EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: TheAmmoniacal on June 17, 2016, 05:45:38 pm
-
Just want to check if an LCR meter (like the U1733C) will double as a water conductivity meter, but I don't have any at hand at the moment (vacation).
If you have an ohmmeter/LCR meter with a test signal at 2 V AC p-p or higher (preferably a sine wave, but not important), could you measure the resistance of your tap water at 100 Hz and 1 kHz? Try to keep your probes about 10 mm apart in the water.
Do you get any sensible values? (Something around 20k to 1.2k ohm per cm)
Anyone know what test signal a professional water conductivity meter uses?
-
I was very lucky once to get this professional probe very cheap on ebay
It is a WTW LTA 01
This one is being used with simple DC voltage measurements.
-
Here is a pdf of this sensor
It can measure from 1uS/cm to 20mS/cm
I have used it on normal tap water and also on dual distilled water with a conductivity of less than 5uS/cm
-
I was very lucky once to get this professional probe very cheap on ebay
It is a WTW LTA 01
This one is being used with simple DC voltage measurements.
269 euros >:( http://www.ebay.de/itm/WTW-LTA01-2-Elektroden-Leitfahigkeitsmesszelle-8511-/291177514193 (http://www.ebay.de/itm/WTW-LTA01-2-Elektroden-Leitfahigkeitsmesszelle-8511-/291177514193)
Still, which test signal do the conductivity meter (Konduktometer) that these plug into use?
-
Just DC volts applied and use Ohms law to calculate conductivity
For this probe the mechanical dimensions are known, making the calculation easy.
Before I had this probe, I placed some metal plates in to water but it was not really repeatable at low uS/cm values.
I also have a professional conductivity meter and could open it tomorrow to take some measurements.
It will be interesting how they do it.
I would suspect also only a known DC voltage applied.
-
Measuring water conductivity should be a lot more complicated than that, try probing some water with your DMM (it won't work). You need a certain voltage to reach the reduction and oxidation potentials of the ions in solution - which are the only source of electrons. But with DC you'd be doing electrolysis and thus changing the ions present, and changing the solution (polluting it and the measurement, making values jump all over)... This is my understanding, and why I'm curious how it's done professionally.
-
But with DC you'd be doing electrolysis and thus changing the ions present, and changing the solution (polluting it and the measurement, making values jump all over)... This is my understanding, and why I'm curious how it's done professionally.
It's a long time since I pulled a conductivity meter apart, but, they do use AC to avoid the electrolysis problem - DC will rapidly polarise and/or poison the electrodes. On the meter that I pulled apart I believe the AC was a simple square-wave (AC coupled) from a 555 timer at about 1kHz. Be aware that water has a surprisingly high relative permittivity (80 ish ?) - this can mess up your measurements at higher frequencies.
The conductivity of tap water is very variable.
-
(...)
The conductivity of tap water is very variable.
[/quote]
It is very variable, but you should be able to tell if the measurement is nonsense or real. My U1272A DMM uses a 0.5 VDC test signal for ohms measurement, it does nothing in aqueous solution no matter how much salt I put in it (~10 Mohm at any probe separation length).
The one document I've found describing a water conductivity meter used a 5 V p-p square wave (AC) at 1 kHz, if this is in fact true/real - I see no reason why the 2 V p-p sine on an LCR meter wouldn't work at 1 kHz. If it does work adequately I have some plans I'd like to work on.
-
Okay, I gave this a shot for you since I am waiting for our software engineer to kludge some code for me...
I have a BK Precision 879B handy so I set it to R, 1kHz test signal connected to 2 scraps of PCB with immersion area of 1.5cm² (1.5cm deep x 1cm wide) spaced 1cm apart and got a reading of 700 ohms. The open circuit voltage of my LCR meter is 1.74Vpp, so not quite 2Vpp, but should be enough "overpotential".
And yes, we have very hard water here in Tampa, FL, USA.
-
I'd say that's within a reasonable range, good news!
More measurements please :-+
-
I'd say that's within a reasonable range, good news!
...
No problem - I thought your premise was interesting and it wasn't much effort to test it out for you.
-
I make PCBs in a laboratory sized setup.
I am not on town water, I collect rain water for most of my requirements ( my workshop is in a shed djacent to my home).
As a part of the overallprocess i recycle my water.
Initially the rain water is extruded through a reverse osmosis membrane. So obtained water is not clean enough for the process.
The water is collected in a vat and polished to the required level.
It is recirculated through a carbon filter, Ion exchange vesel and a one micron filter. The quality is measured by a conductivity instrument f own design.
The design uses CMOS analog switches to implement a synchronous detector across the output of a Wheatstone bridge.Detector output is amplified and fed into a LED bar graph driver scaled and referenced to show water conductivity through a colour coded bar graph display intended to reflect process requirement .
The sensor is made of 10mm thick polypropylene box cross section with stainless steel bolts screwed in such that the face area of bolts and spacing between polypropylene walls produces a unit cube of water in the volume of water flow. Say a 10mm Bolt is used end on area of the bolt is 0.25Pi .The distance of between the plastic walls which carry bolts is 0.25Pi cm.bolts are screwed into o posing walls till they are flush with inside of walls and a unit cube is established.
Bulk conductivity is typically measured in ohms/ cubic centimeter.
By establishing the measurement cell as per above a calibration coefficient can readily be calculated ( volume of the unit cube is converted to volume in cm cubed). any resistance measurement between the two faces of the bolts is readily converted to ohm/ cubic centimeter.
The bridge is driven y a 50% duty cycle square wave and synchronously detected to eliminate effects of electrolysis on the probe surface as well as minimise deposition of metal content into the water. This being a continuously monitoring cell.
-
You get better results using graphite electrodes, with them driven with a 50% duty cycle waveform. Actual waveform not really relevant, it must just be exactly 50% duty cycle, and preferably AC coupled to each sensor electrode. Remember as well that you need to compensate for the temperature of the water as well. I have one at work that does that, using a pair of 8mm diameter carbon rods in a 8mm diameter pipe fitting ( to get the unit volume area) with a temperature compensating NTC in contact with the cell. Then a simple drive that gives an AC signal drive and reads the resistance using the current flow through the cell, and finally displaying on a standard ICL7106 and LCD set to 200mV range. Shows conductivity from 0.0uS to 199.9uS to give the condition of the ion exchange resin bed.
Filtering is a 0.3micron filter before it, and a plain activated charcoal filter after the resin bed. Input water is prefiltered with a 5 micron filter, as the municipal pipes are old asbestos cement pipes ( no lead pipes left , they rotted out decades ago from the nice soft water, AFC piping is nicely coated with silt) and leaves a brown film otherwise on the filter.
I put another 0.3 micron filter in at home, so the drinking water at least has less sediment in it. Hate to think what it will look like in a year though, the first one at work came out with a housing full of brown sludge it had caught, as the tap off it uses is at the bottom of the pipe and I am not going to solder a new fitting on a 2inch copper pipe to correct that, I do not have a big enough acetylene torch to heat up 5m of pipe and boil 50l of water.
-
Perhaps topic started could specify test method more clearly, keeping in mind easiness to reproduce with common goods. I have 4263B available to do tests, but I'd expect random electrodes in random can with water would give random results, which may be just waste of time.
-
Perhaps topic started could specify test method more clearly, keeping in mind easiness to reproduce with common goods. I have 4263B available to do tests, but I'd expect random electrodes in random can with water would give random results, which may be just waste of time.
I don't know how all the different variables effect the measurement results. Steel probes, graphite, platinum, does it matter? This is something I'd like to experiment more on once I get back from vacation. The intention of the thread was to determine if an LCR meter is doing an actual conductivity measurement of the aqueous solution or not (if the test signal is appropriate). As long as the values measured is within a reasonable range (not off by many orders of magnitude) it would indicate that it does.
-
The OP pretty clearly stated that he was asking whether using an LCR meter as-is to measure the conductivity of water gives remotely realistic results. He wasn't asking if it gives standards-lab certified results, just something within the realm of plausibility.
I know my water is fairly conductive and a quick check with my common 879B LCR meter gave what I consider a realistic result; the OP agreed.
