Category Archives: How this project started…

The story of how this project developed. Wordpress generates this list is in reverse chronological order, so SCROLL DOWN TO THE BOTTOM, and click on the post titled “In the beginning…Part 1”

Interview: Cave Pearl Project Team: 2015-09-20

Now that we have recovered from the last round of fieldwork, I finally managed to blow the dust off an old Macintosh in the basement and produce a little promo for the project:

Still needs some polish, but not bad for a few quick clips in the living room. My next task is to update my “How to build a data logger” posts to include all the new improvements.

Addendum 2015-08-24:

The new build series tutorials are now live to help folks get started.

<— Click here to continue reading the story—>

Field Report 2014-08-25: Retrieve & Deploy New Flow Sensors

We started the day with breakfast at Turtle Bay Cafe, and once I had enough caffeine in my bloodstream to engage more than two brain cells at the same time, I reviewed the data on the SD cards from the overnight test runs. They all looked good.  Over breakfast we met up with Monika Wnuk, a multimedia journalist and documentary photographer from Northwestern University, who wanted to interview Trish for a water & development story she was working on. Yesterday, when she heard about our abbreviated schedule, she volunteered to help with the sensor preparation, and to provide shore support for our deployment dive. I was glad for the assistance, as two scheduled days were now being merged into one single operation.

Pre-dive planning with Bill, Trish, Monica & Jeff.

Pre-dive planning with Monica, Bil, Trish & Jeff.

Our diving field work almost always begins with a visit to Speleotech in Tulum, to see our long time friend Bil Phillips. Bil taught me to cave dive many years ago, and I still have much to learn from that remarkable man, who is without doubt one of the most dedicated cave explorers in the world. We also had the good fortune of meeting another good friend, Jeff Clark, who loaned me some equipment I needed for the days dive. The dive community in Tulum has always been generous to visiting researchers because they understand, more than most people, what is at risk with the rapid development that is happening in the region.  We all share a passion for protecting the caves as both a vital water resource, and as areas of natural beauty & wonder.

Checking for rotation, damage, etc.

Inspecting the old units for rotation, damage, etc.

With the kit sorted, we headed out to our main deployment site where I began to adjust the buoyancy of the new sensor units. With the new internal copper ring of ballast mass (45g), and heavier
aluminum battery holders, it only took 2-3 external washers to bring each unit to my target of 15 grams negative. This is slightly heavier than the last deployment but I am expecting any reduction in the tilt angle to be more than compensated by the 14bit 1g resolution of the new BMA180 accelerometers.  With
calibration out of the way,  Trish and I set off on the dive. High tide at the coast meant the system was experiencing very low flow, so we had a relaxed swim, with three new pendulums and a pressure sensor stowed neatly in the mesh bag by my side.

Old vs. New

New  vs. Old

Once at the site, the first task was to do a general inspection of the old units, noting anything unusual in my dive notebook.  After almost five months of submersion, there was plenty of rust on the stainless steel bolts and one of the units needed it’s anchor plate replaced.  Using the checklist I had prepared earlier, we swapped each unit in succession with it’s replacement.  In the calm conditions, percolation obscured our view a bit as our bubbles meandered around the ceiling of the cave, but it was still a very simple operation to exchange flow sensors.

Once the new units in place, we did a final inspection swim:

…checking that the new units were secure, with the X axis of the accelerometers oriented toward north.  While this is not strictly necessary with magnetometers inside the units,  I can use it as a rough confirmation of the compass bearings when I get the chance to do some proper data analysis later. I gathered the old sensors into the mesh bag and we made our way out of the cave.  I am not sure I can fully express the excitement that an inventor feels returning from a dive like this, but it’s very, very cool.

I think there is an ocean and a sunset in this picture. But at the time, we did not even notice it.

There is an ocean and a beautiful sunset in this picture. But at the time, I don’t think we even noticed it. (photo courtesy Monika Wnuk)

Back at the surface we had a chance to do a better visual inspection of the old units, which all appeared to be intact. I had some concern about the hull penetrations, as none of the epoxies were rated for long duration marine exposure. But the indicator LEDs were still piping on schedule, telling us that they were all still running.  Back at the dorms, we were equally thrilled to find complete data sets recorded on the SD cards.  (I will post more on the actual data after we have a chance to work on it.)

