Got my ultrasonic sensor transmitting but not receiving?

Wireless & IoT
1

I’m trying to finish the final leg of my little transducer module but cannot get it to work.

I have a 40kHz transducer transmitter/receiver. I have generated a 40kHz frequency to pulse it and send out the ultrasonic no problem, and have tested this by using a unit I made using a EZ1 sensor that reports distance. When I point my transducer at the EZ1 its distance goes down to 6-7 inches when I point it at it, and when I cut the power it reports the correct distance, so to that effect I am happy to say it is sending ultrasound.

Now to receive is a whole new ball game. From what I can best understand, ignoring the entire transmit then receive circuit, and having just a simple receive only circuit (I will send the sound from my EZ1 sensor), I should just connect one leg of my sensor to ground, and the other leg should pick up the pulses, correct?

So, to that effect what I have is the following:

http://img5.imageshack.us/img5/8105/capturept.jpg

This is a LM324 amp, with a negative feedback loop and a theoretical gain of 1000x, although whether this will do that is another matter as I believe the 1.3Mhz bandwidth means my frequency of 40kHz means the cut-off will be 32.5 gain, not 1000… anywho, carrying on.

With my amp set up like so, if I connect my sensor to the IN pin, and the other leg of it to ground as mentioned, with or without a decoupling cap, the amp always outputs Vcc in all cases.

I thought the idea was if my sensor is connected to ground and to IN, that if nothing was being received the amp should see ground from my sensor, instead I presume it is floating and so amplifying the floating difference by 1000.

I have only had an hour on this bit and will spend more time on it tonight, but just thought I’d ask if I am already doing something obviously wrong here that will prevent my trials tonight from being a waste of time?

My main fall back is I do not know what I am expecting to see on the sensors leg when it is hit with ultrasonic sound, as even with my scope set to 5mV divisions it sees nothing. If I need to amplify the signal 100s of times to get it to 3-5V then I would of thought I would see at least 5mV ripples or something.[/list]

2

Ok done lots more testing. Here are the results:

  • Changed the circuit added an extra res to ground to tie the + to ground until a signal is received. So now the amp stays at 40mV until activity on the + line.

http://img154.imageshack.us/img154/9255/captureag.jpg

  • The amplifier definately works, tested with multiple resistors to change it is amplifying the signal by the percentage. With a R1 of 10k and an R2 of 10k I get approx 2x amplification. With the current 1K and 1M setup I get high amplification from .5mV going to 100s of mV.

  • I have tested that the amp can take 1mV to 3.5V by putting a DC 1mV on the + line and I get 3.5V out of the amp out.

  • Analysing the sensor lines with my scope I get the following results:

With no sound the output is steady at ~0.

When I put the transmitter on front of it I get a wave output at a peak of 12.6mV peak to peak at 40kHz and a 2.7mV RMS, over the 10 pulses, so just as expected.

  • With a capacitor in line the scope only picks up the pulses when it is transmitted (every 50mS). Without the cap the scope picks up an almost continuous pulse wave.

  • I have tried by putting the original 40kHz frequency on the + line and it successfully drives the amp also.

  • I have put the sensor where it should be with the line in the +, and with the scope and everything connected, I still get the 12.6mV peak to peak and 2.7RMS voltage on the + line, so the amp should output 3V+, yet it does nothing.

So… the amp works in ampifying the signal when its:

  • a DC as low as .1mV

  • a 40kHz frequency driven at 3V peak to peak

Yet it doesnt work when driven by:

  • a 40kHz frequency driven at 12.7mV

Anyone got any suggestions?

3

Nobody? I’ve had 3 nights on it now and cant understand why it worked on every signal but the one I want.

4

I’ll send you a detailed response in a couple of hours. I meant to send it earlier but had to go to work…

5

There are several problems I see:

You are asking too much of the LM324. It has a gain-bandwidth product of 1MHz, which means the maximum gain you can expect at 40KHz is about 25. You can’t expect a gain of 1000X from a single stage. You could try using multiple (2 or 3) stages to increase the gain. Also, even if you could get a gain of 1000 out of 1 stage, consider that the Input Offset Voltage (2mV) of the LM324 would lead to large DC errors on the output (the error will be multiplied by the gain).

I would suggest using an op-amp with a better gain-bandwidth specification, and using at least 2 stages. Also, you should build a 40KHz bandpass filter into the 2-stage op-amp circuit, to give better rejection of unwanted noise. There are plenty of websites that deal with filter design.

Another consideration is that you are trying to amplify a signal below the negative supply voltage of the op-amp (0V) and it can’t do this. You would need to either add a negative voltage supply or bias the input voltage using a resistor divider.

