Filling the Antiroll Tank

I have installed the tank and put in the fluid.

I increased the viscosity by adding HPMC, as I said in the last installment.  I also increased the density of the water by adding calcium chloride.  Calcium chloride is a salt (like sodium chloride), but not as corrosive.  I actually added it for more than just increasing density (and therefore kinematic viscosity).  Since density is basically just weight in my application, I could do approximately the same by adding more water.  But I was concerned about whether HPMC was biodegradable, i.e., whether it would rot or turn sour, etc., when sitting in the ART.

I couldn't find anything that really addressed the issue, but HPMC is a type of cellulose, and cellulose is basically wood, and wood can decompose, I thought I'd make the environment a little less hospitable to whatever bugs and/or bacteria might live on HPMC.  Calcium chloride seemed to fit the bill.  It also further increased the dynamic viscosity of the liquid in the ART, thus slowing the weight transfer of the water without using baffles, which studies show tend to not only slow but also diffuse the effect of the internal wave.  Adding calcium chloride really heats up the water (which is why it is so effective when used to melt snow).  Because heat reduces viscosity (as noted in a prior post), I had to wait a few days to finally see the "room temp" viscosity.

I should be saying "internal waves" (plural), as what is going on inside the tank is not a single wave going back and forth.  I found a study where a model ART was built with a glass side so that the authors could take pictures of the wave as the tank teeter-tottered, mimicking a rolling vessel.  Video would be nice, but stills were also interesting.  The study concentrated on commercial vessels (most of the tank dimension calculations given in the study were in the order of 25 meters for a vessel's beam) with water depths in the proposed tank of 2+ meters.  Clearly thousands of pounds of moving water.  The study did also mention the approximate requirements of a "recreational vessel," but the example was of a vessel of over 100 feet.  Not really applicable to my boat, but still interesting.

The various tests included different fill levels, from 10% to over 50%.  Here is the model tank with just a little water at a 3 and 10 degree roll.  Too shallow for a surface wave, the effect is simply the entire amount runs side to side.  Without the video, it is a little difficult to deduce what is going on.  In a properly tuned ART, the water would arrive just after the "vessel" begins its recovery, i.e., its roll back to the other side.  There would be the added transferred weight hindering the return roll as well as the water's momentum suppressing the roll.  The water begins arriving just after the left side of the tank starts to rise and stays there basically the entire time while the vessel tips the opposite direction.  Then, the water rushes to the right to counter the next roll.  The second picture captures this.  No surface wave, just slosh.

Here is the tank with more water in it.

This is at small roll angles (.5 and 1 degree) with the tank half full (i.e., too full).  There is a wave on the right and an approaching wave near the center.  Those will intersect and form a nodal point, all of which has to be timed such that it reduces the vessel's roll.  Not a lot of weight transfer or moment, but then it only needs to offset small rolls.

Here is the effect at larger angles (3 and 10 degree) with the tank still half full.

Here we can see some of the limitations of too much water in the tank.  On a large roll, there isn't enough room at each end to accommodate the water.  Also, we have gone from a wave to a slosh, with some of the attributes of a tidal bore.  The water that is stacked up to the right will be held in place to a certain extent by the water continuing to run left to right. 

The study discusses the two modes that a free surface tank exhibits.  The first is "free surface," where a wave passes back and forth.  This is usually in smaller rolls, say of less than 3 degrees.  And it isn't always a single wave.  It is more like a major wave and some minor waves.  Once the bulk of the fluid is moving side-to-side, it is a slosh, not a smooth surface wave.  At higher roll angles, the entire amount of fluid is transferring, leaving essentially a dry tidal flat at the high end.  Thus, ART effectiveness is a factor of how deep the fluid is in the tank and how much room there is at each end of the tank to accommodate the slosh.  

And, as discussed in other studies, the effectiveness of the ART is not perfectly linear, meaning that the offsetting effect varies.  The ART may have a powerful countervailing effect on a 4 degree roll, less on a 6, then even more on an 8 degree roll.  All that means is that there isn't a perfect synchronization with such a simple system.  Unlikely that anyone would notice (except if using complex instrumentation in a laboratory setting).  I just mention this so that handwringers have something more to worry about. 

