Weightless, The Hopefully-not-vaporware Internet Of Things Chip


Imagine a single chip able to interface with your Ethernet, USB, and serial devices, turn those connections into wireless radio signals with miles of range, able operate off a single AA battery, and costs less than $2. That’s the promise of the Weightless special interest group that wants to put several hopefully not vaporware radio chips in the hands of everyone on the planet.

Long-range wireless networks are a tricky thing; for home networks, Bluetooth and WiFi reign supreme. Venturing into the outdoors, or really any place more than a few hundred feet from a WiFi repeater is a challenge, though. If you’re trying to send data to a fleet of automobiles, track an endangered animal, or make a smart power grid, your only real option is a cell phone tower with very high costs in hardware and battery life.

Weightless hopes to change that with a small radio chip that includes a MAC, PHY, and all the components necessary to turn just about any digital connection into a wireless link between devices. The radio will operate in the spectrum left behind by UHF TV (470 – 790MHz), and the folks working on already have some reference designs etched into silicon.Don’t expect this to replace WiFi, cellular, or Bluetooth, though: according to the getting to know Weighless book, the designers are aiming for a data rate of only a few kB/s.

Still, it’s a great use of now unused spectrum, and would fill a huge gap in what is readily possible with homebrew Internet of Things things.

Tip ‘o the hat to [Mark] for sending this one in.

77 thoughts on “Weightless, The Hopefully-not-vaporware Internet Of Things Chip

    1. Hi Cynar

      You will be likely be able to order your first chip around Q3 this year.

      In the meantime, the core benefits of Weightless technology over alternatives are cost, battery life and range. Click on the three icons in the banner of the home page for comprehensive details of each.

      Or click here:


      Registration gets you access to a detailed summary of the Specification.

      1. Hi,

        I wrote the book…..That’s about right. And that capacity is for the base station. So we’re not claiming anything magical. If all the terminals were right on the cell edge at 10km (outdoors) or 5km (indoors) they’d all be using the maximum 1024 spreading codes and the entire cell capacity would only be about 2.5kbits/s (pretty low for a 6MHz wide channel!). Happily, that sort of distribution rarely happens and so on average you get about 100-500kbits/s total cell throughput. That is shared across sometimes 100,000 machines but each is only generating about 200bytes/day so it all hangs together. And you can always cell shrink or sectorise. I’ve modelled this to death so can provide all sorts of detail if you need it – and there’s some useful stuff on the Weightless website as well (under media/resources).


      1. I have a 1watt 2.4Ghz amp but when I connect it everyone around me start complaining about there dental fillings and the arcs jumping from the sink. So with a high gain directional antenna not only can I connect to networks many Km away but I can inflict remote pain on them.

  1. This is definitely a real thing, but it’s at the very early stages of deployment; the first engineering samples are off the production line. I expect it’ll be about a year before hobbyist-friendly modules are widely available.

    It’s primarily intended for use in smart meters and other distributed sensor applications.

    1. “It’s primarily intended for use in smart meters and other distributed sensor applications.”

      Sure primarily that with a 1 trillion market expectation. You almost sounded informed.

      1. Where are you getting that $1trn number from? Do you think that’s at all a reasonable number?

        Oh, I see, it’s at the bottom of the press release, and covers the entire “internet of things”. Obviously it’s not going to hit that out of the gate, you might be looking at 10 or 20 years to reach that number, across a very large range of manufacturers.

        http://www.weightless.org/about/weightless-a-game-changer mentions the smart meters, and uses them as the default example in its discussion of the tech. It also talks about M2M generally, and why the system is designed for sensor-like uses: small data rates, delay tolerance, infrequent polling.

      2. Hey guys,

        MIT Sloane Report 2012 – stated that the Industrial Internet (which includes M2M market) will open a new market of about USD$70Trillion across both developed and emerging economies. It is a new economy.

        Even if M2M is not as high as this figure, it is widely accepted that is will be USD$Trillions.

        I have graphs from the report, so happy to send it to people who email me, or you can find it for yourself on the web.

        IMO, Weightless is a great open source standard that is the first to address the proper requirements of M2M. Current incumbent technologies have their place, like Zigbee, Wifi, Bluetooth, GSM/CDMA but none meet both the business and technology case for a global M2M solution and truly open up this market to deliver real value to people’s life in healthcare, smart cities, automotive – the list goes on really.

        A tragic story that I heard recently was that 50% of the people that died in the recent Japanese Tsunami had charged mobile phones. They could not make contact for help as the cellular network crashed. Current thoughts with the Japanese government are to coat the country with “throw away” sensors with a GPS and “help me button” which work on a Weightless type network and not GSM. The cost effectiveness of Weightless solution allows for this possibility.

        At Argon Design we are having different conversations with clients wanting to use Weightless in Mining, Water Treatment, Road/Bridge Monitoring, Healthcare, Automotive, Asset Tracking to name a few….

        Weightless Members are from around the world – see the world map on weightless.org. Might be worth finding out more.

  2. This would be good if it actually did what they claim. Being rational, its probably going to do half the range reliably (5km would still be awesome) for probably £20 and the standby battery life may be what they claim, but in use would be far far less!
    Either way, anything close to those figures would be amazing!

    1. They probably can get 5km, although that’s probably line-of-sight rather than with both endpoints at ground level. I know for a fact you can get at least 60km LOS at 1 watt EIRP in the 900MHz band (my employer sells modules that do it); lower frequencies will get much longer range with the same power, or alternatively need much less power for a given range. 5km is a fairly short range, so they’re probably using quite low power. The rest depends on the antennas.

  3. I can’t seem to find anything mentioning USB or Ethernet. It seems that there is some sort of parallel bus for moving data as well as a SPI port for debugging. Still, this could be cool once it’s available.

  4. Lets have a little talk about communication protocols, shall we?

    Everything from Ethernet and wifi to bluetooth hast to conform to certain realities. The laws of physics and all that. Signals get garbled on shared media like a single wire or em spectrum or whatever. There are generally two ways of dealing with it: CSMA/CA or CSMA/CD. That is carrier sense multiple access with collision avoidance (wifi) or collision detection and recovery (ethernet). You can either tell everyone in the area to shut up before you start transmitting or you can recover gracefully when two carriers start sending at once and garble up the message. In order to do either the slot time must be considered. That would be the minimum time that is needed to send a signal from one end of the medium to the other(in this case 10km) and get it back. This is used to create the minimum data frame size based on baud rate. The justification being: how do you know if your signal was garbled if you don’t see the other person’s signal mucking up your own? Therefore the minimum data frame must take twice as long to send as getting from one end of the medium to the other. This allows anyone sending on the far side to get their signal to you and you can detect a messed up communication.

    Now that that is sorted lets check it out:

    assuming afew kb/s means the bit time (or baud) is 4ms, that means one bit travels about 81 cm per ms. Fancy math later and we get a little over 12k in traversal, doubled up to 24k. That would be your MINIMUM data frame on any given channel. That is six times larger than the MAXIMUM frame for ethernet. What that means is this technology as presented will get congested fast. Real fast. I mean just three or four of these devices transmitting normal traffic on one channel within 10km would make that channel unusable.

    This tech would be great for very intermittent transmissions of relatively large continuous blocks of data like integrated sensors platforms. It will also need many many channels. Hundreds to thousands of channels for it to not get congested if the density of these things came within 10% of what wifi is today.

    The conclusion is this will be horrible, absolutely horrible for USB, ethernet, and similar technologies but pretty good for remote sensors, data cashing/sending protocols, and any very low bandwidth applications that don’t need low latency and can generate huge data frames.

