How to narrow ultrasonic range?

bulkhead

Is there any way to narrow the horizontal range of the PING))) or any other ultrasonic range finder? Right now it is picking up anything within 40 degrees of my boe-bot (from 0 to 12 inches). I am trying to scan around and find the positions of objects around, but currently each scan will reveal that any object (say a pencil) will take up at least 80 degrees (of the 180 degree scan area). Any input would be appreciated, thanks.

bjhamltn

Some type of mesh cone might work.
The returning ping would be intercepted by the mesh of the cone.

Robert Kubichek

bulkhead said...
Is there any way to narrow the horizontal range of the PING))) or any other ultrasonic range finder? Right now it is picking up anything within 40 degrees of my boe-bot (from 0 to 12 inches). I am trying to scan around and find the positions of objects around, but currently each scan will reveal that any object (say a pencil) will take up at least 80 degrees (of the 180 degree scan area). Any input would be appreciated, thanks.

Try a shotgun mike approach, put the sensor on one end of a tube lined with material that absorbs us waves, then only what's right in front of the opening will get detected.
You will have to play with length/diameter/material to optimize your wanted results.

Bob N9LVU

Beau Schwabe

I'm not exactly sure how to go about the calculations, but an "acoustic waveguide" would have a similar effect that a lens has for an LED.

I don't think that introducing a material that absorbs ultrasonic waves is the right approach however. I think it’s more a matter of focusing
the acoustic energy rather than ignoring it.

Robert,

You are on the right track with a "shotgun mike approach" , but the cut length is more important than what the tube is lined with. This would essentially
be a "acoustic waveguide" with the length and diameter determining the resonant frequency.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe

IC Layout Engineer
Parallax, Inc.

bulkhead

Ok, so if I understand this correctly...

I extend the tube around one (or both) of the cylinders? And it is the length of this tube extension, not the material that will affect it?

I have tried covering the emitter and leaving only a small 1mm slit, and it does narrow the area of detection but also gives it really random blindspots. It is too unpredictable.

Are there any other range sensors that have a narrower range? I looked at some laser ones that have .3 degree range but cost a couple thousand $$$. Is there anything in between?

Beau Schwabe

The tube should be on the receiving transducer, and yes, the material of the tube will also make a difference.

Instead of putting a dampening liner on the inside of the tube, it may be necessary on the outside.



I am just guessing at this based on some limited experimentation I have done, so someone with more knowledge
in this area should kindly step in.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe

IC Layout Engineer
Parallax, Inc.

Zoot

I'm not sure if these photos show the setup all that well, but the old Hero-1 has sonar where the two transducers are fitted into the ends of two 6" tubes to help narrow the sonar "beam". The tubes are kind of like PVC but more "resiny" -- they're not acoustically super hard, but not soft by any means. The docs for the bot just call them '6" sonar tubes', so I can't really say what they're made of. It might be a place to start for ideas. These transducers are much larger -- I don't think you'd need tubes that long for a Ping.

That said, my own experience is that the Ping units are very narrow in their "beam" -- maybe try angling the unit up and see if you're getting reflections off the floor??

Other Ping users may have wildly different experiences????

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
When the going gets weird, the weird turn pro. -- HST

bulkhead

Thanks, everyone, for all of the replies.

Looking at the documentation from parallax, the results I am getting are consistent with their data for measurements less than 1 foot in depth. The only idea I have come up with is to rotate the sensor 90 degrees and check if the range is any narrower oriented that way.

If it doesn't help, I'll probably try the pvc pipe idea, I'll post again with my results.

Phil Pilgrim (PhiPi)

The wide acceptance angle of the Ping transducers is due to their small diameter relative to the ultrasonic wavelength. There probably won't be much you can do that's simple to narrow the focus, short of using larger transducers. A tube will only offset the spreading from the transducer to the end of the tube, whereupon the same spreading will occur. (I'm not sure that a horn wouldn't work, though, tapering gradually from the transducer diameter to a larger one. 'Might be worth a try.) There have also been experiments done with diffractive focusing using "Fresnel zone plates" (see Google and Wikipedia). For ultrasonics, these "sonic lenses" are chemically etched from thin sheets of metal, leaving open apertures.

I tried the tube-and-absorptive-material idea once. It was in an apple packing plant, where the apples were being sorted optically by size, looking from above with a camera. We were having a problem with Red Delicious apples which, due to their long, skinny shape, looked quite different to the camera lying on their sides versus standing on end. The idea was to measure their heights with ultrasonics to correct the measurements made by the camera. But the conveying machinery was quite noisy and interfered with the measurements. The Polaroid transducer -- even with a snorkel -- couldn't be focused adequately to eliminate the interference, and the plan was ultimately abandoned.

-Phil

Phil Pilgrim (PhiPi)

I just ran across this website: www.falstad.com/ripple/. It's a ripple tank simulator in Java. (You have to have Jave installed and enabled to use it.) With it you can visualize various schemes for channeling the Ping's wavefronts. The "baffled piston" setup is probably the closest to the Ping. Here's what the output looks like. Notice how the waves spread out from the opening:

····

Now look what happens when a snorkel is added. The inevitable spread is merely delayed until the end of the tube is reached:

····

-Phil

John R.

While the extended tube (or snorkel) does indeed just delay the inevitable, this is only the first half of the equation. The second half is the return echo. With the snorkel, the returns will be much weaker, other than "head on". The lengith of the snorkel will affec this. It's been too long since physics class, and don't remember if having the tubes a multiple of the wavelenght helps or hinders the "directionality".

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John R.

8 + 8 = 10

Lawson

um... how about putting the ping sensor at the focus of a spherical or parabolic reflector? a Wak or large metal mixing bowl would be a decent first approximation. This'd also likely improve the range of the sensor too.

my 2 cents
Marty

AIman

Lawson,

You are right on. In the production world parabolic dishes are used to get long distance sounds and focus them on a set point. For example a football game on TV, you can sometimes here the cracking of helments and that is because the mics are in front of a parabolic dish that focuses the sound on the mic.

In my opinion, using a parabolic dish to catch incoming waves and focusing those waves into a reciever is the best bet to get more data.

If you focuse the return waves onto the reciever it will help tremendously.

Question is will it be worth the size and effort.

John Abshier

If you are scanning, could you say the object is at the middle of returned reading. An example:
Angle range
0 max
10 max
20 10 inches
30 9 inches
40 10 inches
50 11 inches
60 max
The object is at (20 + 50)/ 2 degrees or 35 degrees after the math.

RottenJalapeno

the higher the frequency of the transducer, the less the ultrasonic beam spread... I dont know the range you are looking at achiving, or how narrow of beam you are looking for, but if your having problems with the 40khz unit try using something with a higher frequency, but the higher the frequency, the shorter the effective sensing range is.· this website describes the beam pattern vs frequency.· hope that helps..
http://www.ndt-ed.org/EducationResources/CommunityCollege/Ultrasonics/EquipmentTrans/beamspread.htm