When MIL-STD-461 revision “G” was issued, I believed that the answer to many of my measurement equipment calibration issues had been solved. Right there in the “Calibration of measuring equipment” paragraph the requirements clearly stated:
After the initial calibration passive devices such as measurement system antennas, current probes, and LISNs, require no further calibration unless the device is repaired. The measurement system integrity check in the procedures is sufficient not determine acceptability of passive devices.
This change allowed me to take several items off the periodic calibration schedule, saving a lot of money for this service, at an accredited calibration laboratory. Now we did not need so many of the same items to support operations while an item was “out-for-calibration. This was virtually a “no brainer” to adopt because we were doing the system integrity checks anyway.
A secondary benefit was the reduction of “findings” associated with calibration review during each accreditation audit. It seemed that we were always having issues with the calibration program. Based on discussions with other laboratories, I found that many calibration issues were discovered during the audit of most laboratories. A friend that was known to embellish the details, told about his “weather rock” calibration issue. He related that during the audit the reviewer noticed the “weather rock” on his desk. For those of you unfamiliar with a weather rock, it is a stone suspended by a string that identifies weather conditions—when it’s wet, it’s raining; when it’s white, it’s snowing; when it’s moving, it’s windy and so on. The auditor asking what it was, asked “How wet is wet?,” and how could we prove it with a NIST traceable certificate. Listening to him did make all of us realize that we had encountered points like that.
Since this calibration process was available and beneficial, I quickly adopted this approach into our calibration program, listing not only the items identified in the standard, but added all other passive devices such as cables, attenuators, terminators, directional couples, voltage probes, and any other passive device that was used where a system integrity check was part of the process.
The system integrity check results were part of the test report, so we maintained that process and as normal, the equipment used during the test was listed in the test report. This took care of the documentation concerns and at the next audit, the process was accepted until a reviewer that we had accomplished a test per MIL-STD-461F which did not support the lack of periodic calibrations.
So now we were faced with having to maintain the existing process of periodic calibrations and use the expanded system integrity check when a test per revision “G” was performed. Instead of a savings, we encountered added costs in the time spent dealing with the expanded integrity check with each test. As time went on, we started seeing less work to the older standard revisions allowing the formal calibrations to be reduced. Maybe a cost savings would be realized after the transition era was complete.
The Cable Disaster:
Naturally when things start to improve something will raise its ugly head to apply a negative offsetting any gains.
During an RS103 test, an oddity came up where the amplifier output power was unable to achieve the required test level at a couple of frequencies as shown on the chart in Figure 1. The inability to produce the field occurred at ~13 GHz (~8 dB loss) and ~16 GHz (~3.5 dB). Debugging found that the amplifier to antenna cable had a minor pinch, creating an impedance mismatch along the transmission path. A physical examination of the cable did not reveal the pinch point, because the jacket appeared to be normal until pressure was applied while feeling along the cable length after the cable was confirmed to be defective.
This problem was easily found before testing was completed, so there was no issue until we realized that this cable was also used during the RE102 testing a couple of days earlier. This helped us understand that the damage had happened within the last couple of days. However, the frequencies with high loss were at a signal integrity check frequency as pointed out in the chart. This meant that the RE102 emissions measurements were off by at least 8 dB at the worst-case frequency, and we did not have any idea when the cable damage occurred. A review of the RE102 test results showed that the test article complied with the limit when we added the loss to the measurements, but the margin was low.
Now, we had to figure out what tests could have been affected and determine if the results would change the item from compliant to non-compliant or unknown. This challenged our record keeping process because the last time the “passive” cable had a full sweep was almost two years ago. This knowledge limited the number of checks to projects that had used that cable in the last two years.
Our records used the test report listing of equipment used for each test, meaning that each test report had to be opened and searched for that cable ID. The result was 38 projects where that cable was listed. When looking into the specific tests we thought that if that cable was used for RE102 and RS103 we would get the data from the RS103 sweep to identify when the cable damage was present. Well, that did not work out, because if the cable was used between the amplifier and the antenna, we would see the power loss. If the cable was placed between the signal generator and the amplifier, the program would simply increase the signal generator output to compensate. The report did not identify where the cable was connected, and none of the projects showed the drop out during the RS103 test to limit the time span of potential undetected failures.
Examination of the 38 reports revealed that only six showed measurements that could push the RE102 emissions amplitude above the limit when the damaged cable loss was added to the existing data. Although the identification process was tedious, the next step was painful—contacting the customer to arrange for the test item to be re-tested. This had to take place quickly, hoping that a test item would be available and that the sales had not resulted in shipments.
Over the next three weeks, we had completed testing of the six samples with only one producing over-limit emissions. When we compared the new test sample to the original tested configuration documented in the report, we found that the sample incorporated three new circuit board revisions and the emissions were traced to the board changes.
What were the savings obtained by taking the cables off the calibration list? Did the savings pay for the time spent resolving the issue? Did the customer confidence cause them to complain with their feet, by going elsewhere? Although we were fortunate in resolving the issue without having to recall hundreds or thousands of shipped products, it could have been much worse or even disastrous to our financial health. I felt that our team’s discipline to document thoroughly provided the means to uncover the details and resolve the problem.
Changes were immediately made to prevent this long-term exposure—cables are too easy to sweep and confirm that it follows the initial calibration. It may not be classified as a calibration, but it supplements the system integrity checks on a routine basis.
Other Passive Equipment:
The cable disaster work associated with clearing the corrective action report (CAR) incorporated a review of the complete passive calibration process. We needed to be sure that other items could successfully go through the system integrity check and have undetected problems. I will not attempt to create an all-encompassing list, but I provide some food for thought when you look at your processes.