My design has some features that I haven't seen in any of the writings.  First is a modified viscosity/density, i.e., not plain water.  Slightly higher viscosity results in slightly slower wave and slosh speeds.  Second is a curved bottom on the tank, which has three effects: A) the end of the tank has effectively more volume, and B) the initial free surface wave is slowed as it crosses the center "shallow" area, and C)  the tank ends empty at a higher roll angle because of the curved bottom.  Based on the depth of fluid I'm using (less than 20% tank fill), a flat tank would completely empty one side to the other at about a 7 degree angle, while my curved bottom empties at a +10 angle.  How much difference do these features make?  Calling all grad students!

I am purposely using a smaller (shallower) amount of water.  That means that the ART sort of taps out at around +10 degrees with no more water to send to the other side.  The sloshing water does have more momentum at higher roll angles, based on a longer "drop" during a larger heel.  So a larger roll still produces a larger offset, but not a great deal more.  My theory is that the ART will have dampened the rolls that might lead to a 30 degree roll, so no need to actually snuff out a 30 degree roll.  I will be able to show that in the following videos, where I induced my boat to roll with the new ART in place.

First, check out my video in a previous blog where I rolled my boat without the ART.  Three steps aboard on the gunnel is the test and still rocking pretty good 8 rolls later.

https://stuffsax.blogspot.com/2025/09/tank-testing-art-at-dock.html

Here is the same three steps with the ART in the flying bridge with 20 gallons of fluid.  It doesn't rock quite as much to start because the ART is already at work, i.e., it is countering the roll right from the first step aboard.  But what is really noticeable is the damping effect after I have induced the roll.  

In two or three rolls the boat is back to the tiny little motions that it has "at rest" when in the boathouse.  Time to take it out and test it with the Washington State ferry wakes.*

The paint and primer was expensive ($100), the HPMC was $45, the calcium chloride was $50, so the total project is now over $400 total.  Obviously cheaper than hauling my boat out and gluing fins on the bottom.

*  Got to test it out yesterday, 12/2.  I'm not sure what went wrong.  Boat barely rocked.  Need more testing.

Finishing the Tank Construction

 I had done my Beta testing in the prior post and was now ready to finish the tank to blend in better (visually) on the flying bridge.  I finished the holes for the "inspection ports" by epoxying an additional thickness to hold the screws.  Then I painted it.  Because the weather had cooled off for the year, the paint (Total Boat Wet Edge) took over a week to dry.  Grrrrr.  I then put a patch of KiwiGrip on the section that will be stepped on.  The inspection ports were installed and it was raised onto the flying bridge.

As I said in a prior post, the large inspection ports were mainly because I needed access to the inside of the tank to finish construction using epoxy fillets.  Still, it is nice to be able to really look inside and see what the liquid is doing.  I'm hoping to do a video so that people can see the liquid in action.

For liquid, I began with water, of course, but wanted to experiment with additions to alter the viscosity.  I started with about 160# of fresh water (a little under 20 gallons or 70 liters).  This is a little less than I had in my antiroll bag experiment and less than in my first trial with this tank.  I think that the finished tank weighs about 40 pounds (I didn't weigh it), so the total weight of the tank on the flying bridge is about 200# (90kg).

To begin with altering the viscosity of the water, I used some HPMC powder (hydroxypropyl methylcellulose).  This is a weird "food-grade" powder that turns into a slime when mixed with water.  Used in some food preparation, it is also used in cosmetics.  Very slow to dissolve, it then turns from cloudy water to become a clear "suspicious looking" water.  Its origin lies in plant-based cellulose; the same kind of natural fiber found in wood pulp and cotton. These materials are modified by adding hydroxypropyl and methyl, making it more water-soluble.  The final substance is a fine white powder that dissolves in water. 

