s instructions before taking the readingsthat are at issue in this case. A first attempt to take readings had notbeen successful because he neglected to install batteries and unlockreset devices. Therefore, he testified, he restudied the instructionsbefore making the second attempt.Use of dosimeters to measure noise doses minimizes the degree ofexpertise needed. Dosimeters must be properly calibrated and used.However, as we will discuss in greater detail in the next paragraph, theindustrial hygienist testified that he did properly calibrate and usethe dosimeters to obtain the readings at issue in this case, and thereis no evidence in the record of mistakes affecting those readings. Thosereadings are generally consistent with his readings by sound levelmeter, an instrument that even Bonney Forge’s expert agreed \”requireslittle training\” and is \”pretty easy to use.\” [[6]]Industrial hygienist McSherry’s concession concerning calibration ofthree of the six dosimeters provides no support for the judge’s holdingin this case. The industrial hygienist did not make any calibrationerrors that produced readings higher than actual noise levels;therefore, rationale No. 5 has no merit. On direct examination, theindustrial hygienist testified that he calibrated each dosimeter twice,both before its use and again after reading out the results. Oncross-examination, Bonney Forge’s counsel asked about calibrationresults that might indicate a failure to come within tolerable limits.However, his questions did not elicit any testimony acknowledging anyproblem with the dosimeters other than the following, concerning threedosimeters used in the machine shop: \”Looks like a couple of them arereading a little bit low.\” As a concession, this is certainly not adetrimental one, for it meant that the dosimeters registered a lowerexposure than actually existed; thus, any error only accrued to thebenefit of Bonney Forge.Finally, in examining the judge’s sixth rationale, we find no indicationthat any of the microphones connected to the dosimeters were improperlypositioned on the employees at any time during the sampling period. Therecord demonstrates that the industrial hygienist conformed to acceptedOSHA and industrial hygiene practices in attaching the microphones tothe shirt collars of those employees who had collars and to the shirtshoulders of those employees without collars. Two employees testifiedthat they needed adjustments to be made to the wire clipped to theirbacks, which connects the microphone to the part of the dosimeter thatcontains its circuitry, but there is no showing that any microphoneitself slipped or came off. Thus, we see no more than a hypotheticalpossibility and no evidence in the record that a malpositionedmicrophone produced an inaccurate reading. Moreover, the correlation ofthe sound level meter readings to the dosimeter readings suggests that,even if there was a microphone that was not in proper position, theimproper placement had no significant effect on the accuracy of thedosimeter reading.C. _Margins of Error_The last three rationales (Nos. 7 through 9) of the judge are alsowithout merit in light of the factual record as a whole in this case.These three rationales share a common promise: that, for varyingreasons, the dosimeter readings must be discounted by some significantpercentage to allow for an adequate margin of error. Yet, the record asa whole reveals that, even after allowing a very large margin of error,Bonney Forge’s own expert accepted ten out of the thirteen readings asestablishing overexposure. This expert, James H. Botsford, a consultantnoise control engineer, had criticisms of the industrial hygienist’stechniques, but he did not say that dosimeter readings attaining themagnitude of the ones in this case should be disregarded entirely.Instead, he testified that the readings of over 250% could be consideredto reveal overexposure. [[7]]There remain, then, the three dosimeter readings of less than 250%, allobtained in the machine shop. However, even there, we note that expertBotsford’s independent investigation apparently confirmed the industrialhygienist’s finding of overexposure.[[8]] Moreover, we find no basis fordisregarding the three machine shop readings on the basis of anyascertainable percentage for error.The judge referred to a possibility (rationale No. 7) that thedosimeters took readings of noise at levels below 90 dBA. Because TableG-16 places no limit on exposure to noise at levels below 90 dBA, suchexposure should not be considered in determining whether overexposurehas occurred. Specifically, the judge’s concern was based on testimonythat suggests that the dosimeters might have been reading noise from 89dBA upward.[[9]] This does not mean, however, that the readings aremeaningless, simply that they have to be discounted by a larger errorfactor: approximately 50% instead of 32%.[[10]] Only two readings werebelow 200% and even those were well over 150%. Thus, even if thedosimeters were improperly set at 89 dBA, each reading still exceededthe instrument’s inherent error factor by a substantial amount.The judge also submitted (rationale No. 8) that the dosimeter readingsshould have been further discounted for impact noise.[[11\/]] However,the only plant area in which impact noise was generated was the forgeshop, where, as we have previously observed, the dosimeter readings wereof such a magnitude that even Bonney Forge’s expert would accept them asshowing overexposure. In the machine shop and the production grindingarea, the noise was produced by processes such as drilling, cutting, andgrinding–processes that do not generate the high-intensity,short-duration noise peaks that are characteristic of impact noise.Bonney Forge argues that there was a spillover of impact noise from theforge shop into the machine shop, and that, thereby, the dosimeterreadings in the machine shop were contaminated and rendered unreliable.We agree with Bonney Forge that, to the extent that the noise producedin the forge shop consisted of impulses more than one second apart, anyspillover of it into the machine shop would not be properly included innoise measurements taken there. See note 11 _supra. _However, theSecretary’s witnesses uniformly testified that, although the impacts ofthe forge hammers were audible in the machine shop, they were aninsignificant part of the total noise exposure compared to the noisecreated by the machines located within the machine shop.