DIY RF EMC Biconic Antenna

[dazz1]

Hi
My latest DIY project.
I am making a biconic antenna for EMC pre-compliance testing.  I am actually making two identical antenna so I can calibrate them against each other.
With any calibration, a reference is required.  I have recently had my restored Wavetek 2025 output calibrated to within 0.1dB.  My starting reference will be a hard wired (coax) connection between the signal generator  and the spectrum analyser. 
I will then replace the coax with two identical antenna separated by an exact distance, and orientated the same way relative to each other.  I can then calculate the gain compared to a pair of isotropic antenna (which only exist in text books).
I simply compare the measured gain of the antenna to calibrate them.
 
Making two antenna is only a small increment in effort.  Making two identical antenna will be faster than making dipoles etc.    I have spent a lot of time with the detailed design.  Much of the time required to make these is in the setup for each machining operation. 


The approach means I will end up with two identical calibrated antenna.  I only need one.

The first step is to make the top hat.  This is the piece that all of the antenna arms and the centre element connect at. 
Each arm will be held in place with a ball and spring detent.  The arms will snap into place.  This will make assembly and storage very easy.

The cross section shows the interior detail of the top hat. The antenna arms will slide into the top had and be held in place with the detents. 
I took a lot of care in the marking out to ensure accurate location of the holes.
The setup ensures that every part is identical.  I have 24 radial holes to drill and 24 opportunities to screw up.  The setup will reduce the risk of error.

[mag_therm]

Hi dazz,
To make bicycle spoke hubs on the lathe I made a tool post drill spindle with a small dc motor. The indexing was by a 64 tooth gear on the back of the lathe main spindle with a tooth shaped pawl. That gear allows only 2, 4, 8, 16, 32, 64 holes on pcd.

[artag]

When spacing them an exact distance apart, do you measure to the centreline or to the nearest point, biconics being quite large  ?

[nctnico]

Interesting project!
Recently I have been testing with antennas as well for EMC testing (I bought a relatively cheap discone antenna which is also wideband) but I quickly found out that there is too much junk floating around in the air already to make any sensible measurement. IOW: I'm receiving so much noise already that it swamps what I want to measure. I came to the conclusion that I'd need a shielded room in order to have any chance of making a meaningfull measurement. Do you have a shielded chamber?


Also: how large is your antenna going to be? I looked at the $500-ish biconical anntenna from Tekbox but at the lower end (30MHz to 100MHz) it has very poor reception due to the relatively small size.

[coppercone2]

Nice antenna. I have the parts needed for a balun, but I need to make the enclosure (I made a antenna at the wrong impedance, I am not sure I measured it right but it looks like I got like 130 ohms. The balun is for me the hard part of these antennas. I tested them in the basement and it looks like under 1GHz its pretty crowded.

Also have you considered a pre-selector for broad band antenna work? They always have a preselector in the RF labs when they use broad band antennas on the compliance analyzer machine (I think those machines are built with a preselector). Narrow band antenna does the work for you with the filtering.. broad band requires extra hardware I guess

For a hobbyist that collapsible design of the antenna you made in is very important. They are rather imposing on the laboratory.

[dazz1]

Hi dazz,
To make bicycle spoke hubs on the lathe I made a tool post drill spindle with a small dc motor. The indexing was by a 64 tooth gear on the back of the lathe main spindle with a tooth shaped pawl. That gear allows only 2, 4, 8, 16, 32, 64 holes on pcd.

Hi
I looked at using the lathe but I needed precise depth control.  That was easier to do on a mill than a lathe.

[dazz1]

When spacing them an exact distance apart, do you measure to the centreline or to the nearest point, biconics being quite large  ?

I marked the centre lines and aligned on these.  The priority was to achieve 60 degrees between elements. 

[dazz1]

Interesting project!
Recently I have been testing with antennas as well for EMC testing (I bought a relatively cheap discone antenna which is also wideband) but I quickly found out that there is too much junk floating around in the air already to make any sensible measurement. IOW: I'm receiving so much noise already that it swamps what I want to measure. I came to the conclusion that I'd need a shielded room in order to have any chance of making a meaningfull measurement. Do you have a shielded chamber?


