Is 2 taken to the power of one million minus 1 a prime number? NO! and I didn't have to calculate it to find out......I used some LSD

 Mersenne prime numbers are ones that have the form 2^N - 1.  Not all (actually relatively few) prime numbers have this relationship, and of course not all numbers that can be calculated using that formula are prime.  For a simple example, 2^4 = 16.  16 - 1 = 15, which is divisible by 3 and 5 -- so it's not a prime number.

You will have to read on to learn about the LSD.

Prime numbers are important when it comes to generating highly secure encryption codes, so they have been of interest for a long while.

For some reason, perhaps yet another sleepless night, I started thinking about powers of two, in terms of their digits.  More specifically, if the least-significant digit of them has any kind of pattern to it.  Some simple mental arithmetic revealed the answer, and it should become obvious when I write down the first few powers of 2, starting with N = 1:   2, 4, 8, 16, 32, 64,128, 256....and so on.  Looking at the Least-Significant Digit (the LSD, gotcha!!!) of this series we see:  2 4 8 6 2 4 8 6 .... so we have a sequence of 4 digits that endlessly repeats:  2 4 8 6 .... A little more mental gyrations and I came up with a way to predict what the first digit of any power of 2 is.  It does take a little more math, requiring the use of the Residue function.  Residues are calculted by getting the remainder of long division.  It's easier to show by example, like this:  take a look at 10/4.  Long division gives us a quotient of 2 because 4*2 is the nearest multiple of 4 that is closest (but not larger than) 10.  10 - (4*2) gives us the remainder, 2.  That is what the Residue function produces -- the remainder.  So if we examine the Remainder of (any whole number)/4, we find they can only be either zero, one, two or three.

Now let's create an array with the values [6,2,4,8] in it.  The array entries are a little different than what you might expect since the indices into the array are 0, 1, 2 and 3....4 isn't possible because its residue is zero.

Now let's determine what the LSD of 2 taken to the one-millionth power must be.  Some simple math says that the remainder of 1,000,000/4 is 0 (this will be true for any power of 10 greater than 1).  The first entry in the array is a 6, so we know that the LSD of 2 to the millionth power is a 6.

Recall that Mersenne primes have the form 2^n - 1.  If we subtract 1 from 6, we get 5, and all numbers ending in 5 are divisible by 5.  Therefore, it is NOT a prime number.

For the same reason, we also can say that any number calculated by evaluating 2^(10^n) -1 also are NOT primes, as long as n is greater than 1.

It so happens that we can use a similar but slightly more complicated scheme to determine what the next-most-least-significant digit (NMLSD) of a power of 2 is.  I won't go into much of it here except to say that the sequence has a length of 20.  The NMLSD of 2^1,000,000 is a seven.

After that the sequences become ever-longer so the approach becomes less and less viable.  If nothing else, it becomes necessary to accurately calculate some pretty large numbers, just to examine their smallest parts.

Touch Sensor Update

In some ways it's been a rough year, what with Covid and a rental-rehab project we found ourselves saddled with.  Supply-chain and contractor issues caused problems; and in many cases we found that the most expedient way to move forward with the project was to do the work ourselves.  We did a lot of research before embarking on any of the major projects we had to do.  Anyway, all that delayed work on my touch sensor -- but now that we've got the rental fixed up (and rented), I've had time to work on some long-delayed personal projects.  That includes the touch sensor that I designed a PCB for.  I did have time to order the PCB and assemble one, but that's about as far as it got until recently.  I finally was able to hook up my 4-point sensor connectors and test the thing:  and, what a surprise -- it worked, right off the bat.  

The wires connecting the modified battery-charger clips to the circuits are a mess, since I used individual wires but I have some cable management stuff I can wrap around them to make it all a little less like an octopus waiting to snare me when I pass by.

Now I'm working on a really sad antique dresser we bought a few years back.  When we bought it we didn't realize what bad shape it was in, so it needs some work -- to put it mildly.  I think a child may have used some of the drawers as a ladder and stepped through the bottoms.  The dovetails on the lowermost drawers were loose, and the rabbets on several of the side pieces (the ones that hold the bottom piece in place) were split or just plain broken off.  I also had to reinforce the sides for a couple of them.  It has water damage, too -- the oak veneer on one side of the case has delaminated.  I'm not going to try to repair that for now -- it basically was purchased to put in a guest room so visitors should just appreciate having a dresser, however it looks (as long as it is usable, anyway).  The feet are a mess, too -- three have the remnants of some sort of steel foot, and there's nothing at all on one of them.  The steel will be pretty bad for scratching our wood floors so there's some work to do there before the dresser is put into service.  I learned a lot during our rental rehab w/regard to doing stuff like trim work so that will come in handy for this project.

