Sciencemadness Discussion Board - Conductivity cell: DIY ideas needed!

Conductivity cell: DIY ideas needed!

blogfast25 - 28-7-2013 at 07:14

Someone delivered a conductivity cell to me by mistake and is allowing me to keep it, gratis!

Except, I don’t really know how to make it work w/o the usual console. Wiki states that DC [edit: oooops, AC was meant of course] voltage of 1 – 3 kHz is used to measure current across the cell but not much more.

Anyone have decent ideas to make this a working instrument with some DIY?

[Edited on 28-7-2013 by blogfast25]

[Edited on 28-7-2013 by blogfast25]

IrC - 28-7-2013 at 07:39

Actually 1 to 3 KHZ is AC not DC. The wiki page states "An alternating voltage is used in order to avoid electrolysis."

http://en.wikipedia.org/wiki/Conductivity_%28electrolytic%29

The simplest method is to use a conductivity meter.

http://en.wikipedia.org/wiki/Electrical_conductivity_meter

You could build a battery powered inverter circuit which does not rectify the output so that you get a low voltage AC sine wave. If you use say a 555 timer IC you can easily adjust the frequency for best operation. A bridge circuit could be used to measure the impedance of the cell. I do not think it would be too difficult to build your own conductivity meter. Calibrating it may be hard, but not impossible even if you build the meter from scratch.

As a point to start you may wish to study :

http://www.octiva.net/projects/ppm/

http://blea.ch/wiki/index.php/PPM_Meter

While not exactly what you are looking for it should give you ideas plus the information on calibration is most useful.

Attachment: 12chapter12.pdf (55kB)
This file has been downloaded 818 times

Another idea: http://www.practicalmaker.com/products/arduino-shields/ec-sh...

[Edited on 7-29-2013 by IrC]

bfesser - 28-7-2013 at 07:42

<a href="http://www.analytical-chemistry.uoc.gr/files/items/6/618/agwgimometria_2.pdf" target="_blank">Conductivity Theory and Practice</a> <img src="../scipics/_pdf.png" />
<a href="http://www2.latech.edu/~dehall/LWTL/ENGR121/notes/3_conductivity_sensor_intro.pdf" target="_blank">Conductivity Sensor Implementation</a> <img src="../scipics/_pdf.png" /> for the Arduino

Here's a good one you can build with off the shelf components:
<a href="http://www.ullasmann.eu/Doc/Salt_Concentration_Meter.pdf" target="_blank">One-chip conductivity meter monitors salt concentration</a> <img src="../scipics/_pdf.png" />
In fact, I like this circuit so much, that I may try building one.

This product might be of interest:
<a href="https://www.atlas-scientific.com/product_pages/embedded/ec.html" target="_blank">Atlas Scientific Conductivity Circuit</a> <img src="../scipics/_ext.png" />

Related:
<a href="http://www.rsc.org/binaries/LOC/2010/PDFs/Papers/583_0150.pdf" target="_blank">A miniature high precision conductivity and temperature sensor system for ocean monitoring</a> <img src="../scipics/_pdf.png" />

[Edited on 30.7.13 by bfesser]

blogfast25 - 28-7-2013 at 08:43

Thanks bfesser, I meant AC of course.

Now let me scrutinise these links.

blogfast25 - 28-7-2013 at 08:54

Quote: Originally posted by bfesser

Here's a good one you can build with off the shelf components:
http://www.ullasmann.eu/Doc/Salt_Concentration_Meter.pdf
In fact, I like this circuit so much, that I'm going to try building one.

It looks easy enough (even though I'm totally rusty on electronics in general). It will probably be more expensive to build this than buy a half-decent TDS pen-type meter though...

bfesser - 28-7-2013 at 09:26

You asked for DIY. Where's the fun in just buying a pen meter? Where's your sense of excitement for amateur scientific experimentation?