 <— Click here to continue reading the story—>

Field Report 2014-03-22: The next generation of flow sensors is deployed.

The Field Testing Station.

The field testing station. The polished O-ring seats are covered with blue painters tape for protection.

After the successful retrieval, I set to work on scripts for the next generation of sensors. It’s amazing how the kind of focus that coding requires can really mess with your perception of time, leaving you feeling that everything is being done at the last possible minute, though you have been working on it for several days... But after some datasheet slogging (thanks once again to the folks at Turtle Bay Cafe for their patience), the units started to produce reasonable
numbers, and on the morning of the 20th we were “all systems go”. I had three pendulum units (plus one backup) and one high resolution pressure sensor ready to deploy.   The pressure sensor would be stationary to record the water level, and I did not want it swing around on a pendulum until I get a chance to do a bit more homework on the calculations required to compensate for that motion.

As usual, we had one unit misbehave on the bench so badly it needed a complete "brain transplant".

We had one unit misbehaving so badly it needed a complete “brain transplant”, but the modular design of the system meant this was pretty easy to do.

The low power consumption of our bench tests gave me the confidence to set a couple of the loggers to 5 minute sampling intervals, while leaving the third on a more conservative 15 minute schedule. (in case the faster loggers run out of juice before we can collect them). Then we sealed everything up and set out to collect the tanks, etc. from our friend Bil Phillips at Speleotec dive shop in Tulum. On the way there I monitored the heartbeat LED’s.  But unit 3 did not pip, so while Trish sorted the dive gear I cracked it open to find that indeed, it was not logging (I suspect because of a loose RTC alarm/interrupt line).  As luck would have it, a couple of researchers working with a group from Denmark/Austria  (who have done some impressive work ) arrived to prepare for their days dive. They were testing some newly developed 3D scanning equipment, including a flow meter using an optical method based on laser tracking of particles. A good nurtured discussion ensued about the pros & cons of different measurement methods: “How will you calibrate?” “That’s going to be really non linear..” “Yep, but I have no problems with bio-fouling, and no issues with salinity/refractive index…” I will skip the rest of the nerdy details, but let’s just say there’s nothing like a bit of friendly competition to motivate…

I used the deflection of an 8 inch cable tie as a rough field balance. Units were tuned to approximately 10-20 grams negative.

I used the deflection of an eight inch cable tie as a rough field balance. The units were tuned to  approximately 10-20 grams negative.

Once out at the site, I tried to standardize the buoyancy of each unit. The beta’s had significant variation in their response to water flow, and my goal on this build was to achieve a more reasonable amount of inter-unit consistency. Even with stainless steel bolts on the housings, I still had to add about 150 grams of ballast to each logger. (weighted towards the top of the units to offset any torque from the internal mass of the AA batteries)  I am not happy about all that hard iron near my compass sensor, but the data will tell me if it causes a serious problem, as compared to all the other factors, like the batteries, etc. My humble budget will not extend to a degaussed power supply!

They are deployed quite close together, to allow me to assess inter-unit response for this build.

They are deployed quite close together, to allow me to assess inter-unit response for this build.

Low channel flow meant that the deployment dive was pretty easy, and we re-occupied the previous logger location for a continuous data set. The new bungee cord anchors are much easier to attach to the ceiling of the cave than the knots of nylon string used earlier, but of course we don’t yet know how long the rubber will last. Despite my surface testing, I still needed to transfer a few ballast washers to achieve a similar angle of inclination on the pendulums.  During this operation I was promising myself that the next units will be much more compact, and have no metal parts on the outside.  After a final inspection swim, with the capture of a little video, we were done.  Although the whole installation went smoothly, the earlier delays from Unit 3, and my buoyancy calibrations, made for a very long day, so it was well after dark when we finally left the water. After so many months of work, I could finally relax a moment and take it all in – my little cave pearls are starting to feel like a “real” scientific monitoring platform:

(Yeah, shakey cam: but our WG-3 croaked last year and the Heros are not great in low light, so this was captured on a little Powershot D10, that’s nearly 10 years old)

It will be a while before we see data from the new units, but I am confident we will see good numbers from them. (…still have my fingers crossed though!) I think I need to go have a moment on the beach, before my brain starts chewing on all fixes for the next build. I have homework to do before I get a good electrical conductivity sensor in the mix that can cover the entire fresh to marine range (standard electrodes are not designed for this) but I wonder what else I could add to the little loggers?

<— Click here to continue reading—>

Field Report 2014-03-19: The Cave Pearls have landed!

About to surface with the beta loggers.

About to surface with the beta loggers.

All the usual grumbles of getting our kit out of storage and ready to roll delayed the Beta unit retrieval till late afternoon, and I was really chomping on the bit by that point. It was a good haul against the current out to the deployment site, but once there I was very happy to see our little loggers swaying gracefully. After a good visual inspection, to examine the anchors and brush away the sizable clumps of rusty brown goo that had grown on the exposed metal, we popped them into a mesh bag and headed back to the surface.  With my ears above water, I dared a few little shakes to check…had they leaked….?

Breaking the seal.

Breaking the seal.

We cradled the loggers all the way back to Tulum while we waited for them to dry out.  After a quick rinse, we stowed the dive gear and then set to work on the pearls. Both of them opened with a satisfying “ssshhhick” indicating that the seals were indeed good. (they were compressed a bit at depth) I have to admit I have been working on the new builds so intensively, I laughed to see the tape that was holding these guys together, and the leggo I had solvent welded into the battery compartments. I checked the hour: it was 7:45pm…perfect time to cut the power, as the units had been left on a 1/2 hour sampling schedule. Once the power was disconnected, I could breathe a sigh of relief, as we  were then safe from any further calamities that might hurt the precious SD cards. But we did not have a reader with us, and as usual, Trish had filled the evenings schedule with meetings with some of the other researchers in town. In the end it was 11:00 pm before we could look at the SD cards and know if our little experiment was truly a success….

And it was! In fact both units were collecting readings right up to the point where I disconnected the battery. Woot! Trish went to work, and I just sat back, impressed by the super sonic “squiggle wrangling”.

“Temps…no trending..but nothing useful there… offsets…”, she was talking to the screen more than to me. No surprise on the temperature data as the readings were from the RTC, completely trapped inside thermal mass of the housing.

“Can you bring up the two voltage curves?” I asked, “I want to know how we did on power consumption.”

“Right.”…clickety, clickety, click… “How’s this?”

Left: Unit 1    Right: Unit 2 (rubber bottom)

Left: Unit 1 Right: Unit 2 (w rubber bottom)

Quite a difference, but both units had run in the > 4 volt range for three months.  So we had been far too conservative with the 30 minute sampling routine, although I had no way to know that when they went in last year. Pretty much identical components in the build, so for now I am attributing the different power curves to the mix of batteries that were used.

“Z offsets….X and Y as well….” clickity, click… “…easy to fix… and we can do a quick running average for that noise…”

“No, don’t fix it!” I injected, “I want to see the raw data,  side by side.”

Trish’s hands paused, and she dropped out of the excel trance long enough to give me a puzzled look: “Why would you want that?”

I explained that while she was rapidly turning the numbers into something relevant to the actual water flow, I (as the builder) wanted to know how the two units compared to each other, as basic machines…

“Hmmm, Ok”….

Left: Unit1  Right:Unit2  Raw z Axis, sub-sample

Left: Unit1 Right:Unit2 Raw z Axis, sub-sample

The different amplitudes were not a surprise, as our buoyancy control was just best guess approximation. But in theory, the accelerometers were identical, so the offsets were kind of interesting.  Ah well, it will be a while yet before I am at the point of calibrating these things…

I am still amazed that all this makes it through the airport scanners...

I am amazed that we don’t get more grief from airport security when we travel with this kit.

Trish was still talking to the screen  “mmmm…some finer structures here…”, and clicking away, but it was nearing 1:30 am at this point, and I was starting to fade. I uploaded all the data to a Google doc, and suggested that we call it a day.
I knew what I still had ahead of me ->

I hope that Turtle Bay Cafe doesn’t mind if I take up residency for a few days, while I work on the next generation:

“Mass caffecito porfa…”

<—Click here to continue reading—>

The new fleet of flow sensors is ready to sail!

