Single ultrasonic transducer switching to Rx and Tx

[miskol]

ultrasonic transducer driver circuit

Hi,

i am currently trying to figure out how to use a single ultrasonic transducer for transmitting and receiving.

i've managed to do so by using a SPDT switch (CD4053B), however the drawback of using this circuit is that i have to manually do the switching using a microcontroller. Another problem with this circuitry is that it produce a noise that i believe comes from the internal components. It doest matter to me currently but would love to eliminate it for future upgrades. The signal is also somehow amplified even though using the same receiver circuit.

Figure 1(Left) - Using switching circuit

Figure 2(Right) - Direct connection

What i prefer for my project is that the switching for transmitting and receiving is done automatically with the circuit would normally be in receiving mode. When a pulse is given, a switch will automatically route it to the transducer while blocking any access to receiving circuit. At any other time when there is no pulse, a switch will automatically route the transducer to the receiving circuit.


Thank you for any help.

 

[FvM]

ultrasonic transducer circuit

however the drawback of using this circuit is that i have to manually do the switching using a microcontroller

Hard to understand. It should be expected, that the circuit generating the Tx signal can also control the switch?

The meaning of the shown waveforms and the discussion of "noise" issues also can't be well understood without knwowing the circuit.

 

[vsmGuy]

ultrasonic ranging circuit with single transducer

Hi miskol,

I really don't think what you term as "noise" is really that - infact, its due to considerably less loading of the transducer.

I assume you had the sensor setup in a reflective setup with an obstacle very close to the transducer - maybe 1 - 2 cms away (varies on temperature).

Please confirm.

Using switching, you were getting "amplified" signal as there was almost no loading from the circuit on the transducer when compared to "direct connection" where loading was much more.

We should not forget that the analog switches are but buffers and draw very little current from the source!

Hence, in direct connection, you were, in reality, getting an attenuated signal!

As, FvM mentioned though, all this is intution based on assumptions that may/not hold for your particular setup!

For example, I really don't have any idea why CH2 is amplified twice wrt CH1 - what does that trace signify?

Are you afraid to share the circuit because someone may copy it?

I am replying to this thread after almost 3 months because this is very interesting and I would really like to be a part of this project!

I will explain the scope traces (CH1/2 vs CH3) in detail if you are interested and you will be amazed how facts are staring right at you (although you need to know a little about the physical structure of a ultrasonic peizo to make sense of the traces)

 

[miskol]

ultrasonic transducer circuits

i will get back to u soon. yes it has been long left as it is, and it is very interesting i agree.

 

[kender]

impedance matching ultrasound transducer

Pretending to be dumb, I’ll will make some of my own assumptions. After skimming through your post and looking at the pictures, the latter look okay.
Left. The first burst is TX, then a little bit of ringdown, then switchover (not shown where exactly), then RX sees the echo.
Right. I assume that the probe is hooked-up to RX. The little squiggle in the beginning is feed-through from the TX transducer. Afterwards RX sees the echo.

 

[vsmGuy]

ultrasonic transducer

@kender : I am sorry if this does not ring a bell, but did you happen to mail me something back in say 2007 or so about making simlar sonars?

If I indeed recalled correctly, then what happened to the project? Did it work out?

I will take up to build a nice sonar ranger - I have tried and spent some considerable amount of money behing Maxbotics and SRFs. Given, the Maxbotics work better than the SRF and the designer also included a fresnel lens (if you look closely at the black cap on the sensor) it's still way too expensive a pop.

Since you'r in the bay, you can drop by circuitcity and have a look at the physical layout - get a "feel of it".

Now regards the traces, the left oscillation is the ringing of the transducer - this however gets damped considerably (right trace) due to high loading of the "direct connection".

Don't worry much about the noisy traces - I believe the op is a student perhaps and did this on a breadboard or a breakout board.

 

[miskol]

reducing ringing, ultrasonics

CH1 is amplified so that the transmitted pulse sent to the transmitter have stronger pulse. it's like giving volume boost.

The ringing problem that u all have discussed above, can be reproduced when the cable from the transmitter and the receiver is crossed with each other, or twisted together. so to avoid this problem, i've used coaxial cable.

Which is why i suspected that the internal layout of the switching IC, the inputs for both Tx and Rx is very near each other that have make the "noise" i mentioned earlier.


attached, i include more pictures for more clarification.

Please, correct me if i'm wrong. Thank you.

