Friday, May 15, 2015

Use of Commercial Items

The use of commercial items (CI) or commercial-off-the-shelf (COTS) equipment presents a dilemma between imposing military E3 standards and the desire to take advantage of existing commercial systems, and accept the risk of unknown or undesirable electromagnetic interference (EMI) characteristics. Regardless of the pros or cons of using COTS, any procured equipment should meet the operational performance requirements, including electromagnetic compatibility (EMC) requirements, for that equipment in the proposed installation.

Integration of COTS electrical/electronic equipment on DOD platforms is an increasingly common practice for a variety of good reasons. COTS typically offer the latest technology and can be cheaper and more quickly fielded than military systems developed from scratch. Unfortunately, commercial equipment is not designed for the harsh electromagnetic environments (EME) found in military platforms and theaters of operation.

One of the biggest difficulties with integrating COTS products into complex military systems is achieving EMC. EMC is the ability of electrical and electronic equipment and systems to share the electromagnetic (EM) spectrum and to perform their desired functions without unacceptable degradation from the EME and without causing EMI to other systems. Blindly using COTS carries the risk of increasing serious EMI problems within the platform or system.

COTS equipment has typically been designed, tested and fielded to much less demanding commercial EMC standards, if tested at all, than MIL-STD-461 or MIL-STD-464. However, the simple fact that it is a commercial item should not be taken as a reason to accept lower EMC performance. Rather than forgoing robust EMC requirements, program managers (PMs), system acquisition personnel and E3 engineering professionals must first assess the EMC-related risk to full operational capability performance from the use of COTS equipment. They must impose a detailed methodology by which to assess the risk of using COTS and achieving EMC.

To mitigate the risk, an assessment should be performed to evaluate the equipment’s immunity characteristics against the planned EME and ability to meet the desired performance. Factors to be considered in evaluating the suitability of COTS for military applications include:
• Impact on mission and safety
• The operational EME
• Platform installation characteristics
• Equipment immunity/susceptibility characteristics

After determination of the intended operational environment, the risk assessment process starts with obtaining and reviewing existing design criteria (commercial specs), analysis/test data and conducting additional EMI testing (if necessary.) If, after evaluation of the EMI data, it is determined that the equipment would not operate satisfactorily in the intended EME, then the equipment needs to be modified, or it might prove to be necessary to select different COTS equipment with adequate characteristics.

On the whole, most COTS equipment has less strict EM requirements (lower immunity levels, higher allowable unintentional emissions, lax or nonexistent susceptibility limits) than military equipment and could therefore be more apt to be upset or damaged when exposed to high level radio frequency (RF) fields or could interfere with legacy systems. Therefore the use of COTS introduces additional risk of incompatibility and can result in problems, plus associated extra costs, in maintaining performance through life and for re-use in other scenarios. When considering COTS or NDI in an acquisition, it is important to include E3 requirements and obtain and review any existing EMI test and/or analytical data.

-Brian Farmer

Wednesday, May 13, 2015

Beamwidth Loss and Why Test Staff and dB Notation should be Friends

Please Note: - This blog entry continues the thinking behind the creation of the AH Systems webinar on selecting antennas for today’s test requirements.

Supporting notes on the webinar can be found on the AH Systems website.
(Hint - click the ‘Tech Notes’ Tab and simply select the first document listed on the main page itself).

In the last entry we introduced a method that provides superior visualization of the system behavior. We continue on this theme by factoring in beamwidth loss.

Antenna Beamwidth Loss

To fully appreciate the impact of beamwidth when illuminating the test field calibration plane, we must first interpret the data on the polar chart provided as standard by all good antenna suppliers.

Make Up of the Polar Chart

The pictures above show the two main components that make up the polar chart. The first picture shows the 360 degrees surrounding the antenna divided up into 5 degree segments. The black arrow shows the antenna boresight direction. With our higher frequency EMC antennas the maximum power density created by the antenna will always be in this direction.

The second picture shows the concentric rings on the chart that indicate the number of dBs down from the maximum power density. The outer ring represents 0dB down (no reduction from the maximum power density), the first ring in is 3dB down (half the power), the second 6dB (quarter the power), etc.

The picture above shows the complete chart with the dB rings overlaid onto the 5 degree lines.