Hi everyone. I wrote most of this entry on a plane today, as it was almost the first free time I’ve had “away from the workbench” since the initial proof of concept loggers were deployed last year. I have redesigned the Cave Pearl data loggers into a more modular platform that should be flexible enough for quick field repairs, while enabling future development with more sensors.  (I want at least CTD, and my wife has an infinte supply of other suggestions  🙂

The loggers are now assembled in four interchangeable components, which from top to bottom are:

1) Upper housing

It was too cold in the basement for the epoxies to set properly...

It was too cold in the basement for the epoxies to set

Lots of lessons have been learned here about sealing the hull penetrations thanks to the diy ROV crew. Sort lengths of 3/4 pipe form “wells” to protect the sensors, with JB Plastic Weld putty wrapped around the wires as they initially pass through from the inside of the housing. The putty sets on the roughened surface, pluging the hole and holding the sensors in position, but I found that the silicones I tested flex quite a bit after curing, so they are too easily “sheared” away from the pvc surface. As a result,  the current builds use JB weld around the DS18B20 thermal sensors, and Loctite E-30CL to for a transparent seal over the “heartbeat” LEDs which pip when the samples are taken. Experience has shown me that you must have some way of knowing your units are working happily (or if they are in an error state…) before you dive them into the cave.

2) Main electronics platform

The LED is an Octopus Brick because they had a good buckled connector, and the RTC is a cheap DS3231 module from eBay because I wanted the AT24C32 it had on board.

The LED is an Octopus Brick because they had a good buckled connector, and the RTC is a cheap DS3231 module from eBay because I wanted to use the AT24C32 eeprom it also had on board.

I am still quite happy with Tinyduino, as the package integrates the mcu, accelerometer, and now digital compass, with the smallest footprint and the least amount of extraneous wiring. I put riser pins on their new overhanging protoboard, and this jumps out to a grove I2C hub as a central interconnect system allowing me to interchange the logger platform with housings that will sport different sensors in future.  All of the electronic components have had a good bath of conformal coating this time around, so hopefully they will be a bit more robust. (I might try Rustoleums Neverwet next time)

3) Power supply

The gap between the two shells provides room for the interconnect, and some filtering caps, etc. if needed.

The gap between the two shells provides room for the interconnect, and some filtering caps, etc. if needed.

Physically it’s just two pvc knockout caps held together with four bolts & a 1cm “hold down ring” to keep it in place in the lower housing. Electronically there are two versions. The first is an unregulated supply uses two banks of 3 AA batteries, through Shottky diodes to prevent the banks from draining unequally. This supply will drop from 4.5 volts down to a lower cutoff of 2.8 volts before the system stops logging, so it needs fairly robust sensors. The second power supply uses three banks of 2 AA batteries (with three Shottky’s) feeding into an NCP1402 3.3 volt boost regulator which then powers the logger. Several of the sensors I want to use have a strict 1.8-3.3v input range, so they can not be used with an unregulated system. It will be interesting to see if the greater “draw-down” enabled by a boost regulator compensates for the power it wastes (here about 25%). This deployment will hopefully be some months long, so I will find out how the regulated VS unregulated systems actually perform.

4) Lower housing & external weight system

This series needs about 100-150 grams of ballast mass to be neutral.

This series needs about 100-150 grams of ballast mass to be neutral.

The buoyancy troubles we had on the initial deployment showed me that I needed some form of external system to compensate for changes in battery mass, cable buoyancy, salinity, etc. So I have a simple solution using a bolt through a threaded end cap which holds a number of washers as ballast. All stainless steel, but I am curious to see how long they actually last in the near marine environment. The buoyancy mass will be spit evenly on the top & bottom of the units to prevent rotational torques which which would affect the angle readings.

The battery run down tests are still looking very good for one year + deployments!

The battery run down tests are still looking very good for one year + deployments!

So this is the new fleet: Four pendulum units and one high resolution temperature &  pressure sensor that will remain stationary.  Hopefully they will all be underwater logging in a few days. Looking back at the build journal, I should add that there has also been a fair bit of coding, but I will post details on all that later, after I have integrated support for the HMC5883L digital compass & MS5803 pressure sensors into the main logger script.

BUT before we deploy these new units,  we need to go and retrieve Beta’s 1&2 which we left in a cave last December.  My fingers are crossed that they have survived these last few months under water…

<—Click here to continue reading—>

Addendum 2015-01-07

For the DIYers out there, I should mention that this housing style proved quite robust through several deployments in 2014, and probably could go to substantial depth due to the thickness of the 3″ end caps.  But in early 2015 I came up with a new design built with Formufit table caps, which is much easier to assemble provided you can squeeze your electronics package into 2″ pipe.