It forms a thixotropic liquid, meaning it has some odd characteristics.  If mixed in a high enough concentration, it forms a sticky gel which becomes less viscous when agitated.  Weird as that sounds, you've experienced this yourself.  The common example of a thixotropic fluid is ketchup.  Turn the bottle upside down and nothing comes out.  Shake or wiggle the bottle and the ketchup becomes momentarily liquified such that it will pour out.  The same is true with some paints (like the stuff I just painted on my antiroll tank.)  Thick on the brush, but when applied (with a shearing force) it becomes less viscous and spreads easily until the shearing force stops.  Then it sticks to a vertical surface without sagging.  Folks who use an air sprayer love paint with this characteristic.  Anyway, my mixture does stay thicker than water even when agitated, so it will slow down the wave action despite an ever changing viscosity.  

Just as interesting are the viscosity changes in a liquid based on temperature.  Most of us have experienced this with the oil in our car engines.  Started up cold, the oil moves slowly and oil pressure can be high.  Get the oil hot and viscosity drops, sometimes lowering oil pressure . . . sometimes too low.  Water does the same thing.  When you pour ice water into a cup, it is actually five times the viscosity of your boiling-hot water for tea.  Even though there is a measurable change, most of us don't even notice.  For my tank, I am looking for a viscosity change many times that of simple hot-to-cold water, but it is interesting that the viscosity will change from a hot day to a cold day.  Enough to notice?  Probably not.  Based on internet tables for viscosity of various liquids, I'm probably looking for something in the range of olive oil.

No, that's Olive Oyl.

To begin using the HPMC thickening powder, I added approximately 1/4 cup per gallon.  I thought that wouldn't do much, but I was impressed.  I didn't buy an expensive viscometer ($2,000).  Instead, I bought something similar to a Zahn cup ($8), used to compare the viscosity of liquids, but without an exact viscosity measurement (usually measured in centipose or cP).  Dip the little cup into fresh water, lift it out and time the draining through a little 4mm hole (10.4 seconds).  Dip it into my HPMC mixture, lift it out and time (11.3 seconds.)  I could tell the mixture was thicker, but not measured in centipose.  It only measures absolute viscosity, but one can then use that timing to repeatedly prepare the same viscosity.

Before adding HPMC to the tank, I made a little test batch (which is when I first measured viscosity).  Half a spoonful to half a liter.  Mixed, mixed, mixed.  It doesn't dissolve fast, but it does finally dissolve.  It dissolves even slower if just left alone (like hours or days).  When I poured the HPMC powder into the antiroll tank, I could either stir constantly for 9 hours or simply come back in a few days.  I chose the later.  More details on my antiroll "slime" when I visit the boat again.

Here are the viscosity testing cups sitting on top of the tank.  

I simply fill the tank using the inspection ports, lift out a cup full of the liquid, and time how long it takes to drain.  After the mixture cooled down, drain time increased to about 12.2 seconds.  I'll try that viscosity for awhile.

Tank testing the ART at the dock

I took a couple of videos at the dock to compare the motions of my boat without the antiroll tank (ART) and with the ART filled to about 75% of what I thought I would need.  Without that and the graphs that follow I would just be thinking "well, I think that it probably made a difference."  My wife is a true believer, so there is that.  I later modified the amount and viscosity to improve the ability of the ART to decay a roll.  

In an attempt to get an apples to apples video, I stepped on the gunnel 3 times in sync with my roll period.  Not perfectly scientific, but close enough.  I actually pressed a little harder in the second video to create the same roll angle from which to start measuring the decay.  That is because the ART starts decaying the roll even as I was trying to generate it.  Which would, of course, naturally effect the time required to decrease the roll.  To have "apples to apples" I needed to use more force.  So, I guess not really apples to apples in the end.  The ART starts working immediately, even before my measuring.

Here is the action without the ART.

About 8 rolls to the end of the video.  I wish I had longer videos, but Blogspot doesn't allow lengthy videos, so it wouldn't have helped here.  By the 8th or 9th roll, the rocking had diminished by half.

Here is the action with the ART partially filled.  The same number of "boardings" to create approximately the same roll (which required a little more effort on my part) but the roll decays by approximately one-half in 4 rolls after I instigate it.

Here is what the action looks like graphed with a g-force meter.  This is actually a graph of acceleration rather than inclination and was developed when using a bag instead of a tank, but I have found the same effect.  Acceleration is usually what makes the rolling unpleasant.  If my roll was slower, I might not be experimenting with an ART.  But my roll is "snappy" and can be uncomfortable and even dangerous.  

I wasn't careful enough when creating these graphs.  First, I didn't make sure that the recorder (my cell phone) was completely level.  That meant that zero g-force wasn't always aligned with zero on the graph.  Bummer.  Even more confusing was that the graph created by the program decided on its own what metric to use on the Y axis, and I didn't notice that.  The X axis is always in seconds, but the Y axis is different on the two graphs.  Bummer.

Here is the graph created without the tank.

 

As the sine waves are getting bigger (until right after the 30 second mark), they are not symmetrical at the top and bottom because I was rocking the boat (the meter is sensitive enough to register the force I applied when stepping on and off).  After that point is the natural smooth sine wave decay of the roll.  The first Y axis mark is .05G, but the actual g-force isn't what I was interested in (although I would guesstimate from the graph that the max was .1G.)  From about 34 seconds to 61 seconds, the g-force disintegrates by 50%.  Or a 50% decline in 27 seconds.

Here is the graph made with some liquid added to the ART.

It took me awhile to get the boat rocking.  I was trying to get it to approximately the same 7 degree roll.  Also, the graph isn't perfectly centered on zero on the X axis.  And to further complicate, as stated above, the Y markers have now changed to a 1G metric instead of .5 (so the beginning g-force after all my effort is still about .1G).  I tried hard to get the same amount of roll going, but I actually got pooped out before I got the full 7 degrees I was hoping for.  The graph is still instructive.  I had stopped "exciting" the roll at 150 seconds and was only interested in the decay time.  The amount of roll at 150 seconds decreased about 50% by the 163 second point.  With the tank filled, a 50% decrease took only 13 seconds.  A 50% reduction in both time and amplitude by filling the tank.

The ART stifled a similar roll to the same amount in half the time.  Plus, it was much harder to get the roll going in the first place.  It took me 30 seconds with the tank empty and 150 seconds with the tank filled.  It was one of those situations where it felt like somebody must be working against me.  Well it wasn't somebody, but some thing.  Free surface liquid first making a wave and then transferring side-to-side during larger induced rolls.  Amazing what a little water slopping around can do.  Time to finish constructing the tank.

 

Video of Antiroll Tank in Action

Here is some video and photos of our recording inclinometer to show the effect of the roll tank installed in the prior blog.  The videos are short because they are size limited by Blogspot.  

We had already crossed the Straight of Juan de Fuca, but no video there.  The next chance to really take advantage of the antiroll tank were encounters with the BC Ferry and other large vessels.  Here is a little video of a BC Ferry passing us (they tend to travel at about 20kt to our 6.5kt).  When they pass, the wake slowly passes us, making for lots of time to build synchronous rolls.  We usually turn and go directly into the wake of large vessels in order to stop building violent rolls.  Not necessary with the antiroll tank. 

Here is what the approaching wake looked like.  A good ten 2' waves perfectly spaced to get us violently rocking.  Time to batten the hatches.

The antiroll tank handled it just fine.  We could feel the wake, of course, but the rolling never built above 4 degrees.  It would lean us to one side, but the tank reduced the "snap back" such that the effect of the next wake was again fairly mild.  It felt like the tank was constantly frustrating the ability of the wake to really get things rocking.

I later went solo around Cape Caution in order to pick up Beth, who flew home from Bella Bella for a week and was flying back into Refuge Cove.  Here is a little video of "Cape Caution" (which I actually never even saw that day, except on the RADAR).  Very foggy and some long swell on the beam coming in from the Pacific.  Nothing the tank couldn't handle.

Oddly enough, the roughest part of my trip this year was from Squirrel Cove in Desolation Sound, where I spent the night at anchor, across the 3 or 4 miles to Refuge Cove, where Beth was coming in on a float plane at noon.  If I'd had a choice, I probably wouldn't have crossed until later in the day after the tide changed.  This is +20 knot wind against the flood tide and right on my beam.  The ride wasn't pleasant, but also not too uncomfortable once I got used to the fact that the max rolling was all of about 10 degrees and seemed fairly mild mannered compared to similar conditions in the past.  