Oxygen, the master vampire element

 As a preface to this entry, I'm going to write about my first real experience with what I now call the master vampire element, oxygen.  At the time, I was working on a different approach to etching gold.  Since gold is a relatively inert element, it takes some doing to etch it -- basically, turning the metal into a salt of some kind.  I was thinking about gold chloride.  Aqua Regia is a commonly-used etchant for gold, made by mixing nitric acid and hydrochloric acid.  Thing is, the mixture is unstable because the two acids react to form something called Nitrosyl Chloride -- and it quickly decomposes.  It also takes some time for NOCl (its chemical formula) to form so you're running a race between getting the etchant working and then using it before it decomposes.  There also are a number of different ratios given for the ingredients, probably because they come in a number of different concentrations.  So I had some interest in coming up with something that was more stable and more reproducible.  I had concentrated hydrochloric acid available, the same with 30% hydrogen peroxide, so I had the thought of combining the two to see how that would work.  The idea was that the peroxide would oxidize the gold and then the acid would react with the oxide to form its chloride.

Well, my new etchant sort of worked but it turned out to be even more unstable than aqua regia.  The REALLY interesting part was that my mixture quickly decomposed by releasing a green-yellow gas:  chlorine.  Well now, what was that about?  It didn't take long for me to realize that the hydrogen peroxide had done it, using its extra oxygen atom to grab two hydrogen atoms from two molecules of hydrochloric acid (HCl), forming one molecule of water and one molecule of Cl2.  Up to that point, I had thought that chlorine was a pretty strong oxidizer and was pretty safe from being affected by oxygen:  but my little experiment blew that notion right out of the water.  BTW I performed my experiment with just a small quantity of the two materials, under a fume hood so no harm done.

Now I want to talk a little about the idea of "valence".  Fundamentally, it means how many electrons an element in a compound has either gained or lost:  or wants to gain or lose.  Many reactions are all about electrons.  For instance, in the water molecule we have two atoms of hydrogen and one atom of oxygen.  Oxygen has a valence of 2, because it "wants" two additional electrons to fill its outer shell (and each hydrogen atom only has one to provide, so it takes two to form a stable molecule).  And oxygen REALLY wants those electrons, as shown by my little experiment.

It gets even more interesting though.  Looking at chlorine (Cl), it has a valence of 1 when it combines with things like sodium to form sodium chloride, table salt.  In that case chlorine is the oxidizer and sodium is the reducing agent.  But oxygen is such a powerful oxidizer that it can actually wrest electrons _away_ from chlorine, which in itself is no slouch as an oxidizer.  In fact, oxygen is so powerful that it can abduct SEVEN electrons from chlorine, forming perchlorate compounds.  They are used to make explosives in fireworks.  Perchlorate compounds themselves are extremely powerful oxidizing agents, so if mixed with things like charcoal and sulfur they are more than ready to go boom.  Perchlorates are not the only ones that are infected by the bite of oxygen.  Chromium trioxide (CrO3) is notable because it is in a +6 oxidiation state (3 * oxygen's valence-of-2 = 6).  Squirting acetone on a pile of dry chromium trioxide powder will instantly cause the acetone to burst into flame because it's just ripped apart by the combination of hexavalent chromium and oxygen.  Another good one:  the permanganate ion.  In that one, manganese is in a +7 oxidiation state.  By now it  shouldn't be much of a surprise that it also is an extremely powerful oxidizer.   It will react with a sugar solution at room temperature and turn it into black sludge in very short time.  When bitten by oxygen nitrogen suffers a similar fate and as a result becomes usable for things like explosives (think nitroglycerine) and rocket fuel.

In these instances, the base elements -- chlorine, chromium and manganese -- range from being a fairly powerful oxidizer to "not in my wheelhouse" -- but oxygen bites 'em and they turn into vampires themselves.  That's why I call oxygen the master vampire, because it can affect otherwise innocent elements and turn them into monsters, too.

Touch sensor for lathe and mill setup

 Some time ago I learned about a machining web site created by Rick Sparber:  right here.  He has a number of interesting articles about DIY machine accessories and improvements, but one in particular caught my attention -- a simple touch sensor that can be used to set up a metal lathe or mill for machining metal.  

