Weightless, the Internet of Things Chip, Becomes Less Vaporware

weightless

Several months ago, we caught wind of Weightless, a $2 chip that will run for 10 years on a AA battery and communicate to a Weightless base station 10 Km away. Yes, this is the fabled Internet of Things chip that will allow sensors of every type to communicate with servers around the world. It looks like Weightless is becoming less and less vaporware, as evidenced by the Weightless SIG hardware roadmap; Weightless modules might be in the hands of makers and designers in just a few short months.

Weightless is an extremely low-cost wireless module that operates in the radio spectrum previously occupied by analog broadcast television. This is a great place for the Internet of Things, as signals in this spectrum have a lot of range and the ability to go through walls. These signals are sent to a Weightless base station where they are then sent over the Internet to servers around the world.

The Weightless SIG has been hard at work producing new silicon, with the third generation of chips heading for volume production next month. The only thing this chip requires is a battery and an antenna, making Weightless integration for new designs and projects a snap.

There’s one thing Weightless is not, and that’s a free, high-speed connection to the Internet with a $2 adapter. Weightless is designed for sensors that only transmit a kilobyte or so a day – medical sensors, irrigation control, and other relatively boring things. There’s a summary video from the recent 2013 Weightless SIG Summit going over all this information below.

Comments

  1. Caleb says:

    -BEGIN RANT
    Isn’t this the definition of vapor ware? They are showing a proposed road map of a theoretical chip with amazing possibilities. Show me a field test with a prototype unit not a 3D rendering or some stock photo of a chip.
    -END RANT

    Please let something like this actually come out! Getting past hacker wanting to use this for a cheep always on connection to their device this thing could have real uses in the scientific community. A device like this could allow monitoring sensors in decade long experiments be priced like a disposable item.

    • Al says:

      It’s not a *proposed* road map – those stay inside company board rooms. It’s a public declaration (see video). Since this story first saw the light of day on this board more than $25m has been invested in getting from there to here. That they (and other small companies like ARM and CSR and Accenture etc) are talking about Weightless, it is more substantive than a 3D rendering or a stock photo.

    • Al says:

      The chip isn’t theoretical – it’s in fab now, taping out in four weeks, it’s the 3rd generation and the first two gens are already deployed in successful trials. This 3rd gen chip is destined for volume production. The image is the actual chip, not a 3D render and not a stock photo.

  2. Elia says:

    While it’s a nice idea as the targeted applications don’t need high bandwidth connections, Caleb nailed it.

    This still is advertising vapor ware. No prototype unit just some CEOs talking about their fancy brain fart about every water bottle being connected to the internet.

  3. fartface says:

    Wow they are way behind, I have designs for a chip that runs 20 years on a AA battery will cost $1.00 and has a 30Km range. I just need AA sized nuclear reactors to become available and the laws of physics to change. But it’s not vaproware!

  4. lol says:

    Well, the old spec of 10Km range is now, Up to 5Km range in the site….

    • Al says:

      It is testimony to the rigorous real world testing of *not* vapourware products that the range number has been revised in line with actual expected results in urban environments. 10km is not incorrect. It is not even best case. It is just not typical for anticipated environments. In fact the signal propagation characteristics pertaining to range in an urban environment are equivalent to GPRS but with superior penetration into buildings.

  5. Muttitude says:

    Pfft!

  6. az1324 says:

    They say v1 modules available for $12 @ some unknown qty in Q1 2014. The question is does modular certification apply to white space devices meaning anyone can drop these in a product and sell it just like a zigbee module as long as they use an approved base station which enforces the TVBD database?

  7. 0xfred says:

    The article does say *less* vapourware rather that *not* vapourware. If they’re producing some silicon in volume then I’d say that’s some progress in the right direction.

  8. Kerimil says:

    Next time just cover something that actually exists and works like those new rfm69hw modules. From what I’ve heard about them they are really really impressive.

  9. mike electricstuff says:

    The big problem with getting long range on infrequent transmissions is the instantaneous battery draw – most batteries designed for long standby life can’t deliver short high current bursts, so chances are you’d need either an expensive battery or a supercap to provide the current.

    • Al says:

      Long range is not only facilitated by high transmission power (which is rigorously regulated in these bands in any case). Long range can, and is, achieved through the use of spreading and the inherent signal propagation characteristics that are exhibited by transmission in the spectrums being used.

    • Richard says:

      You can lower the peak current requirements somewhat by using slow data rates and narrow bandwidth, transmitting at very low power. To see how some radio amateurs have taken this to the extreme, google QRSS. I’m not suggesting that it’s practical to take it to this extreme here, just illustrating the theoretical possibility of the tradeoff. GPS uses a less extreme method of trading off data rate versus transmitter power to slowly send its almanac and ephemeris data. Of course, the cost of slowing down is that the transmitter must be on for a longer period, so it may not result in a net energy savings, but it lowers the peak power required.

