Ambient noise is commonly dominated by man made noise, and it often arrives equally from all directions. For measurement of such noise, the captured power depends on average antenna gain, and so the calculations below focus on gain averaged over the hemisphere.

Antenna Factor is often very convenient for field strength measurement as it relates the external E field strength to the receiver terminal voltage given a certain antenna (system). In fact, given a short vertical terminated by a high impedance amplifier, Antenna Factor is often fairly independent of frequency over several octaves of frequency.

Whilst it is easy to come up with Rules of Thumb or simple approximations for a short monopole over perfect earth conductor (PEC) either matched for maximum power transfer, or essentially unloaded, the case of a short vertical on the roof of a motor vehicle suspended above natural ground is not so easy.

Some underlying assumptions:

- a vertically polarised antenna will be most sensitive to ground wave noise on the lower HF bands (as horizontally polarised ground waves are more quickly attenuated with distance);
- the system uses an active antenna, the amplifier having a high input impedance and unity voltage gain; and
- reciprocity can be used to infer certain receive performance from transmit performance.

So, the approach is to:

- calculate the power captured by a lossless isotropic antenna, and the Thevenin source voltage for a given E field (1V/m);
- use an NEC model to find the (matched) average gain of the antenna system (so accounting for ground losses etc), source impedance, and radiation resistance (which is used as the Thevenin source impedance); and
- calculate the loaded voltage at the amplifier input terminals, and Antenna Factor at the amplifier input terminals.

An NEC-5.0 model was created for a 1.215m (4′) vertical mounted on the roof of a motor vehicle at 3.5MHz. The model was derived from a sample model supplied with NEC-5, the vertical was lengthened and moved to mid roof, and the unused antennas deleted. The model was changed to introduced real ground (σ=0.005, εr=13).

Above is a 3D pattern plot, the pattern is an almost omnidirectional donut.

Above is an elevation profile.

Above is the azmuth plot at θ=-70° (elevation 30°).

Some key values are extracted from the NEC output report.

The calculated equivalent series source capacitance Cs is given for interest sake, it is not used directly here, but is often estimated for so-called E-field probe antennas.

Thevenin source voltage of a lossless isotropic antenna is found by calculating the available power to be captured by a matched antenna given the excitation scenario (E=1V/m). The available power is given by the product of the effective aperture Ae and the power flux density S (for E=1V/m). Ae is calculated from average gain (unity for a lossless isotropic antenna) and frequency.

Having calculated available power, we can calculate the voltage in a matched load, and the Thevenin source voltage Vth is twice that.

The Thevenin source impedance is radiation resistance Rr calculated earlier from the NEC output.

We are interested in the Antenna Factor when the antenna is loaded by the high impedance amplifier. The amplifier is approximated as some resistance Rp in shunt with some capacitance Cp.

Antenna Factor is given by \(AF=20 log \frac{E}{V_{in}} \text{ dB}\) where E is the electric field strength and Vin is the amplifier input terminal voltage.

AF is quite sensitive to Cp, efforts need to be made in circuit configuration, device selection, and circuit layout to minimise Cp and achieve optimal AF.

The above discussion is based on a unity gain amplifier, if otherwise, the gain in dB should be subtracted from the calculated AF to get AF wrt the amplifier output terminals (ie, the input terminals to the following receiver).

]]>To some extent, the project was inspired by KK5JY’s Loop on Ground (LoG).

This article presents a comparison of Signal to Noise Degradation metric (see Signal to noise degradation (SND) concept) for both antennas, the common elements being:

- based on NEC-5.0 models (as detailed in earlier LiG articles);
- soil parameters used are σ=0.01, εr=20 (calibrated to measurements at the LiG test site);

The LoG models are for 4.6m sides, 2mm wire at 10mm height above ground, and an approximately optimal 450Ω:50Ω transformer.

The LiG models are for 3.0m sides, 2mm wire at 20mm below ground, and an approximately optimal 200Ω:50Ω transformer.

Above is a plot of SND for both antennas over the range 0.5-15MHz.

They are quite different responses.

It can be seen that from about 3-11MHz there is not a lot of difference between the antennas, but the LiG degrades slowly above 11MHz whereas the LoG degrades quickly below 3MHz.

A work in progress…

]]>It has what appears to be a genuine Honda GX160. Now I bought it with I must say a great deal of skepticism, but having worked on many Chondas (Chinese Honda ‘clones’) this is undoubtedly a class above, and I think on all the evidence available, it was a Chinese manufactured Honda destined for the domestic market.

