In a message dated 22/11/2009, barry_lotz@............. writes:
Chris
I had thought of a piezo in the past but got confused with it's output. I
believe it measures the rate of change of movement of the sensor. Since an
accelerometer gives acceleration results when measuring deflection, Does
one have to integrate piezo output results to get acceleration?
Hi Barry,
The MEMs accelerometers give a direct acceleration output which
includes accelerations which are constant with time. The usual way to calibrate
them is to take a DC reading, invert the device and take another reading =
2g change.
The piezo capacitor disks give a voltage output which is ~linear with
the short period stress change (bending), so long as you support them on
two sides and not as a diaphragm supported around the periphery. The disks
quoted have a capacity of ~65 nF in parallel with the amplifier input
resistance. This may be made 50 Meg Ohms to extend the operation out to 20 seconds
- see my circuit. These are basically AC devices with a very extended
decay time. _http://jclahr.com/science/psn/chapman/piezo/index.html_
(http://jclahr.com/science/psn/chapman/piezo/index.html)
The voltage output of a mass loaded disk will be proportional to
acceleration, but this will not include any constant acceleration. If you
integrate the output, you will get a velocity for all periods less than 20
seconds.
Since the amplifier input impedance is so high, you have to keep the
element dry to minimise any leakage currents. But you can also varnish it.
The PZT element has an appreciable voltage / temperature coefficient, so you
need to minimise any temperature changes.
Do always include the voltage protection diodes on the amplifier input
and short circuit a disk before connecting it up. An unconnected disk may
have 100 V on it due to previous temperature changes - this can blow the
input of an unprotected opamp.
I mount the sensor and amplifier inside a glass jar with a metal lid
having a rubber seal. I cut a hole in the lid and stick on a disk of circuit
board having soldered feed through pins for the supply and output
connection leads. I put a small sachet of silica gel inside to keep it dry.
I hope that this helps,
Chris Chapman
In a message dated 22/11/2009, barry_lotz@............. writes:
Chris
I had thought of a piezo in the past but=
got=20
confused with it's output. I believe it measures the rate of chang=
e of=20
movement of the sensor. Since an accelerometer gives acceleration=
=20
results when measuring deflection, Does one have to integrate piez=
o=20
output results to get=20
acceleration?
http://jclahr.com/science/psn/chapman/piezo/index.html=
The voltage output of a mass loaded disk =
;will=20
be proportional to acceleration, but this will not include any=
=20
constant acceleration. If you integrate the output, you will get a velocit=
y for=20
all periods less than 20 seconds.
Since the amplifier input impedance is so hig=
h, you=20
have to keep the element dry to minimise any leakage currents. But you can=
also=20
varnish it. The PZT element has an appreciable voltage / tempera=
ture=20
coefficient, so you need to minimise any temperature changes.
Do always include the voltage protection diod=
es on=20
the amplifier input and short circuit a disk before connecting it up. An=
=20
unconnected disk may have 100 V on it due to previous temperature changes=
- this=20
can blow the input of an unprotected opamp.
I mount the sensor and amplifier inside a gla=
ss jar=20
with a metal lid having a rubber seal. I cut a hole in the lid and st=
ick on=20
a disk of circuit board having soldered feed through pins for the supply=
and=20
output connection leads. I put a small sachet of silica gel inside to keep=
it=20
dry.
I hope that this helps,
Chris Chapman