DIY RF EMC Biconic Antenna

[piotekp90]

I recently opened up my Tekbox TBMA1 antenna and this is what it looks like inside:

[dazz1]

I recently opened up my Tekbox TBMA1 antenna and this is what it looks like inside:

That is interesting.  It appears to be similar to the Ethernet matching network I rescued from a dead motherboard.  When I tested it, the performance was decidedly mediocre so I started looking for another solution. 
You can see the results of my R&D here:  https://www.eevblog.com/forum/rf-microwave/correctly-calculating-impedance-of-a-biconical-antenna-and-impedance-matching/
My results do not necessarily apply to the Tekbox solution.  That could only be confirmed/denied by testing the Tekbox balun.

 I took a completely different path.  One of the key problems I had with a twisted pair design is the difficulty in sourcing wire with the required and known insulator properties.  This is needed to obtain the correct impedance.   The problem of impedance matching a twisted pair was compounded by on-line twisted pair impedance calculators that produce the wrong answers.   That is why I switched to using coax cable.

Coax requires physically large toroids.  The combination has a measured -20dB bandwidth of 20MHz to 300MHz. I think that is close to the limit achievable with coax/toroid approach.   I am about to change tack to look at a different coax/ferrite topology.  First I need some parts.

[Kean]

I recently opened up my Tekbox TBMA1 antenna and this is what it looks like inside:

This got me curious, so I took the cover off my Aaronia Bicolog 30100.  https://aaronia.com/en/emv-antenne-bicolog-30100

Unfortunately there isn't much to see without some difficult desoldering which I'm worried could be destructive.
I can't even see the type of conductor used through the ferrites, but they are somewhat compliant.

[T3sl4co1l]

Probably the same basic design:



The twisted pairs are Zo = 100R, two in parallel is 50R at the port.  Notice the far ends are grounded oppositely, making one an inverting 1:1 autoformer, and the other a straight pass-thru (delay).  The latter doesn't need any choking but I had the binoc cores so I figured why not. The former drops the full input voltage so needs choking accordingly.  The output in series is therefore 200R, a reasonable match to the, in this case wire frame bowtie, but bicons are similar of course.

Tim

[dazz1]

Hi
I have formed the view that emc antenna performance is largely defined by the balun, which is also the item with the least info, test data results and possibly the weakest link in antenna design. 

[dazz1]

Hi
It has been a while since I posted on this thread, but I have been working on developing a wide band balun for the bi-conic antenna.  It has been a journey.

In this post I am displaying (a.k.a showing off) a prototype mount for the two halves of the bi-conic plus bar.  There is minimal contact between the antenna metal and the plastic to reduce capacitance.  The opening at the Tee end is to allow access for a spanner to get to the M6 nut that will hold the conics in place.

The attached photos are development prototypes used to refine the design. The mount will accept a tube of light steel.  That tube will also hold what I call the linear ferrite balun.   The tube will be clamped to a non-conducting tripod.  Too expensive to buy so I will make one.

I anticipate some may call out saying that it should be a shielded metal enclosure.  That 3D printing the mount in plastic is a mistake. To that I would say, no it isn't.  There will be less than 2cm conductor exposed between the apex of each conic, and the shelter of the RF coax cable feed.  No significant opportunity for RF to leak in.  Certainly not enough to justify the time and cost of a metal enclosure.

[dazz1]

Hi
Now I have decided on the type of balun I will use (see this thread for the story of that saga: https://www.eevblog.com/forum/rf-microwave/correctly-calculating-impedance-of-a-biconical-antenna-and-impedance-matching/new/?topicseen#new), I have made the  central enclosure.  I have now got to the point of assembling the antenna so that it looks like a biconic as shown in the picture.

So now I am waiting to receive a cable assembly from Aliexpress to fit the balun and connector.   Then I need to make a non-conductive tripod.

[dazz1]

Hi
I have finished the pair of bi-conic antennae. 
The balun is fitted in a steel tube that doubles as the way to clamp to a stand.
The ferrites are separated by spacers to stop the ferrites touching each other end to end. The ferrite is very brittle.
The spacers also keep the ferrites away from the steel tube.
The colour of the steel tube changed because I needed a slightly larger diameter to fit the hex nut on the rear of the panel mounted BNC connector. 

Now that I have finished the antennae construction, I need to test and calibrate the antenna.  The problem with testing it is that if it doesn't work, I will have two of the largest paper weights I have ever made.

[dazz1]

Hi
Just done some preliminary testing of the Biconic antenna by measuring return loss. 
The attached images show the test results.
Notes along the bottom of the images describe the tests.

These tests were done under far from ideal conditions but experimentation indicates the tests are repeatable, and not greatly affected by objects in the near vicinity. 

The tests included measuring the return loss of a 50ohm test loads.  The aim of the test being to quantify the limits of the test setup.  A 50 ohm load should have infinite return loss.  My DIY balanced bridge is not perfectly balanced.     I found that reflections on the cable between the bal. bridge and the antenna connector had a significant effect.

