A so-called zero test ascertains that all your connections and synchronizations are working properly. This test is especially important if you are using the cloud mode where the devices are not synchronized directly to each other and you rely on the quality of the cellular network. The IAAF Photo Finish Guidelines also recommend that you do a zero test. This post is a guide on how to conduct a zero test on-site, both with a starting gun and without.
A zero test means that you place the start gun at the finish line and fire it. The photo finish should then show 0.0 s at the image where the gun goes off. To achieve this in SprintTimer you can use the fact that you can hand-start the photo finish, and then send a start from Start Sender while recording.
Testing with a starting gun
To get the highest possible accuracy it is recommended to use a sound to start. Place the starting gun on a table and place the start sender device beside it. Connect the starting gun if you are using an electronic one. Place the finish device on a tripod or similar above the gun. Open Photo Finish and sync the devices. Open the Preview and align the LED flash on the starting gun with the finish line (or align the nozzle if you use a conventional gun). To run the test follow these steps:
- Activate the sound start
- Tap the Hand Start button
- Start the recording
- Fire the gun
- Stop the recording when the time label changes from grey to white*
- Scroll to where flash first becomes visible
The time should be close to 0.0 s. However, some starting guns might have a small difference between sound and flash (more below).
*Indicates that the start sender time has been received, this happens very quickly in direct mode, but might take a few seconds in cloud mode.
Testing without a start gun
If you don’t have a gun you need something else that both moves and makes a sound. A simple, but surprisingly accurate, procedure is to take a piece of wood or metal, align it with the finish line and strike it. I used a 10 cm aluminum bar and struck it with a screwdriver (metal generates more sound). I placed the finish device vertically. In the Preview I aligned the side of the bar I intended to strike with the finish line. Otherwise, the procedure was similar to that with the gun:
- Activate the sound start
- Tap the Hand Start button
- Start the recording
- Strike the bar hard enough to trigger the sound-start
- Stop the recording when the time label changes from grey to white
- Scroll to where the bar disappears
Results
To test the procedures I ran multiple “starts”,m both with a gun and by striking a bar, and with both direct (WiFi) and cloud (4G) connections.
In all cases, the results had a spread of +- 0.01 s and with no discernable difference between gun and bar, and between direct and cloud. So the procedures seems to work and start sender gives good accuracy that is in line with earlier tests. However, the starting gun results had an offset of about 0.04 s. I, therefore, tested the gun separately to see if it was the cause.
Starting gun offset
I used two different tests which I each ran multiple times. I first filmed the gun in slow-motion (240 fps), opened the video in an editing program, and measured the time difference between the light and the sound. Secondly, I attached a photo-sensor to the flash on the gun and connected the sensor and the sound output to two channels in an oscilloscope. I then measured the time difference between the two signals when pressing the trigger. In both cases, I got a difference of 0.03 s – 0.05 s. The offset in the diagram above can therefore be explained by the gun. This difference might not matter so much in practical timekeeping since you seldom start on the flash, but it is worth keeping in mind when you do a zero test. I was using a gun from Jex/Gill/Hart (depending on the market) that is pretty common, but other guns will probably give other results. It can therefore be worth the trouble to do a series of tests the first time to find the offset.
Problem-solving
So what do you do if your zero test results in a significant offset? After repeating it and checking that all steps were followed there are a few things you can do. Restarting both devices can solve many problems. The network can also be a cause. In direct mode check the “latency” and after a while the “age”. If they are large, the network might not function properly. That the synchronization takes a long time (>15 s) is also a bad sign. If you have access to the network you might try to restart it or change the network port.
In cloud mode, you can check the “latency” and the “offset”. The latter doesn’t directly affect the accuracy, but if it is > 1 s it is a sign that the synchronization might have problems. However, in cloud mode, it is more difficult to something about the network. But it might be possible to move the start sender device to a better location. It might also be worth testing if 3G gives a better result than 4G (if the 4G connection is patchy). Finally, under really bad network conditions you might have to accept that the cloud mode is not the ideal choice.