 

[arthur0]

transducer tx rx switch

Hi every1,

I worked with (read “built and wrote software algorithm for”) a commercial depth transducer (working on a water depth range of 0.4 – 100m) and I had my share of problems, some of which I still struggle with.
It would be nice to hear and share some of that experience with you guys.
For instance, I can say that the [after-]ringing is the limiting factor of how short a distance I can measure, and to minimize that, I used a couple of anti-phase pulses at the end of an excitation burst (which is normally at the natural oscillation frequency of the crystal, in my case 200kHz).
There’s no TX/RX switch in our (is not really mine) design, but rather a “muting” of the sensitive reception path (that has quite a bit of amplification).
As wsmGuy pointed, having the “load” on the crystal will help reduce ringing.

Arthur

 

[miskol]

ultrasonic transducers rx/tx set pdf

Hi miskol,

I really don't think what you term as "noise" is really that - infact, its due to considerably less loading of the transducer.

I assume you had the sensor setup in a reflective setup with an obstacle very close to the transducer - maybe 1 - 2 cms away (varies on temperature).

i'm using a Sample and Hold method, to sample the ultrasound signal. This method doesnt care about the "noise".

However there's a also a simpler method to sample the ultrasound signal, which is using peak detection(but i'm not using this). This method however must have that "noise" removed or not it will sample the wrong peak on the ultrasound.


yes, the sensor is setup as shown in "Wire crossing" image. A coupling agent are placed between the sensor and a container. ultrasound are sent to the other sensor on the other side of the the container that contains liquid.

 

[kender]

layout do tx e rx

For instance, I can say that the [after-]ringing is the limiting factor of how short a distance I can measure...

I agree with Arthur. [after-]ringing (aka ringdown) should be considered. Here's an example of what ringdown may look like: http://www.prolifictec.com/filehost/Ringdown_waveforms_rev01.pdf. I've recorded these waveforms. I wonder what do you think about them.

@kender : I am sorry if this does not ring a bell, but did you happen to mail me something back in say 2007 or so about making simlar sonars? If I indeed recalled correctly, then what happened to the project? Did it work out?

I've built an tested all components for that project, but didn't integrate them. That project was shelved. I've been overwhelmed with a high-maintenance girlfriend and fast-revenue work for a good portion of this period. She's no longer around. I'm planning a quasi-sabbatical. Maybe I'll pick this project up again.

- Nick

 

[vsmGuy]

sonar signal agc circuits

CH1 is amplified so that the transmitted pulse sent to the transmitter have stronger pulse. it's like giving volume boost.

It's always a better choice to keep the transmitter as simple as possible. Its the reciever that needs to be fiddled with.

Boosting the excitation signal makes matter worse by making the transducer ring, reducing the effective min. distance.

The ringing problem that u all have discussed above, can be reproduced when the cable from the transmitter and the receiver is crossed with each other, or twisted together. so to avoid this problem, i've used coaxial cable.

Which is why i suspected that the internal layout of the switching IC, the inputs for both Tx and Rx is very near each other that have make the "noise" i mentioned earlier.

The main thing to understand is this: All that matters is the coupling of the signal. For best performance, you need to mount your transducers on a felt padding (physical coupling) and make sure that the signal i/p and sensor o/p wires do NOT come near each other(electrical coupling).

The point you raised about the analog multiplxer coupling signals - we need to ask the more experienced members in this forum as I HAVE NO IDEA but think it to be unlikely, though possible because of the high impedance of the traces in the CMOS..

Anyways, can you try the suggestions we have given you and post back what is the:

1. Min distance you can measure reliably
2. Max distance you can measure reliably
3. Resolution

I believe Nick and I would take this project up as well? Nick?

 

[kender]

ultrasonic rx tx

Boosing the excitation signal makes matter worse by making the transducer ring, reducing the effective min. distance.

On the 1st order, the ratio of ringdown power to TX and reflection (RX) power should remain constant with respect to TX power. However, boosting the TX will increase the SNR of the reflection (RX). That's a good thing.

I believe Kender and I would take this project up as well?

I was building building a sonar for water. Something similar to a business end of a fish finder, but more flexible and with a capability to array a bunch of them.

 

[FvM]

ultrasonic rx-tx

I see the problem, that most questions raised in the discussions have quite different answers depending on the application. Unfortunately, miskol didn't mention his application.

- a RX/TX multiplexer switch is only suitable with low transducer voltages.

- transducer Q selection involves a trade-off between sensitivity and bandwidth, ranging from aperiodical to resonant. Generally, low Q is difficult to achieve, requiring a special tranducer design. As already mentioned, the transducer media coupling also has a strong effect. Unfortunately, it can't be influenced at will in some applications.

Some application parameters that should be known:
- distance/time delay measurement range
- intended resolution
- expected echo dynamic range
- media and transducer properties

 

[vsmGuy]

sonar transducer rx after tx

I might be paranoid, but considering this to be your term project, I would like to see the following "cosmetic" changes:

1. Please follow convention while naming nodes/pins. For example, it neither helps me or you if you name Vdd as Vin.
I was, in fact, under the assumption that you were feeding some sort of input signal to this node!

