Test Procedures to Measure BPL Interference
Author: Ed Hare, W1RFI@arrl.org, ARRL Laboratory Manager (with a lot of help
from his friends)
Draft: December 1, 2004
This document outlines information that will be useful to those who are making any type of measurements related to Broadband Over Power Line (BPL) systems. Feedback to W1RFI@arrl.org is appreciated
Measurements may not be necessary
Types of Measurements Overview
How Can Field Strength Measurements Be Made?
Factors Relevant to Specific Types of Test Equipment
Making Measurements Using Conventional Receivers
FCC, NTIA and ARRL Recommended Test Methods
Measurements Using Amateur Station Receiver S Meters
Measurements Using a Step Attenuator
Reporting (work in progress)
Testing of BPL-System Susceptibility (work in progress, drafting group established)
Appendix B: Baseline Measurements Reporting Form
Measurements may not be necessary. In fact, in some cases, they may be counterproductive. At first, this seems counterintuitive, but imagine that BPL comes to your area, and you have S9+15 dB interference (this would be typical of interference occurring at Part-15 levels). Is this an FCC rules violation?
The FCC sets two limits for BPL Part-15 unlicensed radiators. First, because they are carrier-current devices, they must not exceed the radiated emissions limits for intentional emitters. On HF, this limit is 30 uV/m, 30 meters from the source, measured in a 9 kHz bandwidth with a quasi-peak detector. On VHF, the limit is 100 uV/m, 3 meters from the source, measured in 120 kHz bandwidth. The vast majority of amateurs do not have the equipment to make accurate, calibrated field strength measurements.
In addition to the absolute maximum limits, Part 15 has a requirement that a Part-15 device must not cause harmful interference. If it does, it must be corrected (or shut down!). Amateur Radio has similar rules. We have limits on our spurious emissions, but if our spurious signals cause interference to a neighbor’s TV, we may have to add additional filtering to our transmitter.
Does this mean that if we don’t do measurements, we can’t prove interference? If amateur radio were interfering with a neighbor’s antenna-connected TV, would that neighbor have to measure our harmonics to decide if we are legal? Of course not! The fact that there is clearly demonstrated interference would be sufficient. If BPL is operating at the FCC Part-15 limits, S9+ noise to nearby receivers tuned to any spectrum BPL is using is exactly what antenna physics says will be received. And S9+ noise that fills an entire amateur band or two, 24 hours a day, from a system built as large as an entire state is clearly harmful interference. It is not necessary to make measurements of that level of noise to know that the system generating it is violating the Part 15 rules against harmful interference.
In fact, compliance level measurements could be counterproductive. If the noise level were S9, and a measurement showed that the HF BPL system were operating at 25 uV/m 30 meters from the source, the system would be in compliance with the absolute maximum limits. You can bet that if you were to bring those measurements into any interference complaint, the BPL manufacturer and electric utility would say that they are in compliance with the rules, so it is not likely that they are causing interference. It is then necessary to take measurements of signal levels, NOT field strengths, at the operating locations of affected amateur radio operators in order to demonstrate interference.
By focusing too much on measurements, we can fall into a trap that would make it necessary to do calibrated measurements in order to claim harmful interference at all. In some cases, those measurements may support a claim of harmful interference, but they are not necessary to establish a case of harmful interference when noise levels are strong. You may be able to avoid the need to make field strength measurements by making the assumption that the BPL system is operating within compliance levels. In most cases, this is probably true. Part-15 levels are strong enough that harmful interference to nearby receivers is certain on any spectrum that BPL uses. If the BPL system is exceeding allowable Part15 levels, then the excessive signals are helping to prove the case of interference. If the BPL levels are notched or well below compliance, then the documented interference is more proof that the notches are inadequate and/or that BPL truly must not operate in substantial portions of the HF spectrum.
When might measurements be necessary? In some cases, BPL interference may not be strong, but it is still interference. A ham that lives a mile away from a BPL source will not have S9 interference, but may still hear BPL signals in the ham band. A BPL system that doesn’t use amateur spectrum, but operates near it, will produce some BPL emission in the ham bands, but at levels less than the Part-15 limits. But if those signals are stronger than the otherwise quiet noise floor at that station, they will degrade the ability of the station to communicate. In that case, it is not sufficient to say “that weak signals were observed.” It is much better to actually measure the amount of degradation of the station capability caused by BPL emissions. A procedure to do that is discussed elsewhere in this document.
