Measuring high resistance in the range from 10K ohms to 100 Giga ohms?

[Cambo]

Hi Edaboard.... what a busy place

I am not an electronic engineer but I have been trying to design a circuit to measure high resistance for about a year now... which is much harder than I thought, so wondered if I could call upon some help?

The basic aim is to measure the resistance between 2 screws placed into timber - this will tell us what the moisture content is of course - from there I hope to record these values to an SD card but I have the MCU side of things sorted I think - just really stuck on measuring the resistance between the screws in the range typically from 10K ohms to 100 Giga ohms?

- I would like to use low voltage DC somewhere in the range of 9V to 5V.
- I hope to run off batteries for long periods - aiming at 12 months measuring every 5 minutes - 6Ah might be a useful max battery size.
- I would like to use the MCP3424 ADC to take the voltage measurements: ww1.microchip.com/downloads/en/devicedoc/22088c.pdf
- This ADC can measure ±2.048 V
- To avoid the screws suffering too much from electrolysis I would like to reverse the polarity for each measurement.

Using a current sensing resistor and above ADC I can acurratly measure up to about 10K but this method falls on it sface as soon as I enter into the range I'm interested in

Any and all pointers/help apreciated - thanks.

 

[FvM]

Electrical measurements of solid matter moisture content is a widely used method, of course with some variations depending on the application. Besides contact based measurements, capacitive or radio wave (e.g. microwave) absorption measurements are in use.

I notice that you aware of the electrolysis problem, in fact most technical moisture or conductivity measurements are using AC voltage. But the polarity reversal method should hopefully work. To cover the intended wide resistance range you'll need some kind of automatic range switching.

I wonder of you already verified the working of your method with a digital multimeter. The input circuits of commercial resistance meters can be a template for your design, I think.

 

[Cambo]

Hi FvM, I think you are spot on... most of the kit I have looked at has 3 OPamps and a multiplexer or 2. I have pulled one of the meters open here to show you what's inside - I think the right hand stuff is mostly the analog circuit whilst the left hand side is mostly the digital multimeter chip etc.

So the recipe is likely to include some opamps, multiplexer and some resistors etc - I have no idea how to join them all up and have yet to find some examples online.... would you have any idea how to make this work?

 

[crutschow]

Lower resistance values are typically measured by using a constant current source and measuring the voltage across the unknown to determine its resistance. Very high resistance values are often measured by apply a known voltage and measuring the small current.

Note that measuring resistances as high as 100GΩ requires very careful design to minimize leakage currents in the insulators. Teflon wire and standoffs are often used, since Teflon has the highest resistance of common insulation materials. Also guard rings and guard shielding is likely required such as discussed here.

Here is an example of a high resistance measuring system. If you google "high resistance measurment cricuits" you should find other information on this.

 

[Cambo]

Thanks crutschow, really great links - I hadn't seen those before, I will have a good read.

I wonder if either of you might be able to help me pull a circuit together so that I can start playing to see what works in reality, it may be that 100 Giga ohms if really off the useful scale and 100M is more typical for dry material?... in which case presumably some of the problems go away?

It might be too much of an ask as it seems quite complex from where I stand... so completely understand if its not possible

 

[crutschow]

This reference states that the resistance of wood for two pins 30mm apart is 4800MΩ (4.8GΩ) for wood with 8% moisture. If you want to measure a lower moisture level than than, you would need to be able to measure a higher resistance.

I can give you some help on the design but more in the way of suggestions and critique of what you are designing, rather than a complete design.

The simple approach is to measure the voltage across some high value resistors in series with the measurement value. I did a quick check at Digikey and they sell resistors up to 5GΩ in value. The voltage could be amplified by a FET type op amp which has a very high input impedance (such as this), configured as a non-inverting amp.

Depending upon the accuracy you want, you will likely need to switch in different value series resistors and/or change the op amp gain to cover the full dynamic range. The switching would need to be done with very high impedance switches, such as small reed relays. The resistors, relays, and op amp input pin should probably be connected to Teflon standoffs to minimize leakage current.

The measurement process would be to apply the supply voltage to the series combination of the wood terminals and the sense resistor. You then read the sense resistor voltage value and the supply voltage with an A/D converter (could be an A/D built into the µP), and calculate the wood resistance from that. If the measured voltage is too high or too low, then the µP would need to change the sense resistor value and/or op amp gain until the reading is within the desired range.

 

[Cambo]

Hi Crutschow, would it be possible to email you a proposal please - I think you may well be the man for this project. If that's OK perhaps you could 'message' your email address to me? Thanks.

 

[crutschow]

Hi Crutschow, would it be possible to email you a proposal please - I think you may well be the man for this project. If that's OK perhaps you could 'message' your email address to me? Thanks.

Unfortunately I am not able to take on any new projects. I can help you by critiquing your design and offering suggestions through this forum, but that's as far as I can go. :-(

 

[Cambo]

Hi, thanks for getting back to me.... completely understand . I will attempt to pull a circuit together for your thoughts.... It will be good for my learning

 

[devendrakumar]

Hi Cambo

I am doing an experiment in which i need to measure high resistance of wood. So i also need to make the same circuit for which u are working. may we use each other help in making this circuit. If u are ready to work together please let me know.. Thanks

 

[dick_freebird]

At extremely high resistances, you need extremely low leakages
which means clean and simple.

You might consider something as simple as a relay and capacitor
and time-counts; NC relay discharges the cap to zero, NO when
energized connects the other side of the probe-pair to the
voltage, a low-leakage-as-possible comparator watches the cap
voltage rise and stops a timer. The math on the RC exponential
can be done later to derive R from t, C and (Vsupply, Vthresh).
A threshold of Vsupply/2 is cheap and simple. FET followers can
improve input impedance to trivial levels.

 

[Cambo]

Hi, I'm still ticking away on this slowly I came across the simple circuit below that uses a PIC MCU but don't fully understand how it all works so was wondering if anyone else does? I know won't have a huge resistance range but will be a good starting point.

- Two of the analog pins AN2 & AN3 are connected to one side of each of the 2 unknow resistances S1 & S2.
- Another two analog pins AN0 & AN1 have some sort of voltage divider on them and share a connection to the other side of both S1 & S2
- I'm guessing that R1, R2 & R3 are in some way used to measure the current flowing to S1 & S2? Perhaps something to do with covering a greater range of resistances?
- I have no idea what R3 is for - maybe nothing to do with measuring resistance?

I'm hoping to try this out on an Arduino but keen to understand what I am doing first

Any pointers appreciated.



PS - Sorry devendrakumar, just saw your post - happy to collaborate of course, maybe this circuit helps in some way?

 

[crutschow]

Was there no explanation of how the circuit works form wherever you found it?

 

[Cambo]

No such luck - just got an image to work off.

 

[FvM]

I suppose, that the typical electrical characteristics of a PIC (or many other microcontrollers) will allow to determine resistances in a high Megohm range. But the datasheet also tells that I/O leakage currents can be as high a +/- 1 µA. "nano Watt" types are still specified with typical 50 nA input current, which won't allow the intended Gohm range.

A reliable solution needs parts with specified input parameters, like "electrometer" OPs have it. TI OPA129 would be an affordable standard type.

 

[Cambo]

Hi FvM,

The aim is to eventually come up with a more sophisticated design which copes with the large range and high resistances but when I came across this I wanted to understand and try it... just for learning really

 

[chemelec]

This can determine High Value Resistances.

http://chemelec.com/Projects/Gig/Gig.htm