Depending on where you purchase the components, yes, a pen meter could be cheaper. Whereas I have most of the components, if not all of them, on hand already. Looking over the schematic, all I need is a sacrificial voltage meter and a ±15 VDC supply—thinking about using an <a href="http://www.ti.com/lit/an/snoa851/snoa851.pdf" target="_blank">LM311</a> <img src="../scipics/_pdf.png" />. I don't have a conductivity probe, but I do have some <a href="http://en.wikipedia.org/wiki/Electroporation" target="_blank">electroporation</a> <img src="../scipics/_wiki.png" /> cuvettes. I wouldn't expect significant accuracy or precision, but it'll be fun to experiment.

If you just want a dirt cheap ad hoc solution, connect a multimeter to it, and measure the resistance in the first couple seconds before it changes too much. You might destroy the probe through electrolysis, though. You can generate a calibration curve by preparing solutions of known conductivity from DI H2O and a suitable electrolyte. The CRC Handbook (1st Student ed.) has a table of Equivalent Conductivities, λ, of Some Electrolytes in Aqueous Solution at 25°C, and I'm sure you could find others.

[Edited on 30.7.13 by bfesser]

blogfast25 - 28-7-2013 at 11:49

Quote: Originally posted by bfesser

You asked for DIY. Where's the fun in just buying a pen meter? Where's your sense of excitement for amateur scientific experimentation?

Now, now. I'm only saying a cheap pen type TDS meter would be cheaper.

I'm not in urgent need of a conductivity meter, so I'll probably take it slow with the free cell obtained. But I'd like to turn it into a usable scientific instrument, if it's reasonably doable.

Using loads of uncertified DIW as I do, I might even have a REAL use for it too...

Let's just see if some more opinions pop up.

blogfast25 - 28-7-2013 at 13:17

I wonder what the results would be if one used, say 5 V AC at 50 Hz. For that even I can design a 'circuit' that uses the 220 V mains!

watson.fawkes - 28-7-2013 at 15:48

Quote: Originally posted by bfesser

Here's a good one you can build with off the shelf components

That's components that were off the shelf in 1977. There are far more effective ways of doing the same thing now.

The most important change is inexpensive microcontrollers with multiple onboard A/D and D/A, such as the AVR chips used in the Arduino. For the current project, you'd use one D/A pin as an oscillator output, an integrated 4-wire current sense resistor (they're just a few cents in surface mount) and an A/D to sense the current induced. You might need to use the AREF (analog reference) pin to set a calibrated voltage, and there are plenty of one-chip versions of that around. If you've got an AVR programmer (and thus don't need the Arduino bootloader), you could do the whole thing for just a few dollars in parts.

watson.fawkes - 28-7-2013 at 15:54

Quote: Originally posted by blogfast25

I wonder what the results would be if one used, say 5 V AC at 50 Hz. For that even I can design a 'circuit' that uses the 220 V mains!

You don't want 5 V; you want about 50 mV, or around two orders of magnitude less. At 5V, you'll be generating enough voltage to drive electrolysis reactions. At 50 Hz, you're getting relative long drift times for the ions in solutions, enough time to polarize the electrode and start the reaction up. In other words, no.

You want to just measure the electron flow, while the ions vibrate a little in place. Therefore, low voltage, (relatively) high frequency.

bigmol - 29-7-2013 at 03:05

Re Conductivity electrode.
50 cycles works just fine provided that total current through probe is just a few mA. Traditionally a mA meter with a full wave rectifier bridge is used. Modern diodes exibit a fairly constant forward zener effect. Big problen is constant voltagefor this use a simple shunt regulator with back to back zeners. This essentially produces an ac square wave. Put a potentiometer accross the zeners to do final voltage adjustment for distilled water you would typically be using up to 5 volts for conductivity grade high purity water as per modern thermal power stations you might be using 50 volts but for saline only a few mV greater than the zener effect of your meter diodes.
This sort of set up was widely used 40 years ago in commercial instalations for monitoring and lab operations.

blogfast25 - 29-7-2013 at 10:29

Thanks all for your responses so far.

