Ultrasound

Cenlasoft

Hello,
I am trying to use ultrasound with the prop to measure defects in materials. My circuit is simply an waterprooof ultrasonic transducer tied to one prop pin. I have that already contructed and tested with the singnal generator and scope. My question is that in order to migrate the project to the prop I may need to protect the prop pin. I want to:
1. make pin an output
2. send a 40 khz pulse with a width of.5 ms and an interval of 20ms. do I use the synth object or is pwm better?
3. next I make the pin pin an input and listen for the income reflected wave. I will be using an opamp for amplification so do I have to add circuitry to protect the prop pin?
4. I will be using Viewport with my initial experiments.

Any help or suggestions would be appreciated.
Curtis

Mike Green

You're probably not going to get very far with that design for several reasons ...

1) Ultrasound transducers like the one you describe work better with higher drive voltages than the 3.3V available from the Prop pin. Here's one example of a PING-like DIY distance sensor. It uses an H-bridge run from a minimum of 5V to provide effectively 10V Peak to Peak to the transducer.

2) If you want to use a single transducer, you'll have to have some kind of interface circuitry that will allow for relatively high transmit voltages yet have a sensitive input circuit with rapid switching back and forth.

3) One of the reasons for using two transducers is that the transmitter will ring for some time after the transmitted pulse is sent and this will interfere with reception. Solids have a much higher speed of sound than air and you won't have much time from the leading edge of the transmitted pulse until an echo comes back.

4) The transducer still produces its sound pulse in air and that has to be coupled to the test material. You're going to have substantial losses in going through the various interfaces (transducer to air to membrane to gel (for coupling) to test material, then back through the same interfaces. You might do some research first in how commercial ultrasound devices work.

5) You can help with the applied voltage to the transducer by connecting the transducer between two I/O pins, one made high and the other low reversed at a 40KHz rate. You'd have to modify either the Synth or the PWM object to use two output pins.

Cenlasoft

Thanks Mike,
I see what your saying. For my project to work, I must use one sensor as tx/rx because of the area I want tested. How about this:
Use a 555 time (monstable) and clock it with the prop pin. The timer circuit could be tuned to 40 khz and a higher voltage could then be used for the transmitter. The receiver could be another prop pin via an opamp. There maybe timing issues.
Or maybe this:
Use an series LC circuit tuned to 40 khz triggered by the prop pin at 40 khz. I done this before with my eddy current project. The series LC produces high voltage.

Thanks,
Curtis

Mike Green

The DIY ultrasound device described at the link I gave uses an H-bridge running off 5V and I think the PING does the same sort of thing. 10V p-p should be enough unless you need a lot of power. With the H-bridge turned off, you might be able to use an op-amp across the transducer operating differentially. If you need more power, it would be easier to just use a 9V or 12V supply voltage for the H-bridge. You'd need to use an op-amp for your input stage that runs at the same voltage (or higher) and can withstand input voltages equal to the supply voltage.

You could easily use a timer as you suggest and run it at 12V or higher. You'd still need an op-amp that would run at the same voltage. There are lots of them designed for automobile use that would work.

Phil Pilgrim (PhiPi)

A MOSFET driver family, such as the MIC4223/4/5, would be an ideal candidate for driving the transducer. It will handle the capacitive load of the transducer, can drive it differentially (using a single Prop pin with the '4225), and can be tri-stated during the receive phase.

There is zero need to involve a 555. The Propeller can handle the signal generation part with ease.

-Phil

Cenlasoft

Thanks Mike and Phil,
I think I'll use the MIC4225. I just ordered a few of them. I'll do some reseach on the MIC4225. Phil, is there any prop circuits that use this IC so I can see how its used? Mike, Thanks for the link also, I'll try that too. Any other suggestions are appreciated.
Thanks,
Curtis

RobotWorkshop

You may also want to review how the Polaroid (Senscomp) board works. You can see a schematic for their board below:

http://www.senscomp.com/specs/6500%20module%20spec.pdf

Some links that talk about hacking an old Polaroid camera for the sensor and board:

http://www.seattlerobotics.org/encoder/200010/dlcsonar.html

http://www.prc68.com/I/PSOSC.shtml

http://www.micromint.com/app_notes/ti01_data.htm

You'll still want to take precautions connecting the board to the Propeller pins (at least consider other 3.3V/5V interfacing)

Robert

Phil Pilgrim (PhiPi)

Curtis,

I've only used the MIC4469, but it does not have an output enable. There's really not much to the circuit. Just connect a Propeller output pin to INA and INB on the '4225, and another pin to ENA and ENB. You can use a counter output to drive the output pin and just leave it running all the time. Use the enable output to form your pulse, being careful to monitor the counter output so the pulse begins and ends just after an output transition.

One thing that's very important is to connect logic ground and the ground for your transducer supply at just a single point, i.e. the MIC4225's Vss pin. That way you will avoid current surges in your logic ground traces when the transducer output changes states.

There could possibly be a case made for using two Propeller pins for the output function. By setting both to ground after the pulse and before tri-stating, it might be possible to quench the transducer's ringing more quickly. That's just a guess and maybe something you will want to experiment with. The other thing I would do is to bias one of the transducer's pins, using a resistive divider, to half the voltage of your op amp's supply. That way the return signal that the op amp sees will be centered in its common mode range, and you will also protect the '4225 from negative excursions -- or at least prevent it from clipping the return signal to ground.