    1. Hi,

      Can I help as system designer. Most machines actually have very scheduled communications (smart meters ever 3 hours, etc). We set up regular “rendez-vous points” for each device and tell the device where the transmission slot will be. It then responds correspondingly some hours later. In fact we’ve got a rather novel custom-designed MAC that is mostly schedule-based with very limited contended access that we think is pretty good for M2M. So turn-around time isn’t relevant, and scheduling allows us to be more than 80% efficient on the MAC giving pretty good capacity. One cell can readily handle 100,000 machines generating 200bytes/day or thereabouts.

      You’re absolutely right this will be horrible for Ethernet – but happily there’s plenty of other radio systems that do that well!


      1. So it’s hierarchical then. That is good. It is more for infrastructure and monitoring then random chatty consumer goods. How do you handle clock drift on the hosts? Do they intermittently passive listen to keep in sync?

        1. Hi,

          Yes that’s correct. It’s a base station “cellular” type network with terminals talking to it – the same architecture as eg 2G in overview. It’s for terminals to talk to databases, eg a BMW car to a BMW dealer or a smart meter to the electricity supply company. We don’t expect your fridge to talk to your toaster. Clock drift is handled by broadcasting regular sync bursts – the assumption is that terminals have terrible clocks so we have both coarse and fine sync prior to every transmission / reception.


    2. Spreading codes. All GPS satellites emit on the same carrier frequency, and use orthogonal pseudo random codes that have no cross-correlation. 802.15.4 is doing the same. This is also named DSSS. The CDMA mobile phone technology is doing the same.

  5. I am highly sceptical about this. I produce devices for monitoring leaks and flow in water networks and we can only dream about something as good as this. The big boys like Texas can’t come close with any of their chips, either in terms of price or range.

    Firstly to get that range you need both insanely good sensitivity and noise rejection and a power amplifier to transmit at the legal limit. Most manufacturers would charge you 10x what this thing costs for that. The data rate would be <1K baud too.

    Then you have the 10 year battery life. Lithium batteries costing $20 each might get you there, but you will find that after about 5-6 years their self discharge is enough to kill them even if your device uses almost no power. To get to even 5 years you have to use microamperes most of the time, and certainly can't keep the radio on all the time. Say you poll the radio every few seconds for a "wake up" signal, that would work. Then a message takes as long as your polling interval to send every time.

    If this this exists it will revolutionize the world of data logging and smart metering. These guys are fools to bother with hobbyists when they could concentrate on massive volume to data logger manufacturers. The latest, most sensitive chips (IIRC -130db) we looked at had a minimum order quantity of 100,000.

    1. Hi,

      Can I help as the system designer. In terms of battery life, we put devices to deep sleep for 15 minutes. Then they awake for 100ms to rest their clock and listen for network messages. Then back to sleep again. We expect they might want to send about 200bytes/day using 40mW transmission, so very low power levels. The range doesn’t require insanely good noise floors, we’ve got a noise figure of about 4dB – very much industry norm. It’s the use of spreading that gets us the range.

      You’re right about battery self-drain. We do know of some batteries that can go 10 years without complete self-drain but if cheaper ones are used (eg standard CR2302 coin cells) with eg 5 year shelf life then that will be the life of the device too.


      1. Thanks for the reply William.

        So how do you sync up with this 100ms awake period? Obviously sending is no problem as you can keep the receiver base station on all the time. Are you trying to rely on synchronized clocks or something? We looked at that but typical watch crystals are pretty poor and anything better adds a lot of cost, and it certainly sounds like you are targeting the very low cost end.

        1. There are tricks to getting the watch crystal to within the accuracy of your RF crystal if you periodically run an FLL, so for us in narrowband applications we get a time base accurate to +-2ppm using watch crystals. Wide and doesn’t have as accurate RF requirements so they’re not going to be as good.

          Who do you work for mojo? Sounds like we do the same work.

