We operate a LoRaWAN network in the Chilean Atacama desert (ALMA telescopes), with currently two Kerlink iStation GWs and two Abeeway Compact GPS trackers utilizing the Australian sub-band 2 for testing. The Chilean authorities want us to limit the output power to 13 dbm (20 mW). We have a MQTT paho client connected to the TTN server host. I am wondering what the downlink JSON should look like to consider also the TX_POWER that comes with each downlink message. Currently I got something like this: {‘downlinks’: [{‘f_port’: 2, ‘frm_payload’: ‘AwAA’, ‘confirmed’: False, ‘priority’: ‘NORMAL’}], ‘power’: 13}. Suggestions welcome. Cheerio, Ralf A.
If the Chilean authorities limit the output power of the transmitter to 13dBm the gain of the antenna doesn’t matter.
But if they mean 13dBm e.r.p. you have to consider the 6dBi (4dBd) of the antenna. In this case you have to reduce the output power of the TX to 9dBm plus loss of cable.
Any idea what the JSON message of the mosquitto_pub -m ‘{xxx}’ should look like?
Maybe this page can help you: Regional Parameters | The Things Network
There you find a table which gives you the relation between TX Power (0-7) and output power of the TX. First try to understand the difference between e.r.p., e.i.r.p. and TX output power.
Than ask your authority what kind of power they mean.
But I have no idea what your JSON message should look like.
Again, the exact value of the GW power is of secondary prio. My issue is how we can change TX_POWER with each downlink message. BTW, this was even a suggestion by Kerlink’s helpdesk
I’m not aware of any API that allows you to change the settings of a gateway or anything that allows you to set the radio parameters of a downlink for an end device.
From the many discussions regarding radiated power, antennas etc, I only see references to using the gateway console / web interface / remote to adjust its conf file - global_conf.json or local or testing or which ever - or use its entry interface.
Thanks Nick, this would confirm my gut feeling. Consecutive question is the plausibility of the advice of Kerlink’s helpdesk.
Most likely some advanced guessing from well meaning wanting-to-be-helpful first line youngster who has not yet got two gateways and half a dozen devices deployed so hasn’t experienced real-life.
The received wisdom is that if you can setup remote access to a gateway for changes & updates, that’s always a good thing. So if you don’t have that now, when you go to the gateway to make the changes, perhaps consider putting in something that can give you remote access.
You can’t. The one and only valid option is to specify the antenna gain in the packet forwarder configuration. TTN back-end specifies the transmission power for a packet which it calculates based on regional parameters and the antenna gain at packet forwarder level is used to correct it based on the local setup. (Antenna gain, cable loss etc)
Perhaps the Kerlink support know of a setting in the Kerlink system that allows no only to schedule the DL but also set a per message Tx pwr? Problem in a mixed network is how would it work and also how would NS know to only ever send via (that) Kerlink GW if it does happen to support, and what if GW airtime exhausted - drop message or route through another. Drives towards a homogenious network build out - only Kerlink GW’s and only ones that might have the feature? Way out in Atacama Desert maybe not a problem 
I dont know, just speculation…Anyhow why worry about reducing GW Tx power just for one specific message? Doubt means much in overall scheme of things.
@ccatp If it were me I would simply look at either using ineffective Antennas to bring ERP/EIRP down or as most GW’s allow a setting for Ant gain - just populate with exagerated gain figure say 8dbi vs nominal 2.1 / 3 dbi and system should then dial back TX power to limit - so a nominal say 17dbm o/p to allow for 14dbm TX after cable and connector losses etc (to use EU limit as example) would now only be 12 dbm say real world keeping under the limit…but question earlier still holds - is this conducted radio output power (as say spec’d/method used in N America) or radiated - say measured at 1M - as typical in other juridictions. Wrt GPS trackers my experience with most (havent used/seen the one in OP) is that unless using external 2.1/3dbi or better ant on SMA they usually have poor PCB antennas or chip antennas on board or poor pcb antennas on ufl pigtails - getting -5dbi to 0dbi at best so not likely not a problem. The one exception I have is the original SMTC/iMST LoRaMOTE with classic Folded F instance on PCB which IIRC is 3dbi.