[[12]] The onlycontrary testimony came from Bonney Forge’s expert witness, and it wasambivalent. [[13]] Accordingly, we see no basis for discounting themachine shop readings.Lastly, the judge believed that the readings should have been discountedfor high frequency noise (rationale No. 9). As stated previously, atnote 7 _supra,_ Bonney Forge’s expert witness testified that, because ofthe presence of high frequency noise in the workplace, OSHA should haveapplied an error factor of 6.5 dBA in this case; translated to thedosimeter, this would require a reading of nearly 250% to showoverexposure. See _Collier- Keyworth_, 13 BNA OSHC at 1228-29, 1986-87CCH OSHD at p. 36,528; also, Table A-1 of 29 C.F.R. ?1910.95. In_Collier-Keyworth_, however, the evidence proved that the employee wasexposed to _predominantly_ high frequency noise. Therefore, theCommission concluded that his 158.5% reading was insufficient to showoverexposure. 13 BNA OSHC at 1229 & n.27, 1986-87 CCH OSHD at p. 36,528& n.27. In contrast, in this case, there is no evidence that anyemployee was exposed to predominantly high frequency noise. BonneyForge’s expert witness testified that he did notice some high pitchednoise at several locations. However, Bonney Forge did not attempt toprove that this noise was a significant part of any employee’s totalnoise dose. Thus, we find no basis for applying the large error factorto any of the readings in this case, including in particular the threelowest dosimeter readings taken in the machine shop.D. _The Absence of any Evidence of Compliance_Finally, all nine rationales of the judge’s decision are undercut by theabsence of any evidence indicating that noise exposures at BonneyForge’s workplace were within permissible limits. This is not a casewhere the judge was called upon to balance evidence indicatingoverexposure against evidence indicating permissible exposure, andconcluded either that the latter evidence preponderated or that neitherside preponderated. Instead, virtually all of the evidence, in thisenormous and presumably exhaustive record, dealing with the subject ofnoise exposure at Bonney Forge’s workplace, suggests that permissiblelimits were exceeded.Bonney Forge claims that there is evidence showing the noise exposuresto have been within permissible limits, but either the \”evidence\” is notin the record or it does not support the claim. Bonney Forge repeatedlyasserts that its insurance company \”made a noise survey of the machineroom which showed the levels to be within the Table G-16 limits.\”However, any such results of an insurance survey are not in the record.In fact, on that basis Judge Usher sustained an objection by theSecretary to a question by Bonney Forge’s counsel, trying to incorporatean assertion that two prior noise surveys had shown the noise exposuresto be within permissible limits.Bonney Forge also relies on measurements taken by another OSHAcompliance officer during a 1977 safety inspection, which led to the1979 industrial hygiene inspection in this case. During the 1977 safetyinspection, the compliance officer took sound level meter readingsranging from 88 dBA to 92 dBA in the machine shop, which could indicateexposure within permissible limits. However, because of his extensiveexperience in industrial plants, this compliance officer was convincedthat his readings were inaccurate; after the inspection, he turned inhis sound level meter as defective equipment and recommended anindustrial hygiene inspection of the machine shop.III.In sum, we conclude that the Secretary established that thirteenemployees were exposed for impermissible durations to excessive levelsof noise. Accordingly, we reverse the judge’s decision holding to thecontrary, and we remand this case to the Chief Administrative Law Judgefor reassignment and for appropriate disposition of the remaining fiveissues listed in this decision. As we have indicated, the evidentiaryrecord is substantial, and the necessary factual findings and legalconclusions should be in the first instance entered by an administrativelaw judge rather than the Commission. Also, to the extent necessary andappropriate, the parties may be given an opportunity to presentadditional arguments.Edwin G. Foulke, Jr.ChairmanVelma MontoyaCommissionerDonald G. WisemanCommissionerDated: January 14, 1991————————————————————————SECRETARY OF LABORComplainantv.GULF AND WESTERN, ENERGY PRODUCTSGROUP, BONNEY FORGE DIVISIONRespondentINTERNATIONAL ASSOCIATION OFMACHINISTS, LOCAL LODGE 1830Authorized EmployeeRepresentativeOSHRC DOCKET NO. 79-4053_DECISION AND ORDER Appearances:_Howard K. Agran, Esq.Office of the Regional SolicitorU. S. Department of LaborPhiladelphia, Pennsylvaniafor the ComplainantRobert D. Moran, Esq.Washington, D. C.for the RespondentMr. Douglas Navarre, ChairmanInternational Association ofMachinists – Local Lodge 1830Allentown, PennsylvaniaandMr. Thomas Messner, PresidentInternational Association ofMachinists – Local Lodge 1830Northhampton, Pennsylvaniafor the Authorized EmployeeRepresentative_Usher, Judge:_These proceedings were initiated by the Secretary of Labor, UnitedStates Department of Labor, pursuant to Section 10(c) of theOccupational Safety and Health Act of 1970, 29 U.S.C. ? 651, _et_ _seq._(\”the Act\”). The Secretary seeks affirmance of a citation chargingseveral \”willful\” violations of Section 5(a)(2) of the Act and theassessment of a $2,000 penalty.