04 July 2010
ESawdust Speed Timer Headed to USAC Nationals 2010
05/07/10 09:32 Filed in: Speed
Climbing
Synopsis: This is an
overview and technical rundown on the 2010 version
of the speed climbing timing system for 2010 USAC
Nationals. This video was produced based on the
speed timer deployed for the 2010 USAC Southwest
Divisional championship at Movement Climbing and
Fitness in Boulder, CO. The version we will deploy
for USAC Nationals in Atlanta has some improvements
over the divisionals timer.
Here’s a video overview of the 2010 speed climbing timing system in action during the divisional championship and also a technical overview of the parts of the system:
This is a fun video that Sparkfun.com made about the system I built last year for the 2009 climbing season:
The picture above shows the system installed at Movement Climbing and Fitness in Boulder, CO. Since this picture was taken, the master timer (the acrylic covered box that’s in the middle of the two lanes) has been radically reduced in size for 2010 USAC nationals. We may install it on the back side of the wall as well for Nationals.
Below is a full view of the speed climbing wall at Movement Climbing and Fitness. It accommodates 4 lanes of speed climbing. Audience displays were mounted at the top due to space issues at the bottom - as you can see the wall flares wider towards the top.
Improvements based on 2009 Lessons Learned
The 2010 version has all of the improvements made to it that I identified in the article: Speed Climbing Timing Lessons Learned after USAC Nationals last year. The main improvements based on lessons from 2009 are:
1) Easier installation due to subsystems communicating wirelessly now - many fewer cables to run.
2) Less sensitivity to ESD - fewer wires
3) Foot sensors were redesigned so cabling doesn’t come out front (gets mashed against the wall.)
4) Foot present statuses shown on large display and light indicators for climber
5) Better light indicators for climber start (feet) and finish statuses (hands).
6) Better starter console - made it easier to build - made backups
Beyond Lessons Learned
I went beyond 2009 lessons learned on the starter console because it can also display foot statuses and false start indications on an LCD screen.
Furthermore, the new starter console can double as a pickle switch so there can be a manual time for each lane in case the system malfunctions at the top. That way, there can still be an official time and the climber will not have to rerun the race. This follows the model for swimming electronic timing systems where there are touch pads and an official for each lane with a pickle switch to stop the clock on that lane. In swimming, a swimmer can request the pickle time as the official time for an appeal. While speed climbing isn’t quite that far along in terms of rules maturity, I think it will need to go there so I’ve built that into the system.
So, the LED clock may fail, but the judge will still have an official time on the starter console. Here is a shot of the two systems’ starter consoles partially assembled and a less fancy backup starter console. The starter console on the left was used as the primary console for USAC 2010 southwest divisionals:
This year, the foot sensors are also laser based - last year, the hand sensors were laser based but the foot sensors were IR based. This year, the same electronics and wiring pinouts are the same for foot and hand sensors.
On the 2009 system, every sensor was wired to a microcontroller that lived in the main LED display cabinet. This created a cable “tributary” effect where everything converged into one place and by extension, made the system more difficult to install due to cable routing. It also created more trip hazards at the bottom of the climbs where there are a lot of people milling around (judges, climbers, etc.)
Major Improvements in Light Indicator Feedback
2009’s system had a small LED indicator on the hand sensor that would light momentarily when the climber hit the sensor (finished.) It was easy to miss for both the audience and the climber though people honed in on it pretty well. This year, the indicators are microprocessor controlled so when the master timer sees a detect, it lights the indicator for that lane. These proved to be very easy to see when tested at the southwest divisional championships:
They stay lit after the detect and until the system is reset for the next race. They also will light when the climber is staging at the bottom. When the climber’s foot goes into the foot sensor, an indicator light the one above is positioned where the climber can see it and also the indicator at the top of the lane will light so the audience can see the climbers are about to race.
