Class of transmitter: Frequency hopping transmitters
Permitted operating frequency band (MHz) (lower limit exclusive, upper limit inclusive): 915–928
Maximum EIRP: 1 W
Limitations: A minimum of 20 hopping frequencies must be used.
I understand that there are 72 channels (frequencies), 64 uplink and 8 downlink, as defined by the LoRa Alliance Regional Parameters. The Things Network uses sub-band 2, consisting of 9 uplink (8 usable) and 9 downlink channels (8 for RX1 and 1 for RX2).
How does The Things Network comply with legislation if only 18 (17 usable) channels are used?
Secondly, I acknowledge single channel gateways are not compliant with The Things Network, but do these device comply with the legislation, and if so how?
Please note: In Australia there are no legislated duty cycles.
Definitely not a lawyer, don’t quote me on anything in court.
A guess would be that according to spec, devices must be capable of using all 72 channels (RP002-1.0.4 page 50, lines 951-958) - and should do so for the join procedure. So in that respect, they are compliant. That they only use 8 after activation is something I won’t comment on.
Hi Lachlan - you’ve been quite for a while - good to see you back on the Forum
For definitive view & guidance I would refer you to the LoRaAlliance, or perhaps more practically as he has run a commercial network in territory for many years (with both AU915 and AS92x band GW’s & devices installed) at Meshed AU and as a major historic contributor to TTN AU perhaps Andrew @Maj might be in a position to comment?
From my simple view AU regs categorise transmitters under the regulations under different classes and applications/use cases. I think the limit you call out is under (54) for Frequency hopping transmitters (in Frequency Hopping, WiFi and RLAN devices group?).
IANAL so not definitive!:
LoRa Modulation (and hence LoRaWAN) is not a classic frequency hopping mechanism but is a chirp based channelised communications system. Practically this means that where a FH system hops across frequencies (bands/channels) in the course of a message transmission conventional/classic LoRa signals do not - they stay on a channel for the duration of a message - hence reason why e.g. US915 messages cannot practically operate beyond SF10 - they would fall foul of Dwell time limits where the transmission stays on channel without hopping off to somewhere else after 400mS. By channelised I mean it uses a number of potential Tx/Rx channels for each message with a random/psuedo random selection for each sequential message - this is done to mitigate interferers than may exist on a given channel in a band (As you say FSB2 for typical AU TTN deployments), and/or to avoid hogging a channel and distributing it’s own traffic across the available (sub)band and not acting as its own blocker to other users. It ‘hops’ between channels between messages, if you will, but that is not the conventional technical view of frequency hopping in this context. Also, LoRa is a Digital Coding schema (modulation) overlaid on legacy radio capabilities (typically giving a 7-20db Digital Coding gain bringing the underlying legacy capabilities closer to the Shannon communications limit). Perhaps (58) classification (Digital Modulation?) under the legislation is a better consideration? And finally IMHO as a Chirp based symbol Tx you might consider that for any given RF symbol the radio is chirping (ramping up or down) across a wide range of precise frequencies within a given channel over the duration of the chirp - way more than ‘just’ 20 for a given symbol (Higher SF’s of course take longer and cover a greater range of subchannel frequency ‘steps’) - the chirp might be thought of as a fine grained sequential linear ‘hopping’ across various (narrow band) portions of that channel - conventional hopping typically thought of as some more complex algorythm based deterimination of which frq next hop, here it’s simple adjacent stepping until chirp completes.
Pehaps @johan might be in a position to comment further also…
Not a lawyer, somewhat of a pragmatist, and like you, a philosopher!
Consider, LW has been in use overtly in AU for a long time and there are some commercial companies operating whole networks. If it transpires that it has been used illegally, are you likely to deploy a network so large that the feds will go after you before anyone else?
Additionally, TTN doesn’t have to comply with legislation. The user does. And TTN is a not-for-profit based in the Netherlands so not so easy for AU lawyers to get to. And trustees, directors of TTI, have been around LW since the beginning. So I’d consider that they know what they are doing.
