Mini Whip at a noisy urban area - tips for grounding and the de-noising?
I plan to buy a second Kiwi for at home.
Noise levels are increased dramatically at the last 2 years. My home is in a 10-story soviet style block. The MiniWhip is attached to the elevator house on the roof - no other way here. Pic: https://postimg.cc/qzjfLnXp
New motors for the ventilation system, new elevator control system, etc.
The MiniWhip worked with acceptable SNR with grounding 3 points of the feedline - no any common mode choke. Worked good with my Kiwi which is at the country house now (with the dipole)
First grounding point: PCB of the MiniWhip - connected to the lightning rod. Second GND: around the halfway of the feedline to the another lightning rod and finally inside the flat next to the shack connected to the water pipe of the radiator.
So I added a common mode choke ~3.6m far from the Whip, made with FT240-31 toroid with 22 turns of an RG174 thin coax (pic: https://postimg.cc/xq9n30Dy), noises are decreased BUT a "gain peak" appeared around 6 mhz. The 49MB stations overloaded the receiver. First GND remained at the PCPB GND point. The 3.6m coax resonates on 6 mhz? Okay, I put the first GND point BEFORE the CMC (not at the antenna side) and the "gain peak" was gone, but the LW band became noisy. Let's try some thing. I put back the GND to the GND point of the Whip and I move the choke ~ 10m away from the whip - the gain peak appeared again, but on 9 MHz. So the 31 MB stations overloaded AGAIN. I moved ~20m from the whip, next to the 2nd GND, (the 2nd GND is on the TX side of the Choke). Gain peak moved to ~12 mhz. I added the 2nd GND point to the antenna side of the CMC - gain peak gone but LW+MW became noisy! Argh! So the lightning rod became noisy too. Disconnecting the 2nd GND - more noise! CMC removed - loot of noise!
So what is the best solution for me? I must filter the GND too. The whole feed line is ~40 meters long.
Maybe this can work? Whip - 3.5m coax - CMC - 1st GND (connected to the RX side of the CMC) - Coax - CMC - 2nd GND (Connected to the RX side of the 2nd CMC too) - Coax - PSU with GND - Coax - Choke - Shack?
I plan to add FT140-77 with 19-20 turns of thin cable to the whip and to the shack and keep the FT240-31 but with 13 turns of RG58/59 instead of 22 turns of RG174. (Combining them to cover the 100 khz - 30 mhz band...)
What do you think?
Unfortunately Mini Whips and other E-Probe type antennas tend to be 'noise magnets'. The best solution is to use a balanced antenna such as an active dipole or broadband active loop. As these are much more resistant to common mode noise pickup on the feedline.
The unbalanced nature of the antenna, and the very high value of feedpoint impedance, means that it is practically impossible to choke off unwanted common mode signals on the feedline, without also attenuating the wanted signals. This is because the feedline is required to be part of the 'antenna'. If you could provide a perfect feedline choke at the antenna, it would simply not work, as there would be no ground reference.
You have to do two things, which may seem the same as each other, but they are not.
You define the start and stop point by introducing some form of common mode choke (or chokes) at that point. You then need to do everything you can to try and keep the 'antenna' part as 'clean' as possible, by routing it way from noise sources, and providing a 'ground' reference that is also as 'clean' as possible. This is tricky on a roof, but in the past I have used the building lightning earth, metal roof panels, or even a few wires laid out over the roof panels to perform this function.
Once you have established your 'clean' earth reference, you then need to choke off unwanted signals that may have been introduced into your feedline on it's journey from the roof to your receiver. This means adding common mode chokes to the feedline on the 'receiver side' (not the antenna side) of the feedline where you have connected it to your 'clean' earth. More chokes at different points will help, including some at receiver itself.
It's a lot easier if a Mini Whip is used outdoors, but even then it can take a lot of effort to control common mode noise pickup on the feedline.
This diagram may help you to visualise the situation.
Thanks the method on the picture worked perfectly when I had the miniwhip at the weekend house's back-garden :) It was on 5m long wooden pole and GND point+choke near the ground level. So...
For an example MiniWhip +4m the feedline for the "part" of the antenna, and I had to GND and choke there? my plastic-tube pole is 4m long on the roof.
The plastic tube doesn't really matter too much, because it's still the coax that is forming part of the antenna.
The antenna is basically measuring the potential difference (voltage) between the Mini Whip 'plate' and whatever it 'thinks' is the 'ground'.
This potential measurement image from the site I have linked to below, gives an indication of what is going on.
In the previous diagram from my website, I showed multiple connections to ground along the length of the feedline.
If you don't have an actual ground, try to think of a way in which one could be simulated. It doesn't have to be an actual ground connection, but something that is conductive, has a large electrical 'mass', and that is relatively clear of possible interference sources.
