FlightAware Discussions

Antenna on mast performs worse?

I recently was able to move my feeder from my home location to the top of a single story building, which obviously greatly improved performance.

However, my initial install location was quick and dirty. The antenna was zip-tied to the side of an air handling unit, with the lower 1-2 inches of the antenna obscured by the metal frame of the A/C, and little to no elevation above other roof obstacles.
I recently changed the mount to the top of a (relatively short) PVC mast hoping to gain a bit of reception by clearing line-of-sight over the other air handlers on the roof.
After this change, I saw a quite noticeable drop in message rate & aircraft tracked.

I only tested it with the new setup for 45-60 minutes before returning it to the previous install, and saw the message & aircraft stats return to expected levels.

Can anyone explain why this happened, or what I did wrong?

My mounting locations are limited, since the only accessible roof penetration is the one for this AC unit, so to mount it elsewhere would require longer cable, and still need to be attached to an AC unit or vent pipe on the roof.


My setup:

Original install (which did better):

“New” install (which did worse):

The period marked was the new mast installed:


image

You have a signal graph and a range graph for that time period?

It’s peculiar indeed.

Could always be a loose contact in the connector or a coax that is intermittent.

LNA should be in front of the coax. See if you can swap things around. If the fan motor is a BLDC-motor, can cause a lot of broadband-RFI that gets picked by the coax and amplified by the LNB. (Coax shielding is only around 60dB-80dB). If the pre-amp is before the coax, wanted signal is still amplified by the amp, but interfering noise picked up by the coax not.

Not one that’s easy to read since it was more than 48hrs ago, unless there’s a way to generate a 8h or 24h graph1090 report for a specific time that I’m unaware of. I’ve marked the small section of the 7d graph that’s for this test.

image

The better performing mounting location is directly against (painted) metal, where the other location was against PVC. The ADSBx antenna is a CoCo (as far as I know); is there any possible interaction with the bracket against the metal that’s acting as a ground plane or something to “help” the reception?

I’m considering doing this same experiment but with a metal mast, if I can locate some EMT conduit or something at a reasonable price locally.

I didn’t touch anything below the roof level for these tests, so the coax/connectors at the LNA and SDR shouldn’t have been disturbed during the test, especially since it seemed to return to normal reception when the antenna was moved back down. Next time I’m at that building I’ll check it though.

I’m aware, and that’s the goal, it just isn’t practical at the moment. I need to get a water-tight mounting box, and some more coax/adapters to make that happen. I’m also debating having either 2 separate antennas (978 & 1090) or a splitter and 2 separate LNAs up there w/ 2 coax feeds coming down. I may just end up purchasing all of that over time and experimenting which works best, but that’ll be a long term project.

RFI from the motor is an interesting thought, and will probably guide my decision to move the LNA up to the roof. However, this mounting swap was done on a weekend when the building’s occupancy mode was set to off, so I’m certain that this compressor wasn’t running at all on Saturday, so it didnt influence what I was seeing at the time.

You could just put some aluminium tape on the PVC?
I don’t know if there is any difference … i wouldn’t have expected it but can’t say for certain.

The signal doesn’t look worse.
You didn’t show the range graph?
I don’t think the mounting location was necessarily worse … air traffic fluctuates quite a bit and you’d need to leave it at least a couple of days to properly compare.

Also the added height is not gonna make much of a horizon change considering the trees are quite a distance away.
So i wouldn’t expect a drastic difference anyway.

1 Like

Great idea to try out, since I already have a few rolls somewhere :slight_smile:

The range was pretty much unchanged, hard to pick out easily on the graph where I made the swap. It was mainly the messages / # of aircraft that changed.

image

There are a few other air handlers on the roof that are taller than the antenna, so I was just trying to clear those, but you’re right. I just wasn’t expecting things to decrease, lol.

I’ll see if Al tape changes anything and re-try for a longer period.

Thanks!

I guess at this stage you best option is to run a spectrum scan with the LNA bypassed and see what it looks like. First just a broadband one and look for signals that can overload the LNA. If you are sure the LNA is not being overloaded, hook the LNA back in (you should see the noise floor rising by a few dB) and take a higher-res one, with and without the LNA until you can see the actual noise floor. These things can be tricky. You can do the scans on the SDR itself if you dont have a spectrum analyzer available, but only a decent spectrum analyzer will let you see what is going on.

I’d disagree unless you really have a spectrum analyzer.
The “spectrum scans” often described here on the forum usually just autoscale the colors so they aren’t really all that useful.

With the signal graph showing the range being normal and the dynamic range being as good as you can expect from an rtl-sdr … no indication of LNA overload.

If you’re gonna do one of those scans as with rtl_power … i’d do it with LNA and check in the filtered LNA passband how the 1090 signal is vs the surrounding frequencies (which are hopefully pretty quiet).
If those other frequencies are quiet compared to 1090, the LNA is clearly not overloading as that would produce broadband noise.

