Two MR Range Sensors, need to restart CIM

[robgcarlisle]

Has anyone else had any difficulty with having two MR Range Sensors on a robot? We successfully changed the port addresses, put their addresses in the code etc., but when we called them they reported the same range. BUT, following other hints here, we unplugged and replugged the USB from the Core Interface Module, and then they produced distinct outputs. This would be very inconvenient in a competition! Any ideas?

we have also been investigating how much Ultrasound cross-talk we get between the two sensors, because we want them pointing at the same wall at ranges <10 inches. It appears to be manageable: with plates positioned on the robot to act as wave-guides, and some median filtering of the registered ranges, we can get them to act independently.

[mikets]

Where did you place these two range sensors on the robot? In general, it's a bad idea to use two ultrasonic sensors (i.e. range sensor) especially if they are placed close together and pointing to the same direction because they will pick up each other's echoes. So beware!

[robgcarlisle]

We've got the sensors on the same side of the robot, facing the same direction, 15 inches apart. We mount then with metal plates on the side towards each other... like .| |.
This modifies the "cone" antenna pattern so that there's no cross-talk at the ranges we are interested in (<10 inches)... or if there is we filter out highly spurious readings (about one in ten readings are "255").

Still, has anyone an answer to the problem that we need to restart the CIM to get the system to access the two sensors separately? Does anyone see this behavior with color sensors?

[rvansmith]

I used two range sensors on one side and didn't have any problems reading distinct values with both. I discontinued using the two because of cross talk. I was trying to use them at several feet, so the cross talk is not surprising. It would be nice if the range sensors could be synchronized so they didn't pulse at the same time...

I am going to try and use 2 again soon placed perpendicular with each other. I am using SDK 2.61.

[rvansmith]

The second range sensor is back in the code, running perpendicular and not having any issues using them both software wise.

[robgcarlisle]

I'm on an older SDK, that could be the problem. I'll try to update next week.

[rvansmith]

We went with the two MR Range sensors perpendicular to each other, and I believe we were still getting random cross talk. I wouldn't recommend using 2 range sensors, it is not a stable platform in my experience.