As for the relative merits of one electrode over another, I don't see much advantage of carbon over 304 (ideal) or 316 (ok) stainless steel in this application, even for continuous process monitoring. At 2Vpp there will be little, if any, erosion of the electrodes from electrolysis, and, of course, there won't be any metal ion transfer between electrodes due to the AC excitation. Graphite electrodes *are* a better choice for true electrolysis (though platinum plated titanium is the very best, I believe).
-
There was more to it, as someone else might want see test results for different purpose than OP :)
Anyway, here's what you wanted.
100ml glass of distilled water:
Keithley 2001 20Mohm 1.474V open lead 0.01NPLC, 70nA test current = 0.300-0.400 Mohm, voltage drop ~40mV
2MOhm range = ~0.750 Mohm, ~600mV voltage drop (test current 773.9nA)
1Gohm range = ~0.02 Gohm, ~50mV voltage drop (test current 4.56nA)
HP 4263B, R meas, medium integration speed, avg=4
100Hz, 1V = ~228kOhm
1KHz,1V = ~208k
10kHz, 1V = ~195k
100kHz, 1V = ~68k
Tap water:
Keithley 2001
20Mohm 1.474V open lead 0.01NPLC, 70nA test current = 0.700-0.800 Mohm, voltage drop ~60mV
2MOhm range = ~0.39 Mohm, ~310mV voltage drop (test current 773.9nA)
1Gohm range = ~0.013 Gohm, ~61mV voltage drop (test current 4.56nA)
HP 4263B, R meas, medium integration speed, avg=4
100Hz, 1V = ~7.42kOhm
1KHz,1V = ~6.83k
10kHz, 1V = ~6.65k
100kHz, 1V = ~6.68k
-
Thank you very much TiN, the HP 4263B measurements looks well within range ~ what you could expect. Not sure what to make of the Keithley 2001 measurements, I assume it uses a DC test signal?
-
I have opened one of my conductivity meters, a Greisinger GMH 3410
It sends a very slow pulse to the probe
- 5.3 Hz
- 1.6V
- 9ms long positive pulse
-
Here are some measurements with the WTW LTA 01 probe, shown above
German Tap Water
345 uS @ 23.3 °C (Greisinger Conductivity Meter)
540 kOhm with Fluke 87V (DC)
189 kOhm with Agilent 34401A (DC)
356 Ohm Fluke RCL PM6306 at 1 kHz
330 Ohm Fluke RCL PM6306 at 10 kHz
327 Ohm Fluke RCL PM6306 at 100 kHz
Double De stilled Water
2.7 uS @ 23.3 °C (Greisinger Conductivity Meter)
280 kOhm with Fluke 87V (DC)
190 kOhm with Agilent 34401A (DC)
45.3 kOhm Fluke RCL PM6306 at 1 kHz
44.5 kOhm Fluke RCL PM6306 at 10 kHz
44.1 kOhm Fluke RCL PM6306 at 100 kHz
-
HighVoltage :-+
Water conductivity is rather complex, and I still don't understand how the various effects play into the measurement.
When two electrodes with an applied voltage is put into tap water, at least three effects would come into play 1) Ion migration towards the electrode (cations to the cathode, anions to the anode). 2) Charge equalization by ion accumulation around the electrode 3) Redox reactions if the voltage is above or below the reduction or oxidation potential of the ionic species.
We know that a higher salt concentration will measure a higher conductivity even with the same voltage being used to do the measurement, which makes me think ion migration is an insignificant factor. If ion migration was important the reading should rise over time, which it might for very dilute solutions - but I haven't seen it.
If conductivity is mostly a measure of the number of redox reactions occuring at the electrode, I would think the electrode surface area would have a big effect. I don't think it does?
Anyone have a better explanation for how it works on the microscopic level?
-
Here are some measurements with the WTW LTA 01 probe, shown above
German Tap Water
345 uS @ 23.3 °C (Greisinger Conductivity Meter)
540 kOhm with Fluke 87V (DC)
189 kOhm with Agilent 34401A (DC)
356 Ohm Fluke RCL PM6306 at 1 kHz
330 Ohm Fluke RCL PM6306 at 10 kHz
327 Ohm Fluke RCL PM6306 at 100 kHz
Double De stilled Water
2.7 uS @ 23.3 °C (Greisinger Conductivity Meter)
280 kOhm with Fluke 87V (DC)
190 kOhm with Agilent 34401A (DC)
45.3 kOhm Fluke RCL PM6306 at 1 kHz
44.5 kOhm Fluke RCL PM6306 at 10 kHz
44.1 kOhm Fluke RCL PM6306 at 100 kHz
Did you use the WTW probe with the Fluke 87V and PM6306 too? The PM6306 has adjustable output voltage on the test signal afaik, was it set to 2 V AC?