2 Likes

agree with you. I wish there were some good spectrum analyzer software available for RTL-SDR, like the spectrum Analyzer for the SDR-play. but yeah, I normally hook a splitter between my SDR and (Siglent) spectrum analyzer, and have a look. Basically grew up in a RF-lab, so for me a Spectrum Analyzer is just as necessary as a soldering iron, multimeter and an oscilloscope (and of course the knowledge how to see the stuff that you want to see).

This one gives a pretty realistic, wide-band scan covering the entire bandwidth of RTL based dongles.

Spektrum - How-to Speedily Scan RF Noise in band 24MHz ~ 1800MHz

 

The results of scans are below:

The scan 1 shows RF noise picked by antenna and processed without any filter , The scans 2 & 3 were done with filters, and show how filters remove this noise.

Scan 1 of 3 - FA Antenna + Generic DVB-T (no internal or external filter)

Thumb-Generic DVB-T

CLICK ON IMAGE TO SEE LARGER SIZE

Scan 2 of 3 - FA Antenna + ProStick Plus (Only Internal filter of ProStick Plus. No External filter)

image
CLICK ON IMAGE TO SEE LARGER SIZE

Scan 3 of 3 - FA Antenna + ProStick Plus (with internal filter) + External Filter (FA Light Blue)

imageimage
CLICK ON IMAGE TO SEE LARGER SIZE

2 Likes

true, and that is the best on can do with the spectrum analyzer software for the RTL-SDR, but a true spectrum analyzer would let you zoom in and look at the noise floor of around -170Bm in the ADSB-band. (And now the youngsters that have access to very expensive Real-Time Spectrum analyzer will jump in and say they can do better). Bottom line is there is a reason why RF is still considered Black Magic by many. Some basic rules apply, and well, if you need to look deeper, it turns to either try this or that, or having very expensive test equipment AND a deep pool of knowledge. (Will see if my brain works tomorrow, and then I can post comparisons between the RTL-SDR spectrum analyzer software, and measurements taken with a real spectrum analyzer)

3 Likes

What kind of coax do you have, and how long is it? Also, are you using dump1090-fa 6 and if so is adaptive range enabled? I’m wondering if the antenna’s actually performing better on the mast such that your dongle’s tuner is being overloaded.

That was my thought too :thinking:

10’ of S-LMR240, pre-terminated. Not exactly Times Microwave coax, but didn’t expect it to be a big issue. If I get the time to un-pull it through the roof & put it back, I’ll measure the loss through it as best as a NanoVNA can manage. I did this prior to installing it and recall it not being too bad, but I never wrote it down.

No, I’m using readsb, so no adaptive gain, though since my “messages > -3dBFS” number is consistently between 1-3% I don’t think gain or overloading is an issue.

I had that thought as well, but the graphs for “messages > -3dBFS” did not show an appreciable increase. If anything it decreased slightly.
I’ll be experimenting more with this set up later in the week or weekend & keep a close eye on the SNR as well. When I increase the gain manually, I can see the too-strong signal % go way up, the peak signals hit close to 0, and the noise floor (as reported by graph1090) increases dramatically, but didn’t see any of that, making me think it wasn’t just a signal overload issue.

I plan on trying the RF Noise scan from adcd567 with and without the antenna on the mast, and with and without the filtered LNA in-line. Also plan to try the aluminum tape to see what (if anything) changes.

Lacking a spectrum analyzer you could connect the dongle to a PC or laptop with an SDR program and the rtl-sdr driver installed, tune it to 1090MHz and look at the FFT and waterfall displays. Those may give you some clues. Eyes and ears are very sensitive test instruments.

I have already noticed that the vertical radiation pattern of pcb antennas is not 100% circular, any idea if the antenna could have been rotated a bit when fixed to the pvc pole?

This has been discussed here:

See wiedehopf’s comment re: radiation patterns and field strength. It may be relevant to what you were thinking. :slight_smile:

The antenna I have is a CoCo ( ADSBexchange 5.5dBi 1090/978 26"), not PCB strip antenna, so I don’t think it applies. Though when testing this weekend I’ll verify if rotation has any effect.

In the first photo your coax seems to have some slack and gentle curves. I can’t tell for sure but in the second photo with the antenna on the mast it seems like it may be sharply bent around the corner of the aircon unit. I wonder if that’s the issue? If you have a jumper coax that’s a foot or two long maybe try adding it between the LMR240 and the antenna when the latter’s on the mast, or using a slightly shorter mast to leave more slack in the coax. LMR coax doesn’t like sharp bends. They may actually damage the coax if you bend it too sharply.

If that isn’t the problem, all I can suggest is that sometimes antennas and radio waves sometimes behave unexpectedly in real-world installations. Your antenna may have been in its sweet spot for that location and raising it a foot or so higher is what made it perform worse. I remember when I was young our TV antenna fell off the roof and landed on the ground next to the house, and we were getting better reception. My mother called a TV repairman (remember those?) to put it back where it was as she didn’t want it laying in the yard, and he suggested driving a short mast into the ground to attach the antenna to and leaving it there. He explained why it was working better down there but I don’t remember what he said as it was over 50 years ago.

1 Like