What's the best ultrasonic?

[Nathan]

What is the best ultrasonic sensor? We need to buy some... I've been looking at these:

http://www.maxbotix.com/
http://www.parallax.com/Store/Sensor...me,ProductName

I don't think either of those take interrupt ports, a big plus in my mind.

So what do you guys think the best ultrasonic is? Not limited to the two posted above of course.

TIA,
Nathan

[ubergeek5075]

We've used the Maxbotix ones without a hitch. Attach +5V, GND and AN to an analog port, and away you go.

Beware of using two at a time facing in the same general direction. The internal logic gets confused and unstable readings follow because they will interfere with each other.

See Maxbotix FAQ for a solution to that problem. If you control them with the RC (and pulse them at the same time), they won't interfere with each other.

We've had success with them, so I hope that you do,
Good Luck

[EHaskins]

I tested the Matbotix in 2006. They worked great. If I remember correctly an 8bit ADC the value is almost exactly the distance in inches.

[Nathan]

We bought several of the Maxbotix sensors, but are having some serious issues with them.

Sometimes when we have the sensor hooked up and are reading the value in the terminal window the value will go insanely high, and then when we rotate the sensor the insane value returns to a normal value. By "insane value" I mean it jumps from around 13 to 213.

Has anybody else had this issue? We really want to use these ultrasonic considering we bought like 10 of them!!

Thanks,
Nathan

[nparikh]

Agreed that the Maxbotix ones work! We used them on our 07 robot and have a stock of a few of them in case we want to use IR again!

[Nathan]

Any idea why ours return bad values every once in a while?

-Nathan

[gnormhurst]

We're using a Vex unit. There's driver code available on this forum. Search for "sonar".

What you are seeing does not seem unusual. If there is just one or two bad readings in a row, you can write code to ignore sudden changes.

Also, consider that vibrations from the motors, etc, can confuse the sensor. We were getting mysterious readings but only when the motors were running. Unbolting the sensor from the bot and holding it by hand eliminated the problem. We remounted it using foam.

Note that if you use bolts through the foam to the chassis, the bolts will conduct the vibrations past the foam to the sensor. Mount the sensor to the foam, then the foam to the chassis.

[Nathan]

Yes, we could ignore the sudden changes, but since we wanted to use these for collision avoidance that wouldn't help us...

[Phalanx]

Originally Posted by Nathan

Sometimes when we have the sensor hooked up and are reading the value in the terminal window the value will go insanely high, and then when we rotate the sensor the insane value returns to a normal value. By "insane value" I mean it jumps from around 13 to 213.
Nathan

We've used them and they work well.
A few things I have found with them in order of significance....

1) On "power-up" they need a large, clear, wide and obstacle free area for self calibration to work well, if they don't calibrate well you will get odd readings.

2) They don't respond well to rapid large change in distance. For example, it's reading an object 10ft away, you put your hand in front of it and quickly remove it, it will take a few cycles to recover from that. Use an averaging routine or similar logic to get several readings before acting.

3) Isolate them from vibrations as much as possible, vibrations will get picked up and misinterpreted.

4) Use RF shielded cabling to eliminate other noise.

[Nathan]

When I've been turning on the RC (thus the ultrasonics) I had the obstacle about 12 inches in front of it. So maybe that is the issue, they aren't "calibrating" correctly?

I'm starting to like our daventechs more and more here...

Thanks,
Nathan