Hi all
i am working on a project where i have to convert capacitive sensor output to a voltage signal.i know there are several methods to do it but as per my sensors specification i could not figure out what method would be the best one to get the best result
here are some specification of a custom designed capacitive sensor.
capacitance range is from 2pF to several nF
the project require capacitance to work in different medias(like water,air,brine etc)
for now i just want to work out in two medias water an air.
i need output range to be 0-3.3V
please if someone can suggest me a method which can be suitable for my requirements

Thanks a lot for your help.
P.S. I am new to this forum.hope it would be a great knowledge sharing forum

 

Go digital. The usual reason one wants to measure capacitance is to measure something.
Build an oscillator that is dependent on capacitance and measure frequency.
Or get a digital capacitance meter with a digital interface to a computer.

 

Go digital. The usual reason one wants to measure capacitance is to measure something.
Build an oscillator that is dependent on capacitance and measure frequency.
Or get a digital capacitance meter with a digital interface to a computer.

thanks Mr chips for your valued response.
the actual reason for this circuit would not be measuring capacitance i just want that the circuit shows a difference in voltage level as the media change from lets say air to water.later on this voltage will be fed to ADC .
and is there any circuit/sample so i can customize it accordingly to my requirements.

Thanks

 

What are you going to do with the voltage, when you have it?
Are you going to send it to a voltmeter readout, or use it to drive another circuit?

As MrChips says it is easy to build a reliable linear capacitance to voltage convertor based on a few logic gates using the capacitance to determine the frequency of an oscillator.
Personally I would recommend against a 555 based oscillator as I have found them less than reliable.

 

the actual reason for this circuit would not be measuring capacitance.Thanks

The point you are missing is that the thing that changes is the capacitance, therefore, you have to measure the capacitance in some sort of way to get to the next stage. HOW you measure the capacitance is directly related to what you need to drive the next stage (ADC). That's what we are negotiating right now.

 

later on this voltage will be fed to ADC

That is the whole point. A capacitance to frequency circuit is an ADC.

 

That hasn't answered my question of what you want to do with the voltage.

 

That hasn't answered my question of what you want to do with the voltage.

right now i am given the task to convert the sensor capacitance into some sort of voltage level.and yes if there is a method other than frequency oscillator approach.
later on we are going to feed the voltage to a 24bit ADC.
the purpose of the project is just to diffrentiate between different mediums like if the sensor is in air it would have a different voltage level based on the capacitance value(sensor capacitance is 4PF in air)
while in water its capacitance goes up to 60-70PF

 

i already used analog devices AD7745/46 CDC which almost serves the same purpose of changing capacitance to a 24 bit HEX value.
but due to the AD7745 input limitation i am afraid i wont be able to use that. because my sensor capacitance vary from 2PF to several nF.
i hope i am able to clarify my purpose of the topic
i am very new to this field thanks for all the support i am getting.Thumbs up to all the very nice people out here

 

Back in 1977 I had a burned out multimeter and no Capacitance meter, so I built this circuit in place of the original meter circuitry yielding a multirange Cmeter going from a few pf to 100 microfards.

It is still working today.



Basically a fixed oscillator made from a couple of gates in a SN7413 fed a monostable (SN74121), whose period was set by the unknown capacitor.
This pulse train was clamped at a fixed voltage and sent to the meter.

So the meter received a fixed train of pulses of variable width, proportional to the unknown capacitor and integrated these to an analog voltage.

Simple, cheap and very deadly.

I've forgotten how to do smileys again.

 

thanks Studiot
that sounds perfect i will try to discuss this with my supervisor.we have the frequency oscillator approach in our mind but due to some reasons my supervisor says it should be out last resort.the reason i posted here was to get feed back from you guys and if i can get a sample circuit where i can do some calculations and plug in my required values then somehow read the drift in voltage level accordingly to the change in sensor capacitance.

 

The analog meter on the left is the beast in question. I use PP3 9V battery for supply, via a regulator.

http://forum.allaboutcircuits.com/threads/ancient-hioki-analog-meter.14847/#post-91570

These days I would probably use CMOS or LSTTL at least to replace the TTL as it is quite thirsty.

 

Microchip has a couple of interesting application notes on capacitance measurement:

http://www.t-es-t.hu/download/microchip/an611b.pdf
http://ww1.microchip.com/downloads/en/AppNotes/01016a.pdf

The second link refers back to the first, which is hard to find on the Mircochip site per se. Note the capacitance range in the second link, which is pretty close to what you want.

If you only want to tell one medium from another, you might be interested in a capciflector, which is a bit of a spiffy name for a capacitance meter. The student report is attached, as I can't find a link to it ("capaciflector report_ver22"). Note figure 3.1.3. Several years ago, I played with the device and am attaching an old schematic for what I built as well as the "capaciflector" design I used.


My device could be tuned to a specific capacitance using the LM567, so you could get a go/no go result. You can also use a frequency to voltage converter to give an analog output. I used the NJM2211.

John

 

We have the frequency oscillator approach in our mind but due to some reasons my supervisor says it should be our last resort.

Why consider the obvious method as the last resort???
The only way to measure capacitance is to see how the voltage on it reacts to current.
An oscillator is merely a method to keep repeating the measurement so you don't have to push a button every time you want an update.

I can tell you that the wires to the sensor are going to contain more than 4pf of capacitance, so you will have to account for that.

 

some of he alevel guages here use a capacitance probe to measure the level of liquids in a tank. the c robe is connected to a 555 timer as an oscilator, and the output frequency is ran through an integrator ( tachometer type) circuit to get a dv voltage proportional to the level in the tank. I dont have the circuit for this anymore, but they worked pretty well.

 

I would use Studiot's approach but with some changes. I would use a 555 as the clock generator instead of the 7413, (better stability). The 74121 one-shot is a good choice for the capacitance range you are working with. What you have at the output of the one-shot is a fixed frequency oscillator with the pulse width controlled by the capacitance. So, instead of a diode, I would use a simple R-C circuit to average the output. This would make your output voltage directly proportional to the duty factor. (No diode to have to compensate for.)

My two cents plus $4 will get you a cup of coffee.

 

This might be helpful.

It's hard to picture how you could measure capacitance without an oscillation method, with only a single point measurement. I guess you could watch the voltage on a known capacitor and then hook up the unknown (and previously discharged) capacitor in parallel, and note the ∆V.

Could an op-amp integrator be configured to measure the charge that flows during a charge or discharge? It's not exactly single-point but would seem that way to the user.

 

another way would be to use the capacitor in an intgegrator RC circuit, feed constant pulses in and measure the voltage after integrating, more cap = more voltage.

 

i have to convert capacitive sensor output to a voltage signal.i know there are several methods to do it

So basically they know what to do, they just don't want to do it.

 

The paper that wayneh linked to is the kind of circuit I was refering to in post #16. The unknown capacitor controls the pulse width and the meter averages the on time verses the off time. The 555 instead of the 74121 will be more accurate over all, but may not give good results with capacitance in the few hundreds of pf. Simply replace the transistor and meter with a resistor and cap and you are done.

btw: good find wayneh