The art and science of making measurements can cover a lot of ground. Some measurements are fully traceable, using calibrated test equipment whose accuracy had been recently tested at an independent laboratory. Those same measurements can also be made using equipment that it calibrated (meaning that it makes absolute measurements rather than only relative measurements), but whose accuracy has not been recently tested. In some cases, a test engineer may self-calibrate equipment. Other measurements are not relative, but use less-accurate indications, such as an amateur receiver S meter. At the bottom of the accuracy hierarchy are relative measurements.
How Are Field Strength
Measurements Made
One of the harder measurements to make accurately is field strength. Field strength may be measured in terms of volts/meter for the electric field (E) or amperes/meter for the magnetic field (H). Very far away from a radiating source (the far-field region) the E and H fields are related such that E/H = 120* pi, or 377 ohms. Very near a source (the near-field region), E and H are not precisely related, although the point of maximum E field and the nearby point of maximum H field are reasonably close to 377 ohms. Most field-strength probes measure one type of field accurately and have an unpredictable response to the other type of field.
To measure field strength, one must start with an antenna or probe of known characteristics. This can be expressed as either an “antenna factor” or as gain, typically in dBi. If the gain or antenna factor is accurately known, a spectrum analyzer, EMC receiver or service monitor can be used to make a measurement of the received signal level (RSL). (With specialized techniques, an actual receiver can be used to make RSL measurements. See “Manual Testing fo Field Strength Levels Using Conventional Receivers,” http://www.arrl.org/~ehare/aria/ARIA_MANUAL_TESTING.pdf). In most cases, RSL measurements would be expressed in terms of received voltage levels, typically in terms of dBV (dB relative to one volt) or dBuV (dB relative to 1 microvolt) at the instrument input port. RSL can also be expresses as a power level, usually in dBW (dB relative to 1 watt) or dBm (dB relative to 1 milliwatt). A good source of information on making field strength measurements with receivers is found at http://www.vk1od.net/fsm/index.htm.) The help file explains the process in detail.
Each type of test equipment used has specific uses, and limitations. It is important to select the instrument that is appropriate to the measurements being made.
In most cases, field strength measurements on HF are made with loop antennas. These are really measuring the H field. The result would be in A/m or dBuA/m. This is then converted to the E field by adding 51.5 dB to the H field value. This is not always an absolutely accurate method of making an E field measurement in the near field, but if the point of maximum emission is found, it is generally close enough to be useful.
Small magnetic loops are not very sensitive. Their gain is low, so they can be used only to measure strong signals. In some cases, an unamplified loop has insufficient gain to be used with a typical spectrum analyzer to make Part-15-level measurements. Amplified loops are more sensitive, but can be subject to overload from strong nearby signals. In general, a 12-inch amplified loop can make Part 15 measurements above 5 MHz or so.
For measurements made above 25 MHz, a biconical dipole can be used. These generally have sufficient gain to make Part 15 measurements easily.
Most loop or biconical antennas are calibrated in terms of Antenna Factor. They can be used to make traceable measurements for Part-15 compliance purposes if used properly with the correct receiver or spectrum-analyzer instrumentation.
Communication antennas can also be used to make field strength measurements, although they generally will have response to both the electric and magnetic fields. But they have more sensitivity than small loop antennas. Unfortunately, in most cases, they are not calibrated, although their gain in dBi can be used to make scientifically accurate measurements of field strength. Most compliance measurements, however, would require the use of an antenna with a calibrated gain or antenna factor. There is no problem in using communication antennas for measurements in which relative signal levels will be compared, as in the case of taking signal level readings at an amateur radio station.
Many spectrum analyzers are not capable of quasi-peak measurements, and if not, that spectrum analyzer cannot be directly used to make Part-15 measurements for compliance purposes. Most spectrum analyzers can do either a peak measurement or a “sample” measurement. The peak measurement will over-estimate the quasi-peak RSL, although in a 9 kHz bandwidth, that overestimation is not dramatic. “Sample” measurements should not be used to make Part-15 measurements under any circumstances, as they will significantly underestimate the quasi-peak RSL in most cases. Many analyzers default to “sample” mode detection, so this is an easy mistake for the test engineer to make.