Unless someone can come up with a tried and tested circuit based on more modern approaches, I think me and my daughter will attempt the ullasmann.eu circuit (she has some practical experience with actual PCBs). Neither of us are familiar enough with electronics and the theory of operational amplifiers to transform that circuit into a 21st century device.

The circuit calls for a 15 V DC power source, probably the most expensive part of the whole thing. What would you people recommend as such a source?

IrC - 29-7-2013 at 11:02

Actually it calls for a 30 volt supply, +- 15. I would use 4 discrete IC's, and a true Instrumentation amplifier such as the AD8221. The quad used in the PDF would be a poor choice for the cell. You want the input buffers to reduce error from varying cell impedance as well as improving common mode characteristics. A wall wort putting out around 15 VAC using a doubler circuit and regulators such as the 78L15 and 79L15 would be cheap and provide a fairly decent dual supply.

bfesser - 29-7-2013 at 13:37

Quote: Originally posted by watson.fawkes

Yes, and they're still off the shelf. There's nothing wrong with the circuit; he's not working in an analytical laboratory. An Arduino is overkill, and your suggestion is way to complicated for him.

Quote: Originally posted by IrC

Actually it calls for a 30 volt supply, +- 15. I would use 4 discrete IC's, and a true Instrumentation amplifier such as the AD8221. The quad used in the PDF would be a poor choice for the cell. . . . wall wort putting . . .

Really, you're going to contradict him just to say that it's a 30 V supply? Semantics. Also, there's nothing wrong with the chip the circuit calls for. Your suggestion is wasteful overkill. He's looking for an easy DIY circuit, not your unhelpful electrical engineering jargon. Since you're being so pretentious and arrogant; it's wart, not wort.

Excised post, now restored

IrC - 29-7-2013 at 15:03

bfesser "Really, you're going to contradict him just to say that it's a 30 V supply? Semantics. Also, there's nothing wrong with the chip the circuit calls for. Your suggestion is wasteful overkill. He's looking for an easy DIY circuit, not your unhelpful electrical engineering jargon. Since you're being so pretentious and arrogant; it's wart, not wort. "

You have created every shred of 'bad attitude' in my words. I was being helpful and said nothing which can be construed as arrogant in any way. I pointed out he needed twice that voltage because he does. Saves him from buying the wrong supply ahead of time. Yet here you are fucking attacking me for reasons which you made up in your own mind. The input offset and CMMR of that opamp is crap and he will be dealing with very low voltages. I was trying to give him advice to improve it since as he requested he would like to make a useful scientific instrument. You had absolutely no right to 'imagine' emotions and motivations on my part. I was spending my own time and research in an effort to help a member out. What gives you the right. Above all people on this site moderators should go above and beyond the call in being proper in the way they conduct themselves.

By the way bfesser the PDF you posted with the circuit in question I posted first in the first reply to this thread. I saw you come in and post it after moving the thread right in the middle of me writing the first reply, so I edited it to remove the link. That was a kindness on my part to remove the appearance that you did not bother reading the first reply before posting the link yourself. Not much of a mystery since that PDF is one of the first hits on google when searching for conductivity meter. Exactly how I found it. Decency would dictate that you also let blogfast25 decide if he thought I was being rude to him rather than you coming in and making that assumption yourself while defending someone not in need of it.

[Edited on 8-1-2013 by Polverone]

watson.fawkes - 29-7-2013 at 15:05

Quote: Originally posted by bfesser

An Arduino is overkill

What, are you with the discrete components union?

Transistors are enormously inexpensive, and throwing lots of them at a problem is a very good solution to lots of problems, including the present one. Using modern components here is not more expensive than using archaic ones. In the present case, it's cheaper, since a single-ended +5 V supply is far less expensive that dual-ended +/- 15 V one. As a bonus, you get digital reading, integration with logging systems, higher accuracy, greatly reduced need for calibration.