    • Last week there was a demo at the Fraunhofer institute that showed a 1 cubic centimeter transmitter that would use a microscopic switched voltage convertor and small capacitor to slowly drain the main battery into that cap and deliver a very short burst of data (10 bytes or so) over a distance of 1 km.

  10. Aud1073cH says:

    The FCC blocked out the 700 to 800MHZ bands – for this?

    For years, professional audio was able to use the white space in between TV stations in these bands, which has now been outlawed for a few years. This caused pro audio manufacturers and users a ton of money to scramble to replace their now-illegal gear with equipment in legal bands. In addition, legal white space is getting harder and harder to come by.
    So if you happen to hear a professional mic interrupted by another signal, thank the FCC for allowing your angry birds to load faster, and whatever this thing is.

  11. natsfr says:

    Sorry but I’m loling, if by “IoT” you mean long range low powered device, it has been a long time here, with narrow band fsk transceiver (EnOcean,TI,SiLabs,Semtech,…) and more recently with some custom wideband Modulation like Lora from Semtech (SX1272). IoT is more a bullshit bingo blabla from marketing fanboy.
    The main problem is architecture behind all those sensors and on this field there are nothing really outstanding for now, nobody (manufacturers and service providers) wants to work in an open way, in my opinion 6LowPan is clearly too heavy and there are questions still here like “Really each device should be addressable with an IP address ? oO)
    My 2 cents…

  12. ejonesss says:

    10 years on 1 aa battery that is nice except that most batteries have a short shelf life and will probably pop and leak long before that.

    6.2137119 miles that i thought would require an fcc license.

  13. xorpunk says:

    I was thinking of how awesome this was, and kinda still is, till I seen regulations were being created to lock out ad-hoc usage..

  14. Al says:

    Any emerging technology starts with availability of key components (in this case silicon, modules, SDKs, BTSs and deployed networks) in a phased rollout. That the foundation of these, on the critical path for modules and SDKs, is silicon that has been publicly announced as taping out in four weeks should be properly interpreted as significant. Accurate, if at this stage imprecise, dates have been set out by the company. If they prove to be ambitious then we’d have grounds for claiming hype and vapourware. If you take into consideration that earlier iterations of the same silicon have been deployed in successful trials across the world for over 12 months then the vapourware argument starts to look shaky.

  15. How many libraries of congress can it transfer? That’s what I want to know.

  16. Rupert Bear says:

    This technology is a wet dream for the NSA! Imagine … planet wide population surveillance at the water bottle level. How good is that?

  17. zibri says:

    I bet NSA is celebrating right now!

  18. millsley says:

    My prediction is, upon release (should they get there), the specs will be bait-and-sw—- I mean updated to:

    Standby battery life of up to 2 years, daily use reduces it to a month.
    Price will be under $2, for sufficient quantities (10k+) and will be missing functionality that requires a specific external chip.
    Omnidirectional range will be 300 meters in open space with no interference.

    It is just way too much hype, way too outlandish claims, and no objective evidence that this is remotely possible at this time.

    • Bill Stewart says:

      You’re misunderstanding the market. This isn’t something you buy two or three of to automate your house; it’s more like something the electric company or water company buys a million of to automate meter-reading, and uses a fancy base station to make up for a simple endpoint chip. And yeah, it needs functionality from an external chip, such as handing it whatever telemetry data it’s trying to collect (over some bus like SPI or whatever); it’s just the radio, not the sensor.

  19. Al says:

    > “no objective evidence that this is remotely possible at this time”.

    Not true. Real world trials have resulted in a modest recalibration of range (to that equivalent to GPRS) and validated other parameters. There’s a difference between skepticism and cynicism.

    • rasz says:

      >to that equivalent to GPRS

      oh, so 1 Wat while transmitting
      and 10 years on AA

      :)

      • Al says:

        The transmit power from an endpoint is not 1W. Stop for a moment and consider whether companies like ARM and CSR would promote a technology proposition that claimed to re-write the laws of physics. To pretend that such an obvious elephant is in the room but that it’s somehow been overlooked by such companies is frankly a bit silly.

        Weightless endpoints have a very low transmit power. Sun 1 GHz signals have good range and building penetration characteristics. Weightless terminals have very long sleep periods, typically wake only every 15 minutes to synchronise the clock and listen briefly – for about 100ms – to determine whether there are any changes to the system or any messages that need to be acted upon. If not, they will return to sleep again for another 15 minutes. Weightless devices will therefore only be awake for minute amounts of time.

