Feeling interest for a WiFi shield

Hi all,

I'm currently in the process of doing some marketing research on the viability of a 802.11b WiFi Arduino shield. It would be a plug-n-play solution to an Arduino Duemilanove or Diecimila. I'm trying to get feedback and gauge interest in such a part, so any help would be greatly appreciated!

I currently have a working prototype shield that employs 802.11b, up to 2 mbps, works in both infrastructure and adhoc mode, and has a relatively decent range without the need for an external whip antenna. A lot of work has gone into the wireless module to make it low power, so it would be a great fit to a battery powered Arduino project.

There would be some caveats to using the shield and driver code as-is in the beginning, however. First off, anything less than an ATMEGA168 (16K flash, 1K SRAM) unfortunately won't cut it due to the size of the driver and small TCP/IP stack sitting on top. We are projecting about 5-6K or space for user code in the ATMEGA168. We might be able to shoehorn it into a smaller device, but it wouldn't leave much room, if any, for user content. ATMEGA328s would be ideal due to the doubling of space. Packets would also need to be fairly small due to the restriction of SRAM space. We think you should still be able to do small things like blink remote leds, get temperature sensor readings and the like, but you definitely won't be able to stream 720p video content (maybe in the next revision! :slight_smile: ). I've been told by my WiFi module supplier that things will be changing later this year to remove the "small" packet requirement. When this happens, we expect to be able to move full-sized packets over the air.
Initially, we are hoping to sell the WiFi shield in the range of $50-$80, intermixed with all the ZigBee and Bluetooth modules that are out there. The price really depends on you! If there is enough interest, even with the caveats I mentioned, it would help me tremendously to put pressure on my supplier to give me better pricing.

So please let me know if you are interested, and give me any feedback you may have. If there are any questions, I'll try and answer those as best I can in this thread. If all goes well, I'm hoping to have a solution in the market sometime in April.

Thanks!
AsyncLabs

wow... that's exciting. I was asking this couple days back!
will it be a stand alone web server type, like the ethernet
shield? the shield is using UART to talk to arduino? thanks.

Id love to see something like this.

The only problem I could see would be security.
If I was to hook it into my own home network alongside the family pc's and server Id have to expect it to support full wpa security. Or it wouldnt be able to get on my network. To lower my security would be a bad idea.

If I was to have an application that didnt require internet access and was just a closed loop you could get away with no security but theres still the risk of malicious access.

Would it be possible to implement full security in the space available on say a 328p?
and maybe have a light library without security?

Gordon

I would be interested if it cots less than this one
WiFly 802.11b Serial Module - Roving Networks - WRL-08869 - SparkFun Electronics and does more :slight_smile:

My only requirements would be to talk serial, to support encryption and to be plugable into arduino (a shield basically).

Here are some answers:

The shield will support security (WEP, WPA, WPA2) with both TKIP and AES. This will be implemented on the ATMEGA168, so there shouldn't be too much concern there. We are very close to getting everything to work in the 168 (with just the limitation of small packets). The 328 merely opens us up to larger packets. Basically, all the existing Arduinos out there should be able to get secure WiFi access with what they have, and they'll be able to transmit larger packets if they choose to upgrade to a 328 chip.

This will be a full Arduino shield, with the same form factor as the Arduino board. Basically you'll just plug it in on top like any other shield. It should be a plug n play solution (no having to make special connectors or do any additional wire making).

The shield talks SPI to the Arduino (pins 13,12,11,10) and uses INT0 for interrupt processing. The shield will also have a reset pin brought up to make it easier to reset the system. I'll try and take a picture of my early engineering prototype and post it later.

From looking at the datasheet for the WiFly module, I think we'll be able to go toe-to-toe with it. Smiley I believe that our big selling point is that we are going to be a shield, with all the correct dimensions to just plug in right away. It looks like you'll still need to do some work to get the WiFly going (not hard for most guys, but still something that needs to be done).

Thanks for the interest!
AsyncLabs

Getting the security sorted is impressive.
Wont the keys take up a lot of ram?

Ill have one to play with.
I was looking at the wall wart type network adaptors that run through the mains electricity supply. when I saw how much a repeater or wireless access point would cost.
Your price range seems to be on a par with them.
Will you be shipping to the UK?

Gordon

I'm about halfway done with a Matchport Shield, so that'll be one more option available.

I have a working prototype of a shield for the ConnectOne iWifi. Similar to MatchPort and WiFly modules, it does WEP, WPA, SMTP, HTTP, DHCP, etc. etc. Has a small web server on board, and a "SerialNET" mode which just forwards serial data to TCP. $60 from Mouser, but you'd have to make the shield yourself from an Eagle file and several SMD components.

The host driver actually will only store the keys. There is a supplicant on the module which will take care of the security.

It's good to see that options are popping up (even though one of them is by me!). I needed a Wifi shield a while ago and was honestly surprised that it didn't exist.

I am thinking add wifi to homeeasy remote controls...
I want one when it is available. please keep us updated.

Sign me up also as someone who's interested.

Here's my wish list:

  • wpa etc.. as others have said this is a requirement for most networks
  • TCP/IP software interface.. Ideally something that just extends the existing Ethernet software interface. Basically if I can open a port and send things down and read them serially then I am a happy camper.
  • ability to hard reset (or turn off entirely) the device via a digital pin. I think the vast majority of projects don't need a constantly on connection, rather they either poll now and then or even more often only make a request when something changes. Having the board be off during those idle periods would be nice from a power point of view. Also, being able to do a hard reset is always handy.
  • cost. If you are talking ~$50 then sign me up, that's an awesome price point. Definitely can't be over $100.
  • obviously shield format is always nice.
  • 802.11g would be REALLY nice, since the entire network has to back down to b once a b device connects, and most people have g in their homes these days. I could be ok not having this if we have the ability to turn it off when its not actively sending.