The design can detect a small change in already-low resistance.  The basic idea is to measure the resistance between the cutting tool and workpiece being machined.  When the cutting tool is NOT in contact with the work, the current path is through the machine -- the spindle bearings being the major source of resistance compared to the body of the machine.  When the tool comes in contact with the work, that is a lower-resistance path -- and that is the basis of the touch detector.

 It looked pretty good, but being an electronics kind of fellow, I thought there might be some room for improvement.  The idea was to change the design to allow 4-terminal or Kelvin sensing.  Rick's design uses the same wires to force current through the lathe/mill AND sense the voltage change when the tool touches the workpiece.  This means that the design is sensitive to contact resistance at the tool and workholder ends.  The 4-terminal approach avoids this problem by separating the force and sense connections.  A good article regarding Kelvin sensing can be found here .  I hasten to add that Rick has some more-refined designs that DO implement 4-terminal sensing, and interested readers should look into them, particularly if wanting a good milliohmmeter or CNC-compatible touch sensor.  But for various reasons I think my design is a worthy alternative to his simplest design, and I offer it here.

My design schematic:

The design is quite similar to Rick's, including the automatic power-up scheme implemented by Q1.  The major difference is how the inputs to U1-1 and U1-2 are connected.  They are routed to separate sense lines (although my design does permit simpler 2-wire use if that works OK on a particular machine).  Rick's original design used a couple of spade connectors and super-magnets to attach the touch sensor to the cutting tool and workholder, but that's not compatible with the 4-terminal approach.  One of his other touch sensor designs uses two miniature battery charger clips with the sense connections brought in via an insulated contact, and that's what I'm using with my design.  I drilled a hole in the jaw of each clip large enough to accommodate a 1/4" nylon screw, then chucked the screw in my lathe and drilled and tapped a #4-40 hole down the center of it.  A #4-40 brass screw was threaded into the nylon screw, then the nylon screw was attached to the jaw with a nylon nut.  The outside end of the brass screw has a nut to attach a spade connector for the sense line.  If it's not clear from my explanation, the brass screw head forms a Sense contact.  Experimenting with different-sized end mills suggested that it would work better if I added a washer underneath the brass screw, so I also did that.

I haven't had a chance to debug the design yet, but once that's done my plan is to make it open source.  More on that later....

Do-It-Yourself Elemental Analysis

 For some time now I've been working on a system that can produce an analysis of elements that are present in an unknown sample (within limits, explained later).  It's based on a lot of excellent work by an Italian group called Theremino.  They use a combination of parts that can be acquired on ebay, such as a type of x-ray detector called a scintillator, a photomultiplier tube, high voltage power supply, some pulse-shaping circuitry and a computer's audio board (one that can digitize audio).

The basic idea is to use a variation on fluorescence, where some materials will emit visible light when illuminated by ultraviolet light.  The exact term is XRF, for X-Ray Fluorescence.  For ultraviolet illumination, only a few materials will emit light:  and they, as far as I am aware, are compounds, doped or otherwise.  The emission wavelengths are more dependent on the dopant than the compound.  While this might sound OK, the dopants typically are exotic elements like Yttrium, Eutropium, Erbium and the like -- not of much interest if you want to know if your jewelry really is gold or silver.

On the other hand, x-ray illumination can excite fluorescence of all elements except hydrogen and helium.  The key is that the emitted light is characteristic of the element, not a dopant.  So if we had a way to (1) illuminate a sample with x-rays and (2) determine the wavelength of light emitted by the sample's fluorescence, we can determine the elements present and get an approximate idea of their concentration. 

The first problem to solve is a hobbyist-available X-Ray source.  The Theremino group uses radioactive capsules from ionization-type smoke detectors.  They have a small amount of Americium-241 in them, representing about 1 micro-curie of radioactive decay (about 37000 decays per second).  Along with the desired alpha particles used by the smoke detector circuitry, the capsule emits gamma rays with an energy of about 60,000 electron-volts, 60 Kev (visible light is in the 2-4 electron-volt range).  Before anyone freaks out, the gamma ray flux is quite low, and, since smoke detectors typically are installed on the ceiling, a long way away from people.  This reduces the radiation dose even more.  So in normal use they are safe.  The Theremino application uses several of these to produce a higher count rate for better analysis results.  I don't think the total x-ray flux is much of an issue, but, even so, my system uses lead shielding to completely remove any misgivings on my part.