Also: how large is your antenna going to be? I looked at the $500-ish biconical anntenna from Tekbox but at the lower end (30MHz to 100MHz) it has very poor reception due to the relatively small size.

I don't have a shielded room.  For this sort of testing facility, an anechoic chamber is required.  They cost  $ $ $ $ $.
TEM and GTEM cells are more reasonably priced, and can be DIY built, but still lots of $ $ and bulky.  I won't be building one.

I have access to a RF quiet valley about half an hour from where I live.  There is a nearby low capacity cell site, but nothing else.  It is about as quite as I will find.  I does have mains power.

End to end, my bi-conic antenna will be about 1.1m.  That matches the dimensions of the original Roberts design.    Noting your comments about the Tekbox, I can easily make it larger.  The elements will snap into place.

[dazz1]

Hi
Today I worked on drilling the holes that will house the sprung ball detents. 
The holes have to be drilled to a precise depth.  If the hole is too shallow, the antenna element will hit body of the detent.  If the hole is too deep, the sprung ball will not adequately grip the antenna element.

For each hole, the following operations were required.
1.  clamp work piece in position.
2.  spot drill to guide the first drill bit.
3.  drill 2.5mm hole to allow detent to be pushed out if one ever needs replacing.  Also, this hole guides following drills.
4.  drill 5.5mm hole almost to depth to remove most metal.
5.  drill a 6mm hole to 9mm depth to guide the d-bit drill.
6.  drill a 6mm dia, flat bottomed hole for the detent to sit on.

6 ops, 6 holes and 4 work pieces.  That is a total of 144 operations for just one feature.
My setup included a Vee block to locate the work piece x & Y.  A new drill inserted backwards into the horizontal holes allowed me to index work piece rotation.    This setup held the work piece with a very close tolerance. 

The flat bottom hole was formed with a D-bit.  This is a type of DIY precision drill I have made for another project.   

[dazz1]

Hi
I had to precisely drill the flat bottom hole for the detents.  First I needed to know how deep the hole needed to be.  I used a height gauge to measure the top of the hole that the elements will slide into. 
I then figured out how deep the detent hole needed to be.    The setup I used to drill the horizontal element holes ensured the height of the horizontal holes was within 0.1mm measured tolerance.

[dazz1]

Hi
I made a gauge specifically to check the depth of the holes.  I made this so I could use it while work piece was clamped in place.

[dazz1]

Hi
The last operation was to use the D-bit to form a flat bottom hole.

[dazz1]

Hi
The work done on the top hat is now finished.
The next item is the bottom hat.  That is the conical part where all the elements meet.

[dazz1]

Hi
The next part to make is what I will call the bottom hat.  It is the part where all of the elements meet. 
It is going to be tricky because of the shape and the lack of obvious references. 

Most people making something with electronics likely start by turning on their soldering iron.
When I do electronics, I start by turning on my lathes and milling machines.   

[nctnico]

Interesting project!
Recently I have been testing with antennas as well for EMC testing (I bought a relatively cheap discone antenna which is also wideband) but I quickly found out that there is too much junk floating around in the air already to make any sensible measurement. IOW: I'm receiving so much noise already that it swamps what I want to measure. I came to the conclusion that I'd need a shielded room in order to have any chance of making a meaningfull measurement. Do you have a shielded chamber?


Also: how large is your antenna going to be? I looked at the $500-ish biconical anntenna from Tekbox but at the lower end (30MHz to 100MHz) it has very poor reception due to the relatively small size.

I don't have a shielded room.  For this sort of testing facility, an anechoic chamber is required.  They cost  $ $ $ $ $.
TEM and GTEM cells are more reasonably priced, and can be DIY built, but still lots of $ $ and bulky.  I won't be building one.

I have access to a RF quiet valley about half an hour from where I live.  There is a nearby low capacity cell site, but nothing else.  It is about as quite as I will find.  I does have mains power.

End to end, my bi-conic antenna will be about 1.1m.  That matches the dimensions of the original Roberts design.    Noting your comments about the Tekbox, I can easily make it larger.  The elements will snap into place.

The bigger ones from Tekbox (in the US $1k ballpark) is about 1.8m.

If I get motivated enough, I might try to make a shielded tent that fits in my office. The fabric can be found on Aliexpress. A plastic frame from thin PVC tubing shouldn't be hard to make.