Indicator lights were built from 12vdc LED light bars purchased at Sparkfun.com. The mounting plates are blank faceplates from the Chameleon 1 enclosure system.
Laser-based Foot Sensors
The 2010 Speed Climbing Timing system has laser-based foot sensors for false start detection. The climber must place his or her foot on the green line to start the race.
One change between the system deployed at Divisionals and the one which will be at Nationals is there is now a dedicated foot sensor controller that does false start detection. Before, each sensor had a wireless component which was monitored by the master and display for false start detection. Now that is all done at the bottom. When a false start is detected, the controller signals the other subsystems of the false start - the display shows “F1” or “F2” depending upon which lane false started (sometimes both will.)
This has two advantages:
1) If for some reason the foot sensors completely fail or the foot controller fails and there are no backups, the system can still work - it’s no longer a critical component to make the system work. The system can still work without the foot sensors - the judge will have to do false start detection (which he should be doing anyway.)
2) It’s faster and up/down state is more reliable - the sensors are now tied to the INT0 and INT1 interrupt lines of an ATMega8 in the foot controller. When the status of either foot lane changes, the status is broadcast and displayed on the LED and the indicator lights both top and bottom.
False start whistles
At the 2009 nationals, there was no good way to indicate to the climbers when a false start occurred. When a false start happened, the judges and audience could see the display and would start yelling “false start”, but the climbers can’t make out the words and they think they’re being cheered on. The farther they go, the more desperate the audience gets and therefore louder, so the climbers think they’re in a huge race. It was a bad feedback cycle.
At 2010 USAC southwest divisionals, we used a referee’s whistle to indicate false starts. It was very effective, but it requires that adjacent systems (lanes) do not start their races at the same time. We alternated lane set starts so a false start whistle, especially when blowing it right when the “F1” or “F2” is indicated, was very effective and the climbers barely got off the ground before they knew there was a false start.
Speed Climbing Coaches - How to Finish Fast in Speed Climbing
In the days of mechanical switches on speed timers, kids got used to slamming them as hard as they could (usually because they didn’t work.) Those days are over.
Coaches need to tell their kids - “you can slam this sensor as hard as you want but it won’t make a lick of difference except you’ll be slower.” The reason is: all you have to do is break a light beam to stop the clock and a fingernail or tiny touch from a 7 year old can do that. You’ll be slower if you try to slam it hard because it takes more time in the air to take a big swing at it.
Think about how Michael Phelps touched out his competitors in swimming...had he taken a big swing to finish he would have lost the gold. Think about a quick, martial arts style jab to the finish and you’ll be a lot faster than slamming the sensor. You’ll likely touch out your competitor if you don’t slam the sensor.
The sensors can take the punishment - that’s not the issue. One kid during the 2010 divisional championship hit it so hard that it bent the T-bracket (metal brackets used to frame houses with), but the sensor still worked fine. It’s just stupid though because hitting it hard means you finished slower than you could have otherwise.
As the designer of the speed climbing timing system, this is the best advice I can give speed climbing coaches and climbers:
Touch the yellow line as fast as you can, not as hard as you can, and you will finish faster.
2010 Speed Climbing Timing System Subsystems
The 2010 system is composed of the following main subsystems:
1) LED display for audience that communicates wirelessly over XBee to other subsystems.
2) Master timer unit which has two wires to the 2 lane hand sensors.
3) Foot sensor controller which does all false start detection at the bottom
4) Starter console
Gallery of pictures during development and building the speed timer
Parts and components used per 2 lane system:
3 AVR ATMega128 microcontrollers:
1 for starter console
1 for LED display driver
1 for Master timer
Custom ESawdust.com AVR Micro board
1 AVR ATMega8 microcontroller
- for foot controller and false start detection
- Olimex P28 development board
4 pen-light lasers from Sparkfun.com
4 TEMT6000 light sensor breakout boards
8 RS485 breakout boards from Sparkfun.com
4 XBee breakout boards from Sparkfun.com
4 XBee Pro 2.4 GHz Series 1 wireless modules
3 12vdc power adapters
1 LED display with Maxim 7219 display driver
Questions or comments?