Most analyzers allow the user so select the measurement bandwidth. The resolution bandwidth should be set to 9 kHz for HF measurements and 120 kHz for measurements from 30 to 1000 MHz. If the analyzer is capable of only 10 kHz bandwidth, this is generally considered to be acceptable for making HF Part-15 measurements. A bandwidth of 100 kHz will provide a reasonable result above 30 MHz, typically underestimating the value by 0.8 dB or so. However, typical SSB receiver bandwidths are typically in the range of 2.1 to 2.8 kHz. Most spectrum analyzers have a 3 kHz bandwidth filter available. This is a near-ideal bandwidth for closely duplicating the conditions that communication receivers “see” when exposed to the BPL environment.
Most older analyzers are not very sensitive; newer units, especially when used at bandwidths of 3 kHz, or less, are within one order of magnitude of communication receiver sensitivity. Combine that with the performance of actual communication antennas, and you have a system which can be used to document conditions at typical amateur radio installations. When used with an insensitive loop antenna, the analyzer may be just barely capable of making Part-15 level measurements. They certainly cannot be used to make measurements at a typical receiver system or local noise floor. In that case, the measurement would be of the noise floor of the analyzer. There is an easy test to use to determine whether a test instrument is sensitive enough to make a particular measurement. First, set up the analyzer to the desired parameters, then connect the antenna. If the noise seen on the analyzer does not rise, you are measuring the analyzer, not external RSLs. Even if the noise does rise, you need to make sure that it not just the preamplifier noise from an amplified antenna. Experienced radio operators can usually quickly tell by ear whether a particular noise is “band noise” and thus external ambient noise levels, or is the broadband hiss or a preamplifier.
It is also easy to get fooled by “ambient noise.” To an EMC engineer, “ambient” means any signal other than the one you are trying to measure. I have seen BPL testing engineers hook up an antenna, see the strong “ambient signals” from international shortwave broadcast stations and even the stronger signals on the ham bands, then note that their levels are tens of dB below this “ambient” level. To a radio operator, “ambient noise” is the generalized non-specific noise in between the actual signals on the band, and it is this quiet noise level from which interference/noise calculations should be made. If possible, try to use a spectrum analyzer that has the capability of digitally saving its plots, which can then be transferred to spreadsheets for later analysis.
This may be the best instrument to use to make field-strength measurements. Not only can it be used to measure Part-15 level signals, but because it is a real receiver, it can generally be used to measure the quiet ambient noise levels common at amateur stations. Most EMC receivers have a quasi-peak detector option (also known as a CISPR detector) and can be set to common bandwidths appropriate for EMC and other radio-related measurements. Some EMC receivers can be the equivalent of a communication receiver. If the one available has communication receiver comparable sensitivity and bandwidth choices, and perhaps a means to record its data, so much the better.
A service monitor combines the capabilities of a spectrum analyzer and EMC receiver, at least to a degree. Most do not do quasi-peak detection and some have only limited bandwidth capabilities. But they are capable of listening to a particular frequency and showing the spectrum above and below it, and they can be used to demodulate many types of signals. Service monitors usually do not have a means to record data, so manually logging and plotting will probably be necessary.
Communication receivers can also be used – with some limitations – to measure Part-15 level or ambient-level emissions. Their advantage is that they are sensitive enough to measure the ambient noise levels. Their disadvantages are that they are not calibrated; they require specialized techniques to make measurements and, in many cases, assumptions about how their average detectors will compare to a quasi-peak detector limit their accuracy somewhat. They can, however, be used to make measurements that are more sensitive than can be done with any “real” test equipment. Within their limitations, the results, though not traceable, are scientific and can be used to estimate the compliance of a system and to accurately assess its interference potential. The best application for communication receivers will be doing what they do best: receiving signals, intended or otherwise, at communication operating locations. Absolute levels are not important in that application – only signals relative to each other count. Most of the limitations listed for use in Part 15 measurements are not applicable to signal comparison measurements or signal-to-noise measurements, so communication receivers (and the receiver portions of transceivers) can be used effectively to document interference activity. It is even better to provide additional documentation, such as a video with sound, of the signals as received.
