Ah, this is all so great!
This brings up a bunch of interesting issue for me that I'd love to get feedback on.
Naming conventions. Maybe it'd be good to start to figure out a useful naming convention for these various projects; I'm starting to get confused :) For me, I see the Riffle, which has now been used as a basic datalogger board for several projects that don't involve water quality, as a sort of general data logger. I love Mat's suggeston of the 'Sniffle' as the air quality version; and Craig recently made a 'Muffle' version that is intended as a mushroom grow room monitoring device. Maybe we should come up with a name for the basic datalogger that is separate from any of these names? At the risk of generating two many entities to keep track of.
So, Riffle Coqui starts to get even more confusing for me :) But that's par for the course when all of this fantastic development happens so quickly with a bunch of fun and enthusiastic folks!
Just thinking out loud, here: my notion of the Coqui:
had started to move in the direction of a device that "translates a resistance value into an audio-range frequency" -- where the resistance could be a photocell, a pair of electrodes in water, a thermistor, or many other things. I'd come to think of it as separate from the Riffle, which is expressly intended as a device that measures certain standard water quality parameters.
I'm also thinking of the 'Coqui' as more general in application, at this point, because I'm not yet comfortable with what exactly we're measuring when set it up with water as one of the resistors. The 555 sends an oscillating current through the water between the electrodes, and if the oscillation is too slow, the water will become polarized by the measurement, changing the apparent conductivity of the water; avoiding polarization is also a reason to try to use the minimal voltage necessary for the measurement (my understanding is that 3V is on the 'high' side, but we need to look into that). All that said, if all we're making is a comparative measurement (is the conductivity higher or lower in sample A than sample B), then it seems we could get away with not worrying too much about polarization from our instrument. But because we know that the temperature of the sample will affect its conductivity, as you say (high temp --> higher conductivity), we're going to have to be careful about making sure that any differences in 555 frequency aren't simply due to temperture ...
Tap water experiments. Okay, so here are some associated proposed to-dos / questions on my end, to clarify my own thinking, here:
If the conductivity of the water is observed to be different for tap water samples that were obtained through 'hot water pour' and 'cold water pour', and then maintained at 25 C, what might this be due to? Is lead in the pipes the only / most likely explanation?
If lead from old pipes is a concern, what concentrations are typically present in hot tap water due to lead piping, and how much would this affect the conductivity?
What is the affect of temperature on conductivity? I found a nice chart, here, that shows the effect in a graph:
These show the change in measured conductivity of solutions for different solutions of known conductivity, for a range of solutions that are on the high end of potable water in conductivity. Not sure it's reasonable to use this as a rough guide, but from the chart it looks to me like conductivity of these solutions changes by about 30 uS/cm per degree C. Since this temperature effect seems to be larger for higher conductivity solutions, and this is on the high end of potable water conductivity, let's use this as a conservative estimate.
Stumbling onwards: some typical examples for conductivity are in this chart (from the same reference, above):
My conclusion is: a difference of a few degrees Celsius (meaning a change of 100 uS/cm or so) isn't going make tap water (50 uS/cm to 800 uS/cm) appear to be in the conductivity range of industrial waste (10,000 uS/cm), or sea water (55,000 uS/cm); but it seems like it can push us around the typical range of tap water by as much as 15% of the expected range.
For assessing incursion of seawater into a freshwater water source, it looks a difference of a few degrees in the samples isn't going to matter much when making comparisons.
For the tap water - lead question, my next step here would be to find out how much we'd expect the conductivity to change in tap water, in terms of uS/cm, due to the presence of lead from pipes. It might be easy to find that information. This difference will then set our threshold around how close we need our sample temperatures to be. (There are probably a bunch of other confounding factors here that I'm not thinking of .... ?)
Places I'm reading next: