New free eBook to get started with LoRaWAN

We understand that it’s not so easy to get started with LoRaWAN. You can find here a complete eBook on LoRaWAN (from the Device to the IoT platform). It’s a university work and it’s free, available in English and French.
You can download it here.
Hope this document will be helpful. Let us know if you think of any improvement.


Hi Sylvain, this is great!

It looks like a good introduction and overview for many things Radio based IoT, LoRa & LoRaWAN and downstream and to me had instant creds when I saw the reviewers, some of whom I know have been around this stuff for a long time! :slight_smile: (assumming they have indeed read and reviewed all). I look forward to running through it rather than a 5 min scroll through tonight - just as soon as I have some time!

From that scroll through clearly good for a quick dip in to find out some of the LoRa/LoRaWAN fundamentals or to look at a specific topic - possibly as a compliment to TTI’s Johan’s LoRaWAN overview on YouTube :slight_smile: With sufficient 2nd level and even some 3rd level content for the curious or those who need to know or want to know more than just LoRaWAN 101.

You ask for feedback - I have just one immediate concern and potential criticism.

For Gateways, as you say in 5.2.1,

There are many brands of LoRaWAN Gateway. Each model is intended for a particular use (Indoor,
Outdoor, prototyping,…).

It is therefore a pity that from the outset - if I recognise it from the picture - page 4 for the hands on - online training you reference a learning kit…and do I see the STM Nucleo LRWAN2 gateway board?

If so with respect I would suggest a change to one of the many others available as IIRC this is a version that doesnt fully follow the Semtech reference design and as a consequence it has a crippled power output (I need to check the specs but again IIRC it is limited to some 6-8dbm below the nominal 14dbm of a ‘standard issue’ EU868 Gateway - and falls even further behind those targeting other parts of the world where even higher output powers are permitted).

I believe this represents a shortfall on GW range of some half to one quarter of the free space LOS range of a full Gateway. Now whilst this is fine for close in work and for a private network its not good for a general deployment. Potentially I would not recommend it as a learning aid.

A lot of people miss this spec bullet but it does catch people out and can cause problems, even though it has occasionally been called out on this very Forum :wink:

Please tell me I’m wrong and that is not the kit used or refered and I will happily go away… :blush:

If it is the LRWAN2 board the issue for any one reading is that whilst receive sensitivity is ok - and it can therefore hear nodes at relative distance for any given SF, - the problem is that the GW may not be able to respond to the node - e.g. send a join accept and so may cause problems. If deployed into a ‘real’ network such as TTN then the LNS may well choose the GW to send a downlink to the node not knowing that the GW is actually incapable of reaching it…it may keep trying (perhaps in response to repeated join requests…and effectively becomes the vehicle for a mini version of a self inflicted denial of service attack! :scream: Students using the subsystem and not seeing this may struggle to debug field deployments and may not be aware of the range risks

…like I say please tell me I’m wrong… :+1: or perhaps there is something in the text that I have not yet read that explains or caveats the use…

will plow through and will come back with any other comments and hopefully more positive feedback. Like I say great and a valuable contribution to LoRaLEARNING™! :slight_smile:

Maybe the Mods can add this to the list of usual documentation we point users/newbies and questioners to?! :wink: Paging @bluejedi @descartes @kersing

1 Like

I shall divert time tomorrow to review in detail. Apart from the concern so diplomatically put above about the choice of hardware, which I’d extend to the use of the MKR WAN 1310 which is largely abandonware, the first few pages appear to be a thinly veiled advertisement to move from a free book to a a Udemy course or a more formal course (using the borked STM ‘gateway’ or DDOS unit as we know it here).

I am never comfortable when central authority funded organisations tout for business - giving them an huge saving on recouping costs of development as it was funded by students paying for their tuition and with government supplements or direct funding.

It’s “nice” that Semtech luminaries and the CEO of the LoRa Alliance, who definitely needs putting in a room with a B-L072Z to see how she gets on, have endorsed this document, but the main activity that people implementing LoRaWAN get stuck on is the creation of a device, of which there are 8 pages, none of which give any direction on actual real life implementation. The RAK & the Seeed modules only implement the high power output rather than be suitable for EU implementation for more pragmatic low power use.

So at first scan (speed reading course aged 17), a good resource but with some hype and questionable accreditation & progression that breaches competition ethics.

Easily rectified if the university so chooses, in the meanwhile, if you look up at the south face of Mont Blanc, I’m the one in the LS7 virtually ridge soaring all the way to the top in a sailplane.

Everyone is welcome to give his feedback, whatever the criticism. A new version will be released in one month with dozens of english language correction and minor technical improvements.

Yes, that is the P-Nucleo-LRWAN 2, but we never advice it anywhere in the document, this is just the one we use during the training session. It fits our requirements:

  • Cheap (GW+Dev <100€)
  • Ethernet
  • Easy logs of what is going on the GW without external tool
  • Not single channel GW
  • With the possibility to change the targeted NS easily
  • Easy set-up
  • Device with Temp, humidity, accelerometers (and more !) sensors.
  • USB powered

If anyone has other choice, we will be glad to hear. Concerning the RF performance, we really don’t mind as we only use it to learn the basic LoRaWAN specs: Class A, Spreading Factor, Bandwidth, uplink, downlink, ADR, (un)confirmed… and the end-device stands few meters away from the GW.