[[1\/]]The Citation, resulting from an inspection of Respondent’s workplace atAllentown, Pennsylvania, from March 15 to May 9,1979, was issued on July9, 1979. Respondent filed a timely Notice of Contest; a Complaint andAnswer followed in accordance with the Commission’s Rules of Procedure(29 CFR 2200.33); and the issues were tried before me at Allentown,Pennsylvania, on 14 non-consecutive days between April 2 and August 5,1980. The International Association of Machinists elected party statuspursuant to Commission Rule 20(a) (29 CFR 2200.20(a)), and itsrepresentatives participated fully at the trial.Complainant and Respondent filed post-trial briefs on November 17 and20, 1980. The union has not briefed its position._The Issues._The pleadings, evidence adduced by the parties, the oral arguments ofcounsel and their post-trial briefs have served to raise the followingissues: 1) whether Respondent violated Section 5(a)(2) of the Actbecause of its failure to comply with the health standards promulgatedby Complainant and codified at 29 CFR 1910.95(a) and 1910.95(b)(1), oreither of these standards; 2) if a violation or violations occurred ascharged by Complainant, was that violation or those violations \”willful\”in nature, as that term is used in Section 17(k) of the Act and definedby the Commission and the courts; and 3) if violative conduct onRespondent’s part has been proven, what penalties, if any, areappropriate in accordance with the provisions of Sections 17(a) and17(j) of the Act.The Citation issued by Complainant on July 9, 1979, reads in full asfollows:The violations described in this citation are alleged to have occurredon or about the day the inspection was made unless otherwise indicatedwithin the description given below.1a29 CFR 1910.95(a): Protection against the effects of noise was notprovided for employee(s) exposed to sound levels which exceeded thoselisted in Table G-16 of sub-part G of 29 CFR part 1910:a) The employee operating machine no. 299 in the Machine Shop wasexposed to noise, March 26, 1979.b) The employee operating machine no. 198 in the Machine Shop wasexposed to noise, March 26, 1979.c) The employee operating machine no. 398 in the Machine Shop wasexposed to noise March 26, 1979.d) The employee operating machine no. 389 in the Machine Shop wasexposed to noise, March 26, 1979.e) The employee operating machine no. 275 in the Machine Shop wasexposed to noise, March 26, 1979.f) The employee operating machine no. 581 in the Machine Shop wasexposed to noise, March 26, 1979.g) The three Sweepolet grinders in the Production Grinding Area wereexposed to noise, April 4, 1979.1b29 CFR 1910.95(b)(1): Employee(s) were subjected to sound levelsexceeding those listed in Table G-16 of subpart G of 29 CFR part 1910,and feasible administrative or engineering controls were not utilized toreduce sound levels:a) The employee operating machine no. 299 in the Machine Shop wasexposed to noise, March 26, 1979.b) The employee operating machine no. 198 in the Machine Shop wasexposed to noise, March 26, 1979.c) The employee operating machine no. 398 in the Machine shop wasexposed to noise, March 26, 1979.d) The employee operating machine no. 389 in the Machine Shop wasexposed to noise, March 26, 1979.e) The employee operating machine no. 275 in the Machine Shop wasexposed to noise, March 26, 1979.f) The employee operating machine no. 581 in the Machine Shop wasexposed to noise, March 26, 1979.g) The operator of the no. 16 Forge Hammer was exposed to noise March27, 1979.h) The helper assigned to the no. 16 Forge Hammer was exposed to noise,March 27, 1979.i) The operator of the no. 18 Forge Hammer was exposed to impact andcontinuous noise, March 27, 1979.j) The helper assigned to the no. 18 Forge Hammer was exposed to noise,March 27, 1979.k) The operator of the no. 12 Forge Hammer was exposed to impact noiseMarch 27, 1979.l) The three Sweepolet grinders in the Production Grinding Area wereexposed to noise, April 4, 1979.The combination of the above alleged violations (1a and 1b) affected theoverall gravity of possible illness and contributed to the seriousnature of the alleged violations.*Step 1: Effective personal hearing protection shall be provided andused by employee(s) as an interim protective measure.*Step 2: A written detailed plan of abatement leading to the completeabatement of this item shall be submitted to the Area Director. Such aplan shall : a) employ the use of qualified engineering personnel; b)include detailed engineering studies and their results; c) outline theordering of equipment and materials and completion of the design phase;and d) outline dates for the anticipated implementation of the plan.*Step 3: Feasible engineering controls and\/or administrative controls,shall be determined.*Step 4: Abatement shall be completed by the implementation of feasibleengineering controls and its effectiveness at achieving complianceverified. 90-Day progress letters are requested during the abatement period.The abatement requirements set forth in \”Step 1\” were required byComplainant to be accomplished \”immediately\”; those in \”Step 2\” withintwo months; those in \”Step 3\” within three months; and a one-year periodwas allowed for the implementation of feasible engineering controls oradministrative controls, as specified in \”Step 4.\”A total penalty of $2,000 was proposed by Complainant for all theasserted violations.The safety standards referred to in the Citation provide the followingrequirements (in pertinent part):29 CFR 1910.95 OCCUPATIONAL NOISE EXPOSURE(a) Protection against the effects of noise exposure shall be providedwhen the sound levels exceed those shown in Table G-16 when measured onthe A scale of a standard sound level meter at slow response. When noiselevels are determined by octave band analysis, the equivalent A-weightedsound level may be determined as follows:Table G-16–Permissible Noise ExposuresDuration per day, hours \tSound level dBAslow response8 \t906 \t924 \t953 \t972 \t1001-1\/2 \t1021 \t1051\/2 \t1101\/4 or less \t115(b) (1) When employees are subjected to sound exceeding those listed inTable G-16, feasible administrative or engineering controls shall beutilized. If such controls fail to reduce sound levels within the levelsof Table G-16, personal protective equipment shall be provided and usedto reduce sound levels within the levels of the table.Exposure to impulsive or impact noise should not exceed 140 dB peaksound pressure level.The dBA sound levels set forth in Table G-16 are further explained asfollows:When the daily noise exposure is composed of two or more periods ofnoise exposure of different levels, their combined effect should beconsidered, rather than the individual effect of each. If the sum of thefollowing fractions: C_1 \/T_1 + C_2 \/T_2 … C_n \/T_n exceeds unity,then, the mixed exposure should be considered to exceed the limit value.Cn indicates the total time of exposure at a specified noise level, andTn indicates the total time of exposure permitted at that level._Background._Respondent is engaged in the manufacture of pipe fittings and metalelbows at its Allentown, Pennsylvania, facility and produced a grossannual dollar volume of $17,146,000 during the fiscal year ending in1979. (Respondent’s answer to Interrogatory No. 25, Tr. p. 117).On March 22, 26 and 27, and April 4, 1979, noise monitoring wasconducted in Bonney Forge’s machine shop, forge shop and in theproduction grinding area by Michael P. McSherry an industrial hygienistemployed by Complainant’s Occupational Safety and Health Administration(\”OSHA\”). The results of the March 22 sampling were discarded because ofthe faulty operation of the noise dosimeters used on that date. Thus,the results obtained from monitoring on March 26 and 27, and April 4,1979, served as the basis for the issuance of the Citation chargingviolations of the noise standards. (Tr. pp. 834-835)._Summary of the Evidence__Regarding Item 1a of the__Citation._In item 1a of his Citation, Complainant charges that Respondent failedto provide protection for its employees who were exposed to levels ofsound which exceeded the levels set forth in the health standard.The industrial hygienist (hereinafter the \”Compliance Officer\” or \”CO\”)testified that he attached a pre-calibrated Gen-Rad dosimeter to theclothing of each of the employees whom he selected to be sampled fornoise exposure. The dosimeters were attached at the employee’s belt orplaced in his pockets, and a cord was run over his shoulder to themicrophone which was pinned at or near the lapel of his shirt. This,according to the CO, placed the microphone in the employee’s hearingzone. The dosimeter is designed to record all sound levels to which theemployee is subjected during the time it is operating, and it integratesthose sound levels with the duration of exposure. If it is properlycalibrated, and if the integration factor is correct, it produces areadout representing a percentage of the noise permitted by the healthstandard (Table G-16 of 29 CFR 1910.95(a)).The CO took samples of the noise encountered by employees inRespondent’s machine shop on March 26, in the forge shop on March 27,and in the production grinding area on April 4, 1979. The noise exposureexperienced by six machine operators was measured; three grinders wereequipped by the CO with dosimeters; and four forge hammer operators andforge operator helpers were subjected to the testing.According to the CO’s testimony, the dosimeter readings for the machineoperators were:Employee Diaz -298.6% of the permissible levelEmployee FeIguieras -247% of the permissible levelEmployee Joler -293% of the permissible levelEmployee Bachman -375% of the permissible levelEmployee Check -180.8% of the permissible levelEmployee Kuder -172.9% of the permissible levelAnd the forge hammer operators and helpers:Employee Reph -410.5% of the permissible levelEmployee McCormick – 386.1% of the permissible levelEmployee McFarland – 749.1% of the permissible levelEmployee Rodrigues – 725% of the permissible levelThe dosimeter readings reported for the grinders were:Employee Torok – 539% of the permissible levelEmployee O’Donnell – 414.6% of the permissible levelEmployee Wilt – 400% of the permissible levelThe CO testified that he also measured the noise exposure of each of thesame employees using a standard sound level meter (on the A scale atslow response). These readings for Diaz averaged 85 dBA to 103.5 dBA;Felguieras – 88 dBA to 104 dBA; Joler – 88 dBA to 101 dBA; Bachman – 91dBA to 104 dBA; Check -86 dBA to 96 dBA; Kuder – 83 dBA to 96.5 dBA;Reph – 93 dBA to 105 dBA; McCormick – 90 dBA to 109 dBA; McFarland 97dBA to113 dBA; Rodrigues – 98 dBA to 113 dBA; Torok – 97 dBA to 108 dBA;O’Donnell 80 dBA to 110 dBA; and Wilt – 80 dBA to 107 dBA. The 80 dBAreadings were actually listed as \”less than 80 dBA\” and were recordedwhen no one in the area was operating a grinder.CO McSherry used the dosimeter readings to calculate the \”average – orequivalent – dBA measurements for a full eight hour day,\” accounting forthe somewhat limited time periods when the dosimeters were in place onthe employees and operating. He testified that the time periods and thecalculated averages were:Diaz \t-410 minutes – \t99 dBAFelguieras \t-406 minutes – \t97.7 dBAJoler \t-400 minutes – \t99.1 dBABachman \t-395 minutes – \t100.9 dBACheck \t-398 minutes – \t95.6 dBAKuder \t-392 minutes – \t95.4 dBAReph \t-117 minutes – \t110.4 dBAMcCormick \t-339 minutes – \t102.3 dBAMcFarland \t-337 minutes – \t107 dBARodrigues \t-338 minutes – \t106.8 dBATorok \t-390 minutes – \t103.7 dBAO’Donnell \t-385 minutes – \t101.8 dBAWilt \t-390 minutes – \t101.7 dBAThus, according to the testimony, the dosimeters were in place on themachine shop employees for approximately six and one-half to sevenhours. Likewise, the grinders wore them for approximately six andone-half hours. Employee Reph \”left for the day\” after working onlyabout two hours, but the other forge hammer operators and helpers worethe dosimeters for approximately five and one-half hours.In the machine shop, metal castings are bored or drilled as they are fedinto the machines and lathes