I am starting this topic to post the things i learned so for and people can add on this, when placing a gateway on the roof:
If you see a lighting rod rail, connect your antenna frame on to it with a thick cable
If you have lightning rods (bliksemafleiders) you need to be away at least 50 cm, however 80cm or more is better.
after placing check your board (e.g. rPi) temps on certain moments to make sure its not to hot. In most cases it would be smart to add some passive cooling to the Pi, i’ve seen mine hitting 83 C before passive cooling, so i added a heatsink. Also don’t make it ‘air tight’ but only water/rain proof. Add a small ventilation hole on the botton of your chassis.
Try to limit your cable to a single ethernet cable and inject power with PoE (48V !). I can recommend the TL-POE200 its a kit with injector and splitter for less then 20 euro and can up to 100m cable. It has active/gaurded PoE so we cutting the DC cable to add pins for the PI GPIO connection you need to test +/- with a tester and while testing look for a spike to + or - as the splitter will turn off directly because it thinks its shorting. Also dont forget to set the splitter to 5V
Make sure you can access your node remotely. In my case with a rPi3 i added a WiFi AP to it for management in case the ethernet line is broken. If your node is not easily accessable try adding something that will power reset your unit (e.g. add a arduino+wifi power regulator). Saves you going to the roof.
Also If your on the roof placing a gateway, take a moment to watch the view
Why would limiting the wires using PoE be either an advantage or disadvantage?
Besides that I’ve seen different ideas about lighting protection. One of those was that antenna’s that are grounded are more likely to be struck because the resistance is lower. And that even if a strike would be guided to the ground your gateway is likely to die anyway, so then it sounds better not to protect it in any way. I can’t judge the theoretical background though but the explanation itself seems legit.
@1 you can run 5V over longer then 5 meter (accourding to specs from USB) in additional thats on a very good cable, i’ve seen of lot of remarks not to go over 3 meters if you want a stable 5V 2-3A supply. Running 12V+ over a distance will require you to measuer and stabalize the voltage as you will loose over X meters. If have a small cable, its not that hard … heading onto high roofs its gonna be a challenge. Most cases running 220V into a home-brew box is prolly not accepted by e.g. your employer or the roofowner. Also please not that DC interferes less with radio then AC does! And just the simple fact you only need to lay down a single cable saves you time
@2 Prolly the difference in my case then, even if the gateway is directly hit (or the rod nearby) damage is expected, but i rather have a 250 euro GW burn out, then a roof. The data came from the company that did the lightning protection job on my work’s roof and i rather stick with their specifications. But your right that these specifications might be different from other deployed solutions. I can only say the solution we have is a certified one and i would stick to its documentation
Regarding the connect you’re antenna frame to a lightning protections system, I tried to find my books about this topic, but couldn’t find them (there must be somewhere in a box in the attic). But searching for info on the internet I found this document (checkout page 92) and this article (in dutch).
In summary : installing antenna’s on a roof with a lightning protection system is a specialist job
Yeah, specialists can prolly do a good job, sadly bit expensive. Also as a sysadmin, i don’t have a lot of radio experience so i ask around for the clue parts i need /miss to get it all working optimally
Meanwhile I did some testing. To clarify: we have a LPL1 class lightning rod system on the entire roof.
I put a node in my car across the street and took 15 minutes and averaged the RSSI/LSNR, then I did the same with no connection between the antenna frame and protection, finally I also insulated the antenne connection to the frame with a layer or elastic band + foam + duct tape. After each attempt again I waited 15 minutes and averaged the numbers.
My own conclusion is there is no difference at all (but please note that not every lightning rod system is the same). There is only a small difference between these numbers, but that’s so slim it can be accounted for the radio noise.
At least I was able to cross off the lightning protection being a part of my poor radio performance and starting to think either the P2P Wifi or the very big coolant system is generating too much noise, but I have no way to measure that.
@kruisdraad : For measuring (or better to get a indication of the) QRM (man-made interference) you could use a wideband receiver / scanner like the Yupiteru MVT-7100 or of course use professional RF test equipment. The first one I use to pinpoint the source of the QRM. And the S meter gives a good indication of level of QRM.
Regarding the connect you’re antenna frame to a lightning protections system, I tried to find my books about this topic, but couldn’t find them (there must be somewhere in a box in the attic). But searching for info on the internet I found this document (checkout page 92) and this article (in dutch).
In summary : installing antenna’s on a roof with a lightning protection system is a specialist job
It’s mostly a description of a single manufactureres devices, but it includes some information on the workings of different kinds of protection devices (page 10-15) and on how lightning works and how to best protect against it (page 124 onwards).