-Phil

Phil Pilgrim (PhiPi)

Robert,

The Polaroid sensors are electrostatic and need about 300V to operate. Curtis's transducers are of a different sort.

-Phil

Tubular

Curtis, what sort of voltage were you able to get out of series LC tank?

I'm looking at some transformers to do something similar at the moment.

frank freedman

You will probably want to consider two pins minimum for this depending on how much of a dead zone you can tolerate for your project. While the details are sparse, it sounds like you are trying to accomplish the same effect in a material as a TDR would in a wire/fibre etc. Mike mentions a number of transitions in his post, but bottom line the whole transmission will be the propagation through the interfaces and the reflections from each. If you use only one pin for TX/RX, then the setup time will end up giving you a finite amount of dead time between TX/RX. That time coupled with the speed of the pulse through the material will guarantee you a minimum dead zone (depth) in which you will not be able to detect a reflection from a subsurface flaw.

It may be worth considering using multiple pins or if fast enough a serial interfaced A/D converter. If you only look for a pulse of a given size, you may get inaccurate indications from various reflections as the pulse transits and echos off of internal structures. Going analog would allow you to capture the whole stream and enable you to possibly "see" differences in the echos and possibly enable you to code up a discriminator routine to help distinguish the echos produced by a flaw as opposed to all the other mess that would be generated from the pulse.

Cenlasoft

Tubular,
With a series LC and at resonance, I have seen (3.3 V input) 30 to 40 volts AC. @Frank: Thanks Frank, I'll research this methodology. Thanks to everyone thus far. My goal is to create an ultrasonic system to recover hallmarks (worn off inscriptions) on cultural materials. I have suceeded in doing this on metal objects with my eddy current analyser (using the prop) but, I need to do the same thing on non-metalics.
I will certainly post the project with the source code and schematics when done.
Curtis

prof_braino

Mike Green wrote: »

3) One of the reasons for using two transducers is that the transmitter will ring for some time after the transmitted pulse is sent and this will interfere with reception.

Can this "ringing" be dampened by sending an out of phase pulse to the transducer? Or, can it be dampened by shorting the transducer leads together the way "motor brake" is done on electronic speed controllers in RC motors?

Mike Green

I suppose that theoretically you could produce an out of phase signal that would reduce the ringing, but you'd have to be able to modulate the voltage or you'd just introduce more ringing. If you've got that much control of the transducer, you might be able to shape the transmitted pulse to reduce ringing in the first place.

The ringing is mechanical, in the shell / diaphragm / etc. of the transducer, not just in the piezo element, so shorting the transducer leads isn't going to help much. I don't think a piezo element has as much "gain" over an electromagnetic motor (voltage out vs. mechanical force), so you're not going to get as much dampening.

Kevin McCullough

Hi Curtis,

The higher the Q of your tank circuit the higher voltage that is generated between the elements. However, higher Q also means the circuit will take longer to "ring" up to the full potential, and the longer it take to "ring" back down (when turned off).

I would really suggest looking at a small transformer as the drive circuit. The propeller pin (or two pins as a differential signal for extra drive strength) can drive the primary of the coil. This circuit can be more quickly turned on/off and allows you very accurate control of your output drive peak-to-peak voltage (because you know the input drive and you know the number of turns).

Cenlasoft

Kevin,
What type of transformer would you suggest? Would I have to buffer the pin also?
Thanks
Curtis

Tracy Allen

Agreed that a transformer is a possibility, however, transducer impedance (high C) will interact strongly with the transformer impedance. The best transfer will occur at resonance, and needs to be visualized and developed with the help of an oscilloscope. Waveforms output from a transformer can be knarly when driven near/off resonance with a square wave. Little audio transformers from Xicon are rated linear only to 4kHz, but they do work for this kind of thing at higher frequencies. (for example, 42TU048-RC from Mouser, 48Ω to 8Ω, turns ratio 2.45).

Randomly cutting off the current will result in a flyback effect, which might be used to squelch the resonance and would best be timed to coincide with zero flux in the core. Be careful with the flyback and its effect on the driving circuit. All in all, the mosfet driver from a higher voltage supply would be simpler.

Curtis, I really admire your eddy current sensor. Will the wavelength and potential phase and time differences be suitable for that sort of thing with 40kHz ultrasound? I don't know. It seems like the features of the hallmarks will be very small. Medical ultrasound is usually done in the hundreds of kHz range so as to distinguish small features
.

Cenlasoft

Thanks Tracy, coming from you means alot.
I thought that the 40 KHZ would be too low, but I chose it for my proof of concept. I have some "us" 40 khz and 60 khz lying around. If this seems viable, I will go up in frequency.
Thanks for everything and I'll be experiment tonight and will post progress.
Curtis

Cenlasoft

@Phil
Earlier in this post, you had suggested using a mic4225 to drive my transducer and do the switching. I understand most of what you said about the connections, but do I connect one prop pin for each: INA, INB, ENA, and ENB? Also I could not find much on this ic. I have the datasheet but, it does not give much application data. Any help would be appreciated. I have it working with a h-bridge, but since I got these ic's in, I would like to experiment.
Thanks,
Curtis