7 ms·

Lots of misinformation in these comments. For those dreaming of building a mesh Internet alternative, consider that these devices are not being designed for TCP

Lots of misinformation in these comments. For those dreaming of building a mesh Internet alternative, consider that these devices are not being designed for TCP/IP. From the looks of the KS page, they are going to speak a mesh protocol but likely something proprietary, perhaps similar to ZigBee, that is designed as a wire-replacement protocol for Arduino-to-Arduino comms. That's a long way from Internet routing. They are offering a WiFi shield but it's backed by the Arduino. Not really something that you want to mesh-route TCP/IP, in my opinion. There are more suitable choices. I'm also a little concerned about selling a device that is so high-powered by default. I'm assuming that they're using 900 MHz here. Is it really necessary for everybody to transmit with enough power to reach 1 km? In the US, there's a FCC rule that says that amateur radio operators use the minimum power necessary when transmitting. I'm not sure if it applies to a device like the Flutter but it probably should.

TaylorAlexander 13y ago
We are definitely designing these to support mesh networks, and though I haven't written it down anywhere yet, we're not going to default the radios to full on, that will either be automatically configured (probably too difficult, but I'll look into it) or just manually specified. I would really love to make it automatic though, and for anything under constant comms they could easily communicate their signal strength in the ack if needed to coordinate that. The mesh protocols are still under development (that's what we need Kickstarter for!). Our wifi + ethernet shield makes it easy to have an always on device that routes traffic for home Flutter networks, but you can also just plug the board into your PC over USB and feed it traffic that way, you can run arbitrary data of any size over the link and all communications are guaranteed (or explicitly timeout), so I think piping TCP/IP over it will be easy. You can also use a raspberry pi and one of our boards to accomplish what the network shield does (for roughly the same price) but the network shield is designed for exactly this purpose, and many people will find value in that. It's tough to pitch an Arduino with lots of highly technical capabilities - Arduino users just want to know "but can i make a wireless beer brew setup?" That's the kind of information we try to speak to on the KS page. We're also still learning what questions people have!
- chrissnell 13y ago
  Thanks for the reply, Taylor. Perhaps you can explain this to me: What niche are you guys trying to fill with this? I guess I'm just not seeing it yet. You talk about "piping TCP/IP" over the data link but how is this superior that using any of of the dozens of other powerful, low-cost, mesh solutions purposely designed for TCP/IP routing? Maybe you're trying to fill the "talk to my kegerator" niche. OK, but how is this superior to the Xbee shield [0]? That shield interfaces with a number of different Xbee products that offer a wide variety of power outputs and ranges. These Xbees do mesh networking out of the box. How does Flutter do it better? Help me understand. :) [0] https://www.sparkfun.com/products/10854 https://www.sparkfun.com/products/10854
  - Florin_Andrei 13y ago
    I don't think XBee is anywhere near that range. Plus, a Flutter is cheaper than an XBee shield + an Arduino. Anyone bulding remote controlled "stuff" will be delighted.
    - TaylorAlexander 13y ago
      Yeah, exactly this. Used to be, you buy an Arduino Fio and an xbee module[1], and for $50 you get max 100m range? Two nodes then cost $100, which is enough for one project and one device to route traffic to mobile, lets say. With Flutter, for $100 you can have 5 nodes, so you can have the beer brew monitoring, control some lights, measure the humidity of your greenhouse, and build a robotic car you can drive with your phone. We're trying to hit commodity pricing for consumer mesh hardware so people can just toss it in whatever they want to make and easily handle the communications they need. And not relying on another radio module manufacturer means we can design the product to be everything we need and nothing we don't, and only pay for board manufacture once, instead of paying for a module (which is a manufactured board) that you then put on your own manufactured board. The radio and cpu should be on the same board and should be very useful (hence open and arduino). Then we can really build a network of things.
      - 13y ago
adunk 13y ago