The engine has run very well, the only criticism is the sag in speed under full load, that indicates it is a little underpowered, a GX200 would be more appropriate.

A recent leak down test revealed a slight loss in compression, the principal leak being the exhaust valve. (These engines have an automatic decompressor on the camshaft, so old-syle compression tests are not valid, quite a waste of time… other that possibly revealing a broken decompressor.)

So… a gasket set was procured (actually 5 complete engine rebuild gasket sets for about $22 posted on Aliexpress… these are a lot cheaper to repair (in Australia) than Briggs & Stratton engines).

The head was removed (which means removing the fuel tank, extra large muffler and carburettor).

Above, the block after scraping carbon deposits off the piston, cleaning up and fitting a new gasket.

Above, the heat after decarbonising with a small rotary cup brush, gasket removal, valve lapping, cleaning and reassembly.

The valve guides were secure, valve seats were excellent, the exhaust valve though showed small signs of erosion. I purchased a set of valves with springs etc for $12 on Aliexpress for ‘next time’.

The head was installed, torqued down progressively, push rods installed and valve clearances adjusted.

The engine started on first pull, engine speed needed a small adjustment (the speed screw had been back out for disassembly, then screwed back in to approximately the correct setting… but small trim needed.

This work took about two hours, a commercial workshop would not do it much quicker, so a commercial job with genuine parts would render the machine a write-off. Other options like a drop in engine replacement may not be viable as the alternator rotor is probably mounted on a tapered engine shaft (no other support bearings), and procuring an engine with that special shaft end may be impractical.

The frame of the genset used lots of punch’n’tap tabs with M6 screws. Three of the original 11 threads were stripped and the tabs backed with M6 nuts.

Above, a punch’n’tap thread. The sheet is extruded to typically about double thickness and tapped. In this case they were formed in softer steel and several were stripped by machine driven screws.

Above, the punch’n’taps were replaced with M6 rivnuts (nutserts) and a new set of 304 M6 serrated head bolts purchased to suit the deeper threads.

It was delivered with electrical faults as mentioned above.

When it was just weeks old, the die cast fuel cock on the large fuel tank (non-Honda) crumbled into pieces.

Just months ago, the Keihin carburettor needle and seat, and bowl gasket were replaced (a casualty of E10 fuel some years ago).

It has been a fairly good machine, always starts easily, is very quiet (an oversized muffler is fitted), is a low technology machine, but reliable.

]]>An important early step in designing a ferrite cored transformer is to find a combination of ferrite material, core geometry, and number of turns to deliver acceptable core loss at the lowest desired frequency.

Design of a transformer to cover just the 1.8MHz ham band is a relatively simple exercise.

Let’s examine a 50Ω:50Ω autotransformer on a common FT240 core using some popular ferrite mixes (even if the mix is not available in that core size).

The mixes chosen are 31, 43, 46, 52, 61, 73, 77.

First step is to calculate the minimum number of primary turns that delivers InsertionVSWR closest to 1.1 based on published material characteristics. We will then look at expected core loss. That has been done with a Simsmith model that I have published for other projects. The model results are summarised in the following table.

Note that a different number of turns are required for each mix.* means that the mix is not readily available in a FT240 core size. The table is sorted in order of ascending core loss.

Note that a large number of turns and low permeability are not conducive to the low leakage inductance that is needed for broadband performance (though in this specific case, the required bandwidth is quite small).

Above is a column chart of the seven options.

Summary:

- mix 61 is the lowest, relatively low cost and good availability, but way too many turns for this core;
- mix 52 is low loss, moderately low cost and good availability for limited range of cores, but way too many turns for this core;
- mix 46 is low and practical, low cost from the manufacturer, in a limited range of core sizes, and not widely stocked;
- mix 43 is next, low cost, readily available in a large range of core sizes;
- mix 31 has around three times the loss of mix 43, limited range of core sizes and not so readily available;
- mix 73 has higher core loss and again not readily available in a large range of core sizes; and
- mix 77 has even higher core loss and again not readily available in a large range of core sizes.

The darling of the bunch is mix 43, if you are objective you can see why it has been so popular for so long.

Whilst 46 is a good material for a lot of applications, and I have used it in some designs, I tend to prefer 43 in published designs as others may find the cores more practical to source. Mixes 61 and 52 are just too low in permeability for a broadband 50Ω transformer at this frequency.

More turns could be used to drive core loss down for some cores, but more turns increases leakage inductance which is the enemy of broadband high ratio transformers.