Landon Cox
www.ESawdust.com
Here’s a video overview of the 2010 speed climbing timing system in action during the divisional championship and also a technical overview of the parts of the system:
This is a fun video that Sparkfun.com made about the system I built last year for the 2009 climbing season:
The picture above shows the system installed at Movement Climbing and Fitness in Boulder, CO. Since this picture was taken, the master timer (the acrylic covered box that’s in the middle of the two lanes) has been radically reduced in size for 2010 USAC nationals. We may install it on the back side of the wall as well for Nationals.
Below is a full view of the speed climbing wall at Movement Climbing and Fitness. It accommodates 4 lanes of speed climbing. Audience displays were mounted at the top due to space issues at the bottom - as you can see the wall flares wider towards the top.
Improvements based on 2009 Lessons Learned
The 2010 version has all of the improvements made to it that I identified in the article: Speed Climbing Timing Lessons Learned after USAC Nationals last year. The main improvements based on lessons from 2009 are:
1) Easier installation due to subsystems communicating wirelessly now - many fewer cables to run.
2) Less sensitivity to ESD - fewer wires
3) Foot sensors were redesigned so cabling doesn’t come out front (gets mashed against the wall.)
4) Foot present statuses shown on large display and light indicators for climber
5) Better light indicators for climber start (feet) and finish statuses (hands).
6) Better starter console - made it easier to build - made backups
Beyond Lessons Learned
I went beyond 2009 lessons learned on the starter console because it can also display foot statuses and false start indications on an LCD screen.
Furthermore, the new starter console can double as a pickle switch so there can be a manual time for each lane in case the system malfunctions at the top. That way, there can still be an official time and the climber will not have to rerun the race. This follows the model for swimming electronic timing systems where there are touch pads and an official for each lane with a pickle switch to stop the clock on that lane. In swimming, a swimmer can request the pickle time as the official time for an appeal. While speed climbing isn’t quite that far along in terms of rules maturity, I think it will need to go there so I’ve built that into the system.
So, the LED clock may fail, but the judge will still have an official time on the starter console. Here is a shot of the two systems’ starter consoles partially assembled and a less fancy backup starter console. The starter console on the left was used as the primary console for USAC 2010 southwest divisionals:
This year, the foot sensors are also laser based - last year, the hand sensors were laser based but the foot sensors were IR based. This year, the same electronics and wiring pinouts are the same for foot and hand sensors.
On the 2009 system, every sensor was wired to a microcontroller that lived in the main LED display cabinet. This created a cable “tributary” effect where everything converged into one place and by extension, made the system more difficult to install due to cable routing. It also created more trip hazards at the bottom of the climbs where there are a lot of people milling around (judges, climbers, etc.)
Major Improvements in Light Indicator Feedback
2009’s system had a small LED indicator on the hand sensor that would light momentarily when the climber hit the sensor (finished.) It was easy to miss for both the audience and the climber though people honed in on it pretty well. This year, the indicators are microprocessor controlled so when the master timer sees a detect, it lights the indicator for that lane. These proved to be very easy to see when tested at the southwest divisional championships:
They stay lit after the detect and until the system is reset for the next race. They also will light when the climber is staging at the bottom. When the climber’s foot goes into the foot sensor, an indicator light the one above is positioned where the climber can see it and also the indicator at the top of the lane will light so the audience can see the climbers are about to race.
Indicator lights were built from 12vdc LED light bars purchased at Sparkfun.com. The mounting plates are blank faceplates from the Chameleon 1 enclosure system.
Laser-based Foot Sensors
The 2010 Speed Climbing Timing system has laser-based foot sensors for false start detection. The climber must place his or her foot on the green line to start the race.
