Build Your Own BBQ Controller (w/Wifi)

[jdbower]

OK, I give up. I've ordered an Arduino starter kit to play with. I want to see what the software looks like and if I can get an HTTP (HTTPS?) and SSH server running on it.

One of my contacts may be able to build us a custom board once we've cobbled together some prototypes. Most of his experience is in FPGA design and his focus is on high bit-rate telecom systems rather than price so it's not a slam dunk. But if we end up with something the size of a dorm fridge I wouldn't worry too much, we'll fix it in the respin

[jdbower]

I'm actually wondering if a Gumstix-like solution would be a decent choice. Reading a bit about the Arduino it seems a little limited from a management interface perspective - I'll know more once I play with it a bit. What if we had an Arduino as a dumb microcontroller paired with a Gumstix system-on-a-chip providing the Ethernet connectivity and management interface? It does bring the cost up a lot (without volume, I think a WiFi-enabled Gumstix is $200 - there may be cheaper competitors), but it's my impression that the Arduino Uno probably would need custom software to access it which would mean trying to develop cross-platform code (Qt, perhaps? Problem is the future of Qt is in the air with Nokia's troubles...). With a GumStix we could have a web-server and SSH/Telnet server built-in.
http://www.gumstix.com/

[Drunk_J]

Cool, glad you grabbed one to play with, as they are a fun little toy. You can grab their IDE and start playing immediately. Not sure what more is needed with the management software, as you can already host a web server that can control the device and display basic information (just need to add a wifi or ethernet shield for that). Another interesting thing, I had forgot there is another program that can be used with the Arduino called Processing. It was designed for interfacing with Arduino and displaying data....such as a nice temp graph. The coffee roasting guys use it along side their arduino programs to get a nice temp graph of their roasts.

http://www.arduino.cc/playground/Interfacing/Processing

I agree the Uno is limited, but what is needed for a BBQ controller? As long as it has enough inputs/outputs to do all the functions we need. Would be much nicer with an embedded Wifi controller and 1-wire (if we go that route).

Another thought is something suggested to me by another forum member (Dan), you could just buy a Tini board from the same people as Stoker. I believe they were in the $100 range. Not sure if any have the Wifi built into any boards, but that would be easy enough to check. Since they also built the 1-Wire chips, stands to reason interface with them should be built in. Will look into them more when I get the chance.

FYI, not had a chance to do this since I have been working nights all weekend, but next couple of days off I will break open the Stoker probes and try to identify the chips used in the connectors. Also I want to get a better understanding of power distribution via their jacks.

-=J

[Drunk_J]

Here is a cheaper Wifi shield for the Uno provided you could live with 802.11b.
http://www.cutedigi.com/product_info.ph ... ts_id=4564

[jdbower]

Playing around a bit with the IDE I do see a webserver example which is promising. Being security conscious I'd much rather have a secure protocol like SSH or HTTPS which I don't think it can handle, but there are options I'm considering - especially if we go the stoker_mon route where I don't mind a mostly insecure connection from PC to controller because I can add security on the webserver. If I ever get any spare time I'll try to slap together a document or two.

Regarding WiFi, my only real request is WPA2 support. So far it seems like most routers support B, there is a risk that it could be dropped in the future but until then I can't see any issues with network speed. And hopefully there will be a drop-in replacement module (perhaps even cheaper than today's B module).

[Drunk_J]

Great. BTW, only reason I wasn't to happy with the B standard for this project is range. My grill is at the fringe of range with G, which is why I had to use a repeater with the Stoker. FYI, if you really wanted to keep down costs and save on space, we could come up with an integrated shield to handle 1-Wire communications and Wifi (up to the design challenge Syz?).
http://www.mouser.com/ProductDetail/Mic ... mHGDFlC4OI
http://www.maxim-ic.com/datasheet/index.mvp/id/2834
http://sheepdogguides.com/arduino/asw1onew1.htm
http://www.arduino.cc/playground/Learning/OneWire

There is potentially enough room on this Wifi Shield to add the 1-wire controller. I would be willing to pay the $34 more to upgrade from B to G and have the chip already soldered onto a shield. Could even put a tiny riser board on it if necessary.
http://www.sparkfun.com/products/9954

Also, here is the chip used on the Stoker temp probes.....god my head hurts from trying to read those tiny numbers. Interesting they used a LI battery monitor chip. The DS2760A on my probes are no longer in production. The new number is DS2762. Price - $2.48
http://www.maxim-ic.com/datasheet/index.mvp/id/3950

Chip on fan is a switch (DS2405) which is no longer in production. The new chip is DS2413. Price - $1.02
http://www.maxim-ic.com/datasheet/index ... /4588/t/al

Looks like the same two pins are used on the power as the temp probes. So the switch chip just decides whether or not to pass on the power through the chip.

-=J

[Drunk_J]

Some pics of the Stoker Board for anyone interested but too nervous to take apart.