    2. Remember that there’s a price for the chip and a price for the base station. If the chip is cheap and low-power, that means that the base station needs to be more sensitive, higher power, and more expensive. On the other hand, if you’ve got a million endpoints in a 10km radius, the price of the base station amortizes out pretty fast.

    3. You can look at the Semtech LoRa parts, and read the datasheets. I don’t have any arrangement with the Weightless group, but I would bet a fiver that Semtech is the company doing the Weightless silicon. In specific, the Weightless part would be the SX1276, but they make a whole family of LoRa parts.

      Basically, they use a ton of DSSS, plus what appears to be RS coding, plus a high transmit power (17dBm). In any case, read the datasheets yourself!

  6. The laws of Physics and other inviolable constraints are what they are. Shannon’s also applicable as a Hard Limit to what’s real in these schemes. There’s No Such Thing As A Free Lunch. But if you’re not too picky about what you eat or how fast? Keep all that in mind when evaluating these Whitespace plans. It won’t stay white for long But if we plan to defer saturation from the first? .

    Alligators and Jackrabbits was a cartoon in Repeater journals. Meant machines that had more TX than RX and the Reverse. Be mindful of that when building with these chips! Or we’re going to Fratricide faster.

    Meshes in a low contention spectrum realm Vs in a high noise floor *WILL* show degradation in several stacked non-linear fashions. And non-meshes eat noise floor with handshakings etc in different modes than those of mesh overheads. All of these details will result in an Electromagnetic Fug having our data risk slogging like a Greyhound trying to run in cold molasses. Adding noise floor is unavoidable as technology becomes pervasive. Still?

    Imagine the contention of every active “thing” you own seeking a handshake as an event response? And multiplying that to the level of each picture on your walls. Dirac Sea Level?

    I’d hope us Hackers are smart enough to advocate rational deployment so these wonderful creations of RF audacity do not Fratricide our spectrum as Spark did for a time.

    The reality checks are trades between simple-cheap-efficient in use of spectrum. There’s no evading the pick max of 2 that I see absent the modifiers of range or throughput. So I’d see using Shortrange spectrums including optical to keep Longranges free for where they’re the “best” viable ones. That points to a rethink. DC to Daylight was a term I’d been taught. and I was handed a “Dead Tree” copy of a physical spaces explain:


    Teal Deer: For my comment and the Caltech paper- bigger space= more elbow room or the reverse? All we create rubs elbows with all else.

  7. OK Folks, if you are interested in this kind of products, just wait mine. It will basically be the same chip but with 100 years+ self sufficient autonomy, no range limit, all that and more for half the price. Stay tuned for my white paper draft.

    1. …don’t forget to have 219 pages of fluff and charge a low price of 200 bucks just to read the thing and sell it out of an IV league institution for cred. (yes I just used IV, I’m a traditionalist like that)

    1. Hi,

      I’m one of the designers of Weightless, perhaps I can help. It’s not magic (I wish I could do that stuff). Basically we have two things to play with – relatively low frequencies that have good propagation and wide bandwidth channels (6MHz). We then trade off data rate against range. So if I were using this for LTE I’d aim for about 4bits/Hz and so deliver 24Mbits/s per cell per channel. But we don’t. Instead we use direct sequence spreading (like GPS satellites) to trade bandwidth against range. The net result is we only get about 100-500kbits/s per channel in total (not per machine), massively less than cellular. But the use of spreading gives us about 30dB extra link budget. That gets us 10km outdoors or 5km for indoor applications using a 4W base station. And those data rates are just fine for supporting most machines that generate typically 200bytes/day.


  8. Why is it that everyone doing internet of things thinks the problem is with the low power RF link between some ‘thing’ and a central node?
    Please, we had good way to make radio… we had micros, we could make systems draw less current then the self discharge of the battery for many years.

    Want to go diy? you can get a micro+rf good enough for such low data rates for less than $5 in single quantities. And i bet you can write code for reliable transmission within a week.

    I bet you could get remote control via short range radio for simple loads like lamps for less than 5, directly built in them by the manufacturer.