One change between the system deployed at Divisionals and the one which will be at Nationals is there is now a dedicated foot sensor controller that does false start detection. Before, each sensor had a wireless component which was monitored by the master and display for false start detection. Now that is all done at the bottom. When a false start is detected, the controller signals the other subsystems of the false start - the display shows “F1” or “F2” depending upon which lane false started (sometimes both will.)
This has two advantages:
1) If for some reason the foot sensors completely fail or the foot controller fails and there are no backups, the system can still work - it’s no longer a critical component to make the system work. The system can still work without the foot sensors - the judge will have to do false start detection (which he should be doing anyway.)
2) It’s faster and up/down state is more reliable - the sensors are now tied to the INT0 and INT1 interrupt lines of an ATMega8 in the foot controller. When the status of either foot lane changes, the status is broadcast and displayed on the LED and the indicator lights both top and bottom.
False start whistles
At the 2009 nationals, there was no good way to indicate to the climbers when a false start occurred. When a false start happened, the judges and audience could see the display and would start yelling “false start”, but the climbers can’t make out the words and they think they’re being cheered on. The farther they go, the more desperate the audience gets and therefore louder, so the climbers think they’re in a huge race. It was a bad feedback cycle.
At 2010 USAC southwest divisionals, we used a referee’s whistle to indicate false starts. It was very effective, but it requires that adjacent systems (lanes) do not start their races at the same time. We alternated lane set starts so a false start whistle, especially when blowing it right when the “F1” or “F2” is indicated, was very effective and the climbers barely got off the ground before they knew there was a false start.
Speed Climbing Coaches - How to Finish Fast in Speed Climbing
In the days of mechanical switches on speed timers, kids got used to slamming them as hard as they could (usually because they didn’t work.) Those days are over.
Coaches need to tell their kids - “you can slam this sensor as hard as you want but it won’t make a lick of difference except you’ll be slower.” The reason is: all you have to do is break a light beam to stop the clock and a fingernail or tiny touch from a 7 year old can do that. You’ll be slower if you try to slam it hard because it takes more time in the air to take a big swing at it.
Think about how Michael Phelps touched out his competitors in swimming...had he taken a big swing to finish he would have lost the gold. Think about a quick, martial arts style jab to the finish and you’ll be a lot faster than slamming the sensor. You’ll likely touch out your competitor if you don’t slam the sensor.
The sensors can take the punishment - that’s not the issue. One kid during the 2010 divisional championship hit it so hard that it bent the T-bracket (metal brackets used to frame houses with), but the sensor still worked fine. It’s just stupid though because hitting it hard means you finished slower than you could have otherwise.
As the designer of the speed climbing timing system, this is the best advice I can give speed climbing coaches and climbers:
Touch the yellow line as fast as you can, not as hard as you can, and you will finish faster.
2010 Speed Climbing Timing System Subsystems
The 2010 system is composed of the following main subsystems:
1) LED display for audience that communicates wirelessly over XBee to other subsystems.
2) Master timer unit which has two wires to the 2 lane hand sensors.
3) Foot sensor controller which does all false start detection at the bottom
4) Starter console
Gallery of pictures during development and building the speed timer
Parts and components used per 2 lane system:
3 AVR ATMega128 microcontrollers:
1 for starter console
1 for LED display driver
1 for Master timer
Custom ESawdust.com AVR Micro board
1 AVR ATMega8 microcontroller
- for foot controller and false start detection
- Olimex P28 development board
4 pen-light lasers from Sparkfun.com
4 TEMT6000 light sensor breakout boards
8 RS485 breakout boards from Sparkfun.com
4 XBee breakout boards from Sparkfun.com
4 XBee Pro 2.4 GHz Series 1 wireless modules
3 12vdc power adapters
1 LED display with Maxim 7219 display driver
Questions or comments?
Landon Cox
www.ESawdust.com
asdfasdf
