Date: Thu, 22 Dec 2005 22:01:29 -0600 (CST)
Subject: radios and website
X-UID: 113
Content-Type: IMAGE/JPEG; NAME="img2473.jpg"
Content-Type: IMAGE/JPEG; NAME="img2474.jpg"

The last week was spent setting up a website at: http://golem5.org/ .

Robot project "mail blog" updates for the last year are placed in a 
historical timeline at: http://golem5.org/robot1/mailblog.html . I find it 
interesting. In my nine years of industry experience, I've never seen a 
true debriefing post-mortem of any project. The messy details of what 
really happened on any given project (not the official schedule) are never 
documented and soon forgotten with time.

The website indicates a documentation and marketing effort for this 
project in parallel with engineering. I feel comfortable doing this as my 
confidence level on the engineering side is much higher at this point.

I tested the radios. They work. Now I see why they are $150 each. The 
radios are by far the easiest devices I've worked with so far. They handle 
all negotiation and together comprise a wireless null-modem cable of 
sorts.

The radio I/O pins can be connected directly to the computer board I/O 
pins. Even though the computer board is limited to sourcing 3 milliamps at 
3.3 volts, there is no problem. This is in contrast to the serial receive 
pin of the AVR microcontroller that draws enough current to pull the 
computer board logic high down to logic low. Buffering is necessary. The 
radio does this for you.

These Maxstream radios are commonly available in 9600 and 19200 RF baud 
rates. This is the speed at which they transmit over the air. I have the 
9600 version, the slower one. The advantage is about -3 dB greater 
selectivity. I think this means 50% higher gain. I'm not sure how much 
more range that gives me.

I've decided to use the radios for what they were designed for, a low 
speed but long range serial communications link. I'm not going to put a 
network stack on top of them. That would give poor performance and is just 
not what the technology is designed for.

The radios will act as a signaling mechanism for remote kill and other 
management functions. Remote kill is the one necessary signal that must 
work. Everything else is optional.

At close range, the laptop base station will network through WiFi. When 
the robot moves out of range, it will become autonomous. The remote kill 
switch will always be an option. But the robot will have to control 
itself.

Here's one more nice thing about the radios. They are easily queried for 
signal strength. So the robot can know when it is moving out of range.

Tomorrow, I'll see if the GPS receivers work and then - as all electronics 
have been verified - I can wrap that part up.