Ultrasonic distance detection - troubleshooting

[Jester]

I'm troubleshooting a new design and don't understand some of the results. Perhaps someone has some insight?

This is the circuit:


A short burst (3-4 cycles) of 40kHz is applied to the transducer via U4, the mux is then switched to listen mode.



The transducer rings for about 600uS after the burst completes (as expected and seems reasonable)
With nothing for the burst to reflect off, I would expect the receive signal to remain quiet after the initial decay, however a signal is detected about 1.4ms after the initial pulse, I have no idea why this is happening.

Any ideas?

 

[emresel]

how about if you remove the transducer from the cct, is it behave the same?
Is 40khz is switched off from the cpu side? Just speculating may be corsstalk within 4053 and it is amplified later?

Good luck

 

[Veriloger]

What medium is the ultrasonic transducer in? Try calculating the time of flight for the signal to travel down the cable and back? Does it match the 1.4ms? Ultrasound travels in all directions, the housing attempts to dampen it in all directions other than the face. It could be reverberating off the housing or a nearby object that isn't in front of it. Is the transducer focused or unfocused? If unfocused, again, it could be reflecting off something unexpected.

If this is an ultrasound reflection, it's quite a strong reflector and i would hope that you know about the reflector.

What is DG4053? A MUX? Are you using it in place of the usual Transmit/Receive Switch used in ultrasonics? I'd be concerned using an DG4053 before an ADC as i imagine that you'll lose valuable information.


Why does the RX sit at -2.5V after being driven? I would have expected that it would return to 0V.

 

[esp1]

hi,
Looking at the Transducer signal, it creates an increasing negative voltage offset during the ringing period.
When this offset becomes great enough the opa/comparator is biased Off to a point where it no longer detects the damped signal.
At about 1mSec the negative offset reduces and the ringing signal again falls within the opa/comparators switching range, causing the output at 1.4mS.
If you notice after 1.5mS the RX signal is continuing to decay.

As a test I would suggest you ping the transmitter at 1 second rate and re scope and post the waveforms.
E

 

[Audioguru]

A DG4053 is a TPDT Cmos switch made by Vishay. Why don't they use the same CD4053 letters like everyone else?

 

[Jester]

What medium is the ultrasonic transducer in? Try calculating the time of flight for the signal to travel down the cable and back? Does it match the 1.4ms? Ultrasound travels in all directions, the housing attempts to dampen it in all directions other than the face. It could be reverberating off the housing or a nearby object that isn't in front of it. Is the transducer focused or unfocused? If unfocused, again, it could be reflecting off something unexpected.

If this is an ultrasound reflection, it's quite a strong reflector and i would hope that you know about the reflector.

What is DG4053? A MUX? Are you using it in place of the usual Transmit/Receive Switch used in ultrasonics? I'd be concerned using an DG4053 before an ADC as i imagine that you'll lose valuable information.


Why does the RX sit at -2.5V after being driven? I would have expected that it would return to 0V.

What "usual Transmit/Receive switch" are you referring to?

 

[Jester]

What medium is the ultrasonic transducer in? Air Try calculating the time of flight for the signal to travel down the cable and back? Does it match the 1.4ms? , no cable, just 2 cm of wire, so no Ultrasound travels in all directions, the housing attempts to dampen it in all directions other than the face. It could be reverberating off the housing or a nearby object that isn't in front of it. Is the transducer focused or unfocused? If unfocused, again, it could be reflecting off something unexpected.

If this is an ultrasound reflection, it's quite a strong reflector and i would hope that you know about the reflector.

What is DG4053? A MUX?, YES, Are you using it in place of the usual Transmit/Receive Switch used in ultrasonics? YES, what "usual" Tx/Rx switch are you referring to? I'd be concerned using an DG4053 before an ADC as i imagine that you'll lose valuable information. I don't follow, why would I loose valuable information?


Why does the RX sit at -2.5V after being driven? I would have expected that it would return to 0V.

I resolved the "phantom echo"
Thanks for your thoughts

- - - Updated - - -

hi,
Looking at the Transducer signal, it creates an increasing negative voltage offset during the ringing period.
When this offset becomes great enough the opa/comparator is biased Off to a point where it no longer detects the damped signal.
At about 1mSec the negative offset reduces and the ringing signal again falls within the opa/comparators switching range, causing the output at 1.4mS.
If you notice after 1.5mS the RX signal is continuing to decay. Great observation, I decreased gain slightly and the phantom echo is now gone, thank you.

As a test I would suggest you ping the transmitter at 1 second rate and re scope and post the waveforms.
E

Next problem is to try to get the transducer to stop resonating more quickly. I tried a parallel resistor and that helps, but it also reduces the Rx signal level. I'm pondering reversing the phase of the last cycle of the transmit burst, thoughts?

 

[esp1]

hi,
Transducer ringing is a common problem.
It is possible to reduce the ringing time by either damping the transducer or using timed gain control [TGC]
Immediately after the TX pulse period, you could use a timed period pulsed clamp circuit.

The TGC would increase the gain of the receiver with time after the TX pulse, make it an adjustable gain rate.

Another option is to Gate the RX pulse reception after the TX pulse.

What working ranges are you trying to achieve, minimum, maximum and which types of reflecting surfaces

E

 

[chuckey]

This topic has been on this forum, within the last month. Try the search facility.

Frank

 

[Audioguru]

This topic has been on this forum, within the last month. Try the search facility.

Frank

Same class but this student is later?

 

[Jester]

hi,
Transducer ringing is a common problem.
It is possible to reduce the ringing time by either damping the transducer or using timed gain control [TGC]
Immediately after the TX pulse period, you could use a timed period pulsed clamp circuit.

The TGC would increase the gain of the receiver with time after the TX pulse, make it an adjustable gain rate.

Another option is to Gate the RX pulse reception after the TX pulse.

What working ranges are you trying to achieve, minimum, maximum and which types of reflecting surfaces

E

To dampen the resonating transducer, I tried driving the transducer with a signal 180 degrees out of phase with the initial burst. This helps, I reduced the resonating period from ~ 1ms down to < 400us. My initial idea was to simply drive the transducer exactly 180 degrees out of phase with the original signal, however this did not work very well as the countering signal alters the frequency of the resonance (frequency increases), so I simply worked my way through the waveform a bit at a time driving one way or the other to drive the signal back to ground. Each ping of the transducer is not identical, however largely repeatable.

Air transmission, hard surfaces, I'm gating the receiver, works well now and can measure less than 6"

Thanks for your help.

- - - Updated - - -

This topic has been on this forum, within the last month. Try the search facility.

Frank

I did search, I found a few threads related to terminating and shorting the transducer to reduce resonant period, I did not see anything related to canceling the resonating transducer with a phase reversal. So what is your conclusion on the most effective method?

 

[FvM]

I don't recognize at first sight how driving the transducer with a phase-reversed burst could change the resonance frequency. Of course if you change the termination the resonance frequency will be affected.

The main limiting factor for the echosounder operation is the small transducer bandwidth which affects both the trasnmitted and the received signal. For best time resolution, it's essential to terminate the transducer with a resistor according to the real part of it's resonance impedance, usually in series resonance. And don't change the effective termination during TX/RX switching. In the #1 circuit, which implements short circuit termination in TX and RX, one would add a series resistor to the transducer to increase the bandwidth.

I believe that a burst with phase reversal can help to speed up the transducer ringdown, I would consider it as a simple case out of a more general class of complex transmitted signals, which also includes frequency modulated signals like chirps.