- Passive antennas: There is a certain amount of risk with some of the passive antennas typically used for MIL-STD-461 testing because the system integrity check removes the antenna when a known signal strength is applied. The system integrity check supplements the check without the antenna by emitting a signal from the stub radiator. The standard calls for doing a physical inspection of the antenna and radiating from the stub at the antenna highest measuring frequency (translated—radiate at the highest frequency that you use the antenna), and “verify that a received signal of the appropriate amplitude is present.” This may be adequate, but damage could be present that is not obvious to the physical inspection. A fractured junction at the element connection, a loose mount, or minor corrosion could attenuate the signal and go undetected because a signal is present.A way to reduce the risk is the placement of the stub radiator and antenna at the same location each time the integrity check is made. Using the same signal source amplitude should then produce the same amplitude (within tolerance) confirming proper operation. Recording the data allows for frequent monitoring of the antenna preventing a long-term risk of unknown results. Although the discussion in the standard is located in the RE102 section, consider this kind check for the RE101 loop or other passive antennas even though the resistance check should be sufficient. The cost of antenna calibration typically is between $700 and $1,000, so it is desirable to reduce the periodic calibration intervals. In making the decision consider ways to keep the checks adequate to mitigate risks.
- Current probes: This passive device brought a different kind of issue with the revision “G” release. Since the probe was exempt from periodic calibration, the CS114 test method incorporated an expanded integrity check process that brought monitoring probes into the check. This is a good way to know that the probe is functioning properly at the time of use. However, it does add some time to the testing by requiring that the sweep be accomplished twice. This re-sweep with each configuration does take a few minutes eroding the cost savings by not having probes calibrated. A secondary concern is that probes used for test methods other than CS114 are not subjected to this full check of the monitoring probe. This omits the complete check for probes associated with CE101, CS109, CS115, CS116, CS117, RS101, and injection probes. As with the passive antennas, a physical inspection should be accomplished, but a dropped probe could fracture the probe core and not show physical damage. A local integrity check of all probes can easily be accomplished by configuring the probe as indicated in Figure 2 for a sweep of the probe. Set the signal source amplitude at a fixed level and sweep the frequency range. The receiver measurements should match the applied current after the addition of the applicable correction and conversion factors. This relatively short check can alleviate concerns about the probe and when accomplished sequentially on several probes takes a few minutes longer than checking a single probe. Note that this check does not replace the requirement to accomplish the CS114 system integrity check method as specified in the standard.
- Terminators: How often do we check terminators? Do you assign an ID and record them as part of the equipment used during test? They are part of the measurement path and included when making the system integrity checks. This passive device is used frequently, and the typical check is a resistance measurement between the center pin and shell. This check is normally adequate if the physical inspection closely checks for connector damage affecting the performance over frequency. Recall that the coaxial connector is part of the cable shield and the skin effect places the cable return current path on the inner surface so a damaged connector could vary the impedance especially at higher frequencies. Using a terminator to provide for the current probe to oscilloscope impedance matching is necessary for CE102, CS115, and CS116 where external terminators are commonly used.
- Attenuators: Risks are low if the physical inspection is thorough much like the terminators. Again, ID them and list as equipment used.
- LISNs: The expanded system integrity check prescribed by MIL-STD-461 CE102 section does a good check that would identify typical failure modes. The loaded and unloaded measurements used at the lower check frequencies verify the impedance that was not part of the check with previous revisions. Placing these items on the calibration exemption status presents a low risk if the revision “G” method is used for the system integrity check.
- Directional couplers: Much like attenuators, the directional coupler gets a good check when incorporated into the system integrity check. Connector inspection a definite must and remember that without an appropriate load the coupling is negated.
Other passive devices often get pulled into the measurement system and need to be considered if they can affect the measurements and that means virtually all items—if it didn’t affect the measurement we most likely would not make it part of the system. Examine your test configurations for other types of passive devices and consider how the checks evaluate their performance.
Active Items Calibration:
This discussion is about passive device calibration, but I want to include a short calibration cost item. The market for many measurement equipment items is very competitive so we need to consider the cost of the item versus the cost of calibration. As an example, a true RMS multimeter can be purchased for about $175 for a quality instrument that includes a calibration certificate. The cost of calibration for a like item runs about $150 plus shipping costs. This makes one think about simply disposing of the older item and buying a new replacement. In the mass market where hundreds of items are produced daily a calibration station is a permanent fixture and costs per unit are low. Calibration labs are faced with individual items so each item requires configuring the calibration fixture driving more labor and labor costs are significant.
Disposing of an older item does not always mean throwing into the trash. Employee gifts, charitable donations and the like makes good use of the uncalibrated item that can be used where measurement accuracy is not critical.
The cost comparison for the multimeter is just an example, there are several items that can potentially fall into this category.
I am sure that most test laboratories can relate concerns and issues associated with the calibration program. This article is to bring forth a discussion on how important the calibration program is for a test laboratory. It is a significant factor in operations costs, but the absence of a good program brings large risks.
Cables and connectors take a beating in the laboratory environment where test configurations are changed frequently, and the impact may be unnoticed for too long is appropriate checks are not thorough.
Being able to capture where measurement items have been used is a must. Just because a passive item is exempt from periodic calibration, does not mean that the checks are unimportant. The record-keeping should support the ability to track usage in a way to learn the extent of the issue and recall if required.
As discussed herein, the physical inspection may be a critical part of the checks—make that part of the laboratory process and train the staff on what to look for during the inspection.
The thoughts here do not alleviate the required system integrity checks, but a few supplemental examinations can reduce risks that span a long time.
Hopefully, you will find this information useful and I welcome questions. If you have a topic associated with EMC that you would like to have reviewed, let me know and I will try to place it in the queue for future articles.