[artag]

When spacing them an exact distance apart, do you measure to the centreline or to the nearest point, biconics being quite large  ?

I marked the centre lines and aligned on these.  The priority was to achieve 60 degrees between elements.

That's in construction though, right ?

I'm wondering what happens when you come to measure the loss in a system with two aerials. They are a fixed distance apart and that's used to determine the signal lost due to inverse square law.

When making that measurement, the centreline might be an obvious point to use, but the nearest points of the large elements will be somewhat closer. I guess the ideal is to space them widely enough that the error is small, but that might give you rather high loss.

[Jay_Diddy_B]

Hi dazz1,

I have been down this path. There are some pictures of my antenna here:

https://www.eevblog.com/forum/testgear/diy-biconical-antenna-30-300mhz/msg2261550/#msg2261550




Regards,
Jay_Diddy_B

[dazz1]

Hi
I have started to make the piece that all the elements meet at.
It is a complex geometric problem to align the tools to end up cutting in the right places.  Lots of opportunities to screw up.

[dazz1]

Also: how large is your antenna going to be? I looked at the $500-ish biconical anntenna from Tekbox but at the lower end (30MHz to 100MHz) it has very poor reception due to the relatively small size.

The bigger ones from Tekbox (in the US $1k ballpark) is about 1.8m.

It won't be too difficult for me to have multiple sets of elements.  I would also need matching balans.

[dazz1]

I'm wondering what happens when you come to measure the loss in a system with two aerials. They are a fixed distance apart and that's used to determine the signal lost due to inverse square law.

When making that measurement, the centreline might be an obvious point to use, but the nearest points of the large elements will be somewhat closer. I guess the ideal is to space them widely enough that the error is small, but that might give you rather high loss.

Measure centre line to centre line.
The method I plan to use requires that the antenna are identical.  Any difference in performance or reflections, calibration etc will degrade the accuracy of the measurements.

[tautech]

Dazz, have you had to invest in a new collet chuck ?
Don’t remember seeing such a shiny beast when I dropped by.

Looks a cool project. Do you have a reflection bridge for your SA to characterise these antenna ?

[dazz1]

Dazz, have you had to invest in a new collet chuck ?
Don’t remember seeing such a shiny beast when I dropped by.

Looks a cool project. Do you have a reflection bridge for your SA to characterise these antenna ?

Hi
I made the collet chuck quite a while ago but only for the small Denford lathe.
Yes, I have a bridge to characterise the antenna but what I really want to know is the gain. 

If I start with a cable between the Tracking Generator (TG) and the spectrum analyser, I can normalise to that.  Zero error.
If I then split the cable half way along, I can reference the isotropic gain to the cable. 
If I place the two  antenna, separated by the far field (lets say 5m), then I can compare their gain against and isotropic antenna. 

I will have two identical antenna, with unknown gain.  Two knowns and two unknowns.  I just need to halve the measured gain between them.  If I measure 10dB between the two antenna, then each antenna must have 5dB gain. 

[tautech]

Yes understood that from the outset however the bridge will help you impedance match them and determine their bandwidth.
I thought of you when she indoors said she was coming down your ways to the NZ Quilting symposium in October and I am tempted to tag along and go visiting customers and you while she's in the various lectures.
I'd be bringing my SVA1032X of course. 

No promises but we'll see how things pan out.

[dazz1]

Yes understood that from the outset however the bridge will help you impedance match them and determine their bandwidth.
I thought of you when she indoors said she was coming down your ways to the NZ Quilting symposium in October and I am tempted to tag along and go visiting customers and you while she's in the various lectures.
I'd be bringing my SVA1032X of course. 

No promises but we'll see how things pan out.

 

[tautech]

Yes understood that from the outset however the bridge will help you impedance match them and determine their bandwidth.
I thought of you when she indoors said she was coming down your ways to the NZ Quilting symposium in October and I am tempted to tag along and go visiting customers and you while she's in the various lectures.
I'd be bringing my SVA1032X of course. 

No promises but we'll see how things pan out.

 

The only catch is you have just over 5 weeks to have it ready for testing !
Could even go all hardcore and bring down a 4 port SNA5004A VNA !
Would you even have enough N type cables ?