The puck is only about 0.138 in. (3.51 mm) tall. Given that, I'm concerned that I can find a sensor with a low enough profile. Also, the playing surface is 27 in. (686 mm) wide x 54.625 in. (1390 mm) long. I am trying to build a matrix of these and use the sensors in an on/off mode to predict the pucks travel based on how the sensors are triggered. This approach was suggested by a colleague since I will be using a PLC to control the system and he thought on/off would work better than analog range finding.
I did test a Ping))) ultrasonic sensor and the maximum range I could get was about 19 in. (483 mm).
For illustration, I'm attaching a picture I took of part of the air hockey table while testing the ultrasonic sensor just.
Thanks! Neil
Hi Neil Petroff,
Perhaps a more practical solution would be to use a camera with blob detection algorithm to track the movement of the puck.
Regards
Fahmi
Hi Neil,
i'm testing this type of ultrasonic sensors, and i suggest you also HCSR04, with a wide range, but pay attention at low temperatures for speed of sound charaterization.
Regards
Umberto
A calibrated camera would make sense here.. but if the only processing device you have is a PLC it might be a bit tough situation. if you want to go optical and no camera try optical IR sensors but think about the off axis sensitivity so you can come up with a neat workaround.
Cheers
Ganindu.
HI all,
Thanks for the responses. This is for a course development so I'm trying to adhere to some "artificial" constraints. One is that a team can only instrument their half of the "ice". For that reason, I was trying to find some photoelectric sensors that would work in diffuse proximity mode. Due to the length of the board, I'm finding this to be a difficult task. However, I am receiving a couple of sensors to test. In addition, I want to test how the orientation angle of the sensor will affect its ability to "see" the puck.
I suppose a camera could be used as long as it did not interfere with the other half of the ice. What are the issues of using a camera with a PLC? PLC control is another one of our constraints. However, I will probably offer this an an option to the students. In addition, I'm going to look into instrumenting the puck itself and use RFID or acoustics to track the position of the puck.
Hi Nell,
There are Ultrasonic (US) sensors availabe to sense higher range as well (Upto 120 meters). the US sensor needs to be in parallel with the ground. The puck used dosent have enough height to align in parallel with the US sensor . even if it dose it will be a very tight fit. Moreover the puck is a round shaped object, The reflected sound will be quite faint for the sensor to pick up which will give a lot of distortion.
The color camera is a good option, but then you have move to to a more powerful board. A raspberry pi , begleboard or any such simiar board will be required to process 30 images a second. A typical 5MP camera will yield a lot of image data which will require atleast a quad core ARM based board with a GB of memory to process the red puck from the white background (using blob or motion or filter algorithms).
I see this can be done with a Raspberry pi 2 and a 5 MP camera (under 70$, 45$ for raspberry and 20$ for camera). combine this sytem with a couple of accelerometer, wheel encoder and Digital compass (raspberry pi has dedicated sesors which can be bought from ebay, banggood or gearbest websights) to sense its position and inclination. With this sytem you can achieve a refresh rate of 50-60 times a second and position the tracker very quickly. Also this will satify the rule to keep all to sensors to your half.
Thanks
Bishwajit Pal
I would suggest you put the sensors in the playing surface next to the jets. That way you could detect the bottom rather than the side of the puck. You may also want to consider controlling the jets to drive the puck. If it works you could program the puck to travel in a designated path like a drone flying a flight plan.
Good Luck, David F.
Hi David,
That's a very interesting idea. We've actually devised a layout where we do sense the bottom of the puck. Controlling the jets is well beyond our scope at this point; however, it would be fun to try at some point.
I did find a mid-range p/e sensor that can see approximately 3/4 of the length of the ice when mounted on the ice surface. This should allow early puck ID without falsely picking up the striker.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Control Theory
- Control Systems