By the way, it’s clearly written on the GW : “For evaluation purpose only”. But I agree, it depends what you want to evaluate :wink: :slightly_smiling_face:

You’re free to use it as you wish.

Hi, you may not mind but am sure the network operators will as use on open networks such as TTN or LoRiot as you call out would potentially represent a DOS attack on their current or future users! By all means use with a Chirpstack or private TTS instance, just please dont encourage deployment on live networks.

Yep, but a case of you get what you pay for…see above :slight_smile:

Agree - that would not be useful for a LoRaWAN intro or course as well…that isnt LoRaWAN! :slight_smile: OK if you want to use privately or for point to point but NEVER on a live LoRaWAN network, please :+1:

True a useful bundle but would respectfully suggest seperate the GW choice and Device choice and pick each on its own merits…atleast device better that the MKR1310 as Nick called out - sadly the earlier 1300 was chosen for last summers TTSA and several people took the 1310 as alternate - neither a good choice as either a learning platform or for future use - as called out

Finally the old adage

…and you buy twice! :wink: A potential false economy IMHO

Still I wish you luck with this and look forward to the revisions :+1:

I really didn’t want to speak about that but here it is. We haven’t been able to explain everything in a book and we didn’t want the document to be written in a tutorial style. These videos are free for educational use and we encourage any lecturer / professor or anyone engage in a teaching task to ask for a free access in our website. They are not free for companies. That helps to finance the LPWAN project (the book is part of it). Trust me, no one is making a leaving on that.

One year ago we contacted TTN about the video project to see how we could work together. We had a warm reply but were very surprise to see that TTN publishes its own Udemy course one week later. If you like TTN Udemy course (not free), good for you, I don’t. Otherwise we wouldn’t have spent weeks to provide new ones and plan to make them available in English soon. Again, free for educational use.

Please read the second page of the book. I don’t ask you to believe it, but it would be fair not to launch rumors about funds from any entity.

At least one thing we agree. But as a forum moderator you must know the huge lack of knowledge on LoRaWAN. How can people create a device if they can’t even make the available ones working?

In a near future, there will be other content provided by the university on multicast, fuota, Class B, Join-Server and roaming.

1 Like

A university using a device that that the practitioners say is inadequate bordering on being a denial of service attack on the network - if students buy one, they will continue to use it …

How many of you have deployed more than a handful of gateways and hundreds of devices? Is this learning resource aligned with the most important thing that industry needs, which is practical knowledge?

All universities are funded by central government in combination with student fees and I clearly said “central authority funded organisations”. I NEVER said anything about any bias due to being funded by other commercial parties, no rumours were being made about funds from a commercial entity, tread carefully with your comments lest they take on a legal dimension.

I supervise PhD students and have worked with universities since 1990 so I am very familiar with the landscape. My point is that you get funding that allows a department to work on projects. Rather than advancing the art or becoming the go-to university for those in electronic engineering who want to be beyond a beginner at IoT (god knows we get enough students asking for help with their assignments), you have created an offering in direct competition to commercial business at the base level. Perhaps someone in the EU could ask for a conflict of interest discussion? Feel free to own this situation, it would make all the difference.

Urban myths used as a marketing vehicle - again, if you wish to advance the state of the art, perhaps look at the infrastructure that’s required and the limitations of duty cycles to download ~100K to a device.

How can you assuming any bad faith here? What were TTI going to do, suddenly rush out a Udemy course to compete with yours. This is just wrong thinking. As I was providing the content on one of the five days of the Summer Academy as well as being a group mentor, I can assure you that creating a Udemy course was all part of the summer game plan.

Their are two TTI courses, one free and the summer academy an (inexplicable) £19.99. Plus the Learning website that has all you need to know to get the LoRaWAN fundamentals certification. Yours is £39.99, en français seulement, quand il a vente, je vais acheter!

It’s very simple. If you wanted to create a free resource to punt via the forum whilst making comment about how the moderators know stuff, one of whom created a packet forwarder, one has the canon framework of LMIC and one who knows gateways like the back of his hand, perhaps some prior enquiries before launch would have been appropriate.

Just say’n.

As part of a research project we were working with gr-lora project by rpp for demodulating and decoding the lora packets from sx1262 at SF 5.

But we have not been able to do it successfully. Due to following gaps in knowledge:-

  1. SF 5 has a different preamble Len than SF 7 to 12
  2. How do I know whether to use implicit or explicit header
  3. How to use gr-lora for SF5. Presently I have just forced the value 5 which gives some output which I assume is wrong

Your book has given good insight into the lora modulation process and helped alot in our project up till now. But getting stuck with this SF 5 issue as we do not know what all changes are there in SF 5 in the encoding process. Only change we have observed is number of upchirps before SFD being a non integer value. Please guide how to move forward.

I would refer you to fellow moderator Jac @kersing advice to your other post on the same topic.

Pleae do not double post as this splits contributors efforts.

You’ve already been told that this is off topic for this forum - so why (double) post again.