Most importantly, I think you are over-reading the law here. There is no definition of “frequency hopping” provided in the document, something that would tie up the legislative system for years, but by implication it references technologies that hop whilst transmitting in one burst. Whereas LW randomly picks the next channel to use and transmits all of its data in one go on that one channel.
By experience, having a user down the road from Canberra Deep Space Communications Complex with a few hundred devices, LW isn’t a bother and if they were going to hear anything that causes them issue, I’d go with them being well equipped to track them down.
If something is not in spec and not for use with a system, its compliance is a bit moot. I could build all sorts of kit breaching the law and glue a gateway on the side. Doesn’t make it anything particularly related to TTN.
There is, however, the totally official One Channel Hub - but I’d go with the “it’s not frequency hopping”.
Finally, most of us (possibly not me on a bad day) are professionals working to best endeavours - we aren’t trying to mess with someone else’s signals - we are using an accredited, internationally recognised specification that has been deployed in many many countries.
It sure has been some time. Since last being active in mid-2021 I’ve gone down a few different paths, exploring and try my hand at many different things. But I thought now is about time I get back into what got me into technology in the first place.
With a few years of experience under my belt and a fresh set of eyes, there’s a number of things I’m now stumbling across which I missed first time round, hence the recent forum posts.
That section of the legislation (54) is what I was referred to a few years ago and what LoRa sits under. I couldn’t tell you who suggested this to me, I just had it referenced in my notes from several years ago.
Ack. what you mean regarding chirping and mini-frequency jumps within a channel.
If we assume that LoRa falls under (58), how does it comply with the following limitation? It looks like a similar question was made in this forum post which coincidently you replied to as well. the radiated peak power spectral density in any 3 kHz must not exceed 25 mW per 3 kHz.
The reason why I said “The Things Network” as opposed to “LoRa” in the heading of my post stems from the wrong classification I had called out (“Frequency hopping transmitters” (54) versus “Digital modulation transmitters” (58)). Based on my earlier understanding, it was around the number of channels being used (72 defined, but only 18 being used for the sub-band by The Things Network).
If we now assume the legislation the LoRa falls under is under “Digital modulation transmitters” (58), the limitation is the radiated peak power spectral density in any 3 kHz must not exceed 25 mW per 3 kHz. It would be great to be able to understand how this limitation is abided by.
I acknowledge that my question around single channel gateways was not relevant to The Things Network, rather I find this forum useful to understand LoRa and LoRaWAN. If you think there are better forums for such discussions or questions, please do let me know.
What, specifically, are you worried about?
I want to understand how LoRa complies with the legislation in order to understand its limitations and capabilities.
My best guess - without further research, as about to head out of office - would be typical LoRa channel = 125kHz, = ~41.x x 3Khz, 41x 25mW >1000mw, ie. the 1W limit used for LoRaWAN in territory…
I think I speak for most of us when I say: I don’t really care, the LoRaAlliance will have this stuff figured out and I assume that RP002-1.0.4 is safe to use. Which is why @descartes asked what you are worried about.
I only very recently learned why EU868 (my ‘home’ region) uses 869.525MHz specifically for Rx2, that is after implementing LW 1.0.4, LW 1.1, RP1.0.4, RP1.1 and TS011.
However, a little digging on the phrasing “radiated peak power spectral density” leads e.g. to this link which, if I understand correctly, implies that if 3 kHz allows for 25 mW (= 14dBm), then 6 kHz allows for 17 dBm, then 96 kHz allows for 30 dBm, and 192 kHz for 33 dBm. Well, guess what? AU915 at 125 kHz bandwidth allows for a maximum of 30 dBm. So… checks out?
Thank you for your input @stevencellist! Your help is very much appreciated.
Does “radiated peak power spectral density” refer to the concentration of power? As in 25 mW being concentrated into across transmission of 3 kHz in bandwidth and 1 W being concentrated across 125 kHz?