Field Report 2013-12-16: The first long term deployment begins


We discovered a stowaway in the car on our way out to the dive site. Everyone took that to be a good omen.

This was our last day in Mexico, so the flow meters were going in for their first long term installation today.  The over night run trials went smoothly so the last minute rebuild of logger 2 fixed the excessive power drain issue. (whew!)
But all the testing I had done over the last few days (with the units sampling and recording at a furious pace) meant that I had to scavenge the remaining good batteries out of our dive lights for the deployment.  I loaded the loggers with a sketch set to take readings every 30 minutes, and sealed the housings.

Then we loaded up our dive gear and drove to Playa de Carmen, to meet a reporter who had been interviewing Trish over the last few days. She was going to dive with us today to get video of us, and also of the little data loggers, for a documentary she was making about the growing water quality issues in the region.  Unfortunately she was was not a cave diver, so we did a “pretend” deployment on a large mangrove root out in the open water. Once she had captured the footage she needed, Trish and I continued on into the cave.

Because we were uncertain about the weight of the new batteries, we decided to install both units as pendulums for this deployment. The current at this location was pretty strong, so it was a bit challenging to stay in place, while affixing the “ceiling anchors” to the roof of the cave.

After securing the sensors, we did a final swim round to inspect the installation:

Fare well little sensor pods! We will come back in a few months to get you…we promise!

<—Click here to continue reading—>


Field Report 2013-12-06: The moment of truth…

A couple of days into the UNAM research, there was break in the schedule, so we had the opportunity to go back and retrieve the units. When we arrived at the installation site, everything looked exactly as we had left it, with no apparent leaks or other damage to the housings.  But on closer inspection, we did observe that the two units were not exactly behaving the same way:

After catching a little more video, we spent a few minutes collecting the sensors. A short while later we finished the dive, and soon I was carefully cradling the loggers on my lap as we drove back to the CEA dorms. Once there I made sure that the units were absolutely dry before I opened them up to retrieve the SD cards.  I could see from the size of the files that both units ran smoothly, and had logged data. But what had they recorded?

Once the files were on the laptop, my wife, an Excel virtuoso, took over.  Within moments we were starting to see bumpy graphs displaying the three axes of the accelerometer.
A little more adjusting, a few labels, and we were looking at this:

3 days of raw data from the very first deployment.

Raw data from the first deployment: x,y&z axes, but with different orientations relative to flow direction.

“Is that good?” I asked. I could barely contain my excitement.

“Yes,” she replied with a big smile, ” for uncalibrated, first run data, this is pretty good.”

“Why two peaks per day?” I thought there might be a problem with the sensors.

“Actually.” she added, “That’s normal.  This area has semi-diurnal tides, and the velocity curves are often asymmetrical like that.”

“Yaayyyy!” I whooped, “We did it!” And I think I even started dancing.  Months of noodling around in the basement, and combing through forums, had just been transformed from “another one of Ed’s crazy projects…” into two real working prototypes!

It was well into the evening by this point, so we headed out for a late dinner, and a couple of celebratory ‘cervezas’.  We discussed where we might put them next, so that we could learn more about the quality of the data they were generating. I wondered about how we might calibrate them against some commercial units, and Trish said that even without ‘absolute’ velocity numbers, the information would still be useful to her research. But for me,  the real bottom line was the moment when she asked:

“How soon can you make me some more of these things?”

<—Click here to continue reading—>

The “Alpha” build (of an underwater sensor housing)

Although the dive-light forums had convinced me to use PVC tubing as my housing material, I still had some significant design issues to work out. How big were these things going to be? How was I going to seal the unit under water? How was I going to actually install it in the caves? etc. I spent so much time rummaging through the plumbing isle looking at fittings, that the staff at my local hardware store were starting to run the other way whenever they saw me step over the threshold. And the ones I did capture, with my half baked story about what I was trying to do, had a kind of “There but for the grace of God go I…” expression creeping across their faces. Of course, after years of exposure from my own friends and family, I guess I am just used to it now 🙂

Anyway, I started out with three inch pvc pipe, for the simple reason that this had the smallest inner diameter that would hold my “alpha” Arduino Uno datalogger. But how was I going to OringSetuphold it together? Bolts?  Bungees?  I had seen plenty of latch clamps designs on the newer lights from Dive Rite, etc., but they all seemed to use machined rod stock with turned threads and special holder grooves for the O-rings. And this was more complicated than I wanted to go.  Fortunately, while I was working this out, I came across a miniDV housing instructable with a really nice system for backing an O-ring on a pipe. I realized that my housing could use this idea, but I did not need any of the clear windows or other things that complicated his design. Yes!

