If you replaced or modified your CM-5...

[Zoid]

Whether your CM-5 died, or just didn't put out, I'd like to hear from those that replaced their CM-5 with something else, or how you modified your CM-5 to better suit your needs.

I haven't even begun to purchase yet, and already I'm seeing controller options that sound inviting, such as the MAVRIC-IB (for its additional RAM) and Gumstix.

--Chris

[JonHylands]

I use both a PC and a gumstix to talk to the bus directly...

- Jon

[Zoid]

Thanks Jon. I've spent hours following the adventures of Brainbot on this site and your blog. He's a neat project. I read forums a lot so that I can avoid the headaches and expensive lessons of others. Even so, it's difficult to know how much of Brainbot's expenses were the result of necessity versus convenience. IIRC you first tried to leverage the Gumstix' processing power, and then demoted it to mail boy. Might you have found another way to achieve 1Mb wireless if the Gumstix wasn't already in your hand? I've done a fair amount of searching, and not found another 1Mb wireless solution, but you're better connected than I am. I can't quite fathom spending $400 on an invisible cable....yet. Anyway, keep up the good work. It's fun reading.

--Chris

[JonHylands]

Chris,

I really wish there was an easier (cheaper) solution for full speed wireless. We almost had something with an ATmega128 and a wifi module, but we just couldn't get the wifi module to talk over SPI.

However, you could easily use the XM-4 ($129) plus the netwifimico-SD card ($99) plus a console-vx ($25) for a total of $255 (plus tax and shipping).

Eventually I would like to have the "brains" on-board, but it will require a much bigger robot than the Bioloid to do it.

- Jon

[roycepipkins]

Jon,

Would the Pico-ITX stuff VIA just came out with work for you?

http://www.via.com.tw/en/initiatives/spearhead/pico-itx/

Regards,
Royce

[JonHylands]

It would work for the simpler version of the brain I'm building right now, but not the full-blown one I will hopefully be building in a couple years.

The full blown one is probably going to require a quad-core board, with 8 GB of RAM and a terabyte of storage. Of course, in two years from now that might be available in pico-itx form...

- Jon

[Zoid]

Jon, I spent last night reading about the Propeller chip, languages, data sheets, boards, and example code and applications. I'm simultaneously amazed and disappointed. The Propeller certainly has its weaknesses, but I think its strengths may be just the ticket to driving bot-side budget wifi. The Propeller strikes me (roughly) as an ATMega that has all its charms (parallel/hardware I2C, PWM, ADC, SPI, etc) stripped away and replaced with numerous speedy processors, each capable of duplicating one or more of those interfaces (in parallel) if you cared to write it. I wonder, if we had the code and components to duplicate a MAVRIC-IB (for example) with a Propeller, how would our budget look, and how much RAM and processing power would remain for application? Not a serious question, just food for thought.
--Chris

[JonHylands]

I'm sure the propeller is a nice chip for doing embedded stuff, but personally it doesn't have any where near the kind of horsepower I require.

Getting wifi cheaply will be doable with an ATmega if we can find a reliable wifi module that has a 1.0 Mbps serial interface.

- Jon

[Robo1]

Hi

If your after a good wifi module have to looked at the wiport. I'm running one at the moment, but haven't tried any faster then 115200. I think the datasheet said something about 920K or 1M can't remember of hand. You cna set it up so that you have 2 serial ports mapped to to com ports on your computer. Plus there easy to use.

Bren

[JonHylands]

The problem is the WiPort maxes out at 921 kbps, which is fundamentally not compatible with a microcontroller set up to talk at 1.0 mbps (different crystal frequencies are required for each).

- Jon