Post Go back to editing

# CN0359 Conductivity Board Constant Resistance Error

My lab and I are using the CN0359 conductivity board to evaluate the conductivity of solutions. When the conductivity probe (Innovative Sensor Technology) is wired to the Analog Board, the conductivity measurements are all across the board. We have run the following experiment in hopes of ensuring that the Analog Board was functioning properly, yet the data suggests there may be an error within the board itself. How can we ensure the board is functioning properly?

Experiments

The Analog Board is capable of accepting resistances between 0.1 Ohms and 10 MegaOhms. If the resistance is constant, the conductivity of the system should be a constant as well. (The first picture presents the experimental set-up.)

Experiment #1

Pins 1 and 4 (red wires) were attached to an oscilloscope to measure the voltage and obtain the waveform. Pins 2 and 3 (blue wires) are wired to a 1 kilo-ohm resistor. The resistor simulates the resistance that would be expected in solution and is well within the range that should be accepted by the board. Additionally, it is a constant and removes possible error from another source. This should lead to a constant conductivity. The Analog board is then plugged in. The settings are changed to obtain a 0.7 V differential at an oscillating frequency of 160 Hz. The waveform is indeed a square wave, yet it is not completely symmetric (Figure 2). Specifically, the oscilloscope shows a max voltage of 770 mV and a minimum voltage of -870 mV. When we look at the Analog Conductivity board, the conductivity measurements bounce across the board. (Values continuously change from the microS/cm range to the mS/cm range and then shows a sensor error before beginning to jump from the uS/cm to mS/cm range again.)

Experiment #2

Additionally, it is worth noting that when we switch the polarity (switch Pins 1 and 4, the analog board stops working as no voltage drop can be measured (Image 3). We would have expected switching the leads to have no effect as it is an AC current that is generated. This was not the case.

Experiment #3

Here we tested to see if the results were reproducible. We returned to the configuration described in Experiment #1. We unplugged the board, waited a few minutes and plugged the board back in. The waveform that we obtained was shifted completely off center and was not symmetrical by any means. (Forgot to take picture.) We once again unplugged the board, waited and plugged that board back into the power supply. The waveform returned to what we saw in experiment #1 and we obtained the same (in my opinion, not sensible) results as previously described.

If these tests were not sensible, what tests can we run to ensure the board is functioning properly? I believe these tests should have functioned, if so, how can we fix the problem?

Parents
• Hi,

Can you please send more details on your sensor being used?  (Datasheets, websites, user guides, any other documentation)  Also how are you connecting the sensor into the CN0359 board?  The images you shared are nice, but they don't show the sensor and they also don't show what the sensor is measuring?  Also what are you measuring with your sensor?  Do you have conductivity calibration solutions or are you measuring an unknown source?  Its recommended to start with calibration solutions so you can benchmark and reference your measurement to some known value.

Cheers,

Brandon

• Before we even plug the sensor in, we are looking for an experiment to test and ensure the Analog board is working properly. If this experiment was not valid, can you suggest one?

If not, I can plug in the sensor, yet we were trying to simplify everything and leave the sensor completely out of the analog board validation. (The we got the same results when the sensor was plugged in measuring standard known concentrations of KCl salt.)

• Hi ND_Deuterium,

Assuming your scope is a standard benchtop unit (and not a small, battery-powered, floating unit with very low self-capacitance), I suspect the problem is that you're shorting nodes to ground that really only should be connected to the sensor's excitation terminals (or your test resistor.)

In experiment 1, the excitation return is shorted to ground. This node is actually a virtual ground and should not be loaded.

In experiment 2, The excitation drive is shorted to ground, so you would expect no signal.

And another thing to confirm - is your power adapter ground tied to Earth ground? Even it it is "floating" when measured with an ohmmeter, the capacitance to Earth ground can be high (though the adapter's noise reduction Y-capacitor.)

Aside from that - the fixed resistor test is exactly what you want to do, but pins 1 and 2 must be shorted together, and pins 3 and 4 must be shorted together.

So try shorting 1-2 and 3-4, and disconnect the scope entirely and see what happens.

It's okay to probe those signals, but scope ground should only ever be connected to CN0359 circuit ground (pin 5.)

-Mark

• Hi ND_Deuterium,

Assuming your scope is a standard benchtop unit (and not a small, battery-powered, floating unit with very low self-capacitance), I suspect the problem is that you're shorting nodes to ground that really only should be connected to the sensor's excitation terminals (or your test resistor.)

In experiment 1, the excitation return is shorted to ground. This node is actually a virtual ground and should not be loaded.

In experiment 2, The excitation drive is shorted to ground, so you would expect no signal.

And another thing to confirm - is your power adapter ground tied to Earth ground? Even it it is "floating" when measured with an ohmmeter, the capacitance to Earth ground can be high (though the adapter's noise reduction Y-capacitor.)