by the employee operators. The metal piecesare rotated within the machines through a cycle until the drilling orboring operation is complete and the finished product is ejected. Thusthe noise emitted by each machine varies as the process continues (e.g.,more or less pieces being machined at one time, or different sizedpieces involved), and the operator is subjected to the resulting noise,not only from the machine he is tending, but from all other machines inthe shop. Variations in the sound levels within the machine shop are, ofcourse, caused by the number of machines being operated at a given time.In the forge shop metal bars are heated in furnaces, transferred to theforge hammers by the helpers, and struck by hammer dies to form castings(elbows and other pipe fittings). Adding to the noise caused by theforge hammers as they descend and strike the heated metal is the\”constant roar\” of the nearby furnaces, and, not unlike the situation inthe machine shop, the operator and his helper are subjected to the noiseproduced by the operation of other furnaces and hammers in the area aswell as that from the furnace and hammer they are tending.The production grinders use hand-held grinding equipment to removeimperfections from forged castings which are placed on tables or on thefloor of the plant, depending upon the size and weight of the piece. Thenoise is uttered from the operation of the grinding tool itself and fromthe \”ringing\” of the metal casting as it is struck by the tool. Thelevel of the sound depends, to some extent, upon the size and shape ofthe casting and otherwise varies depending upon whether it is placedmore solidly on the floor or on a table. The degree of the sound levelin the area increases as the number of grinding operations increases,and each grinder is subjected to the noise caused by the otheroperations around him, as well as that produced by his own. Grinding isnot a continuous operation.According to the CO’s testimony, employees in the forge shop worehearing protective devices; those in the machine shop and the productiongrinding area did not. Thus a violation of the health standard codifiedat 29 CFR 1910.95(a) was charged, but that charge did not include theemployees who worked as forge hammer operators and helpers._Evaluation of the Evidence __Regarding Item 1a of__the Citation_In order to require that Respondent provide \”protection against the[adverse] affects of [the] noise exposure [of its employees]\”Complainant must first prove, by a preponderance of the evidence, thatthe unprotected employees were in fact exposed to \”sound levels [which]exceed those shown in Table G-16 when measured on the A scale of a soundlevel meter at slow response.\” 29 CFR 1910.95(a). Respondent’s counselargues at length that Complainant has failed to carry his burden ofproof.[[2\/]] I agree.[[3\/]]Respondent’s counsel attacks the \”integrity of the measurements\” of thesound levels made by Complainant’s agent and attacks his credibility onthe basis of his inexperience and ineptness. I agree that grosslyinadequate instruction by his employer and the apparent lack ofconviction on the CO’s part cause a resulting absence of persuasion.[[4\/]]The record supports Respondent’s assertions regarding the CO’s lack ofexperience in the field of noise measurement and control. He attended aliberal arts college and had no exposure to the study of engineering ofany kind. He did not work in the engineering field prior to assuming hispresent position, and has attended no courses specifically designed totreat the subject of noise measurement or noise control. A three-weekOSHA training course which he attended in 1977 might have includedtraining in the general subject of noise measurement, but he \”didn’tremember\” whether it did. He \”didn’t know\” how much training he has hadin measuring sound levels, but it was \”maybe a day;\” he \”really didn’tremember.\” Asked how much time he had spent studying the operation ofsound level meters and audio dosimeters prior to the time of theinspection in question, he replied: \”I really don’t know;\” and whenqueried about whether he remembered \”anything [from the OSHA courses]about taking noise measurements,\” he said: \”I’m not sure.\”[[5\/]]These candid admissions by Complainant’s principal witness lead to theobvious question whether his testimony alone is sufficient to supportthe charge asserted, _viz_., that Respondent’s employees were exposed tosound levels which exceeded those set forth in the cited healthstandard. Clearly that is the predicate upon which Complainant has basedhis assertion that protective equipment should have been provided andthat engineering or administrative controls should have been instituted.Viewing this witness’ testimony critically to determine whether itamounts to a preponderance of the evidence leads to the conclusion thatthere is considerable doubt about Respondent’s failure to comply withthe safety standard — considerable doubt regarding the level of thenoise at the various work stations.In no instance did the CO perform a full shift sampling. Expertacoustical engineers may disagree that an eight-hour sampling isnecessary for accuracy, but the expert witnesses who testified here seemto think that it is. [[6\/]] Likewise, the CO made no adjustments for theimpact noise (as opposed to continuous noise) at the work stations. Whatconsideration should have been given for that factor is unclear fromthis record, but its having been ignored entirely casts further doubtupon the ultimate findings of the CO, and those findings are the solebasis for the issuance of the citation.[[7\/]]According to the testimony, an audio dosimeter records and reads out thesound levels it encounters in percentages of the allowable limit (TableG-16). However, the time-weighted average reported may vary dependingupon whether the integration of the dosimeter begins at 90 dBA, or somelower or higher level. The record here is unclear, but apparently somedosimeters integrate at 80 or 85 decibels. The CO did not know at whatlevel — or levels — the dosimeters he used would integrate. Heconceded that the integration point may well have been below 90 dB. Hestated that \”the integrator drops off sharply below 90 dBA.