See attached circuit that shows what I mean. It uses 2 stages to give a gain of about 1000. Note that I have not incorporated the bandpass filter - you’ll need to work that out if you want to implement it. Capacitor C1 is just there to provide a nice stable “ground” reference value for the positive inputs of the op-amps. C2 couples the signal from the transducer. The circuit may be OK without it, but I’m not sure how the transducer would behave with a DC bias that would result without C2.

http://users.adam.com.au/mnoble/ULTRASONIC1.JPG

6

Just trying your design now but the problem its always nothing.

PS I’ve just entered 50V in the sim so I can see a better range without decimal places:

http://img182.imageshack.us/img182/1559/capturelq.jpg

7

Ah my mistake I left off the capacitor to the sensor prevents the return DC from the amp from being sunk to ground. Solves the entire problem!

So now my output is offset bias. Brilliant. Should be able solve it from here now thanks a million.

8

Where would I be now without this scope!

OK build the circuit up for real, left my scope on the 2 pins of the sensor. Grabbed the first resistors I had to hand to get a voltage divider and for the gain dividers. Soon as I powered up was getting 17mV ripple. Took me a bit to click in but then I realised I had grabbed <1K resistors for the gain dividers 22 ohms and 615ohms. I remembered amp inputs need high impedence, but at least now I’ve seen first hand why. Could anyone explain a bit more about why lower impedence causes noisy feedback?

Anyhoo… swapped them out for 10k and 99k (all I have at mo) and the noise went away. So now I should be expecting 1+(109/10) approx 12 * 12 = 144 gain over 2 stages. So a 3mV RMS should give a good half volt gain + the offset of 2.5v so 3v. Enough to pick up for this test just to show its amplifying.

Spent about 2 hours scratching my head. The signal was going in from the sensor at 3mV RMS or more (sometimes as much as 9mV), yet the output was sitting at 2.5v not budging. Tested with the 40kHz 3V driver, and the DC as I did before and they both worked… then it hit me. I was looking for a voltage gain on the output with a multimeter! Although this would of been fine were I sending sound out constantly, I was using the EZ1 sensor that only pulses once every 50mS, so there would be no chance the meter would pick up the pulses.

Quickly and rather excitedly pinched my scope onto the output line, tried again and woooo! There is was, a good strong 3.1V peak to peak spike over 10 pulses. Fab. So I conclude thats working now as the PIC chip comparator should pick that pulse up fine yes?

Any final comments on changes before I make a board up for this or is it good to connect to a sensor and the output straight to a comparator?

Finally, I will be using a capture and compare module to generate a PWM at 3V with duty cycle to give an output between 0-3V. This will then be sent down 10 meters of cable. Would yo recommend I pass this PWM output through a unity gain amp first, as its only been driven by the PIC?

9

Good to hear you had success.

As discussed, if you incorporate a 40KHz bandpass filter into the amplifier stages you will have better immunity to noise. Maybe you don’t need it, but 10m of cable has the potential to pick up noise (you’re using coaxial cable I guess?).

The PIC will have plenty of drive current to drive the transducer without amplification (the current drawn by the transducer would only be 1mA or so if driven by a 3V sinewave - obviously it will be a bit different for the rounded squarewave that it will actually receive). I would only add an amplifier if you wanted to increase the drive voltage.

I would suggest adding additional protection to the PIC pin that is exposed to the “outside world” (this is always a good idea). A resistor (100 ohms or so) between the PIC pin and cable, with a pair of schottky diodes to clamp any spikes to the positive and negative power supply rails should do the trick. The PIC has internal protection diodes, but they can only hande limited current, so the extra protection is advisable.

If the voltage is at a suitable level, you should be fine to connect the amplified signal straight to the PIC’s comparator. I can’t remember what PICs have, but some TI micros I’ve used have the very nice feature that you can adjust the threshold voltage of the comparator using an internal DAC. You could do a similar thing by connecting a PWM output of the PIC to an RC low-pass filter, and connect this to the negative input of the comparator. Adjusting the PWM duty cycle changes the comparator threshold voltage.

10

I am 27 and I am still studying about ultrasound repairing. I am in Ultrasound Repair Training near in our place. My wife is still ongoing check up with our baby in her womb.Ultrasound Repair Training [SEO Services](http://computerseo.com/)

11

I have found the one site which explain the ultrasonic sensor transmitting system and also provide the solutions of some basic problem, Just check it http://www.sensorsmag.com/sensors/…ul … rement-838 I hope this site will solve your problem. [SEO Expert](http://www.triplezseoservices.com/hire-dedicated-seo-experts.html)