I’ll leave a fresh console open for a couple of hours to see if I can catch one or more downlinks. Will let you know if there’s anything apart from ADR stuff.
This device is only two floors and a couple meters separated from the antenna, meaning that it also immediately took a dive down to DR5.
There shouldn’t be any ADR stuff - if it’s not moved and given the close proximity of device & antenna and it’s been set, I’m hard pressed to understand why there should be more downlinks.
Perhaps @Jeff-UK or @kersing who are gateway masters could comment …
Gateways are not an active component in the ADR stuff. A node and the LNS are. Gateways are just dumb media convertors (to quote someone) 
YOU, Mr Jac Kersing, are THE Gateway Master - it was an honorific, not a reference to the gateway’s involvement, although just for clarity, why is it that LoRa doesn’t log in to the gateway - it works so well for WiFi.
What I was wondering was if you had any observations on why the LNS should repeatedly send Rx1 settings that may or may not have some ADR bits attached when the device is clearly on a good signal.
Although how I type this, it may be trying to get the power output down further and if the device can’t and/or doesn’t understand the power level, it’s not ack’ing the request and carrys on transmitting. Which would may hold true for the Speed LoRa-E5 as the cheap-skates only wired in the HP output.
Whilst we are talking about power, why can’t I get 10km in an urban environment?
Because it doesn’t work well for resource constrained devices (gateways in this case) if the are thousands of nodes in its vicinity. Because of this WiFi AP can only handle a small number of devices. A LoRaWAN gateway thousands.
That would be an excellent example of a reason. Decoding the downlinks could show what’s happening.
Because there are to many absorbers around. Including the kind walking around ingesting alcoholic beverages.
My console is running for seven hours now and I seem to find some interesting behaviour.
It looks like about every four hours (extrapolating based on total number of uplinks now), the LNS sends an ADR request to the AM103. The AM103 responds on the next uplink with an ADR accept. But then on the next uplink 10 mins later, the LNS once again decides to send an ADR request, to which the AM103 properly responds on the next uplink again with an ADR accept. After which the LNS shuts up and the process repeats about four hours later.
Console is still buffering so I can share whatever details you are looking for - I should also have the gateway logs from the ADR-duo that’s happening right now again, but it’s too crowded that the first one four hours ago has gotten pushed off.
As Tesco (the supermarket) says, every little helps!
Is it possible that LNS sends ADR-commands to the nodes to reduce their TX-power?
ADR initially works by stepping up or down the SF’s as needed…only once at SF7, if dialling down node activity, will it potentially command a Tx PWR reduction. As Nick noted sometimes that is moot or not practical to comply with depending on the build of the node (H/w, firmware?) but more importantly the LNS will not command that if signal levels are reasonable and SNR managable for the given SF. If it does request changes this will likely be cyclic in nature depending on recent stats and any changes in receive quality, the cycle-time may be determined by number of messages - hence update rate dependent, or triggered by any link-check/integrity check messages initiated over time (may be time dependent if time a ‘known’ qty or may be set as per x messages IIRC). I note that Steven’s node is txing with an entirely reasonable approx 80-90dbm RSSI with good SNR so at that level I would suspect (at least on TTN) the LNS is unlikely to be asking for further power reduction - as this is pretty much in the sweet spot of device operation (unlikely but not impossible, may depend on stats over a longer period than we have seen, IMO).
@descartes am definetly not a GW Master - perhaps a mere Padawan sitting at the feet of the true masters such as Jac! 
Thanks for the insight - Nick has access to two sets of downlinks & all related info from node & GW. I can send them to you as well.
Hopefully it’ll show why there’s still 2 downlinks about every four hours. (26 downlinks last 50 hours)
The device is hard mounted to the wall and GW is hard mounted to another wall in the same building… so signals should be very stable.
The country selection supports the search feature, for example, enter Neth then you can find Netherlands easily
8 months later … it’s Necro-Thread day!
Regardless of how obvious the search UI is, alphabetical order not only makes sense but also seems the norm for 99.9% of sites …
Just came across this post yesterday, we fully agree alphabetical order is standard and will provide the suggestion to Freshdesk as well. Thx for your guys’ feedback 