2. The reason why your circuit is noisy is because you have to connect all unused terminals to Vss.
One of my friends suggest terminating them to Vdd with a 10k resistor, although I feel that's being too paranoid.

Similarly, why are opamp pins 1 & 5 floating?

Could you rectify these omissions? If you feel I am wrong, do give valid reasons.

The noise should reduce a lot I assure you, once you do this.

3. What is Tx1 (non-inverting input signal) ?

4. Are you using a breadboard/veroboard/etc for prototyping?

 

[miskol]

ultrasound transducer driver circuits

No problem, i'd like to learn more about my mistakes for improvement.

1. I agree with that, it sure gives problem about the labeling. i labelled my circuit Vin is because Vin is actually Vdd for the opamp, and the Vdd are sourced from another development board. the dev board labeled it Vin, so for me to easily troubleshoot things, i labelled it same with the dev board

2. I'll try that and ground all the unused terminals of the switching IC.

Both pin 1 & 5 are for input offset voltage nulling. "If external nulling is not needed, the null pins maybe left unconnected." i followed the quote from the datasheet of the opamp and left it unconnected.

3. Tx1 is the transmitter pulse sent from the microcontroller to the opamp. the output of the opamp will give a higher pulse.

4. Testing the switching circuit, i'm using breadboard. the transmitter pulse circuit have been fabricated on a FR4 board.

 

[FvM]

what is rx and tx

I realized, that part of my questions has been answered with the drawing of the measurement setup. Considering this additional information, I think, that the waveforms shown in the original post are basically representing normal operation. The only detail, that can be most likely easily improved, is the unwanted crosstalk in the direct connection circuit, because it's a pure electrical effect. But as you are targetting to a single transducer design, it's probably less important.

The transducer ringing in switched operation can't be easily reduced. As said, it depends on transducer Q, media interface and (partly) transducer impedance matching.

In the shown circuit parts, no impedance matching elements are present, so there's possibly room for improvements. But the transducer impedance parameters must be known to decide about.

 

[vsmGuy]

ultrasonic

@FvM, exactly! You know your stuff.

This is the point I have been trying to make from the very beginnig - the cuitcuit operates JUST as expected!

However, there are a few things that can be improved:

1. Transmitter drive
2. Reciever AGC based on time (yes, sound degrades at an exponential rate, besides being absorbed)
3. Reducing ringing by driving antiphase

But as you can appreciate, these are EXTREMLY circuit and sensor specific, so unless the OP can find time to markup his schematic in a more conforming way and upload the schematics then, there is very little we can do.

@miskol, there are few room for improvements and a lot more for your specific case.

We need to know:

1. Impedance of the sensor (The range is ~2nF with 550E @ 40Khz, but you need to find this out)

2. Can you design an antiphase circuit to reduce the oscillations of the transducer?

3. Have you modified the schematic to the best you can? If naming a standard signal differently helps you - it's OK, but it's something you need to document on the schematic.

This is going to take a lot of time and effort but the outcome will be a project better than the Maxbotix, so we need you to upload the project to EDABoard as sensors and their conditioning circuits are very specific things - there are no general solutions!

 

[miskol]

switching transducer

1. The sensor's nominal impedance is 1k ohm at 335kHz
Capacitance at 1KHz (pF) = 250±20﹪

2. I've never heard of an antiphase circuit. Is this for the receiver part or for transmitter part. Anyway, i will take a look on it more on the web.

3. No, i believe what i have now is not the best yet. Like FvM mentioned, i have not yet control the impedance matching. this circuit are used for my old transducer which have nominal impedance of 10k ohm @ 40kHz.

thank you for all the help and ideas. i sure will need more time to implement them. will report back when i've understand more on the subject and have produced better results. thank you.

 

[vsmGuy]

circuito tx de ultrasonidos

miskol, what I said was, what you have is "good enough" considering the information we have.

To match impedance we need to know a lot of things, like excitation freq, capacitance of the peizo, power rating, resonant freq etc. It's very sensor specific.

My idea was to take forward this project with you and Nick - I am sure more would have joined along the way. With you having done the basic design, it makes sense for us to invest time and effort in making it better, and that is the only reason why I had been interested in this topic.

Your intial design would be the base for a much better one (we would add SPI, I2C, AGC, tuned LC tank for the peizo at the very least)

Getting a basic sonar setup working does not take much effort, but making it precise/or sensetive is VERY hard work.

 

[kender]

how to match ultrasound transducer lc cicruit

Speaking of AGC. Here's a schematic for variable gain peak detector: http://www.prolifictec.com/filehost/programmable_gain_peak_detector_rev01.pdf. It can be used as a building block for a sonar AGC.