Change in resistance detector circuit

Hi everybody,

I'd like to get your help in designing a circuit that can react to changes in the total resistance of the circuit.
My goal is too have a small voltage battery powered device that will light a small LED if the resistance changes.
The LED should also fade out after a couple of hours.

Let me explain what I mean by changing the resistance - let's say I have a switch and a parallel resistor, if I disconnect the switch the circuit would still be connected but with a different resistance.

Thanks for helping, you and this forum are awesome.

BTW - the topic should be: "Change in resistance detecting circuit".
Couldn't find a way to change it.

Thanks again,
Aviad.

yes you simply use a voltage comparator with a fixed reference voltage. The resistance would have one level and the switch would have another level. This would work for such a requirement as you have described although you have not given much detail. Another approach is what is know as a balanced bridge approach which is much more accurate and is relative to a common voltage source, it has advantages.

Hi Earl,

Thanks for the fast reply!
Sorry for not giving enough details, I don't really know that much about EE...

but the basic idea is this:
I need a circuit that can have several resistance values, let's say - 10, 11, 12,...., 20 Mohm ,
and every time the resistance changes, a LED will light up and fade out slowly.
For example - the circuit is now in 12 Mohm resistance - the LED is off.
I change it 13 Mohm and the LED lights up and slowly fades away (3-4 hours).

I hope I'm clear enough, sorry for not being able to explain it better...

Thanks again!

If the values change quickly then there is a simple dtetector circuit that would work. How quickly do the values change?

They change very slowly, about once a day.

Then I would do it with an MCU with a sample rate of something like 1/2 a day (but you decide). You could use an MCU internal ADC to monitor a voltage level that (I'm assuming) would change as the resistance changes. Sample the voltage, store the value and then compare that value with the next sample.

If an MCU is out of the question, then I would use an oscillator driven divider array (i.e. cascaded counters) that would trigger a sample and hold that is fed to a comparator. The divider chain would also multiplex the sample input to two separate hold circuits. Each of the hold circuits would feed the input of a self calibrating window comparator arranged as a +/- delta detector (so it lights that LED when the resistance goes either up or down--beyond some band gap (otherwise it would, likely, trigger one every sample). You could probably use the same divider chain to time the duration of the LED indication.

Now, if these resistances are constantly changing, then it will be kind of hard to define what "change" means. Because if they are constantly changing, then isn't a change detector going to always indicate a change?!?

But, if the rate of change varies, then how about a slope detector--that detects REALLLY shallow slopes (difficult to design). Think op-amp differentiator with extremely high input impedance and a large low leakage capacitor and lots of gain. OR if using an MPU, set a sample rate that is a multiple of the maximum sample window period and store a series of samples (FIFO style) with a register length equal to whatever that multiple is, and run it through a FIR filter that amplifies changes.

So, it's either watch for a bump in slope (absolute value of), or watch for an extent of change (i,e. "band-gap"), or sample periodically and compare (but even then, you're probably going to want a band-gap for noise immunity).