You like the circuit? Fine. Build one. But don't make the mistake of asserting that it's a better circuit just because you can build it from an schematic and don't know how to think about designing one with more modern device.

watson.fawkes - 29-7-2013 at 15:17

Quote: Originally posted by IrC

The input offset and CMMR of that opamp is crap and he will be dealing with very low voltages.

The original versions of that part weren't particularly great. The recent versions are much better: 1 mV input offset voltage, 4 nA input offset current, CMMR 90 dB. Of these, the biggest problem is the 1 mV input offset, since that's 1% of full scale (at 100 mV) and would limit accuracy accordingly. If two significant digits of accuracy (best case) is acceptable, that part is fine.

bfesser - 29-7-2013 at 15:42

watson.fawkes, I see your point about the power supply cost offsetting any advantage. Perhaps you could explain in more laymen's terms how you would implement an Arduino for this?

IrC - 29-7-2013 at 15:43

watson.fawkes "The original versions of that part weren't particularly great. The recent versions are much better: 1 mV input offset voltage, 4 nA input offset current, CMMR 90 dB. Of these, the biggest problem is the 1 mV input offset, since that's 1% of full scale (at 100 mV) and would limit accuracy accordingly. If two significant digits of accuracy (best case) is acceptable, that part is fine."

Yeah I agree. Laser trimmed resistors is critical in that regard and I think instrumentation devices are best. But not merely the Laser trimmed resistors improving offset, the input buffer stages which are not in the part number called out in the PDF. Those buffers go a long way in making the cell provide better data, ans as well would reduce electrolysis. I did not see why it is a bad idea you can get all 4 of the devices I mentioned for nearly the same amount Radio Shack would overcharge for a single sloppy quad.

[Edited on 7-29-2013 by IrC]

watson.fawkes - 30-7-2013 at 03:14

Quote: Originally posted by bfesser

Perhaps you could explain [...]

Perhaps I'll feel well-disposed enough to continue this conversation with you after you restore IrC's post that you sent to Detritus, the one I that quoted above. You're the one that started it, after all:

Quote: Originally posted by bfesser

In case you are unaware, this kind of post is called a flame. Worse yet, it includes a spelling flame, widely considered utterly petty. The phrase "unhelpful electrical engineering jargon" reflects poorly on you; it's exactly equivalent to the phrase "unhelpful chemistry jargon" used in response to an answer that includes basic terms of art that the k3wl is ignorant of.

And as for the two technical points, they're both wrong. blogfast25 wrote "15 V DC", not "+- 15 V DC", and whether or not it was a typo, a correction is anything but semantics and certainly warranted. And there is too something wrong with that chip, more than one something, in fact. The input offset voltage is rather too high, particularly in comparison with recent op-amps (of the same cost) with much lower values for this figure of merit. And a second thing wrong with is that it requires a split supply at much higher voltage than necessary with all that attendant cost, something absolutely not needed for this application. Even if you want to stay with op-amps, there are single-ended op-amps that operate at lower voltages that would do just fine for this application.

blogfast25 - 30-7-2013 at 04:23

Errrmmm... ok.

So a +-15 V DC is needed, I kind of understood that from the amplifiers on the diagram.

I do have quite a simple question before we start buying components.

As such the circuit only yields, from what I can tell at least, the voltage across the cell and not the current. The voltage is of course indirectly proportional to the conductivity but the latter can't be calculated without the current (and the cell constant).

Any way of measuring the current through the cell?

Watson: you believe this circuit will not work or will not work optimally?

watson.fawkes - 30-7-2013 at 04:36

Quote: Originally posted by blogfast25

The circuit calls for a 15 V DC power source, probably the most expensive part of the whole thing. What would you people recommend as such a source?

If you're going to be doing this kind of thing more than once, the quickest way is to acquire a bench supply. They're not as expensive as they used to be (Chinese manufacturing). If you go this route, make sure you get one with adequate short-circuit (overcurrent) protection, since you will accidentally short it out at some point.

