How to Focus RE102 Testing on the Things that Really Matter

MIL-STD-461 RE102 is probably the most commonly failed test in the aerospace/defense world, with 50-90% of units failing their first pass testing. This is a frequent cause of schedule delays, first for troubleshooting, and then for all the meetings needed to process waivers. There are ways that it can be tailored, even very early in the product development process, to minimize the need for waivers after test failures. Any time a unit fails a test but is allowed to move forward after going through a waiver process, it’s an indication that the requirements were not set appropriately at the beginning of the program.

Ultimately, we want to do the minimum amount of testing that gives us the best assurance of mission success. We don’t want to over test and jeopardize cost and schedule targets. But we also don’t want to under test and miss something that could cause issues on the integrated platform. Understanding the purpose behind RE102 requirements helps us tailor them in a program-specific way.

RE102 exists primarily to protect intentional RF receivers on a platform from stray emissions from onboard electronics. If you look at the limit levels in MIL-STD-461 (Rev G is the most recent version at the time of writing), a typical value might be 69 dBµV/m. That equates to a field strength of 2.8 mV/m at 1 m. Generally speaking, not many non‑RF electronics modules will be sensitive to that level of noise (consider the typical RS103 level of 20 – 200 V/m). Given the 1/r fall-off of field strength over distance, RF receivers not co-located the unit will usually not react to these levels. It’s the RF receivers installed along with the electronics that are most at risk from these high frequency but relatively low amplitude emissions. Thus, our RE102 limits should be tailored to the RF systems that will be present on the platform, if known.

There’s a lot we don’t know at the beginning of a program. We may not know exactly what radios will be selected, or which vendors will provide them. We may not know what antennas will be chosen, what their field of view is, where they will be placed, etc. However, we do know what kind of program we’re working on: spacecraft/aircraft/marine/terrestrial. And we likely know what kind of RF systems will be required: UHF, GPS, S-band, Ka-band, air traffic control, special electronic warfare devices, etc. So even if we don’t know our specific spectrum allocations yet, we can say “We will have receivers that will be using these frequency bands.” Once we know that, we can focus our tailoring efforts appropriately.

Selecting Frequency Ranges

The first thing to look at is the frequency ranges to test. Here are the default frequency ranges from MIL-STD-461 Rev G, RE102:

Testing at the lowest frequency range, 10 kHz – 2 MHz, should be done only in cases where there are onboard systems that are sensitive to very low frequency energy, such as antisubmarine warfare detection units. Otherwise, this frequency range should be omitted. As Ken Javor has noted elsewhere in the pages of this magazine, testing in this range is done in the near field, not the far field, and is messy and hard to reproduce.

The same can be said for the 2 – 30 MHz range (rod antenna) and 30 – 200 MHz range (biconical antenna). If you do not have any VHF receivers, consider skipping these ranges as well. They also have reproducibility problems, plus skipping two antenna configurations saves a good chunk of time. The fewer times you have to change antennas, the better. “Failures” in this range, without an intentional RF receiver in that range, are likely not relevant. (Obviously, there are other considerations that might come into play, such as if the unit or program is also required to meet CISPR 32 or EMCON limits.)

Here’s a supporting quote from MIL-STD-461 Rev G, Section A.5.17:

“The 104 cm rod antenna [10 kHz – 30 MHz] has a theoretical electrical length of 0.5 meters and is considered to be a short monopole with an infinite ground plane. It would produce the true electric field if a sufficiently large counterpoise were used to form an image of the rod in the ground plane. However, there is not adequate room. The biconical [30 – 200 MHz] and double ridged horn [200 MHz – 1 GHz] antennas are calibrated using far-field assumptions at a 1-meter distance. This technique produces standardized readings. However, the true electric field is obtained only above approximately 1 GHz where a far field condition exists for practical purposes.”

If these VHF frequency ranges are important, consider using a common mode current measurement as an alternative, such as that described in GSFC-STD-7000B, Section 2.5.2.1.2. That will give you relevant information without the concerns about near-field vs. far-field antenna measurements. Testing can be done in a plain shield room without the need for a full semi-anechoic chamber (see Figure 1).

This is a test method significantly different from the conducted emissions test methods found in MIL-STD-461 Rev G, which are differential mode measurements. The Goddard standard also calls for the use of an absorbing clamp rather than a typical clamp-on current probe. It has an extended argument for why this is a better approach in Section 2.5.3.3.2.4. I found the argument fascinating, and recommend that everyone give it some consideration. 

Reining in the upper-frequency range is equally important. It is easily possible for the clocks of fast electronics systems to throw harmonics well into the GHz range. But if your highest frequency receiver is an S-Band comms system, do you need to test from 4 GHz – 18 GHz? Again, any “non-compliant” test results will likely be waived, since there is no on‑board receiver that will suffer from the interference.

This is a good place for another caveat: receivers can be susceptible to interference outside their passband, both below and above it. It has been known to happen sometimes, that a spec sheet can misstate the out-of-band susceptibility of a radio receiver by 20 – 60 dB. Depending on the criticality of the system, you may need to start with very conservative assumptions of both sensitivity and susceptible frequency ranges, then relax the limits and/or narrow the frequency ranges when more about the potential victim system is known, either through analysis or test.

You may also need to “future-proof” your system. While you can be fairly sure a new RF receiver won’t be added to a satellite after launch, the chances of new systems being implemented on an aircraft or naval vessel are much higher.

Designing Notches

Once you’ve identified which frequencies have receivers that need protection, you can calculate the minimum measurable field for each, without any knowledge of the specifics of the RF systems selected. (For the derivation that follows, I’m indebted to GSFC-STD-7000B, Section 2.5.3.3.7.4, with a few modifications.) We start with:

Where E is the minimum detectable electric field strength in dBµV/m, V_N is the thermal voltage noise floor (dBµV), and AF is the antenna factor (dB/m). To calculate V_N:

Where B is the bandwidth of the receiver (Hz). This equation is based on the theoretical noise power formulation, assuming a temperature of 290 K and a system with 50 Ω impedance. Then for the antenna factor:

Where f is the frequency (GHz), and G is the gain of the antenna system (dBi). The constant also assumes a 50 Ω receiver system, so it should be adjusted if a different impedance is being used.

To get these equations, we’ve assumed a 50 Ω system at 290 K that’s experiencing a free space plane wave (similar to the nominal test conditions for RE102). To calculate the minimum measurable field then, we need to know the bandwidth of the victim receiver, the frequency of the victim receiver, and its gain. We know at least the frequency range of the receiver from the system specs. We may not know the specific bandwidth of the system, but often there are standard bandwidths associated with things like GPS detectors. A close-enough guess is likely good enough for this initial limit.

Considering gain, do NOT use the main lobe gain of the receiver if it is known. You don’t know much about either the receiver system or your electronics module under test, but you can be reasonably sure that it will not be installed in such a way as to block the main field of view of an RF receiver antenna. Instead, use the worst case of the sidelobes or back lobe of the antenna. Of course, if the antenna is intentionally omnidirectional, use the omni gain. If nothing is known about the system, 0 dBi is a conservative assumption.

At this point, we have drawn some very conservative notches that will flow down to the equipment designers as the program moves forward. This might not make everyone super happy, but you can just about guarantee that those limits will be relaxed up as the program evolves, instead of adjusted down at the last minute. (And no one seems to complain about relaxing limits!)

Relaxing Notches

As the program progresses, there will be plenty of opportunities to revise these limits upward. One of the main ways comes once you have a better understanding of the overall construction of the platform. If all the RF receivers are on the exterior of a metal chassis (as is typical on aircraft and spacecraft, for instance), then electronics that are installed on the interior of the chassis can have their limits relaxed based on what is known about the shielding provided by the chassis.

Other parameters that you can include as they become available: the side lobe/back lobe gain of the specific receiver antennas; the out-of-band rejection performance of the RF receiver, and the noise tolerance of the RF receiver, especially once the link budget has been determined. This initial tailoring analysis assumes that if any detectable signal is present, the RF system will be interfered with. That is likely untrue, and any knowledge about noise tolerance or error correction of the system should eventually be taken into account.

What About the Other Frequencies?

What about the frequencies where we don’t have any receivers? Let’s say you have a UHF receiver at ~400 MHz, and a GPS receiver at 1.5 GHz, but not much in between. The safest thing to do is to use the default limits provided by MIL-STD-461 for your platform type in that middle-frequency range. The cheapest thing to do is skip 500 MHz – 1.4 GHz completely.

Another option would be to significantly relax the limit in the middle-frequency range. That way you still test it and can see if something is drastically wrong, for example, if the unit is throwing off levels of emissions that might interfere with neighboring systems either on or off the platform. But you won’t call something a “failure” unless it is fairly extreme. Obviously, this will depend on the customer and the needs of the project. What the customer says they need and what they’re willing to accept is usually the final word.

Conclusion

Is this kind of tailoring acceptable? Chapter and verse from the MIL-STD emphatically says “Yes”:

“Possible tailoring by the procuring activity for contractual documents is as follows. The limits could be adjusted based on the types of antenna-connected equipment on the platform and the degree of shielding present between the equipment, associated cabling, and the antennas. For example, substantial relaxations of the limit may be possible for equipment and associated cabling located totally within a shielded volume with known shielding characteristics. It may be desirable to tailor the frequency coverage of the limit to include only frequency bands where antenna-connected receivers are present. Some caution needs to be exercised in this regard since there is always the chance [new] equipment will be added in the future. For example, it is not uncommon to add communications equipment (such as HF radio) onboard an aircraft as different missions evolve.” – MIL-STD-461G Sec A.5.17

This gives me an excuse to issue my standard (if you’ll pardon the pun) reminder: Always Read the Appendices! For those standards that have them, informative and normative appendices and annexes often contain golden nuggets of wisdom that help the new user figure out how best to apply the standard. They are included by the standard’s working group specifically to give context and guidance—and those working groups benefit from the collective experience of the careers of their members, sometimes a couple of centuries worth if you add up everyone’s resumes. You’re doing yourself a disservice if you don’t give them a read-through whenever you’re applying a new-to-you standard.

Endnotes

1. General Environmental Verification Standard (GEVS) for GSFC Flight Programs and Projects, GSFC-STD-7000B, 2021.
2. K. Javor, “(Re)Discovering the Lost Science of Near-Field Measurements – Part 1,” In Compliance Magazine, July 1, 2023.
3. Requirements for the Control of Electromagnetic Interference Characteristics of Subsystems and Equipment, MIL-STD-461, 2015.