Wanted: IC Amplifier for Piezo Pickup

See National Semiconductor LMC6081.

piezo pickups produce relatively high voltages with very low current capability, so a very high impedance fet amplifier IC should be chosen and set to amplify the voltage and current to the level required.

Aurizon

Guest:

Depending upon the system layout, the charge amplifier may be a better choice. It looks like a giant capacitive load, sucks up all of the charge produced by the transducer and amplifies it. The advantage is that it greatly reduces signal losses to cable capacitance because input signal voltage in the cable is virtually zero. There is only the transducer charge current flowing in the cable. You do have to start with a high gain, wide band amp.

DickL

I'm not sure what your application is-that is how much gain? What input/output levels? How long a line do you have to drive? What is the temperature? In other words what are you trying to measure?

A word of caution. a 'Charge' amplifier may not be the best choice. They are very susceptible to tribo electric cable noise, capacitively coupled noise and are bothered by leakage resistance i.e. moisture and contamination.

A preferable way to get the signal out of the crystal is by using a high impedance Fet interface with a selected input resistor to set the low frequency corner and driven by a constant current source ( 24 volts DC and 1/2 ma. diodes are commonly used.) with an out put coupling capacitor to block the DC bias component. That will provide a low noise system that has low output impedance and is not susceptible to the problems mentioned. Vibes

http://www.maxim-ic.com/appnotes.cfm/an_pk/1127

imding:

I'm confused by your linked material. Specifically,

"The sensor has high output impedance and requires a high-impedance buffer amplifier. The circuit shown includes a differential charge amplifier followed by a differential-to-single-ended amplifier."

The charge amplifier CIRCUIT, as pictured, does not have a high input impedance, even though the amplifier DEVICE may have. The CIRCUIT, as a load, looks like a very large capacitor.

Judging by other comments, I must have been lucky for all those years of aerospace instrumentation in which charge amplifiers worked very well. Without them, gain and frequency response would have been determined by cable capacitance and would have varied from sensor to sensor on the test stands. A calibration nightmare. As it was, the sensors were calibrated in picocoulombs/G and the amplifiers in millivolts/picocoulomb so it was easy to determine Volts per G. Cable noise was not a major problem, even though large jet and rocket engines were running nearby. I've always been lucky though.

DickL

Hi DickL: No Dick you weren't lucky and you weren't alone. The whole aerospace universe used charge amplifiers. Many still do. Mainly because major vendors were pushing them. It is all a matter of how things are used. Most used special 'low noise' cables that were tied off and imobilized to avoid the low frequency flexing tribo noise. Or they just filtered the tribo noise out. Most charge systems were/are used in electrically inactive test lab environments where high frequency capacitively coupled noise was not a particularly big problem.

The low impedance (50 ohms) constant current(Or constant voltage) technology is not affected by cable length at all. It is basically independent of cable capacity.

Its output is in millivolts and the mv/unit sensitivity is the same at 10 feet as it is at 1,000 feet of cable. The only problem you have to be aware of is 'Slew rate' if you don't have enough constant current capacity at high frequecy. (2ma was usually good enough for 10,000 ft of cable at relatively low voltage swings. You don't need expensive 'special' low noise cable. You can use standard twisted pair shielded cable. Many early aircraft engine monitors used charge conditioning and they had so many problems that many pilots just turned them off. Most new engine monitoring technology uses integrated electronics for conditioning with special high temperature electronics and excellent coupled noise immunity.

I spent years in my early days trying to make users aware of these problems. I must have missed you. Consider yourself lucky.

The same was true in the distributed monitoring of machinery. Early 'charge' systems were a disaster. These days the constant current systems are 'exclusively' used in this field, except where high temperature is a problem. Charge is fine as an internal impedance matching/amplifying front end where conversion can take place locally and may be fine for this application. My point was that if putting the charge amplifier at the end of a long line in a hostile or dynamically active environment you should be aware of the pitfalls. John

Hi John:

I think we're talking about two different things. I was speaking of a passive sensor, in a hostile environment, generating a few picocoulombs of charge and required to drive some variable length of cable. I'm interpreting your description as applying to a sensor, perhaps in a more friendly environment, containing at least one active component acting as a preamp and line driver. Two different system configurations.

The system you've described sounds fine if it can survive the sensor environment, but jet and rocket engines can be pretty tough on electronics. I'm not familiar with the environmental limitations of the sensors/systems you've described.

DickL

Hi Dick: I guess we are both spending our day punching things into the computer! I'm about to call it a day.

The sensors I'm talking about were used for measuring shock and vibration. PCB calls them ICP (Internal circuit PiezoElectrics) A fancy name for constant current. The typical application environment was limoted to temperatures from -65 F to 250 F, although we made some for Pratt for the Raptor and JSF with built-in self test electronics that went over 400 F and one special that went to 500 F. As far as I know that was really the first time low impedance vibration sensors with internal electronics were used on jet engines. We also had some on missles but not on the engine. Most were used for shake and bake Sine/Random vibration qualification testing in the early days but later were used applied in machinery montioring. We found out in a hurry that the charge approach didn't work when running lines all over a plant. They are made commercially by most sensor companies-Endevco, PCB, CTC, Kistler etc. One of the first we made was at MB Electronics in Connecticut back in the early 60's. (God that seems like a long time ago.) It was a constant voltage system called Zero Drive. It was good but it was a little more expensive because it required a voltage regulator at the driving source. We found that Constant Current was the most efficient. It only required a FET chip and a contant current diode or resistor at the driving source. It is the most commonly used version around today. Our company was Vibra-Metrics now part of Physical Acoustics. Now I am rambling! Where did you work in aero space? John

Guest:

This string illustrates the danger of not fully defining your problem. Each responder creates his/her own scenario and proceeds to address the problem as they've imagined it. You may or may not find something useful in the resulting ramblings. I hope your question has been answered.

The "amplification factor" of a charge amplifier is usually really charge sensitivity in mv/picocoulomb and can be changed, within design limits, with one capacitor.

DickL