Now following your levels for the rule:
radiated peak power spectral density in any 3 kHz must not exceed 25 mW per 3 kHz
Formula: “Power in dBm” = 10*log10(“Power in mW”)
Bandwidth
Absolute Power
Logarithmic
3 kHz
25 mW
Approx. 14 dBm
6 kHz
50 mW
Approx. 17 dBm
12 kHz
100 mW
20 dBm
24 kHz
200 mW
Approx. 23 dBm
48 kHz
400 mW
Approx. 26 dBm
96 kHz
800 mW
Approx. 29 dBm
125 kHz (125/96 = Approx. 1.3)
1000 mW (800 * 1.3 = Approx. 1040)
30 dBm
192 kHz
1600 mW
Approx. 32 dBm
The difference between 1000 mW and 1040 mW in dBm is 0.17 dBm which is negligible.
@Jeff-UK, you sure do put it simply by saying that:
125 kHz / 3 kHz = 41.6 recurring
41.6 * 25 mW = 1040 mW
And that’s close enough to 1000 mW anyway (it would be if I listened to my old physics teacher and cared about the number of significant figures (1)).
It’s clear there’s a few smart people on this forum, so I might as well ask if anyone knows how this legislation comes about? How do they make the call that 1 W transmitted across 125 kHz is acceptable rather than 250 kHz or 500 kHz or 0.5 W or 2 W? Is it based on existing legislation of other transmitter types?
Please stop specifically referencing TTN - which like all other legitimate providers, uses the LoRa Alliance specifications. Your questions are universal to LoRaWAN and anyone using it in Australia.
As a SCPF provides a denial of service to a network, we just don’t discuss it because we don’t want them. See Single Channel Packet Forwarders (SCPF) are obsolete and not supported. And because they are a DoS attack, we aren’t going to facilitate finding somewhere to discuss them either.
But more importantly, there is absolutely no need to discuss them. They were created at a time when gateways were impossibly expensive for Makers. Now gateways are relatively reasonable in cost AND there is now an official One Channel Hub specification with is basically a legitimate SCPF because the LNS knows how to manage the gateway so that it does not represent at DoS attack on an network in it’s coverage area.
Generally, I’m sure you aren’t, but when someone starts understanding the limitations, they are trying to warp time & space to achieve the impossible. LoRaWAN has a core use case, battery powered sensors sending in data at intervals measured in tens of minutes.
Assert location via GNSS is a thing. Tracking is marginal. Real time tracking can’t possibly work with legal limits and has gateway coverage to consider.
I’m not sure what you want to achieve by picking at the legislation. A pile of radio experts will be in the federal comms regulation dept who will have been asked about radio comms for the legislation. At the point of creation of statutes there is a wide consultation with the obvious names but generally anyone can comment. If you are taken to court then you can address a jury of your peers on why you think you abided by the law. I’m sure many of them will lose the will to live with the details, the physics of which are off in some very niche corner that intersects with black/white boards full of complicated maths.
If you want to worry about something, worry about some Sheila in Hollywood who invented spread spectrum technology in the 1940’s.
Or tell us precisely what you are worried about.
Or is this just a very deep rabbit hole you’ve fallen in to of rather academic application that you’ve pulled a few of us in to.
If you have an actual use case, do share and we can help out.
@ElectronicallyE you might have noticed most of us pragmatically accept the Lora association has done its homework while creating the standard. Trying to decipher the legal requirements is not easy and often requires expert radio signaling knowledge (and some legal knowledge as well).
Over the past 20 years I’ve tried to interpret the EU/Dutch regulations multiple times for various radio protocols and the only consistent result has been a headache.
These days my main worry is compliance with the ever increasing number of environmental, security and consumer regulations.
I appreciate my original question called out The Things Network who do not comply with the legislation directly, rather comply/follow the LoRa Alliance standards.
My question around LoRa’s compliance to the legislation was to understand where the restrictions/limitations originate from. A 1 W power limitation can appear arbitrary without context and/or justification (at least to me). I appreciate that people may have different opinions but personally believe understanding why the rules are the way they are is important. I don’t have any specific concerns and am confident in the judgement of the LoRa Alliance but wanted to know what the basis of the decision is.
It is my intention to follow and comply with these rules.
I appreciate the effort that everyone has put into their replies and sharing their knowledge.