    But the problems are the same as with home automation. Here are some examples:
    -the benefit over added complexity is not worth it
    -there are sooooooooo many manufacturers with proprietary, non compatible technologies
    -right now is way too expensive for what it does (i don’t want to pay 50 to buy a thing that can turn on/off a lamp for which i paid 10, plus the cost of a central node, plus plus plus).

  9. Understanding everything begins with an idea. White spaces AKA guard bands, they exist for a reason however technology has improved to where they most likely can be utilized while protecting the TV broadcasters/viewers. Personally I’d like to see low power(analog to keep costs low) FM and TV have access as well in the US, but entrenched broadcasters ain’t gonna let that happen For some reason in regards to weightless, the proverb; wish in one hand and crap in the other and see which one fills up first, come to mind. Realistically this is going to require as much power as the lowliest of a cell phone requires, and that cell phone requires cryogenically cooled receivers at the phone towers to work. Somewhere (a comment maybe in all the referred links I read this spectrum as structure penetrating. I wish they would use another phrase because that one creates unrealistic expeditions as to what material are penetrating. Only structure penetrating in the sense as compared to longer wavelengths is reflected by more materials and finds an exit out of the building more easily. People living on a forest or jungle canopy are,’t going to find this useful. Here in the US we where sold that that spectrum was to be use for first responders and public safety communications, and demand is likely to fill all that spectrum including guard bands. This may be /\/\otorala versus Microsoft, but my guess Motorola may be a weightless player in countries where their digital telephony is unaffordable and wireless internet is in demand.

    1. Hi,

      Actually, we don’t need the same power as cellular towers to get the same, or greater range. This is because we trade off data rate against range using spreading. So instead of delivering 10+Mbits/s we spread signals (like GPS satellites) by up to 1024-fold. This gives us an extra 30dB link margin allowing us to operate at less power but we lose data rate so go from eg 10Mbits/s to 100kbits/s. This is a trade off that’s fine for machines (smart meters don’t care whether they send their 100bytes reading at 10Mbits/s or 10kbits/s) but would be awful for people. That’s why cellular systems don’t use spreading – they maximise data rates even if it reduces range. Horses for courses…


    1. Hi,

      Yes, you’re right, there’s many similarities with Sigfox. The big different is that Weightless is an open standard. Also the TV white space bands are about 20 – 100-fold bigger than the 868MHz Sigfox uses so our network capacity (and data rates) are correspondingly higher.


      1. I know a lot of open standards that are open as in “dont sue competitors” , meaning “licensable to other industrial companies only, still requiring an NDA to get specs”, so let be doubt about this :)

    2. Well, as I see it there are quite a few differences between Sigfox and Weightless (although I’m not a Weightless expert so please do correct me if I’m wrong on these points).

      First of all Sigfox is an operator. The connectivity is thus out-of-box without the need to deploy application specific networks.

      The Sigfox network is deployed in the ISM meaning that it can be deployed worldwide today (currently being rolled out worldwide). Sigfox could be deployed in white spaces (and thus use higher data rates), but since white spaces aren’t available in many places, it is not used by Sigfox today.

      In regards to standardization, Sigfox is leading an ETSI working group that is standardizing low throughput networks (cf http://portal.etsi.org/ltn/LTN_List_members.asp). FCC involvement is also planned.

      1. Thomas,

        Many thanks for the useful information.

        If you dont mind, I would like to make a few points clear here.

        Let’s look at one of the reasons why the GSM solution has stifled growth of the M2M market. Terminal costs. A significant proportion of Terminal costs are royalties paid to a few companies. This is why we need open standards, ones that are royalty free. For the M2M market to be truly opened, terminal costs have to be “cheap as chips”. Thus, we do not want companies trying to claim that they own a standard and then stifling the market through forced technology and/or royalties. I am not saying SigFox is doing this, but what I am saying is that we must look at how we can open the M2M market and deliver proper value to the evolution of humanity and this is through open standards, royalty free with global participation – thus, IMO, why Weightless is a good approach and would not surprise me if it became part of 3GPP.