\” If, asmight well have been the case, the integration point of the dosimetersused by the CO was less than 90 decibels — perhaps 85 — the resultingreadouts would have been in terms of a percentage of 85, not 90 decibelsas provided for in the health standard. The error could, of course, becrucial to a determination whether the standard was violated because ofexcessive sound levels.[[8\/]]Complainant’s agent attempted to bolster the accuracy of the dosimeterreadings by the submission of his findings when using the sound levermeter. While the audio dosimeters were in place on the employee’s, theCO held the sound level meter (presumably in the hearing zone of theemployee) and read the \”low point and the high point on the fluctuatingdial.\” He observed these readings for \”three or four seconds\” each timeand then recorded them. He did not record the lengths of time the meterindicated the higher levels or the lower levels. Nevertheless, heconcluded that his findings, made and recorded as the result of thesound level meter readings, confirmed the accuracy of the dosimeterreadouts.[[9\/]]As pointed out above, the sound levels at the several work stationsvaried measurably from time to time depending upon which machines wereoperating, how many machines were in use, what castings or other partswere being machined, whether forge hammers and furnaces operatedsimultaneously, whether pieces were ground on tables or on the floor,etc. Likewise, the frequency of the noise emitted as the differentoperations were conducted — as well as the volume and intensity of thesound — varied considerably.[[10\/]] Despite that fact, the CO did notallow differing error tolerance factors when calculating his findings.He stated that he allowed an error factor of 2 dBA throughout. Themeasurement of high frequency noise requires the application of agreater error factor, according to Respondent’s expert witness.[[11\/]]Complainant’s principal witness was somewhat less than scrupulousregarding the actual sampling of noise exposures at the employees’workstations. He presumably knew how long each audio dosimeter wasattached to the individual employee, but he was quite uncertain aboutwhere the employee spent that period of time. If a machine was down, orthe employee left his work station for other reasons, went to lunch orto the rest room, he made no calculated adjustment.[[12\/]] He purportedto measure the full-shift noise level at certain work stations, –atleast by interpolation– but his uncertainty about where each employeewas during the time he was monitored leaves further doubt about thepurity of his reported findings. Complainant’s expert witness, DoctorClayton H. Allen, conceded as much when he stated that \”their[employees’] positions throughout the day determine the relativeimportance of these [noise] sources.\”[[13\/]] One single fact to begleaned from this record is that the CO could not be certain that thedosimeters he relied upon collected noise emitted by the operationsperformed by the employees at their work stations and that noisealone.[[14\/]]Further indications of a lack of exactitude appear in the record. The COconceded that three of the six dosimeters he used in the machine shopdid not come within tolerable limits upon calibration both before andafter their use.[[15\/]] He was not certain that the dosimetermicrophones remained in position on the employees throughout thesampling periods; he said, \”I don’t remember.\”[[16\/]] Improperpositioning of the audio dosimeter microphone can materially affect thesampling results, according to the testimony of Respondent’s expertwitness.[[17\/]] The CO lacked total conviction about the time of theemployee’s exposure to the offensive sound levels.[[18\/]] This, ofcourse, is crucial inasmuch as the health standard relied upon byComplainant is based specifically upon the excessive sound levels andthe time they are endured by an employee._Denouement _Complainant’s burden of proof is clear. He must convince the fact finderby a preponderance of the evidence that the asserted violation issupported by the facts adduced. The burden of persuasion has not beenmet here. I am not persuaded by the testimony that Respondent’semployees were subjected to sound levels in excess of those prescribedin the health standard promulgated by Complainant. The record affordsmuch room for doubt.As stated above, if Complainant fails to prove a violation of the healthstandard codified at 29 CFR 1910.95(a), _viz_, that \”. . . the soundlevels [at Respondent’s workplace] exceed those shown in Table G-16 whenmeasured on the A scale of a standard sound level meter at slowresponse\” there is no requirement that \”protection against the effectsof noise exposure be provided\” and no need for the utilization of\”feasible administrative or engineering controls.\”_FINDINGS OF FACT_A preponderance of the probative evidence of record, taken in itsentirety, compels the following findings of fact:1. Respondent, a corporate entity, is engaged in the manufacture, saleand distribution of pipe fittings and similar metal products at itsBonney Forge Division, Allentown, Pennsylvania, where it grossed inexcess of $17 million during fiscal year 1979.2. Goods produced by Respondent at the Allentown plant are shipped ininterstate commerce.3. On March 26 and 27, and April 4, 1979, Complainant’s agent monitoredthe sound levels at various work stations in the machine shop, forgeshop and production grinding area at Respondent’s Bonney Forge facility.4. The means and methods employed by Complainant’s agent in thecollection and recording of the sound level data at the various workstations were imprecise, confused, fraught with inconsistencies andconsequently were rebuttably unreliable.5. Complainant failed to sustain his burden of proof, and