PoE (Cat5e) Protection and EMI mitigation

SatNOGS Hardware
1

Hello all,

I’m looking into setting up a PoE-powered, no-rotator, ground station about 50ft from my residence. Current thinking is to keep short runs of coax between Pi/RTL-SDR and the Antenna ({something} fairly omni, but circular-polarized), with a bias-T-powered LNA figuring in somewhere on one end of the coax or another.

I’m still working through ideas on protecting the residential, plenum-rated Cat5e cable from the elements (UV, moisture, etc…) and may eventually go with a metal or PVC conduit burial solution, but for right now I may simply use one of those 50’ flexible (woven sheath w/ rubber core) garden hoses to keep things protected, and temporary (as I sort things out). The weight of the protection will keep the Cat5 from coiling up into a tripping hazard, and should protect it a bit more from careless footwork.

I want to keep the design somewhat flexible and portable because I plan on having a “mobile-demo” version to take to high-schools and other local sci/tech/geek/STEM gatherings to promote the learning opportunities for building your own system. So I like the idea of a semi-protected Cat5e cable that I can string out on a parking lot, or school campus that’ll weather the wear & tear, but also illustrate that cheap solutions (garden hose) work just fine if you think it through.

The PoE injector will stay sheltered, and 50’ of Cat5e is “nothing” when it comes to delivering enough juice to power a Pi + LNA, etc… so I’m less worried about power delivery than protecting the items at either end. While I can more easily write off the loss of a Pi, loosing the PoE network switch would hurt. :wink:

50’ of {anything} will also act like a great big antenna though, so I’m imagining that some ferrite bead protection, or toroid choke(s) are going to be necessary.

Also, having done some cursory research on the interwebs, with folks running Ethernet out to she-sheds, and other IoT devices, lots of folks speak of having had their network gear blown out by the EMI surge from nearby lightning strikes. I’m not even talking about direct hits… just induced EMI coupled to the long run of Cat5, producing a surge that runs out to either end of the cable.

Amazon (and other sources) list quite a few different brands and solutions, but I’m curious what others have done to protect their PoE devices when they run the Cat5e outside.

(obviously depending on where you live, brands and availability will vary, but just looking to get a “feel” for how in-depth others have thought this through.)

I welcome any suggestions or recommendations.

Cheers,

-Lucky

2

Unless you have no other alternative, I would strongly recommend not putting the RPi near your antenna. Keep your pre-amp located near the antenna, but run coax instead of ethernet.

50ft of something like RG213 is only 2.5 dB loss in the 70cm band (and of course less at 2m), and with a preamp at the antenna feed-point you won’t notice the loss.

There are many stations in the SatNOGS network that I suspect are limited by their own noise - noise from switch-mode PoE injectors, noise from RPis mounted too close to antennas, and noise from USB. With all the equipment mounted near the antenna and relying on PoE for power, it’s very difficult for the station owner to test to see what is causing the noise.

3 Likes
3

I’ll defer to @vk5qi about interference. But I’d like to point out that there is CAT 5E you can get that is “direct burial” so you don’t need a garden hose or other conduit.

4

The Pi 4 is definitely a huge cause of interference in my system. (And maybe my LNA power supply or the Pi’s power supply.)

1 Like
5

So, @vk5qi you bring up some excellent points there, and I was afraid that the whole plan would have to be re-examined… Not surprised, mind you… just wasn’t looking forward to the likely possibility. :unamused:
Still, I appreciate the advice.

I was operating on the idea that short runs of coax would limit my loss, but having to deal with a raised noise floor from the “other stuff” is maybe just as bad. I could try to work on minimizing that with shielded boxes, ferrite cores, proper ground bonding, etc… but I could quickly end up at a point of diminishing returns.

I guess I need to expand, where I can, some details in my link budget, which BTW can get as awesomely complex as the one found here (caution: .XLS link) on the amsatuk site, but I’d like to avoid getting that deep into it. It’s kind of why I wanted to poll the crowd for their own experiences, and sorrows, before incurring my own.

I’d given some thought to longer runs of coax for the RF segment of things, then the obligatory SMA-to-N adapters throughout (with their own loss) as all the LNA, RTL, etc… gear is of course, SMA, and the antenna and coax would undoubtedly be N (a much better outdoor-rated connector), or the classic PL259/SO239.

So, on the topic of self-assessing the noise floor, etc… how have folks here typically worked that out ?

Running {rtl_power} surveys & heatmaps for certain bands, over long time intervals & different conditions ?

I wonder if all those waterfalls, from all those observations could be mined for assessing averages for each station’s operating “noise floor”, which could then be compared (if the station shared more details about their setup) to see what seems to be working better than others.

The one thing I have going for me is time. I’m not rushed to get online, so I can spend some time testing a few options (within reason).

What I would prefer NOT to do is “buy my way” down one path, only to have to back-track.

Eventually I’d find a good, cost-effective solution that suits my requirements (both mobile-demo & fixed station), but I’d really like to avoid staring at un-used coils of either Cat5e or expensive coax when I reach the end… :flushed:

Lastly, I’m also leery of starting what could be the RF equivalent of the classic (vi vs. emacs) “holy war” here. :wink: I don’t think that’s likely, but I really don’t want to get caught in the middle, just by asking the trigger question…

Cheers,

-Lucky

6

Assessing the noise floor can be difficult, in particular trying to get any form of ‘absolute’ measurements, as this requires knowledge of antenna pattern/gain, preamp gain, cable loss, and a calibrated receiver system. It’s certainly not impossible to do, but it’s not simple.

However, assessing if anything under your control is affecting your system isn’t that difficult. You need to be able to receive signals from the antenna/system under test, and you need to be able to turn things on and off.

My tests at home generally involve receiving using a SDR, from a battery powered laptop (my Mac Pro is very quiet on 2m / 70cm), turning essentially everything in my house off, and turnings circuits/devices back on one at a time. I keep a watch out across the whole band (or as much as my SDR will let me) for spurious signals appearing, and for general rises in the noise floor. I generally just use something like GQRX to watch the spectrum live. I’ve used this technique to identify a set of very noisy LED lights around my house, which I replaced with some better models.

As you can see, doing this in a situation where the receiver is halfway up a pole can be difficult! Being able to do it at the end of a run of coax is a much easier.

2 Likes