        From my understanding SigFox has raised some venture capital but from the sums I am hearing no where near enough to make a telco type play around the globe as you suggested. In fact my understanding is that SigFox want to enable organisations to develop their own solutions but tied to their technology.

        To properly address and open the M2M market we ALL need to participate in an open standard. Proprietary does not work, never has, never will. Look at the history of the standard fighting for industrial comms between France, Italy, UK, USA in the 90’s with FIP, Profibus, Fieldbus, Foundation Fieldbus – the list goes on. All this arguing over standards stifled the market, wasted pots of money and was primarily down to political agendas of ego-centric companies. I just hope this is not repeated. I want to see value added to people’s lives.

        If we look at the kind of companies supporting Weightless, they are all heavy hitters – ARM for example, a USD $13Bn that specialises in mobile/computing technology.

        If we look at the organisation behind Iceni the first Weightless silicon, it is world class, The Cambridge Silicon Radio founders. These people not only help develop a globally adopted wireless standard, bluetooth, but grew their company to USD$10Bn. These guys know how to develop and bring new wireless standards to market. They have a successful track record and will no doubt be successful again.

        As the VP of a technology and product company, we use all types of technologies in our client solutions from Bluetooth, GSM, Wifi, Ethernet and so on. We feel that Weightless is a good approach for all sort of reasons – see my blog – http://www.argondesign.com/news/2013/feb/15/why-weightless-good-m2m-solution.

        Standardisation is always good, so long as it not proprietary, it is open, it empowers community involvement, it opens up markets and adds value to people’s lives. So, I would assume that in the spirit of “creating adding value to people’s lives” and “not being ego-centric”, Weightless will be invited into the ETSI conversation?

        Finally, from my conversations over the past few years with global heavy hitters out of Silicon Valley, who are game changers, their philanthropic approach is only to do business with those that have the interest of humanity as a driver, laying down core ego-centric agendas. They have amazing strategies/technologies to empower global transformation and their Future of Life vision is truly inspirational.

        We MUST work as a community. We MUST focus on adding value to people’s lives.

        In support,

    1. Hi,

      Just to let you know we designed very robust 128-AES encryption with complex over-arching cipher modes and very strong authentication and repudiation into Weightless. We believe it’s at least as secure as cellular systems, perhaps more so.


  10. I develop firmware for stuff like this for a living and the products we’ve made so far we base around MSP430 MCUs and CC RF transceivers. We easily get 5-8kms on our system and we use a custom TD MAC protocol to keep power consumption low and support multiple hops to extend the range. Obviously there are different tradeoffs between our products and this Weightless proposal but the two most complex requirements of our products have always been the power consumption and keeping accurate time and these requirements fight each other. There’s also a big difference between keeping accurate time at ambient temperature and keeping accurate time at -20C.

    I find there are always additional costs and for our product the cost of keeping accurate time (good quality calibrated crystals) and supporting the wide range of temperatures which exist outdoors (don’t assume your capacitors will still be capacitors at -20C) means the BOM of components required to support the RF chip is about 2-3x the cost of the RF chip itself.

    1. Hi,

      As system designer I might be able to help. Weightless is comprised of base stations and terminals. Base stations have highly accurate clocks (typically GPS-fixed) and mains power. They hold the system timing. Weightless terminals have very inaccurate clocks but sync to the network. Every network frame starts with a course timing sync burst from the base station and every transmission to a device starts with a fine timing sync burst. This means the timing on the terminals can be very weak – about 1s in 15 mins accuracy is just fine. They turn on, listen for a sync burst and then update their clock. They only need to maintain sync for about 100ms. So the external components are not all that expensive at all.