There are further steps to the design process. The results above are a good start to design of a higher ratio autotransformer, essentially you are unlikely to develop a good design on a core and turns combination that does not deliver a good 1:1 ratio transformer.

These results apply specifically to the scenario described.

Attached is FT240-1-1-160m.7z containing a model and material files. The model is an autotransformer model used and described in several recent blog articles, and in this instance is configured for a 1:1 unun which means there is zero leakage inductance, and InsertionVSWR and CoreLoss is due to the shunt magnetising admittance of the transformer primary. Whilst a 1:1 unun might not seem to require the core, this is a first step to designing a different ratio transformer.

Be aware that ferrite tolerances are relatively wide, so measurement of a prototype might not reconcile exactly with the prediction based on the datasheet.

]]>The Loop in Ground project is about a receive only antenna for low HF, but usable from MF to HF. The objective is an antenna of that is small, low profile, and can be located outside the zone where evanescent modes dominate around noise current carrying conductors, like house wiring to minimise noise pickup.

The antenna comprises a square loop of 3m sides of 2mm bare copper wire, buried 20mm in the soil.

This article reports measurement of feed point impedance and a ‘calibrated’ NEC-5.0 model.

Australia is experiencing a La Nina weather pattern this spring / summer, and it has rained and rained and rained. The ground has varied from saturated to nearly saturated, so measurements are a little atypical. Several measurements have been made, and the ones reported here are at a less saturated time, but the same broad pattern has been observed with all measurements.

Whilst the measurements to not calibrate exactly with the NEC model, they are quite close, and the model used here is adjusted for better reconciliation in the range 1-5MHz. The soil parameters used are σ=0.01, εr=20, which are suggestive of very ‘good’ soil.

The calibrated NEC model is probably the best predictor of behavior that other constructors might experience.

Key to the performance is system gain which depends greatly on MismatchLoss, which is quite dependent on the load impedance presented to the LIG. A spreadsheet model was constructed from the NEC feed point impedance and expected ambient noise (per ITU P.372-14) and Signal to Noise Degradation (SND) calculated (Signal to noise degradation (SND) concept).

Provision made for a transformer as part of the system model. So parameters to the spreadsheet model included:

- NEC-5.0 model of feed point impedance and average power gain of a 3m a side square loop of 2mm HDC buried 20mm in soil σ=0.01, εr=20;
- ITU P.372-14 ambient noise precinct ( Rural used here);
- transformer ratio (2:1 turns, 4:1 impedance, with 1dB loss used here);
- receiver Noise Figure (6dB used here); and
- transmission line (10m of Belden 8215 RG-6/U used here).

The 75Ω feed line is chosen as a low cost feed line even though it is not 50Ω, in a real implementation with a buried feed line, flooded RG-6/U or RG-11/U might be very practical choices.

Above is the graph for the Rural precinct as mentioned, SND is in blue. It can be seen that is less than 3dB from 1.0-13MHz.

If you live in a very noisy neighborhood, the Residential precinct may be more appropriate to your environment.

Above is the graph for the Residential precinct as mentioned, SND is in blue. It can be seen that is less than 3dB from 1.0-28MHz.

Most designs of small Loop-on-Ground antennas use higher transformer ratios, and they may or may not be appropriate, but for this LiG in the specified soil, it is clear from the spreadsheet model that choosing other integer ratio transformers gives poorer SND response.

An IC-R20 was used for a listening test as with near zero length feed line, system gain is not significantly affected by common mode feed line contribution. Tests were conducted at 19:00 local time on the MW BC band, 80m, 40m and 20m bands. In all cases, external noise was audibly greater than internal noise (assessed with a 50Ω termination, some very minor contribution from the termination), and the receiver performed pretty much in line with prediction. There was no evidence to question the predictive models for Rural in this location. It was interesting that even an hour before sunset, many MW broadcast stations were heard at very good strength even though this location is not the the formal service area of any MW broadcast signals, stations in Canberra some 200km distant were at very good strength and excellent quality (steady signal, no buzz or other significant intermodulation distortion.

A work in progress…

]]>If we consider a two wire transmission line, we can define currents I1 and I2 flowing in the same direction in each conductor.

These two currents can be decomposed into differential and common mode components:

- Id=(I1-I2)/2; and
- Ic=(I1+I2)/2.

The diagram shows the relationship of the various example phasors, their sums and phase relationships.

Rearranging these, we can write:

- I1=Ic+Id; and
- I2=Ic-Id.