Well said. We will have remote access to the GWs within the next weeks. I was merely exploring some alternative solution, though. As a matter of fact one of the gateways will be at a rather remote location, i.e. the Cerro Chajnator in the Chilean Atacama at an altitude of 5600m (CCAT) and is, most likely, the uppermost LoRa GW. We get signals from Abeeways trackers 45 km apart at S/F=7.
They want it for everything transmitted from this gateway - because where this gateway is rather high up and will be heard for a very very long way - like most of the middle third of South America. If they felt so inclined, I’m sure they could take the LoRaWAN distance record without even trying.
Yep, that is understood from OP. I guess 1dbm extra c/w EU limits not much but dropping from 21-27dbm down to 13dbm in context of US/Americas etc where higher power allowed more understandable 
Still think best option is adapt Ant/cabling infrastructure…in line attenuators or reduced efficacy of selected ants (dont strap on a 8+dbi ant that will help drive that horizon reach and S.Amecican coverage, rather use a 0dbi or even -3dbi ant?) and or software adaptation based on fooling a standard option on most GW’s wrt ant gain selection such that it self limits Tx power seems easiest option and path of least resistance…latter can be assymetric so should not impact reception sensitivity for hearing those remote GPS nodes
There are two main reasons to limit the LoRa signal strengths:
Tipo Norma :Resolución 755 EXENTA
Publicación :25-07-2005
Promulgación :15-07-2005
Organismo :MINISTERIO DE TRANSPORTES Y TELECOMUNICACIONES;
SUBSECRETARIA DE TELECOMUNICACIONES
Título :FIJA NORMA TECNICA DE EQUIPOS DE ALCANCE REDUCIDO
Versión :Última Versión De : 17-10-2017
Inicio Vigencia :17-10-2017
Derogación :17-10-2017
Id Norma :240404
Texto derogado :17-OCT-2017;Resolución-1985 EXENTA
Ultima Modificación :17-OCT-2017 Resolución 1985 EXENTA
URL :Biblioteca del Congreso Nacional | Ley Chile
…
j.3) Operen, al interior de inmuebles, en la banda de
915 a 928 MHz con una potencia máxima radiada de 100 mW y,
al exterior de inmuebles con 20 mW, en ambos casos técnicas
como: espectro ensanchado con secuencia directa o con saltos
de frecuencia, monitoreo previo, selección dinámica de
canales u otras técnicas de modulación digital que
funcionen con un ancho de banda mínimo de 5 MHz y permitan
compartir frecuencias.
…
furthermore, we operate LoRaWAN very closely to astronomical antennas (115 GHz < f < 1 THz) designed for the detection of very, very faint signals, such as the microwave background of big bang …
Total hearsay, red shift is due to too many cocktails at high altitude. Steady state for the win and if you look out the window you’ll see the earth is flat …
-
I guess Google translate may help - Forum, by consensus, is in English
-
Totally get that - grew up in the shadow and with the reputation of Jodrell Bank and did early work on inter array comms (f.o rather than rf) for SKA project definition & feasibility work…and have GW’s & nodes within spitting distance of JB and associated AA’s creating a larger virtual receiver around UK
I just turn on old Analogue TV receiver and look at the snowy image and use my imagination! 
I thought that was Poltergeist (the movie) - “They’re here” …
Here a reply from Kerlink’s helpdesk which I forward to the experts amongst you.
+++
We received the feed back from our radio expert.
Tx Max Power has to be managed through the LNS. It’s not recommended to change the antenna gain into json file. In case of modification of antenna gain, Rx will be impacted too.
+++
Changing the power when an application level downlink is specifically requested is unsupported, but wouldn’t really be a solution anyway since it would do nothing when a network management downlink is automatically generated, or when some other user’s node gets a downlink through your gateway.
Kerlink is right; in the design of LoRaWAN it’s up to the network server to make only transmit requests which comply with applicable RF regulations.
If as you’re alleging TTN is not complying with the general RF regulations of Chile (likely as a result of attempting to do so only indirectly through the regional parameters of a different country) then that needs to be carefully documented and filed as a bug report.
Possible temporary workarounds for general non-compliance, plus any location-specific extra regulations could be any of a custom server install distinct from TTN, modifying the packet forwarder to add a transmit power hard limit, or even a man-in-the-middle server that would re-write the transmit commands to a legal power. However, if some gateways have lower permissible transmit power than others, that’s really something the network server needs to know about so that it can take that into account when chosing the best gateway for a transmission.
Thanks, well said, that’s most likely it.