I can't speak for the creators of Flutter, but I can definitely attest to that TCP/IP with full mesh routing should be doable with the Flutter hardware. We (Thingsquare) have the CC1101 as one of our supported hardware platforms and use it to run full TCP/IPv6 networking over a self-forming mesh. The system is used in a number of commercial products and systems. The chip-level firmware is open source so I guess the creators or backers of the Flutter system might even be able use it pretty much out of the box: http://thingsquare.com/tech/ http://thingsquare.com/tech/ One of the things needed to get full IP routing working is to do channel switching over the sub-GHz connection. Both the FCC and ETSI regulations let you transmit at a higher power and with longer bursts if you switch channels with a regular interval and if you back off if you find other transmissions in your channel. That said, TCP/IP routing over a sub-GHz CC1101 link isn't going to be very fast. It is suitable for automated devices but not a general purpose replacement for WiFi connectivity. For IEEE 802.15.4g-compliance, the raw bit rate is 50 kbit/s which isn't a lot. The CC1101 can be run in faster, non-802.15.4g compliant, modes though.
- stevenrace 13y ago
  Agreed. This isn't going to form the 'alternate Internet' that some people are clamoring for. But there are tons of applications (machine->machine) where a basic serial/ssh is all that's required. We've had similar hardware in the field for 3+ years now, covering ~50 square kilometers in a mesh configuration.
  - TaylorAlexander 13y ago
    Absolutely! There's lots of hardware around, but being able to turn an Arduino project into an emergency beacon or IP link has some benefits too, and $20 is likely cheaper than what's been around. At $20 these are commodity prices, which I don't think we've had for high quality general purpose computing + wireless platforms.
    - stevenrace 13y ago
      It's always good to see cheaper hardware - our early 'proof of concept' hardware were based on an 'Arduino FIO' + 'XBee Pro' which came out to ~$60. Then quickly migrated to a TI MSP430 + CC2500(CC1100) which slashed the price to <$20 all told. We're now on the ARM M3/M4 SoC bandwagon as well. So I understand that angle. I'm sorta curious as to what you plan on using the $80k for. Are you going to farm out PCB fabrication to OSHpark/China/whatever - or get a pick and place machine, CNC mill, and try to do everything inhouse? Anyways, best of luck with the project.
ohazi 13y ago
From the datasheet of the TI radio they're using[0], the maximum programmable output power is +12 dBm, which is about 16 milliwatts. There's no way you're going to see the advertised range with the antenna pictured. For comparison, the wifi hardware in your laptop typically outputs around 100-200 milliwatts. Edit: TI appears to have a "range extender" part (i.e. a power amp) that can kick the output power up to +27 dBm (500 mW) at the band that they're using. I don't see any mention of this part on the kickstarter page, but this is the only plausible way I can see for them to get anywhere near the range advertised. Half a watt seems pretty high to me too, but apparently it's comfortably within FCC limits [1][2], and since ISM devices are designed for consumer use (as opposed to licensed use by a competent radio operator), I don't think the same rules about being clever with output power would still apply here. [0]: http://www.ti.com/lit/ds/symlink/cc1101.pdf http://www.ti.com/lit/ds/symlink/cc1101.pdf [1]: http://www.afar.net/tutorials/fcc-rules/ http://www.afar.net/tutorials/fcc-rules/ [2]: http://www.beagle-ears.com/lars/engineer/wireless/fccrules.htm http://www.beagle-ears.com/lars/engineer/wireless/fccrules.h...
- TaylorAlexander 13y ago
  Hah, thanks for the edit. Yeah we've already gotten the range. I've even checked the output of our boards with a radio expert (google Earl McCune, he's an advisor), and using his many expensive Agilent scopes we've experimentally verified that performance of my design is pretty much exactly on par with the reference designs, taking into account that we are using a custom designed RF balun [1] instead of the discretes they use on the reference design. Our RF Expert confirmed I had made a good choice with that, as the losses are tiny compared to ease of manufacturing it provides. [1] (I think it's this one, I'd have to check the part number, they have a couple similar ones). http://www.johansontechnology.com/datasheets/chipset-specific/0915BM15A0001.pdf http://www.johansontechnology.com/datasheets/chipset-specifi...
  - Dwolb 13y ago
    Are you sure you're able to legally transmit at 1km. Had thought antenna gain + transmit power had to be around 0 dBm, at least for 434MHz. Hadn't seen the rules for 900MHz though.