The inclinometer shows that the maximum roll to port was about 5 degrees more than any roll to starboard.  That is because the wind was coming across the deck from starboard, causing a constant list of about 5 degrees from just the wind, even had there been no waves.  It was howling pretty good.  So the actual wave induced rolling was about 10 degrees.  That is about 1/2 of what I would have expected from these conditions without the tank.

My assessment is that the antiroll tank is definitely worth the time and effort.  I still have some experimentation to do with the amount and viscosity of the tank liquid,* but even with just a guesstimated amount of tap water, its utility was proven.  If I decide to integrate the tank into the flying bridge structure, which would entail some fiberglass work, it would still be only a $1,000 investment.  From what I've heard, it would be more effective than bilge keels (at $15,000).  Might be close to being as effective as active fins (at $40K).  Maybe not as effective as a gyro stabilizer (at +$60K).  And there are the advantages of no maintenance, nothing to snag crab lines, works at zero speed (at anchor), no need to haul out, no decrease in speed, no increase in fuel usage, no need to run a 240V generator (I've removed mine), etc.  As I said in a prior blog, at $1,000 there is also no profit margin for a seller/installer/designer.  That is likely the reason that antiroll tanks are rare.

I have read a couple of places where antiroll tanks are claimed to be dangerous in that they might affect the vessel's stability.  Well, yes, that is possible with any system, although probably less likely with a "gravity activated" system as opposed to an "electric/hydraulic controlled" system.  Gravity is predictable and always "on."  Other antiroll systems have had their failures, but for some reason don't inspire the same level of criticism (and even fear).  Take for example the effects of an active fin system that gets a bit out of tune.  The fins started rotating the wrong direction, exacerbating the rolls rather than calming them.  Here is a picture of seas similar to what I was in, but their fins got out of sync and the system had to be shut down before it sank the vessel.  Gravity doesn't have glitches.

I know what that feels like.  Here is a picture of my inclinometer taken before I had the antiroll tank operational (going through Race Passage on Johnstone Straight).  This is about what the boat above was doing with their expensive active fins. I don't intend to do that again, thanks to my DIY antiroll tank.

* I found the chemical I had used to slightly increase the viscosity of the water in my tank allowed a little growth to form.  It don't hurt the runnin' none, but it was kind of gross.  I have an idea for a sanitary next fill.

Antiroll Tank: First impressions

We have now been cruising with our test antiroll tank for over a month.  I have purposely taken the boat into conditions that I would normally try to avoid.  One of those instances was crossing the Straight of Juan de Fuca from Port Townsend to Sydney, B.C.  It can be calm, but even then, it is likely that there will be swells coming in from the Pacific.  Our crossing was sort of "medium" conditions for the Straight.  2-3' waves on a 6 second period.  Underneath was the swell, but it only made itself known when it coincided with the wind waves.  

We had a mini-disaster during the crossing, but it wasn't really related to the antiroll tank.  I had changed our hollow fiberglass mast to a tabernacle mast so that it could be folded downs when entering our newly acquired boat house.  The hinge was strong enough.  The lock down mechanisms were strong enough.  What wasn't strong enough was the thickness of the fiberglass mast (and I didn't properly reinforce).  

I had some questions about the PO installation of the Garmin radar dome on the antenna (with a custom stainless mount).  I wasn't sure if the original mast was intended to be merely decorative.  I had already reinforced the bottom bolts where the motion of the mast had begun to wear larger holes in the fiberglass.  When I removed the mast for cutting and adding the tabernacle, I was surprised at how light the Garmin dome was, and that lead me to not super-reinforce the newly installed hinge.

In purposely getting into "test conditions" for the antiroll tank, I put plenty of strain on the unstayed mast.  And so the hinge failed in the Straight.  The mast fell onto the solar panel, but not much damage.  Even though the rolling motion had been reduced, the strain was still too much for my tabernacle design (or lack of design).  The top of the mast was then sistered to the stub on the deck.  It was secured with duct tape.  Radar was still functional, but anchor light was not.  We seldom have other boats in the remote anchorages we chose (most don't even have names), but still we left on the aft cockpit light at night.  Sort of like leaving on the porch light.

That's our new Starlink antenna on the deck to the right of the mast.  Mindboggling that I'm typing this while at anchor in a foggy little cove SW of Bella Bella.  I said in the prior post about the tank's construction that I might be able to continue the posts using Starlink, and it definitely is possible.

But the antiroll tank was a success.  Our rolls didn't hardly register on my recording inclinometer.  Below is a photo of the tank installed.  It has "8 inch inspection ports" on each end, not so much for inspection or filling, but because I needed to access the interior in order to do the final epoxy fillets to construct the tank.  Right now, the ports are simply sitting in their openings.  I didn't have time to finalize the tank before we left.

You can also see the hinge on my tabernacle mast in the down position in the lower left corner.  That is the hinge that failed to handle the rocking motion in the Straight, despite the rocking motion being greatly reduced by the tank.  How much it was reduced will be the topic of the next post.

Antiroll tank on a small recreational trawler

My initial testing using an "antiroll bag" could only get me so far.  There is an amazing amount of flexing that goes on when the boat rolls and the available materials for a bag don't seem to be up to the challenge.  A rigid tank would not wear out (hopefully), but presented some additional calculations in order to synchronize with my boat's roll period.  Actually, not quite synchronize.  The idea is that the liquid in the tank will transfer from side to side slightly slower than the vessels roll period.  The liquid (or a wave) will reach the "down side" of the roll tank just after the vessel begins to right itself.  Thus each "snap back" is diminished, making building synchronous rolls difficult or even eliminating the problem.

The issue can be addressed mathematically by measuring the vessel's natural roll period, measuring the moment arm of the tank location, calculating the dynamic viscosity of the water in the tank, adjusting internal baffles to effectuate the proper timing of the fluid, etc.  For those of you who like to nerd out on mindboggling calculations, check out the grad school papers on the subject of antiroll tanks that are on Google.

You will be glad to learn that there is another way of doing this.  Commonly called "tank testing" by those wanting to sound professional, it is essentially trial and error using hunches.  Keep in mind that the Wright brothers were bicycle mechanics, not rocket scientists (or even aviation engineers, as that moniker hadn't been a thing before manned flight).  I'm guessing that Wilbur and Orville couldn't write a graduate thesis on the mechanical transfer of kinetic energy into centripetal force resulting in velocity directed by a secondary rotational mechanism using gravitational friction (grad speak without the Greek letters) to explain bicycles.  Also keep in mind the immortal words: "It won't never get off the ground, Orville."  We can ignore equations, explanations, and diagrams for the moment and concentrate on results.

Scale models of antiroll tanks are easily built and studied in a wave tank, but for full scale testing, it is a little more complicated.  Fortunately, my marina is near a ferry dock and a Washington State ferry leaves about every other hour (my state-funded "wave generator").  The perfect tank test site.  My trials with the antiroll bag gave me some idea of what would be needed.  But it seemed that the best experimentation would require something that allowed ongoing "tuning" of the tank.

The requirements are two-fold.  First, there is the timing of the transfer of the liquid from side to side. The second issue if the amount of force needed to negate the vessel's natural tendency to roll back beyond its upright position.  Keep in mind that this antiroll "force" is a little different than other antiroll systems.  It doesn't impose a drag on the vessel's speed or tendency to roll (as with dragging plates in the water, fins that rotate to lessen a roll, or "rolling chocks").  In fact, it works the same at anchor.  The transfer of weight in the tank merely reduces a return from a roll, thereby reducing the ability of the vessel to build synchronous rolls.  There's nothing that needs routine maintenance, wiring, rigging, hydraulics, etc.  Nothing to snag fishing nets, kelp, flotsam.  No need to haul out to install.

Most boaters have already felt this type of roll reduction when transiting a confused seaway.  At certain points the boat starts rocking and one thinks "here comes a big roll" only to have an unexpected wave "slap" against the hull and negate the coming roll.  Thank you, errant wave.  Well, the effectiveness of an antiroll tank is based on that principal.  In this case, every roll receives its corresponding "slap back" wave.  Except the wave is overhead and there is no slapping noise from my tank.