The other issue with regard to the Americium241 (Am241) is the 60Kev photon.  It can only excite fluorescence in elements if their characteristic fluorescence is at a lower wavelength.  That makes sense, right?  The emitted light's energy per photon CANNOT exceed the incident photons that are stimulating the fluorescence.  Otherwise you have a variation on the perpetual-motion machine, where the energy-out exceeds the energy-in.  So Am241 can't excite the primary fluorescence line for any element past Tungsten, whose k-alpha line is about 58Kev.  Elements with higher atomic numbers won't get "tickled" enough to emit X-rays.  It gets a bit more complicated because inner-shell electrons can get involved and emit much lower energy x-rays, but their energies are much more difficult to detect, out of the range of commonly-available x-ray detectors.

So.  We have a fairly readily-available x-ray source to cause elements up to Tungsten to fluoresce.  Now, how do we tell them apart?  The key is a particular type of radiation detector called a proportional detector.  For each incoming x-ray photon its output is proportional to the photon energy, which is directly related to its wavelength.  Planck's constant at work.  Look it up!  Anyway, there are a few hobbyist-available proportional detectors.  The first is a scintillator.  It's a special crystal that outputs a burst of light whose intensity is proportional to the x-ray photon that generated it.  The crystal typically is optically coupled to a high-speed detector called a photomultiplier tube that can turn the burst of light into an electrical signal, along with a substantial gain (that's the 'multiplier' part of the name).  The photomultiplier tube (PMT) output also is proportional to the light input, so the pulse height coming out of the PMT is proportional to the x-ray photon that created it.  Each element has a characteristic x-ray fluorescence wavelength, so by sorting the pulse height into bins, where each bin represents a small range of pulse heights, we obtain a spectrum that is a fingerprint of the elements present in the sample.

This sounds pretty good, except for a few things I haven't mentioned yet.  The first is that the PMT requires a fairly high voltage to operate, about 800-200 volts.  It doesn't require much current, but that's still a challenge.  The voltage determines the gain of the PMT, so it also is necessary to regulate that voltage, while avoiding instability (oscillations) and noise.  Photomultipliers also are sensitive to magnetic fields so they need to be magnetically shielded, using mu-metal surrounds.  They are very light-sensitive so they also need to be completely shielded from visible light (and if exposed to light when their supply voltage is still on, they can be destroyed!).  Finally, the best scintillator crystals are hygroscopic -- they absorb moisture -- so ones you find on ebay may be seriously degraded.

PMT's can be found on Ebay for not a lot of money, but then you need to add in the power supply and scintillator.

An in-between solution is to use a silicon photomultiplier, also called an SiPM.  It is a type of avalanche photodetector and is more robust compared to a PMT, plus it usually only needs 20-50V to operate.  It still requires a scintillator crystal with all the attendant issues related to the types that offer the best energy resolution (i.e., very hygroscopic).

Finally, there are PIN (P-type/Intrinsic/N-type) diode detectors.  The name is related to their structure, with the important part being the relatively wide lightly doped Intrinsic region that can capture lots of x-rays.  This type of diode can be all over the map, depending on the range of elements you want to analyze, plus the sensitivity of the system.  The most sensitive detectors have an energy range that covers Boron, with a primary x-ray wavelength of 185 electron-volts on up to lead and beyond, in the 10,000 electron-volt range.  These are operated in a liquid nitrogen dewar at 77 degrees Kelvin, operate at hundreds of volts, and would destroy themselves if for some reason they rose to room temperature.  These systems are very expensive and very unforgiving if you screw up.

However, if we're willing to give up something on the lighter-atomic-weight range, we can get by with something that runs at room temperature.  Silicon PIN diodes specially designed as x-ray detectors can be had for something in the $100 range, and, in some cases, for a bit less if already in a consumer application:  PocketGeiger.  In the latter case, the device is not meant for XRF -- it's a simple geiger counter, not sensitive to the incoming x-ray photon energy.  But it appears to be possible to tap into an internal signal line that MAY permit some form of elemental analysis.  I'm hopeful about this and will report results as I get them.  I have already found that the boost switching regulator, used to generate the ~26V bias voltage needed to operate the PIN diode, is a significant noise source and needs to be replaced by an external bias voltage source.  I didn't find an easy way to eliminate the noise and keep the on-board switching regulator.  Moving it off-board might work, but for now I'm using an old linear bench supply to provide the detector's bias voltage and it is working OK.

