Why is the device's output impedance relevant? Surely all you are interested in is its supply voltage and supply current draw?
Why do you want a current limiter? The accelerometer will draw whatever current it needs.
Edit:
Minimum load current for an LM317 can be as much as 5mA, so additional loading may be needed.
All that means is the required voltage is 18-30V. The sensor will draw 2-5mA at whatever voltage you supply.
You could attach this sensor to a 22V 100A power supply, and the sensor would still only draw 2-5mA. Providing of course the supply could regulate with such a small load. Which of course appears to be the issue with the LM317. It has a minimum load of 5mA, so if you get a good sensor that only draws 2mA, the LM317 will not be able to regulate the voltage well.
The simplest way would be to setup the LM317 with 20V out and connect the sensor in parallel with a 1/2W 4K ohm resistor (20V @ 5mA is 0.1W). The resistor will guarantee the 5mA minimum load and the sensor will draw what it needs. The LM317 will supply 20V with a load of up to 10mA or power output of 0.2W.
My point exactly. I don't believe you need a controlled current source; unless you have a very unusual sensor. As Eric says, can you post the accelerometer type/model info?
yes..
that's what i am telling..
the sensor itself is very costly, i cant risk to connect it to supply..
can you please suggest me with some design so that i get the max of 5mA current with atleast 20V .
even if my circuit accidentally gets shorted with ground, I want the current restrected with 5mA only.
Here's one way:-
The graphs show the current limit for a V1 range of 23-30V and a temperature range of 10C-50 C.
Increase R3 to reduce the current limit.
R1,C1 limit any supply spike but may not be needed.
Your sensor is NOT a 100 Ohm load. It gives an output of 10mV/G and has a 100 Ohm output impedance.
(It's a pity the datasheet doesn't include an application example showing the power supply arrangement).
Here is a circuit I have seen used with similar accelerometers. I think I saw it on some application note, long ago.
This circuit simultaneously biases the accelerometer to operate with constant current through it; the acceleration modulates the effective resistance of the accelerometer and the opamp converts that to a buffered output voltage.
Note that by selecting the voltage at the non-inverting input of the opamp (1V in this example), and by picking R2, you can set the current through the accelerometer.
In my example, I choose 10mA (1V/100 Ohm). For the simulation, I vary the effective resistance of the accelerometer to see if the current remains constant, and to see the effect on the opamp's output voltage...
In the second simulation (time domain), I added a soft-start capacitor and plotted the current through the accelerometer at three different values of accelerator resistance (500, 1K, and 1.5K) as the power is first turn on. Note that the current comes up nicely controlled.