Like all antiroll systems, one cannot expect total elimination of the vessel's motion.  I have seen a few systems that cater to land lubbers, costing lots of dollars and diesel, that claim to steady the boat like a dirt home.  Basically, a dirt boat.  I'm not convinced that is necessary.  I hope to do some kind of a "dollars per degree damping" analysis when I get my system tuned.  So far, I have less than $300 into my tank, so comparing it to a mid-level $30,000 active hydraulic fin system should be fun.  I've heard that the +$50,000 240V gyro systems (with their own dedicated generator and required annual haulout for maintenance) are even more efficient.  Well, one would hope so.

But on with the experiment.  Having tried a few things, here is what I've come up with.  

Yes, it looks like a coffin.  I got some weird looks when transferring it to the marina on the roof of my car.  The curvature has a significant effect on the "wave" inside of the tank.  As we all have been told, waves slow down in shallower water.  I needed a way to slow down the wave in the tank, as the natural transfer of a water wave is generally too fast for an antiroll tank.  Naw, you will see the reason in a later blog.  But the curve is the bottom of the tank, i.e., the tank is upside down in the pictures.  It was just easiest to build that way.

The construction is epoxy fillets, no fasteners, similar to a stitch-and-glue dinghy that I built years ago.  I had some old West System epoxy, but not enough to do the entire project.  So I had to buy another gallon of West System ($200).  That's why my project was $300.  Not shown in the picture is my super secret baffle system that makes the transfer of the special antiroll fluid perfectly timed.

Actually, although there is a baffle of sorts, the timing of the fluid transfer relies on something else.  I didn't get enough time to "tank test" my project using the Washington State ferry system, but I did get some understanding on how to "tune" the system without the use of baffles.  Adding or removing baffles from the tank would have been very difficult, so I chose to alter the viscosity of the fluid used inside the tank.

Here are the basics.  The "wave speed" in the tank is partially controlled by the viscosity of the fluid.  The higher the viscosity, the slower the wave.  A wave in acetone moves faster than a wave in fresh water.  A wave in vinegar moves slower than a wave in fresh water.  And a wave in saltwater also moves slower than a wave in fresh water.  Now, you may have never noticed this because we are talking about very small differences, but we are also talking about small differences in the timing in an antiroll tank.  

Just to make things fun, there are two types of viscosity.  There is absolute viscosity (like we were talking about above).  That is basically how fast will a liquid drains through a hole.  Water will be one of the faster liquids, 50 weight oil will be slower.  Simple enough.  But the second type of viscosity is called kinematic viscosity.  It is based on the density of the liquid.  Sort of like how much "punch" does the moving liquid have.  That is a function of mass.  Think of it this way.  Getting hit by a water balloon compared to getting hit by a molasses balloon.  Set aside the stickiness, the molasses balloon packs a bigger punch when it hits you.  It has more mass.  That results is a bigger smack from the same sized balloon.

Water has been the traditional fluid in antiroll tanks.  Not much has been said about whether to use fresh or saltwater.  Understandable because the absolute and kinematic viscosity isn't that great between the two (although saltwater wins on both counts.)  But a greater absolute viscosity doesn't always win.  Take the example of 50 weight oil above.  Thicker than water, i.e., higher absolute viscosity.  But lighter in density, in fact it floats on water.  It loses in the kinematic viscosity battle.  

In an antiroll tank, both types of viscosity matter.  For a small roll, the surface wave crosses from one side to another at a certain speed.  On larger roll, the entire volume in the tank may transfer from side to side.  And it isn't just the weight, but also the "smack" that determines the roll cancelling effect.  Well, that and the moment arm, which I haven't gotten into here.

Turns out that there are fluids that can be adjusted with noncorrosive additives to create the right absolute and kinetic viscosity without using baffles or specially shaped tanks.  That's where I'm at right now.  Adjusting the amount of fluid and the kinematic and absolute viscosity of the fluid in my tank.  Unfortunately, I had already planned to cruise to northern B.C. this summer.  The tank is with me, but I don't have what I need to fine tune the tank.  It works, but I don't know if it is at full potential.  Since I now have Starlink aboard, I may be able to update this blog while "on the road," so to speak.  

* Starlink was quite amazing and I was able to make cell phone calls from remote areas, get better weather updates, and even write a blog about the continued use of the antiroll tank.

Anti-Roll Bag ver. 2.2.

I took a break in my Anti-Roll Bag research and spent several months cruising in northern B.C.  I did take both my homemade and my store bought anti-roll bags.  As I mentioned in the prior blog, ARB ver 2.1, my store bought bag was lightly made and immediately sprung leaks.  Also, it was longer than the width of my flying bridge and therefor was folded upwards at both ends.  I was wondering if this might effect the timing of the water passing back and forth.  I would say yes, in a bad way.  The end of the bag would "inflate" with water, but then drain back out slowly because of the crease at the end of the bag.  It lost some of it "umph" or slosh effect.  Here is a picture (empty) showing how the bag folded at each end.

The $35 store bought bag with about 1 foot extra at each end.

Just as bad was that the bag end folds and unfolds thousands of times a day as the boat rocks.  The material couldn't handle it and developed leaks.  So I went back to my other DIY bag made of stronger material.  But it developed the same issues.  Little leaks and obvious wear points where the bag folded and unfolded thousands of times.  Plus, the force of the "inflation" slowly opened the end seams.  By the time we reached Race Passage on Johnstone Straight, it had just a bit of water in it and was doing nothing.

Here is the bag shortly after a roll to port.  Basically all 150# of water is resisting the roll back to starboard and damping the motion.  I was hoping that the seam being up against the side wall would reduce the stress.  It probably did, but unfortunately, the force of water opened the end seams after days of this.  The material was also starting to delaminate at the creased areas.  You can see that under the bag end and around the coaming is wet from the small leaks.

For a video of the bag in action, look at ARB 1.0.  Unfortunately, the camera was handheld and I didn't include the horizon line.  Still, you can see that the bulk of the water reaches the port side just as the port side is lifting.  Even though a transfer of only about 150# of water, it significantly reduced the roll.  Part of that is the "slosh effect," i.e., the liquid mass has accelerated (for 8 feet across the deck) and has momentum.  The "moment arm" from here to the boat's center of gravity is about 9 feet, thus allowing 150# of water to have more effect than one would imagine.  The boat wants to lurch back, the water says "Not so fast, Buddy."

For those of you who are unfamiliar with this area of B.C., it is notorious for rough water when the wind runs against the current.  We ran it going north on such a day.  Just as we were getting ready to cross to the west to exit by Current Passage, I was hailed by a cruise ship (still out of sight) and asked if I could stay east on the Kelsey Bay side.  The ship was travelling at about 25 knots, so we slowed and let it pass.  But that +20 minutes changed the wave train considerably.  We crossed over to Earl Ledge and rode a fairly smooth laminar flow at about 3.5 knots.  Until the end.  Then the wave train started.  Steep 6 to 8 foot standing waves that were curling over at the top because of the wind.  Even when slowed down, we were still crashing through.  

I decided to move over close to Hardwicke Island so that we could bail out and retreat a little to anchor in Forward Harbor for the night.  But in order to get out of the mess, I had to take several waves on the beam and miss a powerful eddy.  We had prepared for this and everything was stowed properly.  Still, a couple things went flying.  And I got read the riot act by my wife.  What is important to me is to not frighten her and effect her desire to go cruising.  Also important to survive, of course, which we did.  Here is a photo of my recording inclinometer.  It would have been a good time for testing an anti-roll system, however, I don't expect to repeat this.

That's right.  Maximum was 23 degrees to port and 38 degrees to starboard.  Enough of that.

The ARB was proof of concept, but it looks like an anti-roll tank is required for several reasons.  Although the bag slowed the internal wave/water transfer a little without the use of baffles, a bag has other interference issues.  And vinyl coated polyester (or probably any flexible bag material) simply can't handle being creased and re-creased thousands of times under pressure.  Looks like it is time to build an actual tank.  

I've looked for something off-the-shelf that might be modified, but having no luck it will likely need to be completely fabricated.  The interior of my flying bridge cowling is just under 8 feet side to side so plywood is probably the material for a test tank.  I'm thinking that it would probably be about 18-24 inches wide.  As you can see from the picture of the bag above, although a tank would only need contain 3-4 inches of water, the tank (or at least the ends of the tank) would likely need to be 10-15 inches tall to contain all the water that would be transferred in a larger roll.

To be continued . . .