Ultrasonic transducer time of flight - Electronic Circuits Projects Diagrams Free

hash · 3rd July 2009, 02:10 PM

Hello, I have a slight problem using an ultrasonic transducer. Specifically, in a circuit with two of those, one being the emmiter and one the receiver, the amount of time it takes between the initial 40khz pulse from the transmitter and some voltage being generated in the receiver is too great.

For example, when they're set at 30cm from one another, the delay should be around 800us. Instead, I see a delay of about 1ms, which is significant. And yes, I've compensated for the temperature difference and made a bunch of tests at different distances, the delta t is still too long.

Is this normal behaviour on these ultrasonic transducers? I have no datasheet for them and the only thing I can think right now which could be responsible for this behaviour is the inertia of the transducer. If so, I should be able to compensate for it, subtract a constant amount of time from the result each time a measurement is taken.

Any thoughts?

ericgibbs · 3rd July 2009, 02:38 PM

hi,
There are time delays in the reaction of the TX to transmit sound and delays in the RX.
These delays are usually constant and can be subtracted from the time as a constant.

Look at this data, may help with the project.

JimB · 3rd July 2009, 03:29 PM

I cant claim much prior experience, but this sparked my curiosity ans a did a little test using a couple of 40khz transducers which I have had for a few years.

With the transducers face to face, there was a delay of 0.2ms between the excitation of the transmitter and seeing a response in the receiver, the rise time of the signal in the receiver was about 0.6ms.

With the transducers separated by 30cm, there was a delay of 1.1ms between the excitation of the transmitter and seeing a response in the receiver, as before the rise time of the signal in the receiver was about 0.6ms.

So, I guess that my results are similar to yours, as the transducers are quite narrow band (high Q), it is reasonable to expect a propagation delay through them.
As Eric said, you just have to know that it is there and subtact it from your real measurements.

JimB

hash · 3rd July 2009, 06:53 PM

@JimB - your results are nearly identical with mine, that's good

I'll just subtract the difference in a known environment and it should be calibrated. The rise time should be no concern because I'll use a great deal of amplification in order to detect the very moment at which the receiver activates.

@eric - thanks, that pdf is useful, although they calibrate the sensor by taking a measurement at a known distance then extrapolating that value for all their measurements, not deriving the distance from theory. If all else fails I'll use the same method, seems simple enough.

By the way, I'm trying to build an ultrasonic wind speed meter, so the registered time of flight will be very important, a couple of microseconds could mean the difference between a gentle breeze and a hurricane

ericgibbs · 3rd July 2009, 07:04 PM

Quote:

Originally Posted by hash

By the way, I'm trying to build an ultrasonic wind speed meter, so the registered time of flight will be very important, a couple of microseconds could mean the difference between a gentle breeze and a hurricane

hi,
Assuming velocity of sound in air is 330m/s and the ultrasound frequency is 40KHz, the wavelength is 8.25mm.!!!,, I make that about 25uSec.

I would say the chances of measuring the reflected time to within uSecs will not be feasible.

Lets know how it goes.

user_88 · 3rd July 2009, 07:23 PM

It is not clear whether you are using a µcontroller ...:
There was a similar problem topic someplace in a book .... Quintessential PIC Microcontroller ....author Katzen .... You have to actually go through the program line by line and add up the µSec for each command. ....

hash · 3rd July 2009, 08:01 PM

Yes, I'm using a PIC microcontroller for time measurement, with a 16 bit timer @ 16MHz (fosc/2), that'll give me about 62ns worth of resolution... in theory at least

JimB · 3rd July 2009, 08:48 PM

I have a vague memory from distant experience that ultrasonic anemometers use doppler effect rather than time of flight measurements.

JimB

hash · 4th July 2009, 12:25 PM

Flow rates can be determined by using either doppler effect or time of flight:
Ultrasonic Doppler and Time of Flight Flowmeters

ericgibbs · 4th July 2009, 12:43 PM

Quote:

Originally Posted by hash

Flow rates can be determined by using either doppler effect or time of flight:
Ultrasonic Doppler and Time of Flight Flowmeters

The 'time of flight' solution shown your link is not subject to the same delay problems as explained in your original post.

