Thursday, September 22, 2011

F3F Timing System Update: from delusion to improvement

I had much delusion with the wireless base grips and their 2.4GHz radio modems. Although the RF modules are meant for serial data communication up to 1.5km, and although the communication was zero lost signal while developing indoors, we had many lost signals as soon as we were set up in the field.
At first, I thought it was eventually due to the processual program. So I improved reactivity even more by implementing interrupts for the base buttons. It didn't solve the problem.
Then it could be because of the modules being held near the ground, so range would be affected too much. I implemented a relay RF modem on a 3m central pole and limited addresssing of each module to strict destinations. It didn't solve the problem either!
I wanted my system to be wireless, but reliable too! I had to grab the facts and get some 100m wire...
In between, I also thought about the story of a colleague who used dead simple walkie-talkies for F3F training this summer. This would at least be a good solution as backup, if my wireless timing system remains useless because of those bloody RF modems. So I went to the local electronic store, checked out different brands of walkie-talkies, and finally grabbed a trio set of those 446MHz 8 channels 10km range (?) radios. For the price, those were not the fanciest. But I soon discovered the talkies I bought had a nice feature not to be found on most of the other brands!
Back home, testing my new radios with my son, I thought about talkies as F3F base switches... Sound detection of the "Roger Beeps"? No way because the Roger Beeps are spit out after the push to talk (PTT) button is released... What we want is a signal detection right a button is pressed. So maybe the "Call" function? Once pressed, the receiving talkies beeps... for as long as 3 seconds:-( And while beeping, no signal can be transmitted. So if the glider passes base A, the judge pushes "Call", the B base talkie could beep longer than the time for the glider to cross base B, and the judge at B cannot send his "Call" before beeping ends !!! No solution at all. Hmmmm....
Light! There's a LED turning on at the receivers units as soon as one talkie emits! Could this be the way? So how to tell the CD timing system that the CD talkie's green LED just went bright? Photoresistor? Could work, but not practical at all. Eureka!!! Hack the brand new talkie! Screw it open, check out the voltage on the LED when lit: 2 volts. And since the talkie circuitry is not the latest in miniaturization, the LED is a large one, with two nicely soldered legs! NICE!!! Grab the soldering iron, wire the LED legs, check out the current flow, grab the drill, drill a hole in the molded plastic back, push the wires through, reassemble the talkie parts, screw it close, solder some connector and it's almost done.
Next, I adapted the timing system adding an external connection for the talkie with two wires, one to the ground, the other to a free analog input on the Arduino Mega board. Then I modified my code and replaced all complicated wireless base signal parsing with the dead simple voltage detection of the talkie LED.
First test: works flawlessly !!! I just added some timer procedure to avoid double signals and that was it.
So now I have an F3F timing system in which the base are dead simple walkie-talkies on which the judge just has to push the PTT button when the glider crosses his base. Conditions for a "delay free" delivery of the signals with the talkies I have is to use channel 1 or 2 (less delay than channels 7 or 8) and no use of specific "scrambling" codes which on one hand permit "private" communication, but on the other are prone to much delay between PTT button push and signal received due to signal encoding and decoding.

Well, I was so glad with the unexpected positive result of this important improvement that I decided to create a new WindStation module, one with an LCD display on which instant wind and wind direction are updated twice a second, and in the middle of which an uppercase "OK" appears when the wind speed and direction were good for the last 20 seconds. I also added a longer cable, so there's no need to gather around the wind post to read the display!
With these improvement, I am now much more positive toward my timing system than ever before. And just in case it would fail again... there's a conventional stop watch timer on the walkie-talkies ;-)

1 comment:

  1. Congratulations, Reto. It seems you found a good way to progress.

    Next steps should be done towards the automatic detection of the model when it crosses the base, but this is another story :-(