TaylorAlexander 13y ago

We are definitely designing these to support mesh networks, and though I haven't written it down anywhere yet, we're not going to default the radios to full on, that will either be automatically configured (probably too difficult, but I'll look into it) or just manually specified. I would really love to make it automatic though, and for anything under constant comms they could easily communicate their signal strength in the ack if needed to coordinate that. The mesh protocols are still under development (that's what we need Kickstarter for!). Our wifi + ethernet shield makes it easy to have an always on device that routes traffic for home Flutter networks, but you can also just plug the board into your PC over USB and feed it traffic that way, you can run arbitrary data of any size over the link and all communications are guaranteed (or explicitly timeout), so I think piping TCP/IP over it will be easy. You can also use a raspberry pi and one of our boards to accomplish what the network shield does (for roughly the same price) but the network shield is designed for exactly this purpose, and many people will find value in that. It's tough to pitch an Arduino with lots of highly technical capabilities - Arduino users just want to know "but can i make a wireless beer brew setup?" That's the kind of information we try to speak to on the KS page. We're also still learning what questions people have!

chrissnell 13y ago
Thanks for the reply, Taylor. Perhaps you can explain this to me: What niche are you guys trying to fill with this? I guess I'm just not seeing it yet. You talk about "piping TCP/IP" over the data link but how is this superior that using any of of the dozens of other powerful, low-cost, mesh solutions purposely designed for TCP/IP routing? Maybe you're trying to fill the "talk to my kegerator" niche. OK, but how is this superior to the Xbee shield [0]? That shield interfaces with a number of different Xbee products that offer a wide variety of power outputs and ranges. These Xbees do mesh networking out of the box. How does Flutter do it better? Help me understand. :) [0] https://www.sparkfun.com/products/10854 https://www.sparkfun.com/products/10854
- Florin_Andrei 13y ago
  I don't think XBee is anywhere near that range. Plus, a Flutter is cheaper than an XBee shield + an Arduino. Anyone bulding remote controlled "stuff" will be delighted.
  - TaylorAlexander 13y ago
    Yeah, exactly this. Used to be, you buy an Arduino Fio and an xbee module[1], and for $50 you get max 100m range? Two nodes then cost $100, which is enough for one project and one device to route traffic to mobile, lets say. With Flutter, for $100 you can have 5 nodes, so you can have the beer brew monitoring, control some lights, measure the humidity of your greenhouse, and build a robotic car you can drive with your phone. We're trying to hit commodity pricing for consumer mesh hardware so people can just toss it in whatever they want to make and easily handle the communications they need. And not relying on another radio module manufacturer means we can design the product to be everything we need and nothing we don't, and only pay for board manufacture once, instead of paying for a module (which is a manufactured board) that you then put on your own manufactured board. The radio and cpu should be on the same board and should be very useful (hence open and arduino). Then we can really build a network of things.
    - 13y ago

adunk 13y ago

I can't speak for the creators of Flutter, but I can definitely attest to that TCP/IP with full mesh routing should be doable with the Flutter hardware. We (Thingsquare) have the CC1101 as one of our supported hardware platforms and use it to run full TCP/IPv6 networking over a self-forming mesh. The system is used in a number of commercial products and systems. The chip-level firmware is open source so I guess the creators or backers of the Flutter system might even be able use it pretty much out of the box: http://thingsquare.com/tech/ http://thingsquare.com/tech/ One of the things needed to get full IP routing working is to do channel switching over the sub-GHz connection. Both the FCC and ETSI regulations let you transmit at a higher power and with longer bursts if you switch channels with a regular interval and if you back off if you find other transmissions in your channel. That said, TCP/IP routing over a sub-GHz CC1101 link isn't going to be very fast. It is suitable for automated devices but not a general purpose replacement for WiFi connectivity. For IEEE 802.15.4g-compliance, the raw bit rate is 50 kbit/s which isn't a lot. The CC1101 can be run in faster, non-802.15.4g compliant, modes though.