4" rings over 3" end caps

4″ pipe rings over 3″ end caps formed the basis of my housings

In fact all I really needed was two end-caps and a short length of pipe and I would have something that presented a nice smooth profile to the water flowing around it. But to connect the latch clamps I would need much thicker walls, or the screws would puncture the housing.  A bit more noodling around and I made the happy discovery that the inside diameter of 4″ pipe just barely goes over a 3″ pvc endcap, and the two solvent weld together nicely. Of course, hand sanding those matching faces down through to 600 grit took a while, but I was left with a nice smooth polish on the O-ring seats.

Getting down to 600 grit takes allot of hand sanding.

Getting those O-ring seats down to 600 grit takes quite a bit of sanding.

It took ages to find marine grade latch clamps, and I was surprised to find them costing $15 to $20 each. (After a great deal of time reading spec sheets, I found the cheapest clamps and O-rings at amazon – I will post a complete parts list for those later). So I had a basic “latch clamp & clam shell” idea percolating away. But how was I going to suspend this thing in the water column? Initially I had thought that I would simply run a bit of fishing line up to the float, but as I thought more about what the 3-axis accelerometer was actually doing, I realized that I could get much more than a simple tilt angle out of it:

If I could keep the unit from rotating, I would also get the direction of the water flow from the same sensor data! This realization was at the heart of the question of how to suspend the units inside the flooded caves.

So I need 180 degrees of freedom on the anchor points, but no rotation about that axis, or the direction information in the data would become meaningless as the sensor spun around. I suppose I could have just put a compass sensor in the unit and been done right there, but I had this sneaky feeling feeling that the problem could be solved more elegantly if I just burned a bit of midnight oil.

Corrosion had locked up some of our drip sensor tipping buckets a few years before.

Corrosion had locked up some of our drip sensor tipping buckets a few years before.

I started making all sorts of gimbals with hinges, bent wires, rods, bolts, springs, tubes, you name it, and I probably tried it. Most of them worked too, but they tended to be fiddly looking things that depended on one or more bits of metal, and I knew from previous projects that corrosion was eventually going to do them in. I also had to figure out how to attach those pivot joints to stiff rods, of varying lengths, which then somehow connected to the housing itself. On top of that, the whole assembly would have to gracefully fit inside a suitcase. And finally, just to complicate things still more, whatever I came up with had to be easily assembled in a dark cave, with a divers cold fumbling hands.

Well this little nut took me a few weeks to crack, and with all the factors in play, it represented the most complicated thing I had tackled on the project to date.  Especially with “easily repaired in the field” also echoing around inside my head.

Pivot joint with no rotation

Pivot joint with no rotation

But I am happy to reveal here, for the first time, a bodgers masterpiece of simplicity made with two cable ties, a length of pex tubing (cut into a washer), and a threaded pvc cap. I can whip up one of these puppies in about five minutes, from parts at any hardware store, and the pex tubing, which bends easily into a suitcase, is just barely positive under water…and there are no metal parts. With this in hand it was full steam ahead, and as soon as the latch clamps & 3 inch O-rings arrived (341 EPDM 70A) I would be ready to start testing the alpha build.

Boyancy testing

Buoyancy testing

But I ran into a bit of a snag when I started doing dunk tests: I had not really counted on the extra mass of the latch clamps (20g each), so with my rough calculations, I hadn’t left enough internal volume.  By the time I put my calibration mass (for the batteries, the Ardunio, etc) into the clam-shell, it sank like a stone. So I started drilling holes in the outer rings that the clamps were attached to, trying to increase the buoyancy so that the unit would just barely float when the simulated payload (about 220 grams) was inside it. The whole thing started to look like Swiss cheese.

One of the alpha housings, with anchor

One of the alpha housings, with anchor, and the short connector rod I made for the buoyancy testing.