      1. It seems to me the duty cycle of time spent in RX on the remote terminals to maintain or recover the time is going to significantly eat into you energy budget. What size battery are you planning on using to get your 10 year life span, and what proportion of your bandwidth is being sacrificed on time sync messages? Claiming 1s in 15 min accuracy on the terminals should mean that you lose that proportion of your bandwidth to sync and at least that duty cycle must be spent in RX on the terminal to catch it.

        Assuming a receive energy requirement of 12mA for example this equates to 12uA/s to maintain time in a remote terminal which in 10 years requires at least an additional 1Ah of energy.

  11. Hey Europe and Russia, when are we going to replace american microwave like we did their food preservatives? Is one type of “accelerated cell growth” better than the other? Why not just use mutiplexed low frequencies which is efficient for use and keep people from abusing technology and become lazy and worthless as ‘screen slaves’?

    I don’t like that I can get “accelerated cell growth” because the masses are ignorant and uneducated and engineers work for marketing interests

    1. You included the words Europe, Russia, People, Slaves, Masses, Ignorant, Uneducated and you managed to slight the US in the process while adding nothing of merit to the discussion in under 100 words. Quite impressive.

  12. I went as far as to register on the Weightless Website to have a pick at the magical specification. MAN! I must say I’m impressed: the specification is a perfect definition of “weightlessness”. It just does not exist.

    I’m seriously annoyed by this one man team. William Webb who is representing weightless is a professor, a book author, a system designer and its CEO.

    1. You have registered as an Observer. Observers are entitled to immediate access to a summary of the Specification. It is a detailed summary and certainly sufficient to make an informed decision as to whether the technology is suitable for your application. The full Specification runs to well over 300 pages and is comprehensive and detailed. You need to register as an Associate or Core Member to access this document. Registration for full membership requires that you sign a License Agreement.

        1. When you registered for Observer group membership you received an email with login details. Clicking on the link in that email takes you to a Welcome to Weightless landing page. On the right hand side of that page is a book icon which is clickable. That opens up the Specification Summary. Please say if you have any problems.

          1. And incidentally, you can return to this page via the Membership & Login menu, then click on the Welcome to Weightless option. If you are not logged in you will be invited to log in. If you are not registered then there is a link provided to the Registration page.

  13. Sounds interesting, if it ever comes to exist… but a few sticking points:

    >>Yes that’s correct. It’s a base station “cellular” type network with terminals talking to it – the same architecture as eg 2G in overview

    So, this is not “mesh networking” at all, but basically an alternative cellular network to AT&T, Tmobile etc… similar range, but geared toward low data rates (sensors, etc.). I.e. it is not a direct replacement of ANT or BLE where two battery-powered devices may directly talk to one another – a network requires a central wall-plugged basestation (ZigBee-like). The million-dollar question: can you buy your own basestation, or is that what they’re really selling – a connection to their network with a monthly access fee? Subquestion: if the basestation is buyable, is it buyable by hobbyists? (Or does it require FCC license, other maintenance fees or regulatory requirements that would put it out of an average user’s reach?)

    A typical sensor today can make daily, maybe even hourly-or-more dumps to a standard cellular network for years with a small solar cell and rechargeable battery. In the US at least (and likely many other places), it is monthly access fees, not power consumption, that keeps “M2M” an empty buzzword rather than something people actually use.

    Another consideration: what kind of instantaneous Rx/Tx currents will be required to exchange data multiple km at the end device? Pulling 10+mA from a coin cell for more than a few ms is an iffy proposition at best, and that’s at room temperature. Current low-power wireless protocols include various load-shaping tricks to keep the Ipeak down during hungry tasks such as peer/timeslot discovery. What sort of considerations are built into this device for operating from real-world batteries over real-world temperature ranges?

  14. Hi Tim, I will try to help here.

    It does exist.

    Correct, it is not a mesh structure at all. And yes, it is reasonable to think of it as a cellular network optimised for M2M applications in terms of its power consumption, cost and signal propagation characteristics (range and building penetration).