Assuming the two transceivers have the same TX and RX delays these will cancel out.
The measurement is made from the difference in transit time of the two beams.

hash · 4th July 2009, 01:37 PM

The link was just to show you can use time of flight to determine the speed of a fluid.
Yes, by switching the roles of the two transceivers you can overcome the delays and environment properties (such as temperature) assuming they remain constant throughout the measurements. I usually stick with something simple that works -- using a triangle configuration in which there's one transmitter and two receivers at 60 degrees from one another, I'm measuring the time of flight separately for each one and then deriving the speed and direction of the wind with some trigonometry. I'm compensating for the temperature changes in software since I already have a fairly accurate temperature sensor installed.

ericgibbs · 4th July 2009, 02:20 PM

Quote:

Originally Posted by hash

The link was just to show you can use time of flight to determine the speed of a fluid.
Yes, by switching the roles of the two transceivers you can overcome the delays and environment properties (such as temperature) assuming they remain constant throughout the measurements. I usually stick with something simple that works -- using a triangle configuration in which there's one transmitter and two receivers at 60 degrees from one another, I'm measuring the time of flight separately for each one and then deriving the speed and direction of the wind with some trigonometry. I'm compensating for the temperature changes in software since I already have a fairly accurate temperature sensor installed.

hi,
I am puzzled, if you were aware of the delay problem and that it would be cancelled in a dual system.
Also you already have working project.

What was the purpose of your original post.

hash · 4th July 2009, 03:38 PM

I was only aware of environment-induced measurement errors (the air temperature) and didn't want to use a dual system since it would complicate things a bit.
But the time delays didn't make any sense to me, that's why I posted here, to find out if it was the normal behavior of the sensor or I was doing something wrong... Also the project isn't fully functional, I've only described what it will become

Slightly off-topic: What are 'NL' resistors on a schematic? I'm trying to read the microcontroller dev board schematic and they're pretty common, didn't find anything on Google...

ericgibbs · 4th July 2009, 03:41 PM

Quote:

Originally Posted by hash

I was only aware of environment-induced measurement errors (the air temperature) and didn't want to use a dual system since it would complicate things a bit.
But the time delays didn't make any sense to me, that's why I posted here, to find out if it was the normal behavior of the sensor or I was doing something wrong... Also the project isn't fully functional, I've only described what it will become

Slightly off-topic: What are 'NL' resistors on a schematic? I'm trying to read the microcontroller dev board schematic and they're pretty common, didn't find anything on Google...

hi,
Understood.

I'll check the context of NL.?

EDIT:
Here we go.
http://www.firstohm.com.tw/Spec/NL.htm

hash · 4th July 2009, 03:57 PM

Hmm... I don't think it's that, it's like a value or something... Could it mean they're not present, like not connected? I see them in circuits that could have a dual purpose, but I don't have the board at hand to check. I'm attaching a couple of pictures from the datasheet.

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Similar
Title	Starter	Forum	Replies	Latest
Ultrasonic Transducer	rcd316	Datasheet/Parts Requests	7	22nd April 2009 05:18 PM
Looking for Ultrasonic Transducer	pjainmg	General Electronics Chat	13	22nd January 2009 02:05 PM
Pic driving ultrasonic transducer	mlt57sooner	Micro Controllers	1	22nd April 2005 09:42 AM
RF Time of Flight Measurements using a PIC	jfunkie	Micro Controllers	4	22nd February 2005 06:13 PM
IR Time of flight	The Real MicroMan	Electronic Projects Design/Ideas/Reviews	3	25th February 2004 10:19 PM

3rd July 2009, 02:10 PM	#1
hash	Ultrasonic transducer time of flight Hello, I have a slight problem using an ultrasonic transducer. Specifically, in a circuit with two of those, one being the emmiter and one the receiver, the amount of time it takes between the initial 40khz pulse from the transmitter and some voltage being generated in the receiver is too great. For example, when they're set at 30cm from one another, the delay should be around 800us. Instead, I see a delay of about 1ms, which is significant. And yes, I've compensated for the temperature difference and made a bunch of tests at different distances, the delta t is still too long. Is this normal behaviour on these ultrasonic transducers? I have no datasheet for them and the only thing I can think right now which could be responsible for this behaviour is the inertia of the transducer. If so, I should be able to compensate for it, subtract a constant amount of time from the result each time a measurement is taken. Any thoughts?

