Wifi Links.

I haven't made any progress on the Wifi project lately. I've been on a pseudo-vacation for the past week.

In case you've landed here from Google with an interest in the MRF24WB0MA Wifi board from Microchip, I've got  few interesting projects I've run across you may find interesting.

"Jeff", trying(unsuccessfully?) to get the board working with an Arduino:
http://jmsarduino.blogspot.com/2012/01/mrf24wb0ma-not-working-yet.html

"Manis B", got a webserver running on an Arduino with the board:
http://theiopage.blogspot.com/2011/10/connecting-mrf24wb0ma-to-arduino.html
(Also, really nice, pragmatic setup with connecting the pads, made me realize my mount was overkill - but fun)

"Blondihacks", did extensive experimentation with the board with an Atmega 168:
http://quinndunki.com/blondihacks/?p=840
(Also, a fantastically done write-up, even though she deems it "a failed experiment")

And last but not least, one for my favorite blogs, Dangerous Prototypes, recently posted about a MCW1001A chip that is a companion to the MRF24WB0 boards that implements the TCP stack for you and communicates via UART:
http://dangerousprototypes.com/2012/04/04/mcw1001a-tcpip-socket-communications-interface-with-gpio/

--P

MRF24WB0MA Wifi board update.

Success!
(For various values of  'success' at least)

I finally have the board physically put together:

Not pretty, but I think it will work.

After drilling the pcb holes for the .1" headers, I am now certain that surface mounting was the way to go for the daughterboard. I never would have been able to get the accuracy needed with my uber cheapo drill press, it actually moves about 0.5mm when you apply too much vertical force to the press, so about half way through the job I realized that slow drilling was the key to better accuracy. Still, I don't think I could pull off drilling holes for 0.05" headers.

To do a surface mount solder, I "pre-soldered" the pads where the board would be sitting:

Microchip Wifi update.

My wifi board project has moved kind of slow lately. When I last left off, before the holidays last year, I had ordered 1.27 mm headers needed to make my planned adapter board to a more reasonable 2.54 mm pitch. I've been working with the awesome KiCid to design the adapter. While KiCad is amazing in what it lets you do for free, the user interface is not the most intuitive (I common phenomenon it seems in software designed be hardware people). It turns out that I think I may forgo the 1.27 header and try to mount the Wifi daughter board surface mount. I have a low-end pcb drill press for drilling through-holes, but I have serious doubts that I would be able to drill with the precision needed.

Anyways, I took the plunge this evening and decided to attempt a toner-transfer to a copper-clad. The result is this:

Try again!

I find myself trying to remember some Edison quotes about failure and perseverance. When I was ironing the thing I was convinced it was going to come out perfectly. I thoroughly washed clad ahead of time, I found the glossiest magazine paper I could find, and I ironed it for so long that I thought I was going to scorch the paper. Yet you can clearly see in the top traces where the steam-holes in my iron where. 

I was doing this late night, my whole family was asleep, so I tip towed through the house, fruitlessly searching for my wife's nail-polish remover, to remove the toner. 

As far as my final plans for what I'm going to actually do with the board, I'm leaning toward trying to make my own device inspired by openPicus. It is a general purpose standalone Wifi/tcp/http enabled I/O board. My only problem is that the pic chips I have are not nearly powerful enough to run the openPicus firmware, and I'm having a hard time finding dip PIC24's with 256k of flash, so I may peruse the code and see if I can use a subset of it.

--P

Tiny Gecko ^h^h^h^h^h^h^h^h^h^h + Microchip Wifi Project update.

I finally got my 1.27mm headers in today. They were coming in from China after all.... I didn't pay a lot of attention when I ordered.

Anyways, to recap: the guys at Farnel hooked me up with this wifi board. I planned to use it with my Tiny Gecko demo board. The Microchip board is heavily geared toward using it with PIC processors, but they release the source code to a tcp/ip stack. My plan was to port the Microchip stack to run on the Gecko, shimming in Gecko-native SPI and interrupt code in place of the PIC code.

While I was waiting for my headers, I figured I'd start working on the port, so that when the headers arrived, I'd be almost ready to connect it to my Gecko. I got about 1/3 of the way into it, and realized I was doing more work than it was worth. I decided to scrap the idea of porting the code. I'm just going to use a PIC chip, and treat it like a high level tcp/ip "black box". I'll use a UART, or maybe SPI port to send and receive socket data, and let the PIC do all the heavy lifting. Hopefully, this means I can mostly use the Microchip code "out of the box", and make a simple serial messaging protocol. This also means I can also make it portable so that it's not just for the Gecko.

I just ordered a few PIC chips and a PICKIT 3. I had always planned to get around to exploring the PIC line of chips, but now I have a good excuse.

--P