T&MW Goes to the Calibration Lab

-December 01, 1997

12F297.gif (24108 bytes)
Figure 1. Hewlett-Packard recommends a Fluke 5700A calibrator for calibrating the 34401A DMM. (Courtesy of Fluke.)

Maintaining accurate measurements requires that you periodically have your test equipment checked and calibrated. Many firms, lacking the resources for an in-house calibration lab, send their test equipment to independent labs. Cal labs can perform checks, calibrations, and repairs on a range of test equipment including DMMs, oscilloscopes, power meters, and spectrum analyzers. Labs can perform full or partial calibrations to your specifications.

To learn firsthand how to work with a cal lab, I set out to find a lab to calibrate a Hewlett-Packard 34401A 61/2-digit DMM. The meter hadn’t been calibrated in more than five years. What I learned is that there are many ways to calibrate a DMM, and if you don’t specify what you want in a calibration, you may not know what services you’re paying for.

I began by speaking with several engineers, some of whom are metrologists and some of whom are test engineers that use outside calibration labs. From those discussions, I developed a list of questions to ask cal labs before choosing one. I needed to ask about the labs’ equipment and procedures. I sent the list to two labs. (See “Questions for a Cal Lab”)

I also studied HP’s 34401A service guide so I’d know the calibration procedure and which equipment the lab should use. If a lab does not use the manufacturer’s recommended calibrator, it must substitute equipment of equal or better uncertainty. In my case, both labs I contacted had a Fluke 5700A, HP’s recommended calibrator for all functions except frequency. HP also recommends using an HP 3325A counter for verifying the frequency measurements of the 34401A1. Figure 1 shows a typical calibration setup with a Fluke calibrator and an HP 34401A.

Verify Traceability
Just knowing which calibrator a lab would use wasn’t enough. I wanted to verify that the calibrator was properly calibrated with equipment traceable to NIST. A calibration certificate that you get from a lab should state that the equipment used to calibrate your instrument is traceable to NIST or some other national lab. A lab should keep records of its equipment’s calibrations on file, and the equipment’s calibration report number should appear on the certificate. Before choosing a lab, get a sample certificate.

Having received sample certificates and answers to my questions from two labs, I decided to use Essco Calibration Labs in Chelmsford, MA; I visited Essco on September 16. Upon receiving the meter and letting the meter’s temperature stabilize, an Essco technician checked the meter in only those ranges that HP specifies for a quick check. The check told the technician that the AC volts, resistance, and DC current functions were out of tolerance and required adjustments.

The technician then performed a zero adjustment on each meter function, after which he recorded the reading for each function and range. You’ll find those results in the Measured Reading column of Table 1. After performing a zero adjustment, some of those ranges then fell within tolerance. That’s why, for example, the measured values for resistance fall within their tolerances, but the Adjust column indicates that the technician adjusted the gain on those ranges.

12f2tab197.gif (65125 bytes)

After checking the meter and adjusting, where necessary, the meter’s DC and AC voltage, DC and AC current, and four-wire resistance ranges, the technician declared the calibration complete. I asked why he didn’t check the frequency function. He hadn’t checked it because the lab’s datasheet didn’t call for a frequency calibration. Essco does not include frequency calibration in the service I requested but will calibrate for frequency on request. If you do request a check of frequency accuracy, the technician will use the Fluke 5700A, not the recommended HP 3325A counter.

The technician also didn’t calibrate the meter’s two-wire resistance function. To perform calibration on two-wire resistance, you must add 200 mV to compensate for the resistance of the lead wires. Essco performs two-wire calibrations on request and charges for the service. According to HP’s R&D project manager Scott Stever, a four-wire calibration is all you need.

When I was satisfied with the calibration results, the technician placed a calibration sticker on my meter. The sticker’s information includes:

  • calibration date;

  • calibration due date;

  • technician’s name;

  • instrument manufacturer, model number, and serial number;

  • my company’s name and the lab’s customer number;

  • the lab’s internal test report number in both print and bar code (the bar code also appears on a label placed on the lab’s receipt); and

  • the words “traceable to NIST.”

12F2FG2A.gif (58110 bytes)
Figure 2. A calibration certificate should include which equipment was used and give a calibration report number that’s traceable to NIST. The certificate should also include which functions were out of tolerance when the lab receives an instrument.

Who Decides?
Figure 2 shows the calibration report for my DMM that I obtained from the lab. The bottom of the report states NIST traceability and states that the calibration conforms to other quality standards. The list of calibration standards shows a report number for the calibration of each standard. In this case, Essco has its own standards lab and calibrates its 5700A in-house. In addition, the calibration certificate also states the last and next calibration dates.

Besides receiving a calibration certificate, I also received a test report. Table 1 shows the results of the 34401A calibration.2 I found that the meter still met many of its tolerance specs even after five years. I don’t use the meter every day, so finding some ranges still within tolerance was no surprise.

You’ll need to decide whether to instruct the lab to fully adjust the meter at each calibration interval or to leave the decisions to the technician. Letting the technician decide saves you money because you won’t pay for adjustments you don’t need. But not calibrating every range may increase your risk that the instrument will drift out of tolerance before the next calibration date. You must decide what’s important and instruct the lab accordingly.

HP recommends that all ranges and functions be adjusted at calibration time to ensure maximum performance. According to Essco’s quality systems manager David Donnelly, only about 1% of Essco’s customers request a full adjustment of all ranges at every calibration interval. About 50% of his customers request that Essco provide calibration data. The remainder of Essco’s customers just want to know that their instruments are in calibration upon leaving the lab. Some customers use the calibration data to track an instrument’s stability and don’t want to alter an instrument during a statistical process control (SPC) study.

If you use an instrument 24 hours a day in an automated test station and the meter measures DC voltage only, then you may decide not to have the lab adjust every range. You may decide that the other ranges need adjustment only when out of tolerance, but the DC voltage function needs adjustment every time even if it’s still in tolerance.

If you use an instrument in an engineering lab, you never know which functions will be used. Therefore, your best bet is to have the lab adjust all functions and ranges at every calibration interval. If you don’t adjust every range at calibration time, then the calibration clock—the time since last adjustment—will differ within your instrument.

Getting the calibration you need is important, but there’s more. You should know if the lab keeps records of your instrument’s past calibrations. Assume I had brought my DMM to the lab every year. I’d have five years of calibration history that I could use to decide to change the calibration interval. Given the test results, I might decide that a two-year calibration interval is adequate.

Proof in the Documents
In some industries, engineers need documentation proving that an instrument was within tolerance throughout the calibration interval. If you require such proof, you should request that the lab provide you with incoming and outgoing calibration data. If your instrument gets heavy use, you may find that data from a calibration indicates that you change your calibration interval. If you see the instrument getting close to the tolerance limits, you may decide to shorten the calibration interval.

To get the calibration history of an instrument, a cal lab must record an instrument’s incoming calibration data before making any adjustments. Some labs will do that as part of the calibration fee, while others will charge extra for the service. Essco will take incoming calibration data on request, so you must ask for them. If those data aren’t important to you, then the lab will tell you which functions were out of tolerance before making adjustments. That’s the service that Essco performed on my DMM. Because three ranges were out of tolerance following the quick check, the technician checked the OUT box in the certificate and then described which functions needed adjustment.

So what does this calibration service cost? Essco charges $47 for the service, which takes about 30 minutes to complete. I received the calibration certificate and datasheet and my calibrated meter. If you want Essco to check and record the readings of every range before making adjustments, expect to pay another $47 because the lab considers that a second calibration.

Remember that the numbers in Table 1’s Measured Reading column were taken after the zero adjustments and therefore don’t count as incoming readings. In contrast, the other lab I had contacted (Tekserv of Chelmsford, MA) quoted $65 for a calibration that included recording of incoming readings. T&MW

1. HP 34401A Multimeter Service Guide, Hewlett-Packard, Loveland, CO, March 1992.

2. The test report has columns for measured and adjusted readings, function, range, and tolerance in percent. From the range and tolerance values, I calculated the high and low limits. I also listed which ranges required adjustment. On the actual test report, the ranges that didn’t require adjustment are marked “N/A.”

Calibration: Philosophy In Practice, Fluke Corp., Everett, WA. ISBN: 0-9638650-0-5. 1994.


Questions for a Cal Lab

Q: What equipment will you use to calibrate the DMM? Is the calibration equipment traceable to NIST?

A: The lab should use the manufacturer’s recommended equipment. If the lab doesn’t have the exact models, it can substitute equipment that meets the manufacturer’s uncertainty requirements for calibration equipment.

Q: What is the test uncertainty ratio of your calibrator to my UUT?

A: The lab should specify the test uncertainty ratio of the calibration equipment’s uncertainty to the uncertainty of your equipment. That ratio is typically 4:1.

Q: Does your lab have any quality accreditations or registrations?

A: The lab should tell you which quality standards it complies with. Expect to hear designations like ISO 9000 and ANSI/NCSL Z540-1994. Some labs may have accreditations from the American Association for Laboratory Accreditation (A2LA, Gaithersburg, MD) or from the National Voluntary Laboratory Accreditation Program (NVLAP, Gaithersburg, MD).

Q: How much will the calibration cost?

A: The lab should specify what you’ll get for your money. You may have to be more specific. See the next two questions.

Q: Do you normally test the instrument as received as part of the cost? If not, what is the additional cost?

A: Some labs will perform a functional test on the UUT prior to checking the instrument’s calibration. Some will just start checking the calibration. You must specify what you want.

Q: Does the cost include adjustment of all DMM functions regardless of as-received condition? If not, what is the additional cost?

A: Some labs will tell you that they will adjust all functions and ranges for minimum uncertainty. Others may say that the technician decides which ranges to calibrate. If you need all ranges adjusted at every calibration, then tell the lab.

Q: How long will my DMM be out of service?

A: A lab can turn around your instrument in a day or a week, depending on the level of service you require. Take into account any shipping time, too. Some labs have mobile facilities and will come to your location.

Q: What will you do if you find the instrument is out of calibration?

A: A lab may simply test your equipment, or it may calibrate every function. You should specify what you want the lab to do with equipment that’s out of tolerance. If you need the out-of-tolerance data, then request them. The calibration certificate should specify that the equipment was received out of tolerance.

Q: What information do you include in the calibration report? Does the report comply with any quality standards? If so, which standards?

A: The calibration report should include the equipment used, its calibration date, calibration due date, and test report number. It should also include the temperature and humidity of the lab at the time of calibration. The report should include the results of the calibration.

Q: Do you retain the before and after readings so I can track the history of the DMM?

A: The lab should tell you how long it maintains calibration records. If the lab keeps those records on computer files, it should maintain off-site backups. If the lab uses calibration management software, then it should be able to track incoming calibration data and recommend lengthening or shortening the calibration cycle.

Q What are the environmental limits of your lab and what do you do if the lab’s temperature or humidity is out of tolerance?

A: The lab should provide you with temperature and relative humidity ranges. The lab should have a written procedure for how often it checks environmental conditions and what it does when the conditions are out of tolerance. The lab should cease operations and specify how long it waits before resuming.—Martin Rowe

Loading comments...

Write a Comment

To comment please Log In