More information on using conventional receivers to do testing is found in the
ARRL article, “Manual Testing fo Field Strength
Levels Using Conventional Receivers,” <http://www.arrl.org/~ehare/aria/ARIA_MANUAL_TESTING.pdf>.
The FCC has an open rulemaking on BPL. In this rulemaking, they outlined a proposed test method for field strength compliance measurements. While ARRL and others disagree with some of its provisions, the following is the method that the FCC recommends:
Measurements Using Amateur Station Receiver S Meters
Not all amateurs have access to spectrum analyzers or other sophisticated test equipment. In that case, they may have to rely on S-meter readings. S meters are not known for their accuracy, but as a general indication of interference levels in terms that many radio operators will understand, S meter readings will provide useful information. When presented, S-meter data should also include station information, such at the receiver used, its mode, bandwidth, preamplifier or attenuator settings, etc. Also, if possible, S-meter readings should be provided in fractional S units. While the resolution of S-meter readings is not as good as fractional S-meter readings might imply, those fractional readings may be useful in other tests, some of which can scientifically assess the relative amount of noise, expressed accurately in dB.
Baseline Measurements:
If BPL is coming to your area, it will be
helpful to know just what the present noise levels are in your community. It would also be helpful to record the
typical signal strength of standard time and frequency stations (such as WWV
and
Most hams will do this using their receiver S
meters. This can be useful, if the readings
are made in fractional S units and all information about the receiver and
station is logged. Later, when BPL comes to the area, the test can be repeated
and the difference in S meter reading will provide some quantified data on the
degradation of that station. It may also
be possible to calibrate that S meter against a calibrated signal generator,
improving the accuracy of the S-meter readings. If you do, it is important to
do a separate calibration for each band, as some receivers’ S meters vary a lot
from band to band. They also can vary a
bit over time, but that really can’t be helped, so just follow the usual “warm
up for 30 minutes” practice to ensure that your receiver is at thermal
equilibrium.
If you don’t calibrate your receiver S meter against
a signal generator, the approximate relationship between S meter readings and
RSLs is shown as Appendix A to this document.
Appendix B shows a
suggested reporting form.
If the baseline measurements have been made,
they can later be compared to the BPL noise levels after the BPL system is
deployed. That way, as a minimum, an
amateur can say, for example, that the
noise level before BPL was installed was S2.2 (some noise is present), but it
is now S9+10 dB – an extremely strong level of noise. Estimating at 6 dB per S
unit, this is a degradation of approximately 63 dB. If the S meter were
calibrated against a signal generator, an even more accurate estimate could be
made.
Those S meter levels can also be used to
obtain RSLs, accurately if the S meter were calibrated against a signal
generator, or approximately by using the levels in Appendix B. If desired, the RSLs can be converted to
field strength using the formula
described earlier in this document.
A good step attenuator with 1 dB, or smaller, steps can be used to make relative measurements. There are two ways to do this.
The first uses the receiver S meter. Let me explain this with an example. If you took a baseline reading with your receiver and obtained a reading of S2.5, then found S8 noise later with exactly the same receiver settings, you could estimate that you had about 33 dB of degradation. If you inserted a step attenuator in the coax feeding your receiver, you could then adjust the attenuator until the S8 noise read S2.5 on your S meter. The attenuator setting would be the amount of degradation that the S8 noise was giving to your station.
You can also make measurements using a step attenuator, your receiver and an audio voltmeter. This generally must be done over a relatively short period of time, because you can’t touch the receiver volume control during this test. This can be used to measure a signal that is turning off then on, or to compare the levels outside a BPL area with the levels inside the BPL area, perhaps by using a mobile receiver. Here are the steps:
Appendix A: If there
is interference from a new BPL system
Keep
in mind that it is possible that the first few days this is up,
they may not have all of the bugs shaken out. (Think back to the time you got that new
computer system and days later, you were still configuring things to your
liking.)
This
is an opportunity to be reasonable, but firm. If there are problems, report
them, tell them what they need to do to fix them, and let them know that
although you understand it may take "a few days," the BPL industry
has been telling everyone that it is easy to fix interference, so they should
keep that promise and do it quickly.
