Anyone on the forum had any experience with these, especially on 800-950MHz ISM bands? Does anyone know of Mega-compatible libraries to work with them?
Many thanks
Jim
awneil 
alexru 
Am I missing something? What is so good about that LoRa stuff?
Pretty low receive power, claimed "long range". I'd be interested in an Atmel chip with multiple-km range and less than a few mA receive current. It must work at 950MHz or less due to the potential environment (tropical rainforest with lots of obstacles and nasty weather). 430MHz would be better, propagation-wise but I don't know about international ISM licensing, yet.
Appears from the search, above, that the answer is basically NO.
Jim
alexru RF212B has reasonably low receive power. I'm not sure if ISM band transceivers can realistically achieve lower power consumption.
This stuff http://www.atmel.com/products/Wi... can probably run LoRa software, but I have not seen actual specification to tell for sure.
But if you are not fixed on LoRa, have a look at those ICs.
Thanks, Alex -
I'll take a look. I'm not fixed on any protocol at all, at this point. My big concerns are power consumption and range, which are usually mutually exclusive.l
Appreciate your suggestion,
Jim
Hi Jim,
I used to work with LoRa chip. It was WiMOD im880A chip and LoRa Semtech SX1272. Pretty small, about consumption I don't remember well but it was something like 10-120mA in transmission (i have to check on my report, it seems a lot)
Concerning range, we had 6-7 kilometers, but we had tested that with a Kerlink LoRa gateway.
I don't know what kind of modules you were thinking about ?
Hope it can helps you
Regards,
Gael
Thanks -
That sort of transmit power does not surprise me, as some of them are rated at something like 100mW transmit power. I have not thought about any particular module, yet.
Jim
stevech 
The LoRa chips from several vendors are spread spectrum with a low chipping rate and very low bit rates (as options). Once locked, they thusly have lots of post-detection correlation gain as any good spread spectrum long code system does. Vendors claim, at 100 bps or so, and longest codes, they can lock and get -120dBm or better receiver sensitivity (unsaid: for very short packets). They have options for heavy coding (forward error correction bits) - at 50% and more, which reduces the net throughput.
The market for these is rural and some suburban locales - for wireless SCADA and meter reading/control. Saves money with less infrastructure needed.
Several on the radio head forum are experimenting with these. Takes more acumen than most non-RF-geeks have, to choose the right config values for these.
On this page http://www.airspayce.com/mikem/a...
scroll down to
RH_RF95 Works with Semtech SX1276/77/78 and HopeRF RFM95/96/97/98
Anarduino sells an AVR board with these radio modules aboard. The AVR runs the RH datagram protocols - which have to be adjusted for long timeouts on ACKs, and so on.
Thanks for those details. I had not yet dug into the spec sheet in enough detail to figure that out.
My symbol rate can be pretty low and packets can be around 25 bytes plus overhead.
Appreciate the cross reference with the other chips. I'll take a look at the protocol, also. I am somewhat of a radiohead, though more hardware than protocolia.
Again, thanks
Jim
stevech I have some Anarduino boards with other radios, and some LoRa modules.
I don't see a need in my world to struggle to run at -120dBm. And from what I've read, it IS a struggle to get below -100 with these - due to the zillion config items.
Thanks...
I don't have a power budget yet. But, thanks for the warning.
Jim
stevech LoRa radios use spread spectrum with long code lengths, lots of FEC, etc. This lets them demodulate at very low or negative SNRs (due to post correlation gain).
These $5 radios can use very low DC power on average and provide miles of range at VERY low bit rates - like 100bps or so.
Use case: rural wireless utility meters.
Radios like the RFM12, RFM69, et al, are FM/FSK and are used in sub-GHz bands (per regional regulations) for about 100m or so, depending on how line-of-sight the path can be. Lower channel bandwidth and lower bit rates = longer range.
Often these radios are used for SCADA and so on where there's a relatively nearby gateway to a WAN. Typically, they use 9600bps, but the radios can produce a 125Kbps signal. The FM deviation and modulation index you the operator choose governs the range, and less so the antenna gain.Typically FEC is not used, and instead reliable datagram services are used with ARQ (ACK/retransmit for error correction).
Example hardware
http://www.anarduino.com/ lots of use of Atmel AVR MCUs for managing the sensors and radio links
Example open source (Arduino compatible) datagram protocols for many radio types
Thanks
That explains a lot
Jim
To put it simply LoRa modules are capable of extreme range links at very low power.
@434Mhz, 10mW, 1042bps, will go around 130km LOS on simple 1/4wave wires.
Reduce the data rate dow to 98bps, and you will get over 300km.
Some background here;
http://www.instructables.com/id/...
Stuart Robinson
GW7HPW
alexru What planet are you on, that you have 130km LOS? Curvature of my planet screws that up :)
alexru Also, just plugging in your numbers into the link budget calculator gives -117 dBm (-124 dBm for the second case) of required receiver sensetivity without taking into account any other losses, just free space propagation.
Even with some modulation gain, that is highly unlikely that such receiver is possible and can be made cheaply. I would like to see some independent test reports.
And if they do go this far, how will this work if 100 people within 10 km radius try to use the system? It seems like their receivers will be over-saturated all the time.
What planet are you on, that you have 130km LOS? Curvature of my planet screws that up :)
I am on Earth, in fact I got 269km LOS @1042bps @ 10mW, but that was with a 5\6db omni on the roof of my shed so I adjusted the distance down a bit for 1/4wavesy.
Semtech were impressed with the results too.
All you need is a little imagination (often lacking these days I know) to come up with a way of getting LOS distances (TX to RX) to that extent.
The RX cost me about £20 to build, the TX about the same.
alexru Yeah, you can create rigs to get huge distances, that is not the problem, HAMs do it all the time. Let's see how this works when this technology is actually used by anyone but two endpoints used for testing. There is just no way that transmitters that powerful and receivers that sensitive can be useful in a crowded environment.
Pretty low receive power, claimed "long range". I'd be interested in an Atmel chip with multiple-km range and less than a few mA receive current. It must work at 950MHz or less due to the potential environment (tropical rainforest with lots of obstacles and nasty weather). 430MHz would be better, propagation-wise but I don't know about international ISM licensing, yet.
I have done a fair bit of testing with LoRa devices, nothing more sophisticated than an RFM98 module and a simple microcontroller.
When compared to previous ISM band devices, such as the RFM22B, the range of LoRa is at least 10 times and up to 30 times greater.
However the propagation differance between an obstructed view such as in a forest or across a flat Urban area can be as much as 1/1000th of the LOS range, that has been my practical experience.
So whilst you might get 300km with LoRa when there is clear line of sight between RX and TX, that may drop to 1km or less with an obstructed view.
I am dropping this one. It is not going to be particularly useful in my application.
Jim
stevech To put it simply LoRa modules are capable of extreme range links at very low power.
@434Mhz, 10mW, 1042bps, will go around 130km LOS on simple 1/4wave wires.
Reduce the data rate dow to 98bps, and you will get over 300km.
Some background here;
http://www.instructables.com/id/Introducing-LoRa-/
Stuart Robinson
GW7HPW
way too optimistic on that range. Esp. without tall antennas, and the radio horizon. Let's pray for more mountain tops. Flatter earth!
Too much naive hype.
stevech 10's of Km.
Not.
For what ever it might be worth, "radio horizon" is not what it might seem to be.
I have communicated on 440MHz with a few watts from near ground level (and just the whip antenna on the handi-talkie) for distances in excess of 50 miles/80km. I cannot tell if it was a propagation fluke or not, but it was repeated several times. This was using FM voice, which is, at best, mediocre for weak signals. It was important to choose a location free of close obstacles such as buildings and trees; trees a half mile out were no problem.
Thus, I can see the quoted range, perhaps under super-optimal conditions, especially given the processing gain of spread-spectrum.
Jim
alexru There is no question that it is possible to transmit signals that will circle the earth many times. The question here is what we are doing here? Consumer/industrial automation equipment, that must work in all sorts of conditions, or proof of concept long range systems, where you need to be in the middle of nowhere to see them work as claimed?
I'm much more interested how this works in metropolitan, industrial or office environment. You know, places where people actually need wireless networks.
Alex -
I agree, fully.
Jim
The LoRa devices will also go between 10 to 30 times further in 'across a city tests' when compared directly with devices such as the FSK based RFM22B.
One benefit of this performace gain is of course that if LoRa were to replace current systems, considerably lower power levels can be used, thus reducing the local RF noise floor.
And they are also highly resistant to interferance that can kill other FSK based systems.
alexru You sound like a salesman. I'm out, this is no longer a constructive discussion. All I see on this technology is a lot of marketing mumb-jumbo and no technical specifications of any kind.
I have to agree with Alex, again.
Claims of 100+km are meaningless in the context of a city or an industrial plant. Of course, if done properly, something that will communicate X farther in open space (compared to solution Z) OUGHT to communicate X farther in a highly obstructed environment. However, unless SS does something to eliminate multipath and all the other ills of such a place, I'd almost be willing to place a bet that it won't, in fact, behave so fantastically.
All those words sound to me like marketing hype.
I'm done.
Jim