stevenrace 13y ago
Agreed. This isn't going to form the 'alternate Internet' that some people are clamoring for. But there are tons of applications (machine->machine) where a basic serial/ssh is all that's required. We've had similar hardware in the field for 3+ years now, covering ~50 square kilometers in a mesh configuration.
- TaylorAlexander 13y ago
  Absolutely! There's lots of hardware around, but being able to turn an Arduino project into an emergency beacon or IP link has some benefits too, and $20 is likely cheaper than what's been around. At $20 these are commodity prices, which I don't think we've had for high quality general purpose computing + wireless platforms.
  - stevenrace 13y ago
    It's always good to see cheaper hardware - our early 'proof of concept' hardware were based on an 'Arduino FIO' + 'XBee Pro' which came out to ~$60. Then quickly migrated to a TI MSP430 + CC2500(CC1100) which slashed the price to <$20 all told. We're now on the ARM M3/M4 SoC bandwagon as well. So I understand that angle. I'm sorta curious as to what you plan on using the $80k for. Are you going to farm out PCB fabrication to OSHpark/China/whatever - or get a pick and place machine, CNC mill, and try to do everything inhouse? Anyways, best of luck with the project.

ohazi 13y ago

From the datasheet of the TI radio they're using[0], the maximum programmable output power is +12 dBm, which is about 16 milliwatts. There's no way you're going to see the advertised range with the antenna pictured. For comparison, the wifi hardware in your laptop typically outputs around 100-200 milliwatts. Edit: TI appears to have a "range extender" part (i.e. a power amp) that can kick the output power up to +27 dBm (500 mW) at the band that they're using. I don't see any mention of this part on the kickstarter page, but this is the only plausible way I can see for them to get anywhere near the range advertised. Half a watt seems pretty high to me too, but apparently it's comfortably within FCC limits [1][2], and since ISM devices are designed for consumer use (as opposed to licensed use by a competent radio operator), I don't think the same rules about being clever with output power would still apply here. [0]: http://www.ti.com/lit/ds/symlink/cc1101.pdf http://www.ti.com/lit/ds/symlink/cc1101.pdf [1]: http://www.afar.net/tutorials/fcc-rules/ http://www.afar.net/tutorials/fcc-rules/ [2]: http://www.beagle-ears.com/lars/engineer/wireless/fccrules.htm http://www.beagle-ears.com/lars/engineer/wireless/fccrules.h...

TaylorAlexander 13y ago
Hah, thanks for the edit. Yeah we've already gotten the range. I've even checked the output of our boards with a radio expert (google Earl McCune, he's an advisor), and using his many expensive Agilent scopes we've experimentally verified that performance of my design is pretty much exactly on par with the reference designs, taking into account that we are using a custom designed RF balun [1] instead of the discretes they use on the reference design. Our RF Expert confirmed I had made a good choice with that, as the losses are tiny compared to ease of manufacturing it provides. [1] (I think it's this one, I'd have to check the part number, they have a couple similar ones). http://www.johansontechnology.com/datasheets/chipset-specific/0915BM15A0001.pdf http://www.johansontechnology.com/datasheets/chipset-specifi...
- Dwolb 13y ago
  Are you sure you're able to legally transmit at 1km. Had thought antenna gain + transmit power had to be around 0 dBm, at least for 434MHz. Hadn't seen the rules for 900MHz though.