DIY A/C, some measurements

Optimizing any complex system requires good metrology -- making measurements to evaluate how things are working.  A model is a good start:  but without real data you don't know if your model is accurate or not.  So I have been working on that.  Significant measures of performance for input and output air would be the air temperature (of course), and the relative humidity of the air.  To that end, I bought two SHT40 temperature/humidity sensor boards from Adafruit so I can measure ambient conditions and air coming out of the evaporative cooler.  I'd like to include measurements of the water temperature in the recirculated-water loop portion of the evap cooler but the humidity sensor is directly exposed to the environment so that's not a good idea.  As a workaround, I used a cheap thermocouple temperature measurement unit purchased from Harbor Freight.

One suspicion I had was that the air flow through the evaporative cooler is too high, so the relatively-warm air entering the evaporation pad doesn't spend enough time to aborb as much water as it could -- therefore reducing the exit air's temperature drop.  The temperature delta depends on the temperature of the entrance air, so it's necessary to measure the temperature and RH of the entrance air -- as the day goes on, the ambient temperature and RH changes so my setup's operating condition changes along with that.

Anyway, here are a couple of photos showing my current setup and one measurement of the recirculated water temperature:

The firs photo shows my Adafruit Feather NRF69 RF transceiver board with two sensors -- one in the fan's exit air stream and the other is there to measure ambient conditions.  It's powered using a LiPo battery.  This particular board only has one native I2C port so (because the sensor chip's I2C address is fixed) I had to implement a second I2C port using the old-fashioned bit-bang approach.  That took a week or so to get working OK.

The second photo shows the temperature of the water being recirculated through the evaporative cooler.  At the time the photo was taken, the ambient temperature was about 87F so I was getting about a 16 degree temperature drop.  71F isn't too bad for house air, depending on what the interior RH is, so that's sort of encouraging.  But I'd expect the water temperature to increase if it's used (via a water-air heat exchanger) to actually cool the house.  The one saving grace here is that water's specific heat is MUCH higher than air, so a little water should be able to cool a lot of air.  But that part of the system isn't built yet so I can't report any measurements.  Metrology.

I also added a PWM motor speed controller to adjust the fan's speed (and hence the volume of air flowing through the evaporator).  As I suspected, running the fan at 100% of its maximum speed was not optimal with regard to delivering the lowest exit air temperature and highest RH.  Setting the speed at about half its maximum seemed to deliver the best performance for lowest exit air temperature.

However.....I'm not really interested in the lowest-possible exit AIR temperature.  My plan is to circulate the evaporatively-cooled water through a heat exchanger that is in the house, so it is desirable to optimize the system so the water temperature is as low as possible.  In this regard, I observed two things.  First, the temperature of the water that is recirculated through the evaporative cooler doesn't have a strong dependency on the fan speed.  That's a good thing to know.  But the second observation is that the water temperature seems to have some variation that can't be totally explained by the ambient air temperature and its RH.  I suspect it is due to solar heat input, so further experiments (using insulation and light shading measures) are needed.  This summer's temperature extremes probably are at an end -- today's high was only in the high 80's -- so more progress may not happen until next summer.

One main result is that our region's relatively high humidity substantially reduces the effectiveness of an evaporative cooler when it uses un-processed ambient air.  By "un-processed", I mean air that has not been dehumidified before being fed into the evaporation unit.

The Tech Ingredients youtube channel has demonstrated two different approaches to evaporative cooling.  Their first system used liquid desiccant to dry air before it entered an evaporative cooler.  On first examination this would seem to a viable approach for more-humid environments:  but they also have to use an ambient-air "swamp cooler" to cool the liquid desiccant after it's been regenerated.

Their second approach uses liquid desiccant to directly cool room air (and remove humidity), but depends on an ambient-air swamp cooler to get the liquid desiccant cold enough to cool the house air.  I think the second approach is less useful (in a broad sense), since evaporative coolers in regions with higher humidity won't be able to deliver liquid desiccant that's much below ambient.  However, reducing the humidity in a house, independently of the temperature, will make the house feel cooler (because of the evaporatively-cooled inhabitants).

A swamp cooler that uses pre-dried air can output coolant water that is below the ambient dew point, which will remove moisture -- thus improving living conditions in two ways.  In addition, such a system can be used in a "bootstrap" mode*, using the coolant water to also chill the liquid desiccant before it drys the evap-cooler's entrance air.  This IS an additional heat load so pre-cooling the liquid desiccant, using an approach similar to Tech Ingredient's second LD cooling scheme, would be desirable.  For these reasons, I believe a hybrid scheme using parts of their first and second systems would be more useful for most who are interested in a DIY approach.  It utilizes the cooled exit air, which otherwise is unused.

*I call this a "bootstrap" mode because at first the effectiveness of the liquid desiccant (LD) will be relatively poor because it's not cool enough to pull a lot of moisture out of the air.  But as the system starts to work, the temperature of the coolant water should drop and thus help the LD dry the entrance air more -- thus further improving the performance of evaporative-cooling step, dropping the temperature of the coolant water.  And so on. 

A hybrid scheme probably won't deliver dry, cool air the instant it's turned on.  It might actually take a day or more to really get up to speed, depending on the amount of water being recirculated through the chiller and the house's heat load.  Then there's the question of how to deal with night-time conditions, where temperatures can drop to a reasonable level but the ambient relative humidity still is uncomfortably high.  For that, it might be necessary to add some auxiliary heat (rather than solar) to regenerate the LD.  But now I'm getting 'way ahead of myself -- currently being far from any kind of real, practical home-brew A/C system.  

Metrology.

DIY A/C More Observations

Today I took more measurements after my cooling tower had stabilized.  The ambient conditions were 33.4C ( 94.6F) and 41.4% RH.  My psyrometric chart indicates the wet bulb temperature should be 22.8C (73F).  Measurement of the exit air showed it was 30.35C (86F), substantially higher than the wet bulb temperature.  The temperature of the (recirculated) water was about 74.5F so it IS close to the wet bulb temperature.

The reason the exit air is much warmer than wet bulb may be due to excessively high air flow through the evaporation pad; or perhaps the air never is going to get all that close to wet bulb.  After all, the wet bulb measurement doesn't measure the temperature of the air after evaporation, it basically measures the residual water surrounding the thermometer bulb.  Videos done by Desertsun02 suggest that the exit air temperature should be much lower, but his setup isn't exactly like mine.

I'd like to get the temperature of the exit air lower, because I can use it (via a second heat exchanger) to cool the return water from the interior heat exchanger.  That would increase the system efficiency, perhaps by a significant amount.  To work on this, I bought a couple of PWM motor speed controllers to experiment with air flows through the evaporation tower and inside heat exchanger.

I did try throttling down the water pump to see how that would affect the system, but reducing the flow rate by about 50% didn't have a noticeable impact on the measurements.

Examination of my psyrometric chart did suggest a way to improve the system performance by a small amount.  Although it sounds counterintuitive, if the air entering the cooling tower is pre-cooled by passing it through a heat exchanger, the wet bulb temperature of the cooler-but-more-humid air is lower.  To check this out, you need to look at what happens when the air is cooled by a heat exchanger.  It doesn't pick up any more moisture so the "humidity ratio", which is the ratio of dissolved water vs air masses, remains constant.  So the heat exchanger just moves the air straight to the left (i.e., it just moves along a constant humidity ratio line).  Then you look at what the resultant wet bulb temperature would be.  Here's an example.  Looking at the ambient conditions I get a humidity ratio of 14.  Cooling the air down to 25C before it enters the tower should produce a wet bulb temperature of 21.5C, which is about 1C lower.  Not a huge improvement, so I don't think it's worth the added cost and complexity.  It's more worthwhile to get the exit air temperature closer to wet bulb so I can reduce the heat load on the recirculated-water loop.  This will ONLY work if the exit air temperature is lower than the return water from the interior heat exchanger....which, in turn, can't be any higher than the ambient temperature in the house.  Clearly, we want to cool the house down so ideally there will be a substantial change in coolant temperature from inlet to outlet.

I have to say that my measurements don't look all that promising for cooling our house with this setup.  There is a final way to (potentially) greatly improve the performance of an evaporative cooler, but it comes with quite an increased bit of complexity.  It would reduce the humidity of the air going into the cooling tower using something called "liquid desiccant", which will improve operation of the system in regions where the relative humidity is  moderately high to high (and our region seems to fall into that category) .  The liquid desiccant absorbs moisture but must be regenerated on a continuous basis in order to keep working.  This requires some sort of heat source and another "tower" to help extract the absorbed water from the desiccant.  Sounds complicated?  Yes, but fortunately it appears that it can be done using a fairly low-tech approach.  I will leave it there for now.