Initial testing indicates a bandwidth of around 1.5GHz, substantially greater than the 100MHz-300MHz I have seen on commercially available bi-conic antenna.  Early tests of baluns indicates that about 300MHz is the limit of common wound ferrite cored baluns.   

Even allowing for a wide margin of error, the measured results are looking positive.  More testing is required.

[uer166]

Of great value would be to somehow know the field (V/m) to output (dbuV) factor output, since it's for EMC after all. I'm not sure how you'd be able to characterize this outside of a full anechoic chamber lab with a known reference antenna though. Return loss measurements don't include antenna efficiency loss, which was always vexing for me for RF experiments.

[dazz1]

Of great value would be to somehow know the field (V/m) to output (dbuV) factor output, since it's for EMC after all. I'm not sure how you'd be able to characterize this outside of a full anechoic chamber lab with a known reference antenna though. Return loss measurements don't include antenna efficiency loss, which was always vexing for me for RF experiments.

Hi
I am getting there.  These initial tests are intended to answer questions like, "does it work?" and "how good/bad is my instrumentation setup?".  I have already learned some important things.

I never expected the antenna to work anywhere near the GHz range.  My target was around 500MHz, but I thought that was ambitious given the highest frequency for this size/style commercial biconic I have seen is 350MHz. I should probably be looking at changing the connector from BNC to sma or N-type.

The SWR appears to be typical for a biconic but the length of the cable between the balanced bridge and the antenna makes a significant difference to the measured frequency response.  I think cable resonance is responsible for the periodic response seen in the plots.

From the start, the plan was to calibrate the field to output.   I don't have calibrated dipoles so I have built 2 identical antenna.  Both with unknown gain, but both are identical.   I will be able to measure the total gain, then halve it to get the antenna gain.  The procedure is  included in version of MIL-STD 416.

I don't have an anechoic chamber but I do have access to a substitute for an OATS.  It is a deep valley about 30 minutes outside of a significant population.   It has power available.  Before I go there, I need to refine the test setup. The antennae are still prototypes so I will make changes if I need to.

 

[uer166]

MIL-STD 416.

If you have the specific clause/appendix number handy, would be good to know. It's a big standard!

[dazz1]

MIL-STD 416.

If you have the specific clause/appendix number handy, would be good to know. It's a big standard!

Hi
I will have to find the right section in the right version.
Basically, it requires a (tracking) sig-gen connected to one antenna, and a spectrum analyser connected to a 2nd identical antenna.  The antennae are place a known distance apart.
The test begins by substituting an antenna pair with a cable.  The zero loss reference.
With the antenna pair connected and separated, the total signal path loss is then:  G1 + air path loss + G2
where G1 + G2 = 2G = antenna gain at each frequency.

The flaw with the test method is the standing waves along the Tx/Rx coax cables, and their lengths, will affect the results.  I am thinking I need to fit a 50ohm attenuator at the antenna to suppress the effects of the SWR.

[tautech]

Getting a low SWR across a wide frequency range is very difficult IME.

I need arrange a road trip down your ways with a few goodies in the wagon.
Do you have any long N type cables Dazz ?
I might have some a few metres long I can leave with you.

[dazz1]

Getting a low SWR across a wide frequency range is very difficult IME.

I need arrange a road trip down your ways with a few goodies in the wagon.
Do you have any long N type cables Dazz ?
I might have some a few metres long I can leave with you.

I think it is a given that such a wide band antenna is going to have a significant SWR so it is something I will have to deal with.  It should be noted that the angle of the cones sets the impedance.  My biconics are designed to be 50ohm, but that is clearly not perfect.

I haven't yet brought/built everything I need to make use of an EMC antenna.  I have been incrementally creeping along that path.  The initial testing indicates the antennae work so the next step is to design/make non-conductive tripods.
I haven't yet purchased the cables I will need to do actual EMC testing.

[tautech]

My target was around 500MHz, but I thought that was ambitious given the highest frequency for this size/style commercial biconic I have seen is 350MHz.

Yet none of your screenshots show a Marker set anywhere near 500 MHz.   

[dazz1]

Yet none of your screenshots show a Marker set anywhere near 500 MHz.   

Not a lot happening at 500MHz. 

I suspect a lot of what I am seeing in the plots is cable resonance in the test setup.   The 50ohm test loads don't have cables and don't have a periodic frequency response.  I need to do some more testing.

[tautech]

Yet none of your screenshots show a Marker set anywhere near 500 MHz.   

Not a lot happening at 500MHz. 

I suspect a lot of what I am seeing in the plots is cable resonance in the test setup.   The 50ohm test loads don't have cables and don't have a periodic frequency response.  I need to do some more testing.

If that is your target frequency more tweaks are required.

Never lose sight that an antenna is a transformer, 50 Ohms > 277 Ohms and SWR of 1.

[dazz1]

MIL-STD 416.

If you have the specific clause/appendix number handy, would be good to know. It's a big standard!

The double antenna calibration procedure is contained within MIL-STD-461A dated 1 August 1968, section 5.2.7.

The important equation is attached:

[dazz1]

Hi
Did some more testing of the biconic antenna.  The balun is enclosed in a steel tube that I will use to clamp the antenna to a stand.
These tests measured the return loss of the antenna up to 1GHz of the complete antenna with and without the steel tube.  The aim of the test was to determine if the steel tube within which the balun is installed affected the frequency response. 

The steel tube is 0.45m long, so if it was going to couple with the balun and resonate, there would be a peak/dip at around 333MHz.  There is a little blip at about 330MHz with the steel tube, but nothing that really stands out.  The test setup is far from perfect.  The antenna response changes with my movement near the antenna so no expectation of a perfect match between the two plots.
 
None of the commercial Biconic antenae of this size and style advertise a frequency response greater than 350MHz but this antenna and balun combination are good up to at least 1GHz.

If the steel tube is shown to be a problem, I always have the option of changing it to something else. 

[dazz1]

Hi
Did some more testing tonight.
Standard biconic antenna include an element down the centre, and a shorting element connecting the corner of an element to the centre element.
I tested the effect of removing these additional elements. 

The testing focused on the ~80MHz resonant frequency.  The natural loop frequency of a finite cone.

Note that the test setup is awful.  No anechoic chamber, no OATS, lots of nearby conductors etc.  This testing was qualitative so don't read too much into the data.

I did 4 tests across a 200MHz frequency range measuring return loss:

• The standard biconic config with all elements in place.
• Removing the shorting elements.  On my antenna, these just clip into place.
• No.2 and removing the centre element that is along the axis of the antenna.
• No. 3 and me standing `2m away.  I noticed that if I stood about about 2m (1.9M = half wavelength at 80MHz), the return loss jumped significantly.

Conclusions are that the shorting and centre elements help reduce the natural resonant frequency of a finite sized biconic antenna.
The enhanced return loss when the antenna is 1/2 wavelengths from a conductor are relevant to 2x antenna calibration.  If the antenna are placed at about 2m apart, the frequency response will be significantly affected around 80MHz.

[tautech]

Dazz, I don't remember if with your analyzer you can select a Marker with a Continuous valley Peak ?
They can be very useful when testing antennae.
Marker 2 can be set to the desired resonance to see the difference.

[dazz1]

Hi
I haven't looked too hard at the Marker functions but I will.

Today I tested the return loss with a somewhat better setup.  I call it my home OATS.  I supported the antenna on my EMC rated clothes basket (means the basket is now ten times more $ valuable  $), and then put that on top of my front fence.    I had some crappy RG58 coax to connect to my balanced bridge and spectrum analyser.    I don't have enough cable to do a back to back antenna calibration.  I did have enough to measure the return loss.

In this fence top position, the antenna was a reasonable distance from the nearest reflectors.   That included aluminium window frames, internal wall house wiring, steel roof etc.  My home OATS is a long way from a real OATS.

I recorded the return loss of the antenna out to 1GHz when mounted on the EMC rated basket.  The unedited plot includes lots of periodic response that I interpret as reflections along the cable.  With a non-zero VSWR, reflections along the coax cable are reasonable to expect.  My DIY balanced bridge has proved to be useful but somewhat untrustworthy as the frequency rises.  The first plot shows lots of peaks and dips at about 850MHz. I can't be certain if this is the antenna, antenna coax reflections, or reflections between the spectrum analyser and the balanced bridge.   

The second plot includes a hand drawn red line that I think represents the actual return loss of the antenna.  If this is reasonably accurate then the antenna VSWR is reasonably flat from 100MHz to about 600MHz.  With compensation, it looks to be usable up to 1GHz.  I haven't tested beyond 1GHz because I wouldn't trust my balanced bridge.   I am looking at making a new and better DIY balanced RF bridge shown here:  http://www.dicks-website.eu/return%20loss%20bridge_part1/part1.html  The plan is to make it so it fits directly to the Siglent Spectrum Analyser. 

So the results from testing in my front yard OATS are positive.    I suspect the results would be even better with improved sensors. 

I still plan to do two antenna calibration and I also want to measure the beam pattern.  The balun had 5% imbalance and I want to know by how much this distorts the beam.

In the mean time I have started to design a non-conductive antenna stand/tripod.

[tautech]

That's a fine 84 MHz antenna. 
Rather than need to get the red pen out, engaging some Averaging might help. 

[dazz1]

Hi
I will try the averaging feature, but my understanding is that averaging smooths noisy time variant signals.

For comparison, attached is the VSWR of a commercial biconic antenna.    Note the chart for this only goes to 300MHz.
I think it would be reasonable to say that my biconic antenna displays significantly better performance over a  much wider frequency range.