-
Did you use the WTW probe with the Fluke 87V and PM6306 too? The PM6306 has adjustable output voltage on the test signal afaik, was it set to 2 V AC?
Yes, WTW probe on all experiments: 87V, 34401A, PM6306
Sorry, the PM6306 was set to 1 VAC
I could repeat the experiment with 2VAC, if you like
-
Perhaps this paper could help?
http://www.jumo.de/en_DE/support/faq_education/literature/Themes/Conductivity_and_ultra-pure_water/FAS624.html (http://www.jumo.de/en_DE/support/faq_education/literature/Themes/Conductivity_and_ultra-pure_water/FAS624.html)
-
128er, will take a read!
HighVoltage, just worried why your values are so far off from what I'd expect. Could you try a measurement with your PM6306 at 1 kHz, same 1 VAC signal, but using regular DMM probes in the water at 10 mm apart? I assume the Greisinger is doing some math with the cell constant and the electrodes might have a different separation?
-
I could repeat the experiment with 2VAC, if you like
No real difference between 1.00 VAC and 2 VAC
-
Water conductivity depends on the water temperature and is typically measured on opposite sides of a 1 x 1 x 1 cm cube of water (or the result is rescaled to that volume of water). Using random size water contact plates can provide relative conductivity readings, but not accurate conductivity readings in mS/cm units.
I have an Analog Devices CN0359 conductivity measurement system. It is relatively low cost ($131.25 from Mouser) and measures conductivity with an AC signal which is adjustable from 100 Hz to 10 kHz at 100 mV to 10 volts. It doesn't come with a water sample vial. Information on it is available at:
http://www.analog.com/media/en/reference-design-documentation/reference-designs/CN0359.pdf (http://www.analog.com/media/en/reference-design-documentation/reference-designs/CN0359.pdf)
Attached is a picture of my CN0359 after I mounted it in a case with it measuring a 1,000 ohm 0.01% resistor for accuracy verification. I used similar resistors of 0.1 ohm to 10M ohm resistance for other tests. The accuracy of the 1,000 ohm resistor shown is typical.
-
Perhaps this paper could help?
http://www.jumo.de/en_DE/support/faq_education/literature/Themes/Conductivity_and_ultra-pure_water/FAS624.html (http://www.jumo.de/en_DE/support/faq_education/literature/Themes/Conductivity_and_ultra-pure_water/FAS624.html)
Read it all, and it's a good summation of water conductivity measurements. But it glances over the physical and chemical aspects way too easily. The only explanation it provides is that conductivity is due to ion migration, and pays no attention to the interaction of the ion with the electrode. An explanation I do not understand at all.
-
The interaction with the electrodes seems to be quite irrelevant. When measuring with stainless steel electrodes or graphite electrodes gives no discernable difference.
The most important thing is that conductivity has to be measured with alternating current to prevent polarisation effects in the solution.
-
The interaction with the electrodes seems to be quite irrelevant. When measuring with stainless steel electrodes or graphite electrodes gives no discernable difference.
The most important thing is that conductivity has to be measured with alternating current to prevent polarisation effects in the solution.
My question is, how do you get electron flow through your circuit? (measurement apparatus) What mechanism?
-
The electrons are passed through the liquid you are measuring. The ions in the liquid take care of that so to speak.
Conductive conductivity measurements are not that difficult. Inductive conductivity is another beast ;)
-
Maybe this can clarify it a bit.
https://youtu.be/5qxenj3NpE0
-
Another variant of the conductive measurement is the 4-electrode sensor.
https://youtu.be/sVcG65dMZfk
-
Maybe this can clarify it a bit.
video here
That video also ignores the one confusion I have, if you see the video at 02:00 .. The ions are migrating back and forth due to the applied AC voltage, but again it doesn't show what gives rise to a current flow in the apparatus.