Probably cheapest to build one. Split supplies aren't nearly as common as they once were, so you're unlikely to find a wall-wart version. If you need a complete schematic, the best place to look is in the hobbyist audio world (for example). A simple linear supply is pretty basic: a center-tapped transformer, a bridge rectifier, filter capacitors, and a pair of three-terminal regulators. I'm not familiar with UK suppliers, but I found this part TF006, a 30 V CT transformer that would allow you to build a +- 12 V regulated supply with. (Count on at least 2 V drop on each side for diode and regulation loss.) The op amps should run OK on that lower voltage. For regulators, a pair 7812 and 7912, one for +, one for -.

watson.fawkes - 30-7-2013 at 05:48

Quote: Originally posted by blogfast25

As such the circuit only yields, from what I can tell at least, the voltage across the cell and not the current.
[...]
Any way of measuring the current through the cell?

Watson: you believe this circuit will not work or will not work optimally?

The second op-amp in that circuit is configured as an AC amplifier for the signal at point A. The gain is determined by the ratio of resistances: the input resistance between A-B and the feedback resistance P₂ + R₁₅. The cell resistance is in the denominator, accounting for the hyperbolic shape of output voltage in figure 2(b). Think of it as linear in conductivity (inverse resistance). Since conductivity has such a large dynamic range overall (think distilled water vs. ocean water), this device is only able to look at a relatively narrow range of it at any given time. You'd want logarithmic response if you want a wide-range device.

There are lots of ways of measuring current directly. You could use a four-wire sense resistor, for example, as I mentioned before, to convert current into voltage and sample the voltage. There are others, but I'll decline to just rant about the topic in general, barring a more specific goal.

As to optimality, it all depends on what you want to accomplish. If you're looking for a learning experience, you'll certainly get one. If you care about instrumentation electronics at all, it certainly won't hurt to get some bench experience with op-amps. They remain the standard technique for signal conditioning, even with otherwise-digital devices. If you want an analytical device, consider that circuit a version 1.0 level of quality.

blogfast25 - 31-7-2013 at 09:52

Now armed with much greater wisdom re. this project, I think the pay off is too small for the effort (by two people very rusty on electronics) and cost involved. I'll probably invest in a £20 pen type conductivity meter.

Here's a nice resource on the theory of conductivity (solutions) measurements:

http://www.google.co.uk/url?sa=t&rct=j&q=&esrc=s...

Anyone wanting to place a bid on the conductivity cell, please U2U me (suspicious minds: NO, that wasn't the 'master plan'). [IF NO BIDS THEN ESHOP]

Thanks all, including Irc, for their contributions.

Mildronate - 21-9-2013 at 11:05

how you calibrate it?

IrC - 23-9-2013 at 20:23

Not much activity on this subject lately. Thought I would post a few PDF's which are useful. For my photodiode gamma detector experiments I believe in extreme precision, excellent performance, meaning I like the LMP7721 3 femtoampere input current Precision OpAmp. So here is a little info on this plus a couple documents about designing instruments useful in our experiments. So much can be gained by not choosing poor quality opamps. Edit to add a Pspice model and a selection guide, plus a couple useful pages.

http://www.planetanalog.com/author.asp?section_id=3140&d...

http://www.planetanalog.com/author.asp?section_id=3140&d...

Attachment: lmp7721.pdf (992kB)
This file has been downloaded 812 times

Attachment: LMP7721MAX-NOPB-Texas-Instruments.pdf (664kB)
This file has been downloaded 1594 times

Attachment: snoa514c.pdf (62kB)
This file has been downloaded 583 times

Attachment: snoa529a.pdf (116kB)
This file has been downloaded 517 times

Attachment: snom096.zip (4kB)
This file has been downloaded 478 times

Attachment: sloc264a.pdf (716kB)
This file has been downloaded 818 times

[Edited on 9-24-2013 by IrC]

Mildronate - 11-10-2013 at 05:54

actualy its very interesting, because with conductivity measuring you can done some research on organic reactions kinetic and mechanisms, using simple condutivity electrodes and data loger or chart recorder, but how to calibrate it for mS and mmS, with KCl?