Aside from that - the fixed resistor test is exactly what you want to do, but pins 1 and 2 must be shorted together, and pins 3 and 4 must be shorted together.

So try shorting 1-2 and 3-4, and disconnect the scope entirely and see what happens.

It's okay to probe those signals, but scope ground should only ever be connected to CN0359 circuit ground (pin 5.)

-Mark

Children
• Sorry for the extremely late reply to this message.

1. I went ahead and performed the fixed resistor test that you previously suggested. I shorted pins 1 and 2 together and then shorted pins 3 and 4 together. I used a 1000 ohm resistor and upon turning the conductivity probe on, I was able to obtain a stable reading. Fantastic. I did plug in the scope to probe the signals and everything looked good.

2.  We are now trying to have the conductivity board probe a standard solution that is being continuously stirred. We are using an LFS 1107 conductivity sensor from Innovative Sensor Technology (IST). We know that the LFS 1107 probe is working properly. As instructed by IST the probe was isolate with a non-conductive epoxy. A 200Hz square wave AC current was driven with a function generator. The voltage drop across the electrodes was measured with an oscilloscope. With the known cell constant, we were able to measure the CORRECT conductivity of the calibration solution. Numerous calibration solutions were tested, the voltage drop varied between 700mV and 60mV (depending on the molarity of the calibration solution). The LFS prove held constant at all calibration solutions and provided the correct reading.

We then plugged the four wires of the probe into the Analog conductivity board. We are not able to get a steady signal. The board settings were indeed changed to provide a 200 Hz current with a voltage drop held at 700mV. The measurement is not constant, no matter the concentration of the calibration solution. The conductivity board measurement jumps from uS/cm to mS/cm to kS/cm all while in the same standard solution. We attempted to change the desired voltage drop to numerous settings yet the results were consistent, the probe continue to jump around and cannot provide a steady measurement.

How can this be addressed and what is likely going on? Please let me know if the description of these experiments is unclear or if it would be helpful to have more information on the equipment we are using.

• So it looks like you may be encountering an issue that recently came up with another customer. As with any conductivity measurement device, there will be a small, nonzero DC component to the excitation signal, and a finite bias current on the sense leads. This is normally not a problem with larger conductivity probes, but appears to cause issues with small, planar probes, like yours:

We don't have a schedule on when updated hardware will be released, but the fix is pretty straightforward - capacitively couple the excitation and sense leads, ensuring zero DC content. (Noting that "zero" is an impossibly small number - there will always be some leakage through any capacitor.)

Here is the original discussion:

https://ez.analog.com/circuits_from_the_lab/f/q-a/120649/cn0359-conductivity-measurement-problem/360220#360220

And another more recent issue:

-Mark

• Mark,

Just to make sure I properly understand. The suggested solution suggests the problem can be fixed with a bread board with the appropriate connections between the LFS 1107 conductivity probe and the analog device.

Yes?

• Indeed, all connections can be made external to the board, no modifications / trace cutting necessary. Refer to the markup in the discussion:

• Mark,

Thank you for this information. We have implemented the revised circuit and are able to correct for the extra error that is introduced. The problem we are running into now is that we are unable to get a reading from the probe/board above 50mM MgCl2. This is a relatively low concentration and we should be able to get readings at several hundred mM. (When we use an oscilloscope and function generator we are able to get easily read 200 mM MgCl2.) The board should also be able to obtain this reading which would read in the 10-50 mS/cm range.

What has changed for us not to be able to obtain this reading? Our experiments require us to get into the 100mM MgCl2 range.

• What fixed resistance does 50mM MgCl2 correspond to, and what excitation frequency / level are you setting the circuit to?

- can you take a look at this once ND_Deuterium provides this information? We'll want to understand any limitations (if any) that the capacitive coupling imposes, and / or any settings that need to be modified from the defaults.

-Mark

• 50mM MgCl2 corresponds to a conductance of 11.8 mS or a resistance of  0.085 mohms. As I mentioned previously, we are able to get a reading here, but not at a concentration of 80 mM MgCl2, Please let me know if you need me to explore the phase space between these two measurements to provide you a more exact cut-off.

The board is set to a frequency of 300 and a voltage of 0.7V.

• Do you mean 0.085 milli-ohms? That's very low. Wouldn't it be:

1/0.0118S = 84.7 ohms?

That's right in the middle of the operating range - when used without AC coupling caps. What is the value of the capacitor in series with your excitation? We had recommended 10uF previously - but the idea is that the AC coupling cap should have a much lower impedance than the cell at the operating frequency.

1/(2*pi*300Hz*10uF) = 53 ohms

which might explain what you're seeing.

First suggestion - double check the math for the equivalent fixed resistor, and verify that you get accurate readings for perhaps 100 ohms and 50 ohms.

Second suggestion - bump the value of the excitation cap by a factor of 10 (100uF).

-Mark

• Yes, you are correct, sorry about that error on my part.

We will go ahead and begin looking into your two suggestions. Thank you.