hisconclusions regarding violations by Respondent at the times and placesasserted are rejected as speculative._CONCLUSIONS OF LAW_1. Jurisdiction of the parties and of the subject matter herein isconferred upon the Occupational Safety and Health Review Commission bysection 10(c) of the Act.2. At all times relevant hereto, Respondent was an employer engaged in abusiness affecting commerce within the meaning of Section 3(5) of theAct and as such was subject to the requirements of Section 5(a)(2) ofthe Act.3. Respondent did not violate Section 5(a) (2) of the Act on or aboutMarch 26 and 27, and April 4, 1979, by its failure to comply with theprovisions of the health standard promulgated by Complainant andcodified at 29 CFR 1910.95(a), as charged by Complainant in hisCitation, dated July 9, 1979.4. Respondent did not violate Section 5(a) (2) of the Act on or aboutMarch 26 and 27, and April 4, 1979, by its failure to comply with thehealth standard promulgated by Complainant and codified at 29 CFR1910.95(b)(1), as charged by Complainant in his Citation, dated July 9,1979, inasmuch as the measured sound levels at the various work stationsdid not exceed those levels listed in Table G-16 of subpart G of thestandard codified at 29 CFR 1910, and feasible administrative orengineering controls were thus not required._ORDER _Upon consideration of the foregoing findings and conclusions, it ishereby ORDERED that the Citation issued to Respondent by Complainant onJuly 9, 1979, is VACATED, and. accordingly no penalty is assessed.BENJAMIN G. USHERJUDGE, OSHRCDated: May 13, 1982New York, New YorkFOOTNOTES:[[1]] Section 1910.95(a) and Table G-16 provide, in relevant part:Protection against the effects of noise exposure shall be provided whenthe sound levels exceed those shown in Table G-16 when measured on the Ascale of a standard sound level meter at slow response. . . .* * *TABLE G-16–PERMISSIBLE NOISE EXPOSURES [[1]]Duration per day, hours Sound level dBA slow response8…………………………………………………………..906…………………………………………………………..924…………………………………………………………..953…………………………………………………………..972………………………………………………………….1001 1\/2 ……………………………………………………1021………………………………………………………….1051\/2……………………………………………………….1101\/4 or less……………………………………………..115[[1]] When the daily noise exposure is composed of two or more periodsof noise exposure of different levels, their combined effect should beconsidered, rather than the individual effects of each. If the sum ofthe following fractions: C_1 \/T_1 + C_2 \/T_2 C_n \/T_n exceeds unity,then, the mixed exposure should be considered to exceed the limit value.C_n indicates the total time of exposure at a specified noise level, andT_n indicates the total time of exposure permitted at that level. . . .[[2]] Section 1910.95(b)(1) provides:When employees are subjected to sound exceeding those listed in TableG-16, feasible administrative or engineering controls shall be utilized.If such controls fail to reduce sound levels within the levels of TableG-16, personal protective equipment shall be provided and used to reducesound levels within the levels of the table.[[3]] There were no background noise readings obtained for employeeBachman. We have rounded off the dosimeter readings to the nearest wholepercent, and the equivalent dBAs to the nearest dBA. The industrialhygienist’s calculations of the equivalent dBAs were confirmed at thehearing by an expert witness, an occupational health engineer employedby OSHA (see note 6 _infra)_.[[4]] At the time of the inspection, OSHA’s Industrial Hygiene FieldOperations Manual required full-shift sampling. Full-shift sampling maybe needed to establish overexposure in marginal cases, or to delineatethe full extent of a noise problem, or to shed light on the feasibilityof engineering or administrative controls. However, as the record inthis case indicates, full-shift sampling is not needed to proveoverexposure when overexposure is shown by a partial day’s measurements.[[5]] The judge was also concerned that the industrial hygienist did notremove the dosimeters or adjust the readings for noise encounteredduring lunch or rest breaks. However, the standard limits total dailynoise exposure and there is no apparent reason why exposure during lunchor rest breaks should not be counted. _Cf. Gilles & Cotting, Inc.,_ 3BNA OSHC 2002, 2003, 1975-76 CCH OSHD ? 20,448, p. 24,424 (No. 504,1976) (a finding of employee access to a hazard can be predicated onaccess during personal comfort activities).[[6]] The Secretary buttressed her industrial hygienist’s qualificationswith testimony from OSHA’s leading noise expert, Dr. John Barry, theoccupational health engineer who wrote the noise measurement chapter ofthe Industrial Hygiene Field Operations Manual (IHFOM). See also_Collier-Keyworth Co.,_ 13 BNA OSHC at 1221, 1986-87 CCH OSHD at p.36,520. Dr. Barry testified that he had worked with industrial hygienistMcSherry and believed that his noise survey in this case was valid andadequate for compliance purposes.[[7]]He stated that the readings \”greater than 250% are probably validindicators of overexposure . . . taking into consideration the type ofnoise . . . and the tolerances of the instruments for that type ofnoise.\” Again, at another point in his testimony, he proposed reducingthe readings by a large error tolerance factor of 6.5 dBA to allow forhigh frequency noise. The usual factor of 2 dBA is applied bydisregarding readings of 132% or less. A 6.5 dBA factor would be appliedby disgarding readings of 250% or less. (See 29 C.F.R. ? 1910.95, TableA-1, where a 250% dose is equated with an 8-hour TWA of 96.6 dBA).