So the component currents Ic and Id fully account for the current at each terminal, I1 and I2.

I1 and I2 can be measured directly by placing a current probe around each of the wires, and I1+I2 (I12 or 2Ic) can be measured directly by placing a current probe around both of the wires.

Measurement of the magnitudes of these three currents I1, I2 and I12 can be resolved into components Id and Ic.

The calculation is not very difficult, it used no more than high school maths.

The measured values I1, I2 and I12 are the magnitudes of three phasors, and for some of the calculator results, we need to find the magnitude of the phase angle between I1 and I2. The Law of Cosines provides the solution, \(|\theta_{12}|=acos\frac{(I_{12})^2-I_{1}^2-I_2^2}{-2 I_1 I_2}\).

We can then calculate using the Law of Cosines \(2I_d=I_1^2+I_2^2-2 I_1 I_2 cos \alpha\) and since \(\alpha=\pi-\theta_{12}\) we can write \(2I_d=I_1^2+I_2^2-2 I_1 I_2 cos(\pi-\theta_{12})\).

Now to find magnitude of the phase angle between Ic and Id, |θ_{dc}| using the Law of Cosines \(|\theta_{dc}|=acos\frac{I_2^2-I_c^2- I_d^2}{-2 I_c I_d}\).

Using the lengths of the phasors in the figure above, we can calculate the components.

Above, the example to scale.

Above, the calculated results.

A graph is not presented, it will be impractical to read.

I12 varies from 10-30mA depending on recent weather, seasonal vegetation changes etc.

Low Ic is achievable with good design, good symmetry, and an effective common mode choke.

]]>But can you trust what you read / see online?

Here an author gives his take on Return Loss.

Errr, no!

Using his terms, \(Return\_Loss=\frac{P_{sent}}{P_{return}}\), a result greater than unity, or we can express it in dB (not db, the B is for Bell, a person’s name, so it must be capitalised in the unit) \(Return\_Loss\_dB=10 log_{10}\frac{P_{sent}}{P_{return}}\) which will yield a positive dB value.

If we use a VNA and it reports the magnitude of S11, we can calculate \(Return\_Loss\_dB=-20 log_{10} |S11|= -S11\_dB\).

So it is quite wrong to speak of, or worse to label a plot of S11_dB as Return_Loss.

But “everybody does it, it must be right”! No, but it is true that on social media, popularity is taken to determine fact.

]]>Above, the two styles of baluns purchased. Given the label, nondescript as it is, there is quite a possibility that the internals are the same.

Above is a plot of common mode impedance for a sample of both styles. They are near as dammit identical.

They have moderately high common mode impedance to 5MHz, perhaps adequate for some applications to 10MHz.

]]>The article uses Rigexpert’s Antscope as the measurement / analysis application, the techniques will work with other good application software.

To demonstrate the technique for matching such an antenna, let’s use NEC-4.2 to create 80m feed point impedance data for a 12m high vertical with 8 buried radials (100mm) and centre loading coil resonating the antenna in the 80m band for simulation of measurement data.

An s1p file was exported from 4NEC2 for import into Antscope, to simulate measurement of an example real antenna.

Above is the VSWR curve displayed in Antscope. Note that the actual response is dependent on soil types, antenna length and loading etc, but this is a good example for discussion. It is not real bad, another example might be better or worse.

Above is a plot of the feed point impedance, ie the serial components of feedpoint impedance R and X. The possibility of a shunt match does not jump out, other that seeing that R<50.

Above is a Smith chart presentation, and this is more revealing. We can ‘see’ that the curve crosses the undrawn circle where G=1/50S. It is undrawn probably because the software author did not foresee the utility of G circles etc.

So, at the cursor, the admittance Y=1/49.5+j1/64.5S (the unit of admittance is Siemens), but they write it in a hammy way that Zpar=49.5-j64.5 ohms and it is quite flawed, algebraically it is wrong, they perhaps should have written Zpar=49.5||-j64.5 Ω to mean 49.5Ω in parallel with -j64.5Ω. A lot of people cannot handle admittance, and talk in the parallel impedances equivalent.

Talking in hammy parallel impedance, the capacitive element of Zpar can be ‘tuned out’ by an equal but opposite parallel reactance, leaving Zpar=49.5… pretty much right on target.

Above is a presentation of the parallel equivalent impedance components.

An important characteristic of this antenna is that at Rp passes through 50Ω, so it may be a good candidate for a shunt match at the frequency where Rp=50Ω. Rp does not pass through 50Ω, it is not candidate for a shunt match.