Eventually, I ended up shaving most of the outer rings off the unit, which lead to several adhesion failures once the latch clamps started to apply pressure. But I just re-purposed those old shells into anchors with a rubber end cap. I had made a few shells, but I still had a pang of regret for the lost time, as I had spent more than an hour hand-sanding each of those O-ring seats. But the alpha housing build, minus electronics, was ready for testing. And just in time too, as this was early summer, and my wife had an undergrad about to leave for some fieldwork in Mexico. I made a few one meter support poles, stuffed the rest of the parts into a ziplock bag , and Trish passed this on to the student who flew out the very next day. I knew eyebrows might go up at the airport, but hopefully my weird collection of plumbing parts would not give the student too much grief from the airport security scanners. And even though the student was keen to help out, I knew that like anyone doing fieldwork, they already had a to do list that was larger than their available time. So there was a good chance they were not going be able to throw my contraption in the water and to see how it behaved.

I went back to developing the electronics side of things over the next couple of weeks, but I felt like a penny waiting for change each time I asked if the housing had been tested, and found out that, no, they had not had a chance to put it in the water. Not yet.

Then one morning over coffee, my wife says: “Oh, yeah. (the student) is back from her field work.”

My eyes widen, “And? Did the units go in? Did the floats respond to the water flow?”

“She put them in at one of the outflows along the coast” she replied, “And they seemed to respond to the direction of flow pretty well….”

“Mmmm, why do I hear a “But” coming…What actually happened?” I asked.

“She says that they wobbled.” and then she added,”Sort of wiggling around as they tilted in the direction of the current. But the flow’s pretty strong there, so it could have just been regular eddy currents. You see that in the seaweed along the bottom all the time…Then they sank.”

Trish wasn’t too worried about this news but I was a bit stunned. I had been expecting something like “it sank”, or “moved slowly”, or “no response at all”, but I was not prepared for “wiggly & wobbly”.  I spent that morning in front of Google, trying to learn something about fluid dynamics, and specifically, the phenomenon of: “Bluff Body Vortex Shedding“. My heart was sinking with each new read because this had the potential to introduce so much noise in the accelerometer’s signal, that the data would be useless.

So we had run into a piece of fundamental physics that might kybosh the whole project. I was pretty bummed out that day, because even when I did start to understand the math, sort of, I still could not see any way around the problem. Fortunately for me, I was about to get some really good news on the electronics side of things, which had me doing my  ‘happy dance’, which, on reflection, was probably a bit “wiggly & wobbly” too.

<—Click here to continue reading—>

The original float/pendulum idea

Once I had seen the old fashioned air speed indicators, it did not take me long to find references to an even older idea: the hydrometric pendulum. (mentioned in books as old as 1884) So a simple pendulum (or float!) would indeed work if I was able to measure the angle of deflection.  I scribbled a few doodles down on a piece of paper, and ran in to my wife’s office at Northwestern University to discuss it with her. At the time, she was conducting a lab session with her instrumentation students, so they also got an earful of my enthusiasm, which amused everyone.

I don’t have any of those original “back of the envelopes”, but I did find one of my early concept drawings from some time in February 2011:

Original concept sketch from 2011

Original concept sketch from 2011

As you can see from the picture I figured I was going to need a really big battery in a heavy enclosure resting on the floor of the cave, so I initially conceived of the device using a float rather than a pendulum. I was still trying to figure out some way to actually measure the angle of deflection without breaking the integrity of the underwater housing because I knew each gasket or o-ring was just another potential point of failure. I had some idea of just attaching the bobber to a joystick mechanism, and one friend suggested putting magnets on the string and using hall effect sensors in the case.

But I was still trying to figure out how it might be possible to put everything, including the sensors, inside the float itself. I soon discovered from the Arduino forums I was rummaging through that there were plenty of people using accelerometers in tilt sensing applications very much like this, for robots, quad copters, and even one fellow who put a datalogger with an accelerometer on his garbage can.

One of the prototypes from 2011. (held together with hot glue.)

One of the prototypes from 2011. (held together with hot glue.)

So I set to work building prototype data logger / accelerometer combination, bootstrapping myself on the arduino micro-controllers with the many helpful tutorials at Adafruit Industries , and of course the Arduino playground.  I was, and still am, beholding to the many people who share their expertise so freely in the open source hardware community. And I managed to cobble together a couple of dry prototypes (that actually recorded data) near the end of 2011, which I dragged around at Christmas of that year, doing show & tell sessions with my more technically able friends. My wife, bless her, put up with “Ed’s latest project” evangelism, even though I had melted my way through most of the drinking cups in the kitchen, and the house was beginning to take on the distinct bouquet of poor soldering & burnt plastic. At the time I was using a vanilla Arduino Uno, with an Adafruit data logger shield.  To that I had added the MMA7361 from modern devices as the accelerometer, and the whole thing gave me a whopping 24 hours of run time out of 6 AA batteries.  Not exactly the year’s worth of data we were hoping for but I had managed to cram all of that into a hunk of pvc pipe from Home Depot, so the “all in the float” approach was looking like it might just be possible.  I still had not given anything a “dunk” test yet, but at least it was a start.

<—Click here to continue reading—>

In the beginning …Part 3

So I dug my teeth into the problem.  This was around the beginning of February 2011, and my first task was to find out just how many different ways there actually were to detect water movement. I had some sense of this from the units we had already used in the past:

  • Mechanical methods, such as rotating impellers, veins etc. (like the RCM7’s Trish had used in her PHD work)
  • Electromagnetic flow sensors (like the InterOcean S4 or the Ott Nautilus) that detected the movement of ions in the water, or flux line deflection.
  • Ultrasonic systems (like the Falmouth Scientific ACM & the later model RCMs)

And of course there were many, many more listed on wikipedia including:

  • Optical flow meters, using reflection of light from particles.
  • Thermal mass flow sensors, using heat conduction.
  • Vortex flow meters, using turbulence and vibration.
  • Pressure based systems, that rely on the Bernoulli principle

I even thought of a method based on inductive coupling (that still might work) based on building a transformer with a hollow core through which water flowed. I hoped the inductance of the core would change (and therefore the voltage on the secondary winding) with the water velocity. But it seemed that I would also need a conductivity reading because the inductance would also change with whatever was dissolved with the water (probably much more so than flow) so I would have to separate the changes caused by salinity from those caused by water flow. This might even be possible, but it just seemed way too complicated.

And I was realizing that many of these approaches were simply out of my league, as my electronics background was fairly modest. So then I concentrated on the apparently simpler physical and mechanical methods, using a process of elimination based on major criterion Trish had outlined to find the the most promising one.

A propeller based system seemed pretty simple to build, and modifying one of the many hall effect magnetic flow sensors already on the market seemed like the easiest approach. But I had enough experience diving those caves to remember all the organic “ick” that could accumulate near the Cenote openings, so I was not really keen to use anything with a little propeller on it.

The thermal loss systems involved heating an element to a constant temperature and then measuring the electrical power that was required to maintain the heated element at temperature. So they seemed pretty easy to make, as they were basically just a resistor in line with something measuring the temperature.  In fact, there were plenty of cheap anemometers already out there based on mass flux that I could modify, but they were physically pretty small so I worried about fouling, and I had a sense that just throwing electrical power away as heat would draw too much juice for long term deployments, especially in an underwater environment.

Then I came across the Salamander Sensor Project. They were working on flow sensors too! And they had already open sourced a velocity sensor design based on a really simple flexible resistor strip. I was pretty excited, and thought that I had reached the end of my quest right there. But as I read into it, I discovered that they had been concentrating on streams, and other surface accessible water flows, with wireless data transmission. And I had a feeling that those flex sensor strips were just not going to like being put under significant pressure all the time. So none of it was really suitable for complete submersion in flooded caves, possibly thirty meters down, and hundreds of meters away from any surface access.

The Johnson Air speed indicator (1935)

The Johnson Air speed indicator (1935)

However, the fact that some other group had already come up with a water velocity measuring device was really encouraging, and the simplicity of the flexi sensors appealed to me, as they had so few moving parts. So I dug a little deeper into the concept, and found the same physical principal being applied all over the place like, for example, in the old Johnson Airspeed indicators, used by barnstormers back in the 30’s. All I had to do was figure out how to use the same “flow-pressure causing displacement” principal in a design that was small, robust, and most importantly, completely submersible for long periods of time….hmmmmm.

<—Click here to continue reading—>