-
The voiceover tells that the current increases with the conductivity. So a higher concentration of ions in the liquids gives a higher measured current.
Around the 2 minute mark it is told that a higher conductivity give a higher current.
-
The voiceover tells that the current increases with the conductivity. So a higher concentration of ions in the liquids gives a higher measured current.
Around the 2 minute mark it is told that a higher conductivity give a higher current.
That's what conductivity means.
-
Electrons from the electrodes are being either added or removed from the ions, so they are still being used as charge carriers. Thus you get electron flow through the cells, carried by the ions, and at the electrodes the electrons are either being added to an atom or removed, but overall there is no actual reaction taking place, as the 2 half reactions cancel out with time, leading to no nett chemical reaction occurring.
A similar thing occurs in P type silicon, where the charge carriers are hole defects in the lattice, which appear to move by swapping electrons in a field gradient. Just here the charges are a little freer to move around as they are not fixed in a crystal lattice.
-
Electrons from the electrodes are being either added or removed from the ions, so they are still being used as charge carriers. Thus you get electron flow through the cells, carried by the ions, and at the electrodes the electrons are either being added to an atom or removed, but overall there is no actual reaction taking place, as the 2 half reactions cancel out with time, leading to no nett chemical reaction occurring.
A similar thing occurs in P type silicon, where the charge carriers are hole defects in the lattice, which appear to move by swapping electrons in a field gradient. Just here the charges are a little freer to move around as they are not fixed in a crystal lattice.
If we can agree that reduction and oxidation of the ions occur during conductivity measurement, I'm good. But that necessitates a certain minimum voltage required to account for all possible solutes. Say you want to measure the conductivity of lithium chloride (LiCl), the reduction of Li+ requires 3.07 V, so unless your test signal is 3.07 V or above you won't be able to measure much. And oxidation of chloride would give a neutral solvated gas molecule that won't migrate. (These are standard reduction potentials, measured at 1 M concentration, at 25 C and 1 atm, not /real/ values, but still representable)
A more benign example is NaCl, Na+ reduction requires 2.71 V, can you measure conductivity of NaCl with a 2 V p-p sine or square? How? Why do many conductivity meters use a lower voltage? Are they selectively measuring ions with lower redox potentials?
-
What do Redox potentials have to do with it? The conductivity measurement isn't reducing or oxidising anything.
-
What do Redox potentials have to do with it? The conductivity measurement isn't reducing or oxidising anything.
If it's not, can you please explain how it works? :)
-
Well, the ions in the solution conduct electricity.
You apply a voltage to the electrode (I just checked mine and it gives out 2.28V @ 128Hz).
The cations will then move to the negative electrode and the anions move to the positive electrode the solution acts like a electrical conductor.
Different ions have different activity (i.e. they move less or more easily) and thus they give different conductivity values.
A conductivity sensor is not selective it basically doesn't care which ions there are in the liquid as long as they are mobile. The conductivity is also highly affected by the temperature in which a higher temperature will cause an increase in conductivity due to the higher mobility of the ions.
-
If an ion takes up an electron, what happens to it? What value determines how easily a given ion takes up an electron? Or how much energy you need to add one? Or remove one? If a cation (positively charged ion like Na+) takes up an electron, it becomes neutral, and thus should stop migrating or moving in the electric field, why doesn't it? (according to you)
-
Ok, I will look into it the theory a bit deeper and will come back here.
But I think that it is a misconception that ions are neutralised by the conductivity measurement otherwise the conductivity of water would become zero over time which it doesn't.
To my knowledge the charge is pased through the solution, if according to your idea a Na+ ion would take up an electron you would end up with metallic Sodium in the solution.
I got my scope out to check the signal from my conductivity meter. It gives 2.25 V RMS and 5 V P-P. The signal is a 128Hz square wave.
-
To my understanding your conductivity meter has the expected test signal (5 V p-p would cover all redox potentials).
In order for current to flow in a circuit, you need a closed circuit -- this is obvious. What I'm still struggling to understand is how a conductivity measurement makes a closed circuit through the aqueous environment. I haven't been able to find an explanation that is consistent with both the chemistry and electronics involved.