[[8]] After the citation issued, Bonney Forge retained expert Botsfordto evaluate the noise situation in all three areas of the plant and todetermine whether engineering controls would be feasible. In performingthis study, the expert made some noise measurements at variouslocations, including the machine shop. His purpose was not to determineemployee noise doses, but his testimony on the issue of feasibilityreflected a belief that the employees in the machine shop were exposedto excessive noise levels. Specifically, Bonney Forge’s expert testifiedthat certain engineering controls for the machine shop would not befeasible because they would not reduce the noise to within permissiblelimits. He also stated that controls in the forge shop and theproduction grinding area would not reduce the noise to within Table G-16levels.[[9]] Although the industrial hygienist understood the operatinginstructions to say that the instruments began reading at 90 dBA, OSHA’sleading noise expert Barry testified that dosimeters of the type used byOSHA have a \”threshold\” or \”gate\” of 89 dBA. This expert and expertBotsford differed as to how significant this was, but both indicatedthat counting noise from 89 dBA upward could produce elevated readings.[[10]] Dosimeters set at 90 dBA have an inherent error factor of 2 dBA.Thus, a 90 dBA reading of 100% means that the actual noise level couldbe as high as 92 dBA or as low as as 88 dBA. To compensate, OSHA issuesa citation only if a reading exceeds 132%, which corresponds to 92 dBAand places the actual noise level no lower than 90 dBA. See _Collier-Keyworth ,_ 13 BNA OSHC at 1228, 1986-87 CCH OSHD at p. 36,528; also,Table A-1 of 29 C.F.R. ? 1910.95. If any dosimeter in this case was setto begin reading noise at 89 dBA rather than 90 dBA, the instrumentshould be treated as having a 3 dBA error factor (1 dBA higher thanusual). Therefore, to justify a citation, its reading should have beenslightly above 150% (which corresponds to 92.9 dBA, according to Table A-1).[[11]] Impact noise is noise produced by the impact of two objects, suchas a hammer striking metal during forging operations. The standarddefines as \”continuous\” (and therefore includes) noise from a series ofimpacts having peaks less than one second apart. See_Collier-Keyworth,_13 BNA OSHC at 1223, 1986-87 CCH OSHD at p. 36,522. The impact noisethat is excluded from the standard is either individual impacts or aseries of impacts that are more than one second apart.[[12]] Expert Barry noted that the impacts from the forge shop did notregister as background noise in his graphic depictions of the machineshop noise. Similarly, the Secretary’s other expert, an acousticalconsultant, Dr. Clayton Allen, testified that the hammer impacts did notadd even a single decibel to the sound level meter readings he took inthe machine shop; and that there was no measurable difference betweenreadings taken while the forge shop hammers were operating and while thehammers were stopped. The testimony of two witnesses, the industrialhygienist and machine shop employee Kuder, also supported the view thatthe hammer impacts did not contribute significantly to noise exposuresin the machine shop.[[13]] Initially, the expert posited that impact noise from the forgeshop was not \”too big a factor in exposure\” in the machine shop.However, he almost immediately revised his testimony. He testified thatthe impact noise levels spilling over into the machine shop while it wasshut down were \”certainly sufficient to elevate dosimeter readings.\” Oncross-examination, the witness changed direction a second time,testifying that impact noise \”may affect\” readings but the noise createdin the machine shop itself was the \”cause of any overexposure.\” Thewitness again shifted on redirect examination, testifying this time thatimpact noise would elevate the dosimeter readings perhaps \”two to threetimes.\”[[1\/]] Section 5(a)(2) of the Act requires:Each employer-***shall comply with occupational safety and health standards promulgatedunder this Act.[[2\/]] Commission Rule 73(a); 29 CFR 2200.73(a). See also, for example_B & B Insulation, Inc._ v _OSHRC,_ 583 F.2d 1364 (5th Cir. 1978)[[3\/]] This conclusion is reached after reading, again and again, the3,000 plus pages of the transcript and almost 100 pages of briefs.[[4\/]]There is no question raised regarding his truthfulness, and Ispecifically note here that he testified to facts which he verilybelieved to be true.[[5]] See Tr. pp. 956, 966, 967, 1065, 1066, 1336, 1337.[[6\/]] See Tr. pp. 520, 521, 522, 524, 835, 1113, 115[[7\/]] See Tr. pp. 844, 1076, 1205, 1206; 1177, 1178, 2858, 2859[[8\/]] See Tr. pp. 1192-1193, 1361, 2913-2916, 2926-2927.[[9\/]] See Tr. pp. 898, 1341-1342.[[10\/]] See Tr. pp. 854, 1067.[[11\/]] See Tr. pp. 2863, 2866.[[12\/]] Perhaps the lunch area and rest room were quieter, but therecord does not disclose that, and it cannot. be presumed.[[13\/]] See Tr. pp. 554-555.[[14\/]] Respondent’s counsel argues that there may have been adeliberate tainting of the CO’s findings:\”He [the CO] could not verify that the noise levels shown on thedosimeters which were used to make the measurements during thisinspection recorded only the noise in the Respondent’s work place (T-1169) and excluded horseplay or tampering with the instruments becausehe was too busy with other things to watch the employees (T- 997, T-1074to 1076, T-1133, T-1318). He admitted that compressed air was used inthe plant (T-1071), that ‘it could happen’ that employees would takethese devices and make noises in the dosimeters (T-1072) that ‘it couldhappen’ that employees might want to make the company look bad (T-1073),and that one thing which would do this would be an OSHA noise citationagainst the company (T-1073), but he could offer no assurance that itdidn’t happen while his measurements were being made because: ‘I didn’twatch all the employees every minute of the day’ (T-1072). Indeed, herather candidly conceded that ‘I wasn’t there for the entire shift thatthe employees were there’ (T-1113).\” (Brief, p. 28)There appears to be no support in the record for such a conclusion, andcounsel’s intense cross-examination of Complainant’s witnesses failed toestablish it as fact. I would discount it as a probability.[[15\/]] See Tr. pp. 1171-1172.[[16\/]] See Tr. pp. 1069, 1071, 2960.[[17\/]] See Tr. p. 1307 and Ex. R-3.[[18\/]] See Tr. p. 1079-1080.”