The next picture is a real polar plot provided on the AH Systems website. Note that the 0 degree marker now marks the boresight direction. The data on the chart shows the performance of the antenna at three frequencies. The red trace is the radiation pattern at 1GHz, the blue is at 9GHz, and the green is at 18GHz.

The webinar was all about better visualization of the system performance, so it explains how to mark the chart with the sector of interest, how to mark the 61000-4-3 Standard’s plus 6.0dB limit as a ‘Do Not Cross Within Sector!’ line, and finally goes on to explain how to transpose key points on the calibration plane onto the chart

Reminder: - A proviso in the webinar is that the test field is created under ideal anechoic conditions. Why? – because if we struggle to achieve a compliant field under ideal conditions, and even with the option to discount 4 of the measurement points, it will be difficult to achieve a compliant field one under practical chamber conditions, and we may need to resort to the worn out excuse called ‘room effects.’

The picture above is a snapshot of the marked up chart used in the webinar.

In the snapshot the calibration points and red crosses could have easily (and more correctly) been positioned on the outer circle, but some audience members might have run away with the idea that any reflection at the plane would all be directed back through the zone of interest. This is not so, so the plane is portrayed as a straight line. As long as the calibration points are at the correct angle, and as long as the data on the chart is read from the outer circle (0dB) along this angle, the straight line portrayal of the plane is just as visually informative as one on the outer circle. Some might argue even more so.

Why call it Beamwidth LOSS?

A generalized block diagram of a loss in a system is shown below

We are quick to relate this with cable loss where the system needs more power to overcome the loss. Let’s put a lossy cable in the block and assume that for an input power of 100W the output power is 80W. Then the loss in dB is 10log10 [80/100] = 1dB. We conclude we need 1dB more power from the amplifier to overcome the cable loss. No arguments here.

Likewise, due to the beamwidth phenomenon, we need more system amplifier power to create the greater power density required at the center of the calibration lane. Instead of putting the cable in the block, we can just as easily put the beamwidth effect in there. If we accept the need for more power to overcome cable loss, it is completely rational to accept the need for more power to overcome beamwidth loss.

Why Test Staff and dB Notation should be Friends

We are about to interpret the data on the snapshot chart, so this is a good point to make the case that dB notation is not out there just to annoy and confuse test staff, but rather it is there to aid computation and visualization.

We start with the ease of use and interpretation of the marked up chart. Irrespective of the actual maximum power density (could be a high value, could be a low one), the radiation pattern fall off in value is always with respect to this 0 degrees boresight maximum. In other words, the value at any angle on the chart as you move away from 0 degrees is automatically referenced / compared to this maximum in dBs.

Let us read the chart data to determine the beamwidth loss at 1GHz. We follow the red trace from the boresight in a clockwise direction until it intersects the brown line defining the periphery of the calibration plane. The intersection is around one third of the distance between the 0dB and minus 3 dB circles, making the beamwidth loss at 1GHz around 1dB. This says the calibration points on the periphery of the calibration plane will be 1dB down from an imaginary point at the dead-center of the plane. Conversely, the power density at the boresight is 1dB greater than that at the periphery, so we need more power from the system amplifier to create this extra boresight power density.

We can do the same thing with the blue trace (9GHz), and determine the beamwidth loss at 9GHz as 2dB or so. Our system design goes to 10 GHz, so we extrapolate this to 2.5dB.

Because we are using dBs we can now simply add this loss to the cable loss (also in dBs) to get the combined loss at any particular frequency of interest. And we can carry on doing this with all forms of loss to get the overall system loss. We then combine the overall system loss with the antenna dBi gain to obtain the resultant system wide gain GSYS. This in turn is then combined with the base system power to determine the amplifier power required by the system. And there ladies and gentlemen, is the beauty of dBs. There will be more on this later.

Teaser Question

Here is a teaser for the more experienced EMC Test Staff out there and for those that relish a challenge. For higher frequency EMC antennas beamwidth loss dominates antenna mismatch. For lower frequency EMC antennas VSWR loss caused by the poor antenna match dominates beamwidth loss. When expressing VSWR as a loss, I took the time to derive the equation:

Where LVSWR is the VSWR loss and ρ is the reflection coefficient presented by the antenna.

How did I derive the first equation, and more importantly, why I did I take the time to do it? It is all there in either the webinar, the supporting notes or these ‘thinking behind’ blog entries. Or maybe it takes all three.

To be continued...