        Weightless uses spreading. Spreading multiplies the data by a pre-defined codeword so that one bit of transmitted data becomes multiple bits of codeword. The receiver uses correlation to recover the codeword at lower signal levels than would otherwise be possible. In the most extreme case, spreading allows an extra gain of 30bB on the link budget. Most buildings have a penetration loss for signals entering them of around 15-20dB, so spreading provides indoor coverage to machines where only outdoor coverage previously existed. Variable spreading factors are a core part of the Weightless Specification, ensuring deep coverage inside homes and buildings.

        Weightless has been designed to be very simple to implement from the endpoint side. All of the complexity is retained in the network, and so the terminal is not required to carry out any complex calculations. This means a Weightless terminal has no need for powerful, quad-core processors. Even when the terminal is awake its lightweight processors mean that power consumption is very low.

  20. LS says:

    AA cell ….. hackers should get these things right I reckon.

    • Galane says:

      Yup. No such thing as an AAAA, AAA, AA, C, D or other “battery”. They’re all cells. A battery is composed of at least two cells. The 9 volt with the round clips is a battery, some of which have six AAAA cells inside. The big 6 volt blocks with the two spring contacts, also batteries, with four separate cells – configuration varies but there’s always four.

      In schematics, a power cell is a symbol of one short and one long line with a gap, and optionally the – and + for them that don’t know which line is which. A battery is a pair of those short and long line pairs.

      I see it often with the battery symbol used on schematics that are designed to use a single cell.

      The battery symbol can also be used for two or more separate cells in series or parallel, if there’s a note about the configuration. If you want to be sure there’s no confusion, diagram the battery configuration with cell symbols.

  21. static says:

    From what I can tell from the article for those who will be the manufacturers or those who will be providing for fee service using this technology is this no longer vaporware. For hackers it’s still vaporware, and there’s a chance it will always be that. I could be wrong but I really don’t see unlicensed operation being allowed in the US using broadcast guard bands and the 700 MHz spectrum. Unless there are dollars in this for broadcasters I do see the North American Broadcast Association fighting this in guard bands. White space is an euphemism to mask why the spaces even exist in the first place, the certainly will apposed unlicensed use, as well as those planning use the Weightless to provide service for a fee

  22. smeagol says:

    “Weightless modules costing around US$12 would be available in 2014, with both price and component size planned to drop substantially over the next couple of years, costs falling to $4 or less by 2016.”

    Yeah and I have a plan for a chip that will cost 5 dollars and can handle 20 terraflops on just 3 watts.*

    What a bunch of vaporware lies. This is as real as flying cars. Technically they exist, and technically you can buy one. But no. They don’t work the way you think they do, they don’t cost what you think they should, and they are completely impractical for the function ‘a car that flies’.

    *Due to ship Q2 2025.

    • Al says:

      > “What a bunch of vaporware lies”

      Let’s be charitable and call this an opinion. If you examine the claims made by the company (listen explicitly to the presentation made by Neul’s CEO) it’s clear that the silicon being taped out next month is the third generation chip. The first and second generation silicon has been deployed in the field in multiple trials around the world for months. The chip plot is not a CAD image mock-up of what a chip might look like, it’s a chip plot. Of the actual chip. That exists. To pretend that this photo is some sort of hoax is in UFO conspiracy territory.

      The chips and modules and SDKs and BTSs and working networks in North America and Europe will be available in the first half of 2014. Look again at the video – the first generation modules and BTSs are visible in plain sight. You could, if you were so inclined, get yourself a module, SDK and BTS and have a commercially viable product ready for market next year. It’s not a flying car. It’s not a plan. It’s a publicly declared roadmap from a credible company. But feel free to stay in denial – everyone is entitled to an opinion – while others are placing their orders for development kits.

  23. Galane says:

    With that range there’s lots of uses these could be put to, especially in agriculture. Fruit orchards and vineyards would love to have soil moisture sensors on electrolytic probes poked into the ground by every tree or vine – and have that data automatically adjust every sprinkler head to ensure the entire crop gets evenly watered.

    The same thing would work for row and field crops. Poke a grid of sensors into the ground and as the pivot sprinkler swings around, the polar coordinate control system varies the amount of water sprayed.

    Have the sensor housings a bright color and shaped with a handle so that before harvest they’re easy to collect.

  24. Al says:

    The Weightless SIG is proposing to negotiate a discount and/or kit deal on modules and SDKs (including Arduino and Raspberry Pi based boards) from Member companies. Discounts for Makers on BTSs are also possible. No details have been made available to the SIG at this time but if Maker/Hacker communities register their interest this will be arranged. Observers (no cost, no license Weightless SIG membership) will be notified via email of availability and discounts during Q1/Q2 2014.

  25. LeSigh says:

    If you think this is great, wait til you see this swampland in New Orleans I’ve got for sale.

    It comes with access to free electricity from the local perpetual energy farm that runs on smurfberries and farts..

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s