-=J

[Syzygies]

Someone needs to ask this, even if there's a good (negative) answer: With mini-ITX motherboard/cpu/vga combos selling for $45-$60 on Newegg (add memory, power), why spent countless hours trying to learn new languages and debug the whole sort-of-a-computer-on-a-chip route? Arduino sounds like an awesome olive oil, but I'm not sure it's the computing environment of choice.

If we went this route, the unique contribution for BBQ would be a PCI-e card for the barbecue-specific IO. Perhaps interfacing with some analog discrete components for scale, but as we all know, analog electronics is kindergarten stuff, at least for what we'd need.

Just asking. If I went this route, I'd get to code the controller in Haskell.

Along these lines, aren't some $30 wireless routers running full-blown, open-source Linux? Hmm, memory, power, all the wireless hardware is already there, just a question of patching in some more IO.

[Drunk_J]

Someone needs to ask this, even if there's a good (negative) answer: With mini-ITX motherboard/cpu/vga combos selling for $45-$60 on Newegg (add memory, power), why spent countless hours trying to learn new languages and debug the whole sort-of-a-computer-on-a-chip route? Arduino sounds like an awesome olive oil, but I'm not sure it's the computing environment of choice.

If we went this route, the unique contribution for BBQ would be a PCI-e card for the barbecue-specific IO. Perhaps interfacing with some analog discrete components for scale, but as we all know, analog electronics is kindergarten stuff, at least for what we'd need.

Well for me, not having a computer (laptop) sitting outside by my cooker was the whole reason I started the arduino research for an interface for the CyberQ. I want as small as possible, all encompassing controller. So that is out for me.

Along these lines, aren't some $30 wireless routers running full-blown, open-source Linux? Hmm, memory, power, all the wireless hardware is already there, just a question of patching in some more IO.

I have a WRT54G (old Linksys Router) that runs DDWRT which is an open source firmware based on the Linux kernel. But there is not much processing power to act as a controller and very little memory (as the firmware pretty much fills it up). Don't see how it is possible to add IO in a usable form for our purpose?

-=J

[Drunk_J]

Forgot to add, if you do want something Linux based, there is always the Chumby.
http://www.adafruit.com/index.php?main_ ... cts_id=278
http://www.chumby.com/

-=J

[jdbower]

My hope is that a smaller board will be easier to put into a weatherproof case and would survive more temperature extremes than a full motherboard.

I've been mulling over the connector options, I wonder if we should consider something like automotive weatherproof pigtails instead of 1/4" jacks? Especially if we're going to need identifier chip adapters for existing probes/fans this would let us tape up the 1/4" to weather-proof connector as we see fit. And we'd be able to have a much higher density of them to boot since we wouldn't need surface mount receptacles.

[Drunk_J]

I've been mulling over the connector options, I wonder if we should consider something like automotive weatherproof pigtails instead of 1/4" jacks?

Those plugs are usually solid injection molded...right? How would you use the chip?

Especially if we're going to need identifier chip adapters for existing probes/fans this would let us tape up the 1/4" to weather-proof connector as we see fit. And we'd be able to have a much higher density of them to boot since we wouldn't need surface mount receptacles.

Not sure I understand....maybe I am still half asleep.

Every time I start thinking about the waterproofing, it usually gets really expensive and complex or really cheap and simple. Not a lot of middle ground. On one hand the zip lock bag or plastic glad container is cheap and effective. On the other, some nice waterproof micro jacks and plugs would be great (the chip can be inline in a tiny box or simply expoxied in place). But the cost of those would probably be more than the rest of the project combined.

One thing about this part, everyone can go their own route if they so choose.

-=J

[jdbower]

I'm thinking dirt cheap. My thought on the waterproofing is that you could take the connector I had linked and cut it in half. The bare wires on one end would be connected to the board (soldered directly or via a header). They'd pass through a hole in the enclosure that you could patch up with silicone or a gasket (or both).

The other end of the connection would have the bare wires soldered to a small board with the ID chip which would, in turn, be connected either to the probe directly or (my preference) to a 1/4" female connector. The 1/4" female to automotive connector pigtail should be able to be waterproofed either with an off-the-shelf enclosure or some Plasti-Dip/silicone. Once you plug in the 1/4" jack you can make it waterproof using some stretch tape.

Of course, getting new cheap probes and soldering directly from auto pigtail to ID chip to probe with a bump in the middle for the coupling would be cleaner, but since I've got Stoker probes I may need to salvage them in the future. Obviously if we use the Stoker components it may obviate the need for the ID chip.

[Drunk_J]

Guess for me, not real fond of those automotive connectors. They are fairly big and ugly. Really don't want a bundle of wires poking out of the controller housing either. If I had planned on making my own probes (but wanted to use either my Guru or Stoker probes), then I would want them as absolutely small as possible. But good part on the connectors is you can pretty much use what you want.

-=J

[jdbower]

Guess for me, not real fond of those automotive connectors. They are fairly big and ugly.

Preaching to the choir! They're cheap and accessible but I'm hoping to find something else before we go live.