DIY RF EMC Biconic Antenna

[nctnico]

Hi
Something I haven't yet nailed down is the choice of balun. 
I don't yet know the frequency range of the antenna, nor do I know the characteristics of the antenna.

An easy way is to simply connect a coax cable directly with a ferrite clamp around the cable near the antenna feed point. That way the shield of the cable does not become part of the antenna.

[dazz1]

An easy way is to simply connect a coax cable directly with a ferrite clamp around the cable near the antenna feed point. That way the shield of the cable does not become part of the antenna.

I think what I will do is make a Roberts balun driving a 50ohm load to characterize the balun. 
Then connect the antenna.  I need to finish the antenna first.

A VNA would be really useful, but I don't have one.  I do have a Siglent spectrum analyser with a reasonable DIY RF bridge.

[tautech]

If your VSWR/Return Loss option is active you can do return loss and SWR
Activating Markers will bring up real measurements.

[dazz1]

Hi
I have all software options enabled so I can do VSWR/Return Loss OK.
Attached is an image recording the rf bridge return loss with a DUT = 50 ohm.     If the bridge and 50ohm loads were perfect,  the reflected RF would be zero.    My DIY Aliexpress bridge can get down to -44dB, which is good enough for me. 

Many of these bridges have a factory fault.  See this link: 
 

Mine did not have this fault.

It would still be nice to have a VNA.

[tautech]

Hi
I have all software options enabled so I can do VSWR/Return Loss OK.  It would still be nice to have a VNA.

Sorry no plans to visit anytime soon and currently I'd be little use anyways as all VNA's I owned are now sold....even the one I had arrive to replace my demo SVA1032X. 
Only a new SVA1015X left that I could cut you a deal on Dazz if it's of interest.

Enjoying your fine work.

[dazz1]

....

Only a new SVA1015X left that I could cut you a deal on Dazz if it's of interest.

Enjoying your fine work.

As much as I would like to take up your offer, I just couldn't justify the $$ for the use it would get.  Even the SSA is an extravagance for me.

[nctnico]

Problem is that VSWR doesn't tell you the whole story when an antenna that is used to receive signals. It only shows how well an antenna is matched to the receiver / transmitter but doesn't necessarily means power is radiated / received most efficiently. Measuring real antenna performance in an absolute way is an art in itself.

For EMC antenna testing, a comb generator with a known output level at the frequencies it is producing, is a much better way to determine its performance.

Something like this one from Tekbox:
https://emcsupplies.com/products/tekbox-radiating-comb-generator-tbcg1

[dazz1]

Problem is that VSWR doesn't tell you the whole story when an antenna that is used to receive signals. It only shows how well an antenna is matched to the receiver / transmitter but doesn't necessarily means power is radiated / received most efficiently. Measuring real antenna performance in an absolute way is an art in itself.

For EMC antenna testing, a comb generator with a known output level at the frequencies it is producing, is a much better way to determine its performance.

Something like this one from Tekbox:
https://emcsupplies.com/products/tekbox-radiating-comb-generator-tbcg1

That is why I am making two identical antenna.   I can normalise the SSA with a cable, then compare the signal sent via the two antenna.  That doesn't tell me anything about matching or impedance.  An SSA won't give me a Smith chart to stare at.
My ancient sig-gen is calibrated to within 0.1dB, and it has a GPIB interface.  I can synthesize a comb generator by stepping across the frequency range to produce an output that looks like a comb generator.  Better still, I can just sweep the sig-gen across the freq range.

It would all be a lot easier with the VNA I don't have.

[tautech]

An SSA won't give me a Smith chart to stare at.

Yet just VSWR and Return Loss measurements are valuable and even getting SWR low for the frequency/s of interest can obtain good performance and provide enough info to get each close to matching.
They are very similar to the inverse of the other.

[nctnico]

Problem is that VSWR doesn't tell you the whole story when an antenna that is used to receive signals. It only shows how well an antenna is matched to the receiver / transmitter but doesn't necessarily means power is radiated / received most efficiently. Measuring real antenna performance in an absolute way is an art in itself.

For EMC antenna testing, a comb generator with a known output level at the frequencies it is producing, is a much better way to determine its performance.

Something like this one from Tekbox:
https://emcsupplies.com/products/tekbox-radiating-comb-generator-tbcg1

That is why I am making two identical antenna.   I can normalise the SSA with a cable, then compare the signal sent via the two antenna.  That doesn't tell me anything about matching or impedance.  An SSA won't give me a Smith chart to stare at.
My ancient sig-gen is calibrated to within 0.1dB, and it has a GPIB interface.  I can synthesize a comb generator by stepping across the frequency range to produce an output that looks like a comb generator.  Better still, I can just sweep the sig-gen across the freq range.

It would all be a lot easier with the VNA I don't have.

I don't think that will work to calibrate your antenna.
The problem with using a signal generator and a random antenna is that you don't know how well the antenna emits the signal coming from your signal generator. The same for using 2 identical antennas. You don't know how much power from the generator ends up being transmitted, you don't know how much energy that is being transmitted gets received and you don't know how much of the received power ends up in your receiver. Too many unknowns, too many variables. Bottom line: using a comb generator with a known output level takes all of those unknowns away.

To show the effect of using an antenna to antenna setup, I refer to the test I did to compare a self made log-period antenna to a self made bowtie antenna. I used a 27MHz antenna to transmit the signal from my network analyser:

https://www.eevblog.com/forum/rf-microwave/log-period-antenna-for-emc-measurements/msg4785857/#msg4785857



You can clearly see the signal level (purple and orange traces) jumping up & down as the transmitter antenna goes in&out of resonance. As my test is relative, the absolute levels didn't matter to me. I'm going to get myself a comb generator (the one from Tekbox or another one with a knows output) to calibrate my log-period antenna and TEM cell. From the result I can create a correction table that I can use together with the receiver. That way the mismatch between the antenna and receiver is also calibrated away.

You could opt to add a balun but you'd need to make relative measurements as many points to see if the balun gives an improvement or not. Keep in mind that for a receiving antenna the mismatch doesn't matter much because the cable is terminated at the receiver and thus you won't get reflections. It works similar to a low-Z probe.

[dazz1]

Hi
The method only works with two identical antenna, and it is not my method.  It is described in one of the MIL-STD461 versions.  Probably other references as well.
I can apply a known signal level from my sig-gen.  I normalize that via a direct cable connection to my SSA.  This effectively calibrates the SSA against the sig gen.
Setting the antenna a known measured distance apart referenced against calculated signal between two (theoretical) isotropic antenna. 
If the antenna are identical, then gain of each actual antenna will be the same.

[nctnico]

But you still don't know what gets lost between generator - transmit antenna and receive antenna - receiver. For that you'd need to measure the SWR at your test frequencies and compensate. A while ago I read a report from R&S where they use 2 different antennas in 4 combinations in order to retrieve antenna gain versus frequency.

[dazz1]

But you still don't know what gets lost between generator - transmit antenna and receive antenna - receiver. For that you'd need to measure the SWR at your test frequencies and compensate. A while ago I read a report from R&S where they use 2 different antennas in 4 combinations in order to retrieve antenna gain versus frequency.

I have a rf bridge so I can measure SWR.   If the sig-gen sends 10dBm to the antenna and 3dBm is reflected, then I know that 7dBm is being transmitted into the ether.  Both the Roberts balun and the antenna are entirely reactive (no power lost to resistance).   The same applies to the signal path from the antenna to the receiver (spectrum analyzer).

The unknown I am trying to find is the gain vs frequency of my antenna.   In an ideal world, I could mount an isotropic antenna (gain 0dB) on one end and my antenna on the other (at a known distance separation), and measure gain directly.  No one will sell me an isotropic antenna, not even R&S, so the next best alternative is to make the connection between Rx and Tx with a cable (gain = 0dB, no separation loss).    That provides a reproducible reference. 

So the problem is then that I don't have a calibrated antenna as a substitute for an isotropic antenna.   If I mount two identical antenna (known separation) with identical gain vs freq, then I can simply measure and calculate the gain.
If the matched antenna pair measure a gain of 10dB, then each antenna must have 5dB gain.  That is it.  I can then produce a measured, verifiable gain/freq curve for my antenna.

So I don't need a VNA to do any of that.  It would be useful to measure the impedance of the antenna so I can match the antenna to the balun. 

I do plan to use a 3rd antenna (any old coat hanger would do) specifically to look for any difference in the gains of my individual antenna.  If they are identical, the gain/freq curves will also be identical.   I am putting a lot of effort into ensuring the two antenna are identical.  I am working to far tighter tolerances than are necessary to build a biconic antenna.

So a VNA would be nice and I want one, I just don't need it for a one-off antenna project.

[dazz1]

Hi
Another ants step forward with the antenna.
I glued in the spring detents with LocTite.

The photo show bits of element inserted to apply pressure to the spring ball detents to keep them in-place while the LocTite sets.  If they ever need replacing there are holes in the back of the bracket to push out the detents.

[dazz1]

After waiting for the LockTite to set, I assembled the antenna.
It just snaps together so easy peasy to assemble/break down.  No threaded elements to turn into place.

The centre element is made slightly too long so it is clamped in place by the 6 elements pulling down on it.
All the parts will fit flat in a cheap Chinese gun case.

I still need to make the loop shorting bar that will also snap in-place.
I also need to figure out what balun to use, and then make it.  That is not a simple thing to do.

[dazz1]

Hi
The next step is to make the shorting bars and brackets.   These damp oscillations within the loops formed by the 6 elements. 
This could be easy or it could be complicated.  I chose complicated.
I have extended the snap-together theme to the shorting bars. 

The brackets will connect the shorting bar to one element and the central bar.  They will be held in place with the sprung ball detents.

The centre bar is offset so the ball is over-centre to keep it in place in the slot. The shorting bar has a flat that the ball rides in. This will allow the distance between the brackets to vary a few millimeters without having to stress anything.

Simple really.   Now I just have to make them.

[dazz1]

Hi
I have started to make the over engineered, overly complex brackets for the biconic shorting bars.
I am only making 4 brackets, enough for one antenna, as prototypes. If they meet expectations, I will make another 4 for the 2nd antenna.

The picture shows a setup for drilling holes in a precise location and depth.

[coppercone2]

the real antenna costs thousands of dollars so I am sure you should feel bad about not using coat hangers to short it out

its OK

you won't get people off your back unless you figure out how to do it with bailing wire. but then someone will tell you that your wasting bailing wire because you don't need an antenna.

[dazz1]

Hi
A rational person would pay the thousands for a tested, proven, calibrated antenna, but I never pretended to be rational.

[coppercone2]

ehhhhhh I bet it will be pretty close. That kind of stuff is when someone thinks a business deal or law suit problems. I bet its more brand name then anything (vouching). cover your ass by over spending so to speak. like board room upper management paranoia.

RF sales + federal authorities + lawyers + statements = insanity with a wallet

Its kind of funny with RF because it all changes if you just move it physically. Something that stays in place is way more justified for high cost/precision/checking (i.e.power company connectors)

[tautech]

Hi
A rational person would pay the thousands for a tested, proven, calibrated antenna, but I never pretended to be rational.

Yet you are in person....or we're both nutters ! 

[dazz1]

Hi
The thing I like about RF is that I can start with raw materials I have in my workshop and turn them into something that does something. 
I can't do that with regular electronics where I have to buy components and assemble them to do something that probably needs programming.   

You can't buy snap-together EMC antenna (at least I haven't found any) like I am making.  That's the other thing I like about RF. I can make something not quite like anything that anyone else has made before.

[dazz1]

Hi
Another ant sized step forward with the completion of the shorting bars, with the over engineered and overly complex brackets.  I would have preferred something that looked a little more organic in shape but I'd really need a CNC mill for that.

These shorting bars attenuate the resonance that happens inside the antenna. 

The next thing to tackle is the housing that supports the antenna.   I also need to make and test the baluns.   

[tautech]

When you you use the shorting bars Dazz ?

[dazz1]

When you you use the shorting bars Dazz ?

All the time.  They are only required because the conic is approximated by the 6 elements. These elements form loops inside the antenna.  The shorting bar reduces this unwanted resonance.