3rd July 2009, 03:29 PM	#3
JimB	I cant claim much prior experience, but this sparked my curiosity ans a did a little test using a couple of 40khz transducers which I have had for a few years. With the transducers face to face, there was a delay of 0.2ms between the excitation of the transmitter and seeing a response in the receiver, the rise time of the signal in the receiver was about 0.6ms. With the transducers separated by 30cm, there was a delay of 1.1ms between the excitation of the transmitter and seeing a response in the receiver, as before the rise time of the signal in the receiver was about 0.6ms. So, I guess that my results are similar to yours, as the transducers are quite narrow band (high Q), it is reasonable to expect a propagation delay through them. As Eric said, you just have to know that it is there and subtact it from your real measurements. JimB __________________ Experience is directly proportional to the value of the equipment ruined.

3rd July 2009, 06:53 PM	#4
hash	@JimB - your results are nearly identical with mine, that's good I'll just subtract the difference in a known environment and it should be calibrated. The rise time should be no concern because I'll use a great deal of amplification in order to detect the very moment at which the receiver activates. @eric - thanks, that pdf is useful, although they calibrate the sensor by taking a measurement at a known distance then extrapolating that value for all their measurements, not deriving the distance from theory. If all else fails I'll use the same method, seems simple enough. By the way, I'm trying to build an ultrasonic wind speed meter, so the registered time of flight will be very important, a couple of microseconds could mean the difference between a gentle breeze and a hurricane

3rd July 2009, 07:23 PM	#6
user_88	It is not clear whether you are using a µcontroller ...: There was a similar problem topic someplace in a book .... Quintessential PIC Microcontroller ....author Katzen .... You have to actually go through the program line by line and add up the µSec for each command. .... Last edited by user_88; 3rd July 2009 at 07:25 PM.

3rd July 2009, 08:01 PM	#7
hash	Yes, I'm using a PIC microcontroller for time measurement, with a 16 bit timer @ 16MHz (fosc/2), that'll give me about 62ns worth of resolution... in theory at least

3rd July 2009, 08:48 PM	#8
JimB	I have a vague memory from distant experience that ultrasonic anemometers use doppler effect rather than time of flight measurements. JimB __________________ Experience is directly proportional to the value of the equipment ruined.

4th July 2009, 12:25 PM	#9
hash	Flow rates can be determined by using either doppler effect or time of flight: Ultrasonic Doppler and Time of Flight Flowmeters

4th July 2009, 01:37 PM	#11
hash	The link was just to show you can use time of flight to determine the speed of a fluid. Yes, by switching the roles of the two transceivers you can overcome the delays and environment properties (such as temperature) assuming they remain constant throughout the measurements. I usually stick with something simple that works -- using a triangle configuration in which there's one transmitter and two receivers at 60 degrees from one another, I'm measuring the time of flight separately for each one and then deriving the speed and direction of the wind with some trigonometry. I'm compensating for the temperature changes in software since I already have a fairly accurate temperature sensor installed.

4th July 2009, 03:38 PM	#13
hash	I was only aware of environment-induced measurement errors (the air temperature) and didn't want to use a dual system since it would complicate things a bit. But the time delays didn't make any sense to me, that's why I posted here, to find out if it was the normal behavior of the sensor or I was doing something wrong... Also the project isn't fully functional, I've only described what it will become Slightly off-topic: What are 'NL' resistors on a schematic? I'm trying to read the microcontroller dev board schematic and they're pretty common, didn't find anything on Google...

4th July 2009, 03:57 PM	#15
hash	Hmm... I don't think it's that, it's like a value or something... Could it mean they're not present, like not connected? I see them in circuits that could have a dual purpose, but I don't have the board at hand to check. I'm attaching a couple of pictures from the datasheet. Attached Thumbnails