I'm in the process of building a LoRa end device, and am also designing protocol. I plan on using the channels specified in the AUS915 ISM Band as outlined in the LoRaWAN regional parameters (though i'm not using LoRaWAN):

  • Upstream Channels 0 to 63,125KHz, DR0 to DR5, 915.2MHz-927.8MHz, 200kHz inc.
  • Upstream Channels 64 to 71,500KHz, DR6, 915.9MHz-927.1MHz, 1.6MHz inc.
  • Downstream Channels 0 to 7,500KHz, DR8 to DR13, 923.3MHz-927.5MHz, 600KHz inc.

I'm confused about the typical operation of an end device. Does this channel plan suggest that to use an end device effectively and this channel plan, I will need to send data from it on any of the 72 upstream channels, but have to switch configuration entirely to then recieve packets in one of the 8 downstream frequencies?


From the Radiocommunications (Low Interference Potential Devices) Class Licence 2015:

  • All transmitters, 915–928MHz, 3 mW max EIRP
  • Frequency hopping transmitters, 915–928MHz, 1 W max EIRP, A minimum of 20 hopping frequencies must be used
  • Digital modulation transmitters, 915–928MHz, 1 W max EIRP, The radiated peak power spectral density in any 3 kHz must not exceed 25 mW per 3 kHz
  • \$\begingroup\$ This is too broad to fit here. Generally you have a regulatory requirement to frequency hop, but how you organize that is more a matter of design. Probably you want to look at some specific existing examples. \$\endgroup\$ Aug 14, 2018 at 14:34
  • \$\begingroup\$ I have updated the post to reflect the regulations with regard to frequency hopping. I hope this makes the question more manageable. \$\endgroup\$
    – Joe
    Aug 14, 2018 at 22:22

1 Answer 1


I'm not sure about AU, but in the US, typically the gateway will listen on a set of 8 consecutive upstream channels called a sub-band. The end device will try transmitting join requests on random channels across all sub-bands until it gets a response. The device will then frequency-hop within that channel's sub-band, changing frequency after every packet.

Join responses, like all downstream packets, are sent on one of the downstream channels. Usually it's channel associated with the upstream sub-band, i.e. downstream channel 0 goes with the first 8 upstream channels, downstream channel 1 goes with the next 8 upstream channels, and so on.

Your end device's transceiver will need to transmit on the upstream channels and receive on the downstream channels. LoRaWAN inverts the signal polarity for downstream channels, so the end devices normally don't receive each other's packets.

  • \$\begingroup\$ The AU and US regulations are actually very similar. I believe the AU regulations are based on the US. Thanks for the info. In a situation where only 8 devices and a single gateway are implemented, is there any need for frequency hopping? In this situation, each end device could use its own downstream and upstream channel. I assume this would also cut down on cost since the gateway would only need to listen to 8 upstream channels. \$\endgroup\$
    – Joe
    Aug 14, 2018 at 23:48
  • \$\begingroup\$ The frequency hopping is to comply with regulatory limits on the time on air. My vague understanding is that it's more important in the EU where they have a 1% duty cycle limit, but I am far from an expert on RF regulations. The gateways I'm familiar with (MultiTech Conduits based on the Semtech SX1301 IC) only listen on 8 channels anyway. :-) \$\endgroup\$
    – Adam Haun
    Aug 15, 2018 at 0:12
  • \$\begingroup\$ As far as I know the max time-on-air for AU is 400ms, and my transmission are likely to come in well under that for most data rates. Additionally, there'll only be one transmission every 2,3,4 hours or so. I'm guessing this then means that frequency hopping is not needed for my application. \$\endgroup\$
    – Joe
    Aug 15, 2018 at 1:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.