Reduction of the sodium ion to sodium metal is exactly what you get when you do electrolysis of aqueous sodium chloride (brine):
$$ Na^{+} + e^{- }\rightarrow Na$$ The Na atom spontaneously react with water: $$ Na + H_2O \rightarrow 2 NaOH + H2$$ which then subsequently solvates $$NaOH \rightarrow Na^+ + OH^-$$. With the oxidation of chloride to chlorine gas Cl2 at the anode.
There is nothing preventing this from happening during a conductivity measurement if the peak voltage is at 5 V, even if you're using AC this should occur as long as there are ions at the oppositely charged electrode. I don't see any means of reversing this electrolysis concerning Cl2/Cl, but the OH- can be oxidized to water again, and the Na+ reduced again to give OH-.
I want to know, does this occur during a measurement? If you only measured long enough, would all the chloride disappear? Even on AC? How far does a given ion migrate during the 1/100th of a second?
Sorry for turning the thread into a chemical soup..
-
Here is a quite from the McGraw-Hill Encyclopedia of Science and Technology.
The transport of electric charges, under electric potential
differences, by particles of atomic or larger
size. This phenomenon is distinguished from metallic
conductance, which is due to the movement of
electrons. The charged particles that carry the electricity
are called ions.
Positively charged ions are termed cations; the
sodium ion, Na+, is an example. The negatively
charged chloride ion, Cl?, is typical of anions. The
negative charges are identical with those of electrons
or integral multiples thereof. The unit positive
charges have the same magnitude as those of
electrons but are of opposite sign. Colloidal particles,
which may have relatively large weights, may
be ions, and may carry many positive or negative charges.
The ions physically move to the opposite electrodes Na+ moves to the negative electrode, Cl- moves to the positive electrode.
The movement of the ions constitute the electric current flow through the electrolyte.
-
This might give some additional insight as well.
http://www.currentseparations.com/issues/18-3/cs18-3c.pdf (http://www.currentseparations.com/issues/18-3/cs18-3c.pdf)
-
More papers like that please, very informative. Still didn't answer the question though.
You say that current is the movement of ions, but ions don't / won't move into the metal electrode. At every point of a closed circuit electrons must be able to move. Unless it's a capacitor?
-
You are not the only one with this question :) There is a good anwser in this link as well.
http://chemistry.stackexchange.com/questions/4979/conducting-current-in-electrolytes
-
Just to conclude this thread. I think I've finally understood what's going on here, I'll try to summarize it below - please correct me if you see anything wrong.
A water conductivity measurement is analogues to measuring across two capacitors in series, with the metal electrode being one plate and the adjacent electrolyte being the other plate. I've also found sources giving ion migration speeds of 0.0001 mm/s if exposed to a field strength of 1 V/cm. If I'm not mistaken, a measuring voltage of 2 V would give a 1 V field strength at 0.5 cm (middle of the two electrodes if separated by 1 cm). Which tells me ion migration is insignificant for ions not very close to the electrode. The electrolyte in the solution is akin to the electrolyte in capacitors.
A conductivity measurement would be related to how many counter-ions (counter to the electric field) there are and how fast they accumulate at the electrode surface.
What would be the physical interpretation of conductance if we imagine the measurement as two series capacitors like in the illustration below?
-
Conductivity is another name for total dissolved solids content of water.
So yes as contamination increases there are more ions in solution therefore IONIC equivalent current increases and resistance accordingly measures as a lower value.
And NO it is not the same as displacement current in a capacitor.
-
For what it's worth, I'm a volunteer water sentinel with the Sierra Club in Ohio, and we use Extech EC400 meters to measure temperature, total dissolved solids, salinity, and conductivity. We also use calibration fluid and the meter's calibration setting before we head out. The fluid is available in various values, but we all use 1413 ?S. But as the container shows, that's nominal and varies by temperature. These are consumer grade testers, but a fellow volunteer checked it against his lab's meter and said its surprisingly decent after calibration.
After seeing this, I checked the meter using my oscilloscope and found that the meter puts out a sine wave with 1 volt peak-to-peak at 500 Hz in all modes.
-
1.413 mS/cm is the most used standard because it's easy to make and therefore cheap.
It is a 0.01M KCl solution so you can make it yourself quite easily.