piezo transducer signal amplification

Discussion in 'Electronic Design' started by desufnoc, Jan 20, 2008.

1. desufnocGuest

I wish to use piezo transducer 273-073A to pick up and amplify insect
activity. What kind of simple circuit would amplify enough for ear
phones?

2. D from BCGuest

The gain will depend on how much noise you can tolerate.

Noise sources can be from induced EMI(ex:hydro lines), intrinsic
transducer noise and intrinsic op amp noise...
Possible power supply noise too..

Most basic would be to use an op amp circuit + head phone driver amp
module.
circuit.

A circuit might pop up with some Goggling..

I don't know the transducer you're using but that part number reminds

D from BC
British Columbia

3. Martin RiddleGuest

http://www.thefreedictionary.com/goggling

4. D from BCGuest

Oopps...
Maybe I can rescue that..

Got some subs ideas...
Googlizing...
Googlization
Googlyzing
Googification

D from BC
British Columbia

5. WinfieldGuest

To understand piezo microphones, it's useful
to think of a piece of ceramic, that's going
to be flexed by an impinging sound wave. The
first problem is that the acoustic impedance
of ceramic is so much different than air that
it'll not interact very well, and only a weak
signal can be picked up that way. In fact,
using Google, it appears this transducer is
meant for use as a contact pickup, e.g. with
a guitar, piano, etc.

Second, in case you do want to work with the
piezo transducer anyway, to analyze its use,
you only have to realize it appears to be a
capacitor with its voltage signal in series.
Resistive loads on capacitive sensors make a
low-frequency rolloff, so to design your preamp
you need to know the piezo capacitance and your
desired low-frequency limit. For example, if
the capacitance is 150pF (I didn't find the RS
part's spec), and you wish it to work down to
50Hz, then the "load" will have to be above
22 meg-ohms. The answer is to use a JFET-type
preamp and keep the necessary dc-bias resistor
at 22M or higher. You can lookup JFET.

You'll want lots of gain, of course, but we can
tell you the noise level will be determined by
a parameter we call voltage noise, e_n, which
is lowest for JFETs with a large die area.

There are JFET opamps that are pretty good, but
discrete JFETs can do much better. If you look,
you'll find lots written about the subject.

6. MooseFETGuest

If you need to go descrete:
www.interfet.com and look at the IF1801 JFET
http://www.linearsystems.com and look at the LSK170
You can also parallel JFETs to get better noise performance. The
improvement goes as the sqrt() the number of JFETs.

Depending on the capacitance of the crystal, the current noise of the
amplifier may also matter. The current noise times Xc gives an

As op-amps go, the LT1169 may be about as good as you can do for low
noise voltage and current.

7. WinfieldGuest

The LT1169 is a nice amplifier, but it's got a
rather small JFET die area, as evidenced by its
capacitance, only 2pF. This shows, with its
moderately-high noise spec, e_n = 6nV.

As we point out in AoE's lengthy low-noise chapter,
optimum low-noise performance is had when a JFET's
Ciss is near to the sensor's capacitance. Sadly,
it's often hard to reach that goal. For example,
no opamps are available with anything like 150pF
of input capacitance, but one of the higher ones
is Analog Device's AD743, which is 20pF and has
e_n = 2.9nV, twice as good as the LT1169.

As for discrete parts, NXP's (Philips) bf862 at
10pF and 0.8nV is pretty good.

It's interesting that the bf862 JFET has half the
of sqrt-2 more noise, it has 3.6x less noise. My
suggestion, get a lifetime supply of bf862 JFETs.

8. JoergGuest

Lifetime supply? Now please don't tell me that they rang the bell on
that one ...

<writing note to self: Get some more BF862>

9. MooseFETGuest

There are other JFETs out there:

The IF1801 is 100pF and 0.5nV/sqrt(Hz)
The IF3601 is 300pF and 0.3nV/sqrt(Hz)

The LSK170 is 20pF and 2nV /sqrt(Hz)

The bf864 has the disadvantage that you can't find it on the NXP web
site unless you know what to look for.