Unwired input measures output signal

[David_Pastorino]

Hello !

I am using a myDAQ system and found that the structure of my program may be wrong to go on with the project:

I am emitting waves with a piezoelectric. Originally, I had connected the output AO (AO 0 and AGND) to my first piezo and can emitt succesfully my waves. The idea now is to record the signal emitted by piezo1 and the signal received by piezo2, 10cms away from the first one.

I wired piezo1 to the first analog input (AI 0+ and AI 0-) and piezo2 to the second analog input (AI 1+ and AI 1-). I received many interferences and strange behaviour.

For instance, by hitting the finger piezo2, I record the vibration well, but hitting piezo1 generates a signal in piezo1 (as it should be) and piezo 2 also (even if the piezo is not emitting).

Thus, I believe I have a wiring issue.

The next step was to wire only the output. I generated my waves (short pulse every 2s). Reading the analog input 0 or 1 allows me to see the pulses (and noise) without wiring anything. Is there a solution to avoid this interference?

Finally, wiring piezo2 to analog input1 allows to record any finger "hit" on piezo2 but strictly no signal of my waves (6kHz).

Do you have any advice on how to manage this issue?

Thank you,

David

[johnsold]

David,

You are using differential connections with an essentially capacitive source. Do you have appropriate bias resistors connected? Read the manual for your device on how to connect differential sources.  Make sure you have this done before trying enything else.

Piezo devices are usually high impedance (and capacitive). Many DAQ analog inputs use multiplexers to switch inputs to a common A/D converter.  When driven from high impedance sources these multiplexed inputs can experience what is called "ghosting" where a signal from another input appears superimposed on the signal from the high impedance source. From your description I suspect that this may be what you are seeing.

The best solution is generally to provide a low impedance driver for every input. A unity gain buffer between the piezo device and the DAQ input may work.  Something which sometimes works (but may not in your case) is to scan four channels and ground the inputs on two of them.  Wire Piezo1 to AI0, ground AI1, wire Piezo2 to AI2, and ground AI3.  Note that you must scan (read) all four channels for this to work.

Lynn

[David_Pastorino]

Thank you for the hints. I´ll inform about this, try it and come back to the forum ! 

🙂

[David_Pastorino]

Hello again,

I took a look to various solutions to measure my signal, and by curiosity tried the solution you suggested:

Something which sometimes works (but may not in your case) is to scan four channels and ground the inputs on two of them.  Wire Piezo1 to AI0, ground AI1, wire Piezo2 to AI2, and ground AI3.  Note that you must scan (read) all four channels for this to work

It seems to work until now as I get a 6kHz signal and no interferences between one and another. As strange as it may sound, is it then normal to let one of the two contacts of the piezo free? (not connected)?

[johnsold]

David,

No. You still need to use both sides of the differential measurement (AI0+ and AI0-) for Piezo1.  Ground AI1+ and AI1-. ...

The bias resistors allow a small amount of DC current (strangely enough called bias current) to flow.  This keeps the transistors at the input of the amplifier working properly. The best value depends on the particular DAQ device and the nature of the signal source.  1 megohm is a good starting point.  For each piezo connect a resistor from AIx+ to AIGnd and from AIx- to AIGnd in addition to the piezo to AI connections.

Lynn

[David_Pastorino]

Ok,

It looked suspicious to let an unconnected wire.

I´ll try with a first value of 1MOhm and get back here to feedback. 

Thank you !

[johnsold]

David,

Please indicate what DAQ device you are using when you post again.

Lynn

[David_Pastorino]

Good morning Lynn,

I am using a myDAQ device.

I followed your advice and mounted resistor of 1 megoOhm between each Ax +/- and AGND.

The unexpected behaviour of the signal does not appear anymore. It is apparently working then. Thank you!

I still have a doubt on the measurement I get. The reason for that doubt is the following:

I emit a 4,5KHz, 1Vp-p sinus during 50ms with piezo0 and record both signals from piezo0 (what is emitted then) and piezo1 (what I actually receive). The two piezos are separated by a distance of 5cm (with air in between).

Naively, I am expecting to get a decent amplitude from the piezo1 signal. Here comes the doubt:

- To check that the system is working, I used a function generator (thus generating a continuous 4,5 KHz 1Vpp sinus) and checked with the oscilloscope the signel received with piezo1. I do not get any signal. I record a signal of 4,5KHz and very low amplitude (3mV) if I stick the two piezo together.

- However, if I use the Bode analyser, I see a 4,5KHz signal clearly when the two piezos are separated by 5cm of air. Moreover, the power of the 4,5KHz peak seems to change with an amplitude change of the "stimulus" signal.

- Finally, a small hit to piezo1 gives a clear signal, confirming that I am actually measuring the good thing.

I am afraid this may be a concept problem from my experiment or a circuit/component bad election as I know few in that field. Fisically, the wave gets to the piezo without any doubt, so , what I am doing wrong here?

Thank you in advance !

[johnsold]

David,

What is the trasnducer factor for your piezo devices? By transducer factor I mean the relationship between applied voltage and sound pressure output. Is the realtionship reciprocal? That is, does applied sound generate voltage as efficiently as applied voltage generate sound?

Is the output actually a voltage or is it charge? Fundamentally piezo devices convert stress to charge.

What is the acoustic coupling factor between the two devices? If the transmitting device emits sound in a uniformly spherical distribution you get the ratio of the surface area of the receiving device to the surface area of a sphere whose radius is the separation.  If the surface area of the transducer is  1 square cm, the receiving device only gets 0.3% of the emitted power.  The devices are probably more directional than that, but you may still have a fairly weak signal.

What is the resonant frequency of the piezo devices? They tend to be more effective at resonance.

What is the signal strength reported by your Bode analyzer?

Can your DAQ device drive the capacitive load at that level?  Some amplifiers are not very "happy" with a capacitive load.

Lynn