General solutions:
Under the FCC rules, if they are 20 dB below the FCC limits, the Commission will not require them to do anything more for mobile stations. In practice, if hams expeience interference to mobile stations, they should report it anyway, as hams can't make the measurement. The BPL operator can either fix it, or measure it to see if it is 20 dB below the limits. This does NOT mean measuring the notch depth at -20 dB. They need to measure the unnotched level, then measure the notch depth, or measure the levels in the notches.
|
S unit |
Received Signal Level (dBW) |
Received Signal Level (dBm) |
|
S9 |
-103 dBW |
-73 dBm |
|
S8 |
-109 dBW |
-79 dBm |
|
S7 |
-115 dBW |
-85 dBm |
|
S6 |
-121 dBW |
-91 dBm |
|
S5 |
-127 dBW |
-97 dBm |
|
S4 |
-133 dBW |
-103 dBm |
|
S3 |
-139 dBW |
-109 dBm |
|
S2 |
-145 dBW |
-115 dBm |
|
S1 |
-151 dBW |
-121 dBm |
This shows the approximate correlation between receiver signal-strength-meter S units and received signal level. The actual levels vary from receiver to receiver. The signal strengths reported by amateurs on the 3.5 MHz band will typically represent the levels in a 2500 to 3000 Hz receiver bandwidth. The S meter in most receivers becomes less linear toward the low end of the scale. S9 is 50 uV across 50 ohms, and each S unit is nominally 6 dB.
Appendix C: Baseline Measurements
Reporting Form:
To better assess the impact of BPL in a
neighborhood, ARRL is soliciting baseline measurements of the present levels of
man-made noise at amateur stations. This
form can also be used to document BPL interference
Please send the following information to ARRL
HQ, Attn: Ed Hare,
Per band:
Band:
Date: Time:
Station Call:
Station Location:
Coordinates (optional):
Name and call sign of the operator making
measurements:
Is BPL being heard at the time these
measurements are being made?: Y/N
Receiver:
Mode/bandwidth (use SSB if possible, about 2500 Hz bandwidth)
Preamp: on/off
Antenna type or gain:
Antenna height:
Feed line length/type or known loss:
Distance of antenna from any house electrical wiring and/or drop line:
(own house or neighbors)
Distance of antenna from any part of the
electrical distribution wiring:
Is the electrical wiring overhead or
underground near antenna?
S meter readings per band (fractional S
units, if possible):
S meter
in the "quiet" parts of the band:
Note
any discernable specific man-made noise and the S meter reading of same
Ignore any very temporary noises, such as a
1-minute burst from a hair dryer, etc.
More is generally better, but poring over the
bands trying to catch every birdie may be time consuming, so even less than
100% complete data is still useful. It
is best to make these measurements when the atmospheric noise levels (QRN from
lightning storms) is low. If that is not
practical, ensure that the S meter levels provided represent the reading
between discernable lightning crashes.
Measurements
of Low-Band VHF
What you would need to make field-strength measurements would be:
Spectrum analyzer or EMC receiver with Quasi-Peak capability. One of the more economical ones for this task may be the Rohde and Schwarz FSH-3 with EMC receiver option
Biconical antenna
for measurements above 30 MHz - nearly any type will do, but you should ensure
it has a current calibration. Liberty Labs in MO can do this for you. See http://www.ce-mag.com/suppliers/co/18/1893.html.
ETS-Lindgren does sell biconical antennas. See http://www.emctest.com/.
24" inch loop for measurements below 30 MHz. Use the EMCO 6502, available from ETS-Lindgren
Measurements above 30 MHz
For >30 MHz, the measurement should be done quasi peak, in a 120 kHz bandwidth. Set the analyzer to read in dBuV, with a reference level of about 70 dBuV. Select a point across the street from a BPL coupler injecting a BPL signal onto the line. This should be about 10 meters from the source. Set the biconical antenna 1 meter off the ground and make a measurement with the antenna in the horizontal and vertical positions. Determine the height of the power line in meters, and calculate the "slant range" (hypoteneuse) distance between the center of the test antenna and the power line carrying the BPL signal.