    Correct, as you say, it is not designed for enabling two terminal devices to communicate. The terminology of ‘machine to machine’ is actually misleading here because in almost every case it is ‘machine to network’ or ‘network to machine’. Of course, if you wanted to intelligently connect two terminal devices on the same network then that is perfectly possible to do.

    You can buy your own base station and they are less costly than a, say, 3G base station. You also need fewer of them to cover a significant geographical area – about one third as many in practice – so the cost of deploying a network is significantly reduced. There are several other factors that substantively reduce the cost of owning and running a Weightless network but I won’t go into those here since you asked specifically about ‘hobbyists’. Whilst a Weightless network can be realised for considerably less than a traditional telephony based network, it is not in hobbyist territory. For guidance – and this will reduce over time – the typical base station cost is around $10k.

    There are many ways to maintain power to a device – connect it to the grid, provide a renewable energy source as you’d just described using a solar example, energy harvesting, or a battery. For many applications however you want fast, easy, low cost, fit and forget, reliable, compact, low maintenance solutions. A solar cell and rechargeable battery could be a solution for some but in reality the cost and added complexity would simply rule this out in many applications. An intelligent vehicle proximity sensor, for example, in an underground car park in an Smart City scenario clearly would not be practical using solar power.

    You are correct re the limiting factor – depending on the application. Some applications are cost sensitive, others are limited by the technology – power consumption for example. A traditional telephony based hardware and network cost is simply not a sustainable commercial proposition for the majority of M2M applications. Very short data packets transmitted infrequently do not fit well with the large costs associated with deploying, operating and maintaining a 3G or 4G network. The hardware costs at the terminal end are typically an order of magnitude higher – say $20 compared to $2. The network cost is derived from both base station hardware (a function of the number needed for a given land area) and spectrum cost. 3G and 4G spectrum is not free. It is very not free. And the operator has to appropriate the costs of acquiring the license to operate a network across all users. White space spectrum is free. So in every sense, Weightless costs a fraction of a 3G or LTE alternative. To put this into context, in round numbers, network access costs are an order of magnitude lower – say $10 per year rather than $10 per month.

    Instantaneous power is limited to 100mW, and the actual Rx/Tx time is reduced to a fraction of that required by other technologies. This means in practice that the battery life is equivalent to its shelf life – it is defined by the leakage characteristics of the cell.

    Weightless technology achieves its extremely low power consumption through a number of innovative techniques that take advantage of the specific requirements of an M2M device (as opposed to a human). A simple terminal architecture that shifts the complex (and therefore power hungry and expensive) processing from an onboard processor to the cloud increases battery life. Scheduled wake-up and Tx/Rx times enable the duty cycle to be optimised for the application. A highly efficient, fit for purpose, protocol ensures that the data packets are in the same order of magnitude as the user data being transmitted – low overhead. Weightless devices are quite lazy – they spend almost their entire lives snoozing (or more precisely almost entirely asleep). They will wake for around 100mS to synchronise their clocks and check for status updates every 15 minutes before slipping back in a deep slumber. And, depending on the application, an interrupt line can be used to wake them according to an externally determined schedule or event occurrence.

    And, in case you were wondering about how the 100mW and 10km range numbers can be reconciled, it is through spreading. Spreading is the same technique used by GPS satellites to transmit over very large distances using very little power. It is a technique that trades off data rate for power consumption and can allow an extra gain of 30bB on the link budget. Again this is a fit for purpose design decision. The utility company does not care if the few bytes of data from the meter are transmitted over a longer period of time but humans get frustrated by low data rates when call quality degrades or their YouTube video is constantly buffering. Weightless is optimised in every respect for lower cost, longer battery life, longer range, better in-building penetration.

    More information about these three elements of the game changing technology are explored and explained on the following pages.


    If you’d like more information, I’m happy to oblige.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.