DPRG List  

[DPRG] Measuring Speed of sound in a liquid

Subject: [DPRG] Measuring Speed of sound in a liquid
From: Kipton Moravec kip at kdream.com
Date: Fri Sep 3 23:48:36 CDT 2004

Thanks to Rick I found exactly what I was looking for, a 180 kHz Teflon 
transducer for only $39.  The maximum voltage pulse to make it move is only 
250V p-p :)  The range specified is 4" to 4'  (for a reflected wave) and 
since I am interested in a sensor and a receiver the maximum range is 
actually 16 feet.  Of course the longer the distance, the better the 
accuracy.  I thought 1 meter was good, (better than comparable systems) I 
am now going to have to look at the sensitivity as I investigate a longer pipe.

What you want to do is send a defined pattern, say 8 sine waves at 180 kHz, 
and sample with an ADC at a much higher frequency (in my case 18 MHz)  Then 
do a correlation of what you sent and what you received.  I believe you can 
get sub sample accuracy.  This is better because you have more a priori 
information about the wave you are looking for than the simple 
"spl-thunk"  Which probably has components in all different frequencies.

The speed of sound in different liquids is interesting.  You can find one 
with a google search.  The other thing to notice is that is very 
temperature  dependent, so I have to measure the temperature of the liquid 
also.  Fortunately it is typically a concrete room and the temperature does 
not fluctuate much.


>Let me preface this with a clear warning that I only have two
>semesters of college physics, that it was 20 years ago, and that my
>interests were less to do with mechanics at a macro scale that wasn't
>at least hundreds of light-seconds worth of "macro" than the
>dark-haired woman in the third row. I remember as much about her as I
>do about physics, and I can't even recall her name. YHBW.
>That being said, how's this for a brainstorm: when all's said and
>done, a "sound wave" is nothing but a bunch of compression waves
>back-to-back. In its basic form, just one oscillation has the same
>properties as a sound wave. It seems to me that you could have a
>chamber with a fixed end and an end that can could move, and a blade
>that chop down. The blade goes spl-thunk, the liquid goes sloosh,
>then the movable end moves a finite amount of time later. The
>distance from the blade and the initial position of the movable end
>is known. Measure the amount of time that it takes for the movable
>end to move, divide by distance; et voilá, the quotient is the speed
>of sound. All of the pieces involved in this method are pretty
>minimalist and can be as durable as you like without sacrificing
>The shape of the tank and initial stillness might throw slop into the
>measurement, but a little more sophistication in tank and
>wave-generator can help dampen errors and increase accuracy. The
>reason this might be better than measuring real sounds is that the
>waveforms for sound are necessarily going to be more complex and, to
>get the maximum gross movement to match my "single wave" method in a
>sound pulse, the energy of my single wave would have to be the
>highest energy in a waveform that rose from zero energy a rapid
>to/fro sequence with a net energy near zero.
>As previously noted, I am not a physics guru and the two books that I
>would resort to in hopes of getting further inspiration are in
>storage. If you have access to a copy of _The Way Things Work_, there
>is almost certainly at least two designs that would meet your needs,
>and the things in TWTW are always concrete,
>they've-been-built-and-tested-in-non-laboratory conditions. Another
>good general reference to check for ideas on this stuff is Van
>Nostrand's Scientific Encyclopedia, and if you have the option, look
>first for a copy that is pre-'85. They stopped using English around

More information about the DPRG mailing list