Pieter, PA3FWM or websdr fame, has some good notes on the subject of how Mini-whip antennas work, which may help provide some further ideas.
Thanks for these infos, now everything is more clear :)
Now the Choke is ~ 25m away from the MiniWhip PCB (and the GND) and I have noise/gain peak on 12 MHZ = 25 meterband and 6 MHz = 50 mb, 4000 khz = 75 mb, etc.
Now the only question is... What is the better option? (I have limited budget now...)
MiniWhip, GND on the PCB and CMC next to the antenna? Or leave 4m of coax, and do the GND and choke on the "ground level"? (This worked in the garden with clear GND and 5m of coax)
I'd try the choke at 'ground' level, or just after the point at which the the 'ground' wire and coax go in different directions. If they run close to each other they will be at the same RF potential, so you need to isolate the cable carrying the noise, from the 'ground' connection, by means of the choke.
But as always some experimentation will be required to find the best solution for any particular situation.
I finished with the tests with the following relsults:
Connecting the GND BEFORE the chokes (on the RX side at the bottom of the pole) = almost clear SW bands but S10 noise on LW+MW.
Connecting the GND AFTER the cokes (on the ANT side at the bottom of the pole) = LW+MW are almost clear, but S3-S10 noise levels on SW bands, especially around 5,4 MHz. The GND wire (lightning rod) is cca. 40-45m long and running on side of the buldinding into the earth. Maybe this collecting noise.
So there is any solution to filter the GND? Or I must accept these noise levels?
You can see the lightning rod wire here: https://goo.gl/maps/Z51XXtFLxP2zyd5m6
And my MiniWhip is here: https://goo.gl/maps/GF92a6289vYv4Jbx9 (the little "bug" on the center of the picture)
#1 FT140-77 with 19 turns of RG-174
#2 FT240-31 with 13 turns of RG-58.
This is tricky.
It's probably because the lightning rod is a 1/2 wave some where around 4MHz and 1/4 wave at around 2MHz.
The only real solution is to use a balanced antenna, but you may be able to find a configuration that works well enough for your purposes.
Unfortunately you are now in the territory where only experimentation can help find a solution for your specific installation.
You can't really filter the ground connection, but you could experiment by adding a small clip-on ferrite bead onto the ground wire and see if you can 'tune' it to find a compromise between the two configurations.
For example, if you connect the ground wire so that LW & MW are good, then by adding a choke that operates mainly on the HF bands, you may be able to achieve some form of 'selective' filtering.
Good luck with your experiments,
So. I spent a lot of time (and coax) to trying different setups. Now:
The 1st gnd point is directly on the PCB GND pin. 3,5m from here there are two chokes:
19 turns of RG174 on FT140-77 toroid and 13 turns of RG59 on FT240-31 toroid.
Noise almost gone from 6 mhz, good on LW and MW.
BUT... when I finished my work noticed, that an S9-S10 noise appeared on the 11-13 MHz band and it almost wanishes the stations from there.
So now works acceptable on 100 khz - 10 mhz and from 14 to 30 mhz. but only noise on 11-13 MHz.
Unfortunately that's how it tends to go, you fix one noise source and then find another.
I seem to spend a lot of my time trying to track down and fix individual noise sources, so that I can obtain the lowest noise floor possible but it's not easy.
It's a case of diminishing returns, once you have got rid of the really bad stuff, much more effort is required to clear the remaining low level noise.
I'm currently developing another KiWi site on a remote hilltop. I've spent a lot of time and effort screening the building with aluminium mesh, bonding all the feeders together and adding a big mains filter box.
In testing it looks very good, and is (IMHO) is likely to be among the best KiWi's in terms of SNR rating, but it still has some low level interference remaining that is proving hard to completely eliminate before I'm happy to make it public.
Good luck with further experiments.
In general mono-poles, whether large or 'whips', tend to be difficult to use because of common mode noise current ingress. See a note I wrote about this and other SNR issues Here.
While semi-solutions of the form "use this ferrite, with this many turns in this way" are common, in reality the problem is severe enough that the ~20 dB (or so) reduction in CM noise current is simply not enough to allow a receiver to achieve the propagated noise floor.
The other "solution", that of identifying and quenching noise current sources can be very frustrating and amounts to playing "Whack-a-Mole". In my opinion it is far better to remove the mechanism that is allowing ingress of these currents into the receive system. Mono-poles rely on a return path that does not include tese currents, down to a very low level. Radial systems, "grounding" and other image planes generally do not accomplish this as well as a symmetric antenna, perhaps a dipole rather than a monopole, which provides its own return path. Even to make this work requires very good symmetry, a.k.a. "balance" and there again simple ferrite solutions such as broadband "baluns" do not suffice.
Creating a broad band low noise antenna system is not easy, it seems almost like a career, but attention to ingress mechanism(s) and careful analysis of the problems can really pay off.