That's my list, looking forward to it!

Sorry I haven't been on the thread in a while. It was a busy weekend! :slight_smile: Thanks for all the great input. Here are some thoughts I have on what my solution will have:

  • definitely support for wep and wpa encryption, so it should work seamlessly on all existing networks
  • we will definitely have some type of ethernet interface. We are putting some more thought into making the interface as easy as possible.
  • hard reset via a push button will be on the shield. This will reset both the arduino and the wifi shield. Alternatively there will be a method exposed that will do a shield reset from code.
  • we have a couple of power savings techniques in mind. We can probably support low power polling and a form of standby/disconnect for times you want to maximize battery longevity.
  • price will definitely be closer to $50 than $100. I'm trying my hardest to work my suppliers down, and hope that we'll be at the very bottom of the $50-$100 range. :slight_smile:
  • definitely will be a true shield, with the same form factor (even the fancy cuts on the side opposite the USB port). Our shield will only cover the pin portion of the arduino.
  • for now, it'll be b only. I don't have the power to change that. From a compatibility standpoint, I can see the concern of having to mix b and g together on the network. However, from a pure data standpoint, most applications of our shield won't come near the maximum throughput of even b, so there shouldn't be any concern regarding performance.

This is a rough idea of what is coming down the pipe.

Note that this is not final and is actually our engineering prototype. Things will be slightly different on our first production board. We are hoping to make a small run of these boards, get them out to people, get good feedback, and make the next revision of the board even better.

I hope to have a video up in a few days showing a demo of at least a simple web page being served that can trigger some leds to turn on and off. We have some other interesting demos we are working on as well, so you can stay tuned for that. :sunglasses:

Thanks for all the help and feedback!

AsyncLabs

Nice work.
Will you be extending the pins through to female on top so shields can be stacked?

Gordon

good job. Do you have a website or email list that I can pre-order
or get notified when it is available? thank you.

I would be extremely interested in a WiFi shield for the Arduino. I've been thinking about controlling the Arduino with an iPhone/iPod Touch both with commands and perhaps scripting.

Maybe I'm crazy, but the idea of spending 2 to 8 times (Arduino at 34, raw parts cost $10) as much $ for the shield as for the base makes me makes me think that any such application would be better served by a much more powerful and slightly more expensive microcontroller. Of course makershed is selling an OLED shield for $150, so I guess someone is buying the stuff.

I think that you will have tremendous pricing pressure; I think that what's cool about the Arduino is that at 30 bucks it can be an "impulse buy" for a lot of working engineers, and is also an affordable uC in educational settings.

Even XBee is pretty expensive once you buy both the module and the shield...

What I'd love to see is an inexpensive solution, less than $25. Just a guess but maybe using the same technology as radio TV remote controls -- those are probably produced pretty cheaply. It probably wouldn't be standards compliant so then also sell an "ethernet dongle" that converts the radio signal (coming from multiple arduinos) to IP.

About the price, I agree with vxir - it has to be reasonable for more
people to adapt it. If it costs you $10, then selling it less than $30 will
make me happy.

But if you look at ethernet shield, it is sold on sparkfun for $45, and
you could guess material cost be around $10, so sparkfun is too
agressive on the pricing ? I don't think so - it all depends on volume.
If sparkfun sells million units of this, you may even see price less than 10.
I don't need Greenspan to tell me this. LOL

I don't want to judge the price of wifi shield, market will tell.
In the other way, I am happy to see wifi shield starts to appear -
even if someone makes a shield using a $69 wifly. Our community
needs more cool staff.

I would be interested in at least one, how many of the Arduino pins are still open for use?

Dale

I agree with everybody that price is a concern, and believe me, it's heavily driving my design decisions as well as my profit margins. I want to do this project so badly that I'm willing to sacrifice the amount of money I can make just to get it out to the community.

Unfortunately, since we are only talking low volume sales, it's hard for me to get the rock bottom pricing for the components. I'm sure that if I went to a supplier with a million unit order every year for the next 3 years, they would give me a really, really good price. Trying to only purchase one or two thousand components, while still significant to me and my severely depleted savings account, is not significant to the supplier, and thus, the price isn't that great.

There are a lot of hidden costs that go into a lot of wireless devices. You have to pay money to have your device certified by the appropriate regulatory agency in multiple regions of the world. With regard to WiFi, you pay additional money to have your device WiFi certified. All these costs get included in the cost of the ASICs and modules, and are then passed down to the regular guys like you and me. Some of the bigger companies in WiFi (Atheros, Broadcom, Marvell) don't even want to talk to you or show you the spec unless you pay big money up front and are able to commit to have huge purchase orders. So that only leaves the smaller guys as being "Arduino friendly" suppliers.

Anyway, my point is not to create a sob story. I'm willing to take a hit and get my shield out for little to no gain on my end. I just hope that it's helpful to others, and that my sales funnel will turn from a drip, to a trickle, to a gushing stream later. :slight_smile: If that happens, I can guarantee there will be more useful wireless products coming out at very competitive prices.

As far as my shield, I'm hoping to have some demo videos created this week, along with more information on the operation and specifications of what you can do. I'm hoping to have the shield ready for general availability by the middle of April. And I apologize in advance if I don't get it done this week. Currently, I'm CEO, CFO, CTO, VP of Engineering, head of IT, and individual contributor for this venture, not to mention my normal full time engineering job, as well as being dad to my 18 month old running around my house at home. So bear with me and stay tuned... :sunglasses: