Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 


Forgot your password?
Close
typodupeerror
×
The Internet Communications Network Software Wireless Networking Hardware

Kickstarter Bets On 'Wired' Arduino-Compatible IoT Platform 138

Posted by BeauHD from the always-connected dept.
L-One-L-One writes: Most IoT home projects today are based on Wi-Fi, Bluetooth, Zigbee, and friends. But this is not always the ideal solution: you end up swapping batteries frequently, which becomes annoying quite quickly. You also have to deal with signal strength issues and interferences. To address this problem, a new Kickstarter campaign called NoCAN is proposing an Arduino-compatible internet-of-things platform based on wired connections that combine networking and power in one cable. The platform uses a set of cheap Arduino-compatible nodes controlled through a Raspberry Pi. The network uses CAN-bus and offers a publish/subscribe mechanism like MQTT and over-the-network firmware updates. It can also be controlled by a smartphone or tablet. Even with such features, can it succeed in going against the all-wireless trend? We'll know in a few weeks.
This discussion has been archived. No new comments can be posted.

Kickstarter Bets On 'Wired' Arduino-Compatible IoT Platform More

Kickstarter Bets On 'Wired' Arduino-Compatible IoT Platform

Comments Filter:

  • Slashvertisement brought to you by.. (Score:5, Informative)

    by viperidaenz ( 2515578 ) on Tuesday June 19, 2018 @03:04AM (#56807784)

    Some Kickstarter campaign marketing company!

    • Re: (Score:1)

      by Anonymous Coward

      Who gives a shit. All I wanna know is...Will it Blend?

    • Re: Slashvertisement brought to you by.. (Score:5, Insightful)

      by Zero__Kelvin ( 151819 ) on Tuesday June 19, 2018 @06:15AM (#56808194) Homepage
      This is 100% in the spirit of the original Slashdot I assure you. The fact that you have no interest in such things suggest that you might prefer other news sources.

      • What makes it worse is the title implies that Kickstarter is backing this thing, not a Kickstarter campaign.
        Campaigns don't bet on anything. To bet on something you need to use your own money. The point of a crowd funding campaign is to convince others to bet on your idea.

    • Doesn't the 1-wire bus or the SPI bus already do all this already? Arduino and Raspi have these libraries. So is this just adding a convenient molex connector?

      • Re: (Score:1)

        by Anonymous Coward

        Not exactly; this is also using CAN instead of naked SPI which shifts the basic signalling from raw high/low to a mixture of 0V, 2.5V, and 5V (DC offset differential signalling) so it's a lot more resilient against signal loss issues.

        The separate networking chip they're using keeps the individual CAN nodes easy to re-flash without fishing them out of wherever they're installed as well, so you can literally develop something in place like a backyard lighting system instead of having to build it on a workbenc

        • thanks. though I don't understand what's no robust about 1 wire line levels. I agree that having having a separate network address has potential resiliency advantages. but 1-wire chips already have that too.

          • differential signalling is much more robust in terms of interference suppression.
            1 wire combats this by using slow speeds.

        • Sounds great from a security point of view.
          All you need to do to gain access to a network is tap in to a couple of wires anywhere on a property and reflash one of the devices.

      • CAN is a lot more robust and tolerant of noise. It also allows longer cable runs. IIRC SPI is meant for short distance only.

      • SPI, no, 1 wire, yes.
        1-wire can go up to 500m, depending on how many devices are on the wire. It doesn't provide much power though and it's extremely slow. You're lucky if you get ~10kbits shared between all slave nodes.

    • CAN is just horrible. It was designed for high noise environments, but get it slightly wrong and it is a real pain to diagnose. Interoperability can be touchy.
      They include *2* 32bit arm cores on the end nodes why? I am guessing they are using some off the shelf code on one of them for the networking, but come on, no real device would function this way.
      Their prices are rather high for the level of functionality.
      And they are asking for a very small amount of money in total - looks much more like someones pers

      • They chose to use the SAMD21 as the main processor for Arduino compatibility - it's the same processor that is used in the Arduino M0. But it doesn't support CAN, so they put in a cheap STM processor to act as a dedicated CAN controller. It's rather like how the Arduino Uno uses a second ATMega chip as a USB controller - because it's cheaper than the FTDI chip that they used in previous Arduino products.

  • CAN-bus is patented (Score:4, Insightful)

    by lkcl ( 517947 ) <lkcl@lkcl.net> on Tuesday June 19, 2018 @03:25AM (#56807828) Homepage

    question: why is this company seeking funding based exclusively round a *patented* interface? https://en.wikipedia.org/wiki/... [wikipedia.org]

    • Indeed, why not use a simple open protocol on top of a physical RS-485 interface. Cheap and versatile.

      • Re: (Score:1)

        by Anonymous Coward

        because CAN is way more safe , the collision avoidance is built into the controller and is less noise sensitive.

        • because CAN is way more safe , the collision avoidance is built into the controller and is less noise sensitive

          Just because the collision avoidance is done in the controller doesn't make it more safe. It only makes it better suited for hard real time applications, because the highest priority devices have guaranteed access to the bus. The price you pay is annoyingly short packets. If your software is good, RS-485 is just as safe, a bit less real-time, but more flexible.

          The high impedance during the recessive bit makes CAN more noise sensitive.

        • because CAN is way more safe

          Shit! We better get right on ripping RS-485 out of oil refineries, chemical plants, and other hazardous industries around the world then!

          Facetousness asside, if we can happily run RS-485 at speed with many multi-drop components along several hundred meters running under HV power lines, I think our Internet of Shit devices will be just fine.

          • Re: (Score:3)

            by OzPeter ( 195038 )

            because CAN is way more safe

            Shit! We better get right on ripping RS-485 out of oil refineries, chemical plants, and other hazardous industries around the world then!

            Facetousness asside, if we can happily run RS-485 at speed with many multi-drop components along several hundred meters running under HV power lines, I think our Internet of Shit devices will be just fine.

            While raw RS-485 is fine, many of the container port cranes I am involved in building actually run CANbus for particular I/O runs.

          • Re: (Score:2)

            by MobyDisk ( 75490 )

            Shit! We better get right on ripping RS-485 out of oil refineries, chemical plants, and other hazardous industries around the world then!

            In this case, engineers are testing and qualifying each device. It's not like an end-user just buys an RS485 device and plugs it into the bus. Although CAN suffers from the same problem. Compare them to USB, where everything is sufficiently robust and standardized that you can buy an off-the-shelf USB device and connect it. Although USB doesn't support the cable length we would need for home wiring.

            • In this case, engineers are testing and qualifying each device. It's not like an end-user just buys an RS485 device and plugs it into the bus.

              Oh wow, have you got an over-inflated view of engineering at these facilities! No it's quite literally buying off the shelf parts and plugging it into a bus and hoping for the best. There's a special place in hell for those people who leave it up to the commissioning technician to identify a common set of speeds, parities, and other settings that devices support. Bonus points for using something like modbus and from a vendor which used a non-standard way of storing data.

              On a consumer level these devices wou

      • Re: (Score:2)

        by AmiMoJo ( 196126 )

        What kind of IoT device would benefit from RS-485 over, say, WiFi or low power long range radio link?

        I'm struggling to see any good applications.

        • Serial makes sense because sometimes you don't want a battery nor a mains connection. For sensors and the like, of course. PoE doesn't daisy chain, serial does.

        • Re: (Score:3)

          by AHuxley ( 892839 )
          Security as in anything that can be jammed on consumer wifi. So wired is getting interesting again for low cost security products.

    • Re: (Score:2, Informative)

      by Anonymous Coward

      It was developed in the 1980s. The patents have expired (except on the new CAN FD parts of the system).

      Also, you realise that Bosch Automotive is a components supply company? From its inception, CAN was always available under quite reasonable licensing terms, as Bosch had an interest in spreading the technology throughout the automotive world.

    • CAN Patents are expired (Score:1)

      by Anonymous Coward

      From the link you gave, the patents are expired.

      I don't know about CAN-FD, but this project does not use CAN-FD.

      Or your meant trademarked? You can call it ISO 11898 instead of CAN...

  • Power and open interface its called ethernet (Score:4, Insightful)

    by johnjones ( 14274 ) on Tuesday June 19, 2018 @03:48AM (#56807892) Homepage Journal

    Use ethernet or you could go with a wireless standard like Wi-Fi for high bandwidth or Bluetooth for low power local

    dont give these fools any money...

    • Comment removed based on user account deletion

      • It's usually also easier to find an outlet for a wall-wart than installing network sockets everywhere.

      • Re: (Score:3)

        by maxrate ( 886773 )
        An ethernet network is generally star-topology - everything has to go back to switch or switches. A CAN bus doesn't need all the connections to come back to a central location on an individual/device basis. This saves on having to install 'home run' wiring back to a central point per device (or group of devices). This could save a lot of ethernet ports too - saving space/power in a central location. If someone created a similar looking device with two ethernet ports per device (to support a non star-top

    • Re: (Score:1)

      by Anonymous Coward

      The problem with ethernet is it's WAY too high speed for most uses, and more expensive than CAN or similar low-speed bus interfaces. This avoids the complexities of rolling your own network for transmitting between half a dozen or more items, and is meant for things like roll-your-own lighting control systems around your house or other larger-scale projects where permanently installing cabling (to specifically get away from the interference issues of wifi/bluetooth) for them makes more sense.

      Also it's WAY e

      • Agreed. In addition, this solution is a bus which greatly simplifies wiring. Wiring your home with Cat5/RJ45 on the scale demanded by typical home automation installations - with potentially dozens of devices per room - is going to be a huge pain in the rear end. Running a couple of 4 wire buses around the house (1, maybe 2 per room) is going to be much easier.

        Most of the stuff in my house communicates by Z-Wave, and reliability, range or interference never have been issues. Battery life on these dev

      • Re: (Score:2)

        by jabuzz ( 182671 )

        Just because Ethernet can do 10Gbps over Cat6a does not mean you have to. If you want to run it at 10Mbps that is just fine and there are plenty of switches available that still support it.

        You could also use 802.3bw if you wanted, with the advantage of it actually being still ethernet but only needing a single pair.

      • Comment removed based on user account deletion

    • Which requires additional equipment, and instead of running things in a chain you have to run them to the router. This is better in that say you have 2-3 nodes in a room, you run 1 wire from the master Pi to the room, then run wire between the 2-3 devices. With PoE you're running 2-3 times as much wire.

      • Re: (Score:1)

        by Anonymous Coward

        You can totally hook multiple ethernet devices on the same cable, or did you forget about how splitters and hubs worked?

        • "splitters and hubs" extra equipment. as I STATED, or did you NOT READ? Just what I want hubs all over the house...

          • Re: (Score:1)

            by Anonymous Coward

            You know what splitters and hubs are physically? Hooking some wires together... So it takes the same amount of effort to daisy chained ethernet as for the bus you propose, with ethernet wiring and connectors being a lot cheaper. Please, learn about the physical implementation of these technologies first.

    • Re: (Score:2)

      by PPH ( 736903 )

      Does CANbus have an infrastructure cabling standard? I've got Cat5 pulled to RJ45s all over my house. Am I going to have to crawl around in the attic and pull more wire?

      Ethernet works. I just plug stuff in and my router provisions it with an address. It would be great if CANbus used the same cabling. And someone were to develop a smart hub that would sense the type of connected device and switch it to the appropriate bus. Although I suspect that CANbus might not work on a star topology with long runs.

  • C'mon Slashdot! You can do better than this!

  • *two* microcontrollers per board?!? (Score:4, Interesting)

    by ecloud ( 3022 ) on Tuesday June 19, 2018 @04:52AM (#56808064) Homepage Journal

    STM32F0 and SAMD21... so it's not cheap and can't be.

    Is the connector standardized for CAN? Otherwise they could have picked one that's easy to install by crimping, like RJ11. We've already been there with the Dallas one-wire networks: either use RJ11 to have power, power ground, signal and signal ground, or RJ45 because those connectors are more readily available and some extra pins are there just in case. Or maybe use an audio connector, for convenience and robustness, although those are more trouble to make up your own cables.

    But there are other standards for a reliable low-speed low-wire-count low-compute-power network. But differential signaling is a must, and higher voltages help to make it more robust too.

    A worthwhile next step would be to get an open core design for one of these incorporated into a next-gen Risc-V based microcontroller. Then all the makers could get behind it, just to support the open-IP ecosystem.

    Remember when RepRaps used RS485 between components? (e.g. https://reprap.org/wiki/Extrud... [reprap.org]) And there have been smart stepper motors. I kindof thought that idea was going to take off, early on, but most seem to have decided it's cheaper to centralize the logic and the stepper drivers on one board. But that doesn't scale to larger machines. If CAN has an advantage over RS485 for that, it might make some sense; but I still think one micro ought to be enough to implement it; and if it's not, then CAN is probably the wrong choice.

    Wireless is popular, but every device needs power so nothing can really be disconnected for the long term, unless it runs from solar power. (Batteries either have to be plugged in to recharge, or else they are environmentally unsustainable. Or both.) And there is the ongoing suspicion that RF exposure might cause health problems too. Whenever that risk finally hits the majority's radar, which technology is going to be in position to be the next contender? LiFi could be fairly easy I think.

    I had an idea years ago to incorporate optical fiber into every power cable and every power outlet (simply standardize the position on the plug, relative to the other 3 prongs, assuming a choice of fiber technology such that precise alignment isn't necessary), so that when you plug anything in, you get networking at the same time. But that's a chicken-egg problem.

    Alternatively, find a way to make one of the powerline networking standards cheaper. We can't get away from in-wall wiring to power stuff; so, one way or another, the network and the power wiring ought to be combined, IMO.

  • Anything else, like adding wires in the walls is nonsense.

  • Hard wired and behind a firewall is the ONLY way (Score:4, Insightful)

    by geekprime ( 969454 ) on Tuesday June 19, 2018 @05:09AM (#56808098)

    Hard wired and behind a firewall is the ONLY way to use any device that interfaces to my real life.
    Call me a luddite all you wish, but you can't fuck with my locks, lights, freezer or whatever else I might care to control.

    • Hardwired does provide improved security, but if done correctly wireless can be almost as secure.

      • Re: (Score:2)

        by OzPeter ( 195038 )

        Hardwired does provide improved security, but if done correctly wireless can be almost as secure.

        That depends on how you define secure. With one powerful transmitter you can execute DOS attack on a wireless network fairly easily.

      • Re: (Score:2)

        by AHuxley ( 892839 )
        Re done correctly wireless can be almost as secure.
        Not if the bad people jam all consumer wifi. That is very bad for security that will use wifi.

    • All about trade offs, and I don't think most IoT is worth the trade off. Someone is almost always home at my house. I can get up and switch on/off lights and other stuff in my house rather than doing it from my couch with my smartphone. I'm willing to pay people to come into my house to do things, except for the rare more expensive item like a replacement AC, so a physical key is all I've ever needed to get in and out of my house. I also pretty much never forget to turn off the lights or similar things

    • That would be my interest as well— wireless is just subject to too many unknowns even with a fairly robust design. I also cringe when people talk about powerline— I expect 24VDC to start taking over in the next 10 years for both commercial and residential power, with a power supply on a room-level basis. Power consumption per device simply no longer warrants 120VAC (much less 220VAC). The lower energy consumption, ability to provide backup power, and lower installation cost/ease of modificatio

    • Comment removed based on user account deletion
  • Why didn't they call it the CANduino? Because it's already taken...
  • Regardless whether this is a slashvertisement or enthusiast submission, it is quite an interesting product - apparently you can make 300m / 1000 feet long networks with it, using the one cable that transports both power and data. I find it quite a cool solution for makers.

    • Good old RS485, two wires offering mutliple drops, can do way more than 300m. When pulling wires anyway it's trivial to add power wires, so no problem there.

      The main problem of bringing power that far is the low voltage so relatively high currents and with it high cable losses.

  • CANbus was developed for the automotive environment and it has some worthwhile features there, but for a flexible network where you want some amount of bandwidth? No way.

    In the automation world we went through a decade or more of using DeviceNET, which is just a proprietary version of CANbus. Yes it's nice that power goes through the same cable, but you have to set your throughput at 125, 250, or 500 kbit/sec based on cable length and you have to add terminating resistors at both ends. There's only one m

  • Over the past year, I started a few projects and tested several microcontrollers.

    There are already several alternatives that are very capable. There is no point in using Arduino anymore with these options available.

    First, you have the STM32 Blue Pill [slashdot.org] and its cousin, the Black Pill. Each sells for less than $3 on eBay. They are 72MHz ARM Cortex M3, with 20KB of RAM, and 128KB of Flash storage. I am using it to make a Goto telescope controller with USB and WiFi based on OnStep, and ported it to the STM32 plat

  • It's about time the updated X-10... I installed an X-10 system in my bedroom back in the late 1980s. All wired, no batteries to change (except the backup battery in that brown faux-wood alarm clock controller thingy I picked up at Radio Shack), and it worked wonderfully as long as there wasn't a lightning storm that sent surges through the power lines. Then my TV would turn on, my lights would all come on...

Slashdot Top Deals

One man's constant is another man's variable. -- A.J. Perlis

Close