From there it gets a bit complicated. The antenna will have an "antenna factor. This should be added to the measured value in dBuV to obtain dBuV/m (with the understanding that adding a negative number will reduce the value of the measurement.) This then needs to be corrected to 10 meter distance by adding 20*log(slant-range distance/10). Most slant range distances will be around 12 meters or so, so this would increase the measurement by a few dB. You also need to add 5 dB to the measurement to correct the measurement for height.
Compare this to the limit of 39.1 dBuV/m at 10 meters above 30 MHz. If the value is more than this, whether or not there is interference, there is a rules violation occurring.
Measurements below 30 MHz
For <30 MHz, the
measurement should be done quasi peak, in a 9 kHz bandwidth. Set the
analyzer to read in dBuV, with a reference level of about 80 dBuV. Select a
point across the street from a BPL coupler injecting a BPL signal onto the
line. This should be about 10 meters from the source. Set the loop antenna
1 meter off the ground and make a measurement with the plane of the loop orinted to be parallel with the power line. Make
another measurement with the loop oriented perpendicular to the power
line. Determine the height of the power line in meters, and calculate the
"slant range" (hypoteneuse) distance
between the center of the test antenna and the power line carrying the BPL
signal.
The antenna will
have an "antenna factor. This should be added to the measured value in dBuV
to obtain dBuV/m (with the understanding that adding a negative number will
reduce the value of the measurement.) This then needs to be corrected to
30 meter distance by adding 20*log(slant-range distance/30). Most slant
range distances will be around 12 meters or so, so this would decrease
the measurement by about 10 dB or so.
Compare this to the
limit of 29.54 dBuV/m at 30 meters below 30 MHz. If the value
is more than this, whether or not there is interference, there is a
rules violation occurring.
Interference
Having said all that, it is quite likely that interference would occur at levels much below the FCC "legal limits." To that end, it is more useful to actually assess interference. Unfortunately, this is not always an easy thing to do. On VHF, with FM or digital voice operation, the effect of additional noise would be to decrease the range of communication (significantly at "FCC legal limit" levels). Well within that range, FM quieting could mask the effect of noise, and a signal that was full quieting would still sound fine even with noise. But that noise would mean, as an example, that a signal from a repeater that would normally have a range of 15 miles would have a significantly reduced range -- perhaps 5 miles, if there were BPL at the FCC limits deployed in the entire area.
This can be difficult to test. Many BPL systems are small, and they may or may not be deployed near a VHF repeater site. Even if they were, it can be very difficult to know after the fact whether that apparent reduced range is due to BPL noise, or some other terrain aspect in that particular direction. Unless the BPL system can be shut off for testing purposes, this is a test that may be hard to interpret.
One easy way to test the impact of BPL can be done with two identically equipped simplex mobile stations. Set one station at a fixed location, in continuous transmission with a voice, and send the other simplex mobile driving away from the station. When the signal just starts to get noisy, the two stations are about at the edge of their useful range. Because the stations are identical, if the noise levels at both locations were similar, they should both still hear each other fairly well. Do this a few times as a "control" in areas where there is no BPL, in different directions, to establish the validity of the method and to give the operators a bit of familiarity with the test method.
Now, repeat this with the unmoving mobile station located in the BPL area. If the BPL system is using the spectrum used by that mobile station, if it has any impact on the communications range of the system, when the second station drives away to the point where it can just hear the station inside the area, the station inside the area will probably not be able to hear the moving station, or if it can, it will be much more noisy than it would be without BPL.
OSHA-recommended
methods to measure field strength
[1] The slant range distance would be calculated using the formula for the hypotenuse of a right-angle triangle. For example, if a power line were 10 meters in height and the measurement were made a 1 meter height, 10 meters horizontally from the power line, the slant-range distance would be sqrt(10^2 + 9 ^2) = 13.45 meters.
[2] Most amateur antennas are located at greater height than the power lines. NTIA has also performed all of its skywave calculations based on the premise that the BPL emitters are radiating no more than Part-15 levels skyward. Aeronautical, Inc. and Boeing have both expressed grave concerns about BPL interference to aeronautical communications, so it is important that the emissions above the power line be accurately determined.