It can be seen that around the cursor, Rp is frequency sensitive, and at the cursor, Rp is almost exactly equal to 50Ω, so we just need to ‘tune out’ the parallel reactance with an appropriate inductor (one with X=64.5Ω).

You could then calculate the inductance of the shunt coil, make a coil and measure / adjust to minimum VSWR. In this case, \(L=\frac{X}{2 \pi f}=\frac{64.5}{2 \pi 3.538e6}=\text{2.9e-6}\), so we could design an inductor for 2.9µH.

This section shows simulation in Simsmith.

Above is a Simsmith model of the match. The s1p file is imported into the L element. The frequency is dialed to the point where the L element impedance crosses the G=0.02 circle, and the shunt inductor adjusted until the impedance at G is approximately 50+j0Ω.

In this case, a 2.9µH inductor could be prototyped as 10t of 2mm wire on a 38mm diameter form, stretched out to 33mm in length. (Design from Hamwaves RF Inductance Calculator.)

Above, the VSWR=2 bandwidth is about 130kHz.

Now the frequency at which we achieve this match might not be what we would prefer.

To do this, we go back to Step 2 and adjust the antenna (length; loading coil inductance, position; etc) to move the frequency where Rp≈50Ω.

Having done that, we calculate the inductance of the shunt coil, make a coil and measure / adjust to minimum VSWR at the preferred frequency.

The sweeps above were used to show what is happening in a wider context, but they are not essential. You do not need a scanning analyser… but scanning might well provide more confidence that the antenna is behaving as you assume.

You could simply set your analyser to display the parallel equivalent components at the frequency of interest, adjust the system to achieve Rp≈50Ω, note the required parallel opposite reactance, calculate the inductance of the shunt coil, make a coil and measure / adjust to minimum VSWR at the preferred frequency.

]]>Above is the prototype transformer wound with 14t of 0.71mm ECW tapped at 2t. The mm rule gives some scale. The turns are close wound, touching on the inner diameter of the core.

Leakage inductance is the enemy of broadband performance, so it is important to minimise it.

Leakage inductance is affected by winding geometry. It is important to avoid opportunity for flux inside and around conductors that does not also ‘flow’ in the core, such flux does not ‘cut’ the other turns and is flux leakage. The winding should hug the core, so winding with wires with thick insulation, thick inflexible wires, and topologies like the Riesert cross over may compromise leakage inductance. High core permeability and high ΣA/l help to minimise wire length which helps in minimising flux leakage.

The transformer is configured as an autotransformer rather than separate primary and secondary windings to again minimise leakage inductance.

Above is a chart derived from s11 looking into the transformer primary with the 14t secondary short circuit showing the equivalent series inductance. The value will be taken as a total value of 236nH, or 118nH a side in the split model.

Note that the inductance at low frequencies is almost independent of frequency even though core permeability changes at these frequencies (see the #43 material data sheet), showing that, for the most part, leakage flux exists elsewhere than the core itself.

A simple model works quite well for predicting nominal load performance on low ratio transformers but less well for high ratio transformers, so best to proceed to measurement of the prototype with nominal load. A load was made from two small resistors in parallel having a combined DC resistance of 2480Ω, quite close to the nominal 2450Ω.

Measurement of the uncompensated transformer hinted that some 50-100pF was the likely optimum compensation capacitance. The transformer was measured with 47pF and it was under compensated, 100pF was perhaps more than needed but InsertionVSWR at the mid to high frequencies was not compromised so since there was a 100pF silver mica capacitor on hand, it was committed to the prototype.

Above is a plot of InsertionVSWR with 100pF silver mica compensation. InsertionVSWR is less than 2 from 1-30MHz. As this type of transformer goes, this is a very good result.

Note there is some contribution of the connecting wires to this response, it is just not possible to use zero length wires to connect the secondary load circuit… but then that applies to its application circuit as well.

Loss modelling ignores conductor loss. Even though the conductors are relatively small, the effective RF resistance referred to the primary side is very low and insignificant compared to core loss. Compensation capacitor loss is modelled, Q=1000 assumed for the silver mica capacitor, but his can be quite a deal lower and very significant for ceramic capacitors.

Above is a plot of the expected loss in dB, magenta on the left scale, and watts @ 50W continuous, red on the right scale. Maximum dissipation is less than 5W which should be accommodated within safe temperature rise for an unenclosed transformer.

Next step is to measure the transformer performance under power, capturing thermographs to confirm the predicted dissipation and power rating.

]]>