-Tom Mullineaux
Lionheart Southwest

Tuesday, May 5, 2015

A Simplified Method of Calculating the Amplifier Power Required by an RF Immunity System

AH Systems recently sponsored an EMC-Live webinar on selecting antennas for today’s test requirements in which I introduced a simplified method of establishing the amplifier power required by an above 1GHz RF immunity system. Existing methods, such as slotting numbers into a spreadsheet, do not provide the necessary visualization of the test system behavior, and provide little or no understanding of the system dependencies.

It will aid understanding if you watch the presentation before going into this thinking behind the webinar’s creation. A recording of the webinar can be found here.

The webinar supporting notes can be found by clicking the ‘Tech Notes’ tab on the AH Systems website

And so, here is the thinking that led to the simplified method.

Taming the Field Strength Equation

A first glance at the beast that is ‘the field strength equation’ can create a feeling of foreboding. But studying and classifying its component parts not only tames the equation, it also gives superb visualization on how the system behaves across the band of interest. The equation is:

Where E is in v/m, d is the distance (set by the standard so a constant), G is the linear gain of the antenna, and P is the RF power injected in at the antenna connector.

Regular blog followers will know we showed how this equation was derived in a previous blog posting

We start the taming process by recognizing and pulling out the obvious and hidden constants within the equation. We then take a brand new step where the field strength E itself is made a constant. In order to do this we introduce a new compensatory factor called ‘Beamwidth Loss’. More on this later.

Obvious and Hidden Constants

The obvious constant is √30 / d. We will call this constant k1

The less obvious constant is √(PG). Note, we are not saying that P or G is a constant, we are saying for a particular fixed field strength E, that P multiplied by G is a constant.

Let’s say P = 10 and G = 10. Therefore P.G = 100 which means √(PG) = 10
If at another part of the band the antenna linear gain G was to drop from 10 to 5, then for P.G to maintain its constant value, the power P would need to compensate by increasing to 20, keeping P.G = 100 constant and √(PG) = 10 constant.

And so P and G do a sort of dance across the band, where as G changes, P changes in the opposite sense to maintain the constant. We could allocate this constant a letter, but for our purposes we will keep it as √(PG), we just need to keep in the back of our minds that this has a constant value for a fixed field strength. So we will stick with

This says that for a fixed field strength (let’s say the minimum allowed field strength stipulated in a standard such as 61000-4-3), any change in G is compensated for by a change in P. All makes sense and stands to reason.

Now we have a better feel for the equation dynamics (the dance), we turn our attention to its real use in the webinar, which is to determine the power required to obtain a particular field strength.

First we rearrange to make P the subject of the equation.

For fixed field value Ek

Taking 10 log10 of both sides puts the equation in dB notation

With logarithms, a linear division is represented by a subtraction, so we can write

What this says is that the power required from the amplifier (assuming it is connected directly to the antenna, that is with no cable loss) is a power level dictated by constants, minus the antenna gain in dBi. In other words, the better the gain of the antenna, the less power required from the amplifier.

Again, this all makes sense and stands to reason.

Note that for the case where the linear antenna gain is 1 (equates to G = 0dBi), the amplifier power level is entirely dictated by the constants. For this reason we will call this power level the ‘base power level’, and we will regard GdBi as a modifier to this level.

So what is the amplifier’s base power level for a 61000-4-3 RF immunity system where the test distance d is 3m, and the minimum permitted field level (our Ek) at distance d is 18v/m?

We will round [32(18)2 / 30] = 97.2 up to 100W

Amplifier power is more commonly stated in dBm, that is the power is relative to 1 milliwat, not 1 watt. The conversion from dBW to dBm is simplicity itself – just add the number 30.

So our base power level is 50 dBm or 100W. We now have a starting point for our system.

Why did we go the dB route? Answer – because antenna manufacturers supply gain data in dBi, cable loss per meter is given in dB, and the soon to be introduced ‘beamwidth loss’ is in dB. The beauty of everything being in dBs is we can just add these algebraically to get the overall system Gain and then simply add this algebraically to the base power level to get the power required from the amplifier.

Don’t forget that the webinar and the supporting notes offer further explanation, particularly regarding visualization of the system wide performance.

Next time we will introduce ‘beamwidth loss’, rationalize its introduction, use it to determine the overall system gain GSYS, and finally use the worst-case GSYS and the base power level to determine the power required from the amplifier.

To be continued ………

-Tom Mullineaux
Lionheart Southwest