Tow Craft Prototype

I've switched back to the "tugboat" to get it ready for a conference, "Marine Propulsion and Design: Inspirations from Nature" in Norfolk, VA, next week. This is a video of an air test of the revised hull. The wiring works!

This is an isolation pod which holds the motor for my next PHISH prototype, a diver propulsion vehicle (see next entry, below). Shown is the iso-pod and motor. A battery pack and electronic speed controller will encompass the motor, inside of this pod. This is the only sealed component. A detail I like: M3 countersunk bolt holes.

Video of next prototype: Phish as Diver Propulsion Vehicle! Hit your space bar to pause the video so you can read the text.
The battery and electronics are internal and used as the inertial mass driven by the torque reaction engine. This one will steer left-right and up-down, controlled by a diver.

Another rainy Sunday afternoon. I have a new fluke with more flexure. It is going faster! I also vary speed, which you can hear in the audio.
The fishBOAT has a piece of lead on its nose to keep it level. This is a bad place to put extra weight, as it dampens the torque reaction. 
The next version will have adjustable displacement, fore and aft, so its trim can be adjusted without dampening the motor. I'm holding up
the wires so they don't drag on the seats (we're in a hot tub).

The fishBOAT swims in a hot tub on a rainy Sunday afternoon in January. 

fishBOAT 2.2, out for a first water test. This trial was interesting because I wasn't using a harness, just a catenary power cable as a tether. 
The fluke was a bit too stiff, but I'll be able to put a more flexible fluke in it. I made it so that I can replace the fluke without too much effort.
 First "free swim" for fishBOAT v2, in a pool. Power use is somewhat less than 17 watts (this total at the motor, not including about 8 watts used by the Arduino, ESC, and power-measuring board -- this DOES NOT factor in power generated during the braking phase), bollard pull is 25-30 grams, speed is about 1+ mph (maybe faster). Compare this to v1, where power use was 15 watts (at the motor), but bollard pull was 3 grams and speed was about 0.28 mph.

fishBOAT 2.2, partially assembled. How fast will it go? Will the linear actuator really change tail flexure at different speeds? 
Who wants to buy one?

fishBOAT for dinner

First "free swim" for fishBOAT v2, in a pool. Power use is somewhat less than 17 watts (this total at the motor, not including about 8 watts used by the Arduino, ESC, and power-measuring board -- this DOES NOT factor in power generated during the braking phase), bollard pull is 25-30 grams, speed is about 1+ mph (maybe faster). Compare this to v1, where power use was 15 watts (at the motor), but bollard pull was 3 grams and speed was about 0.28 mph.

fishBOAT v2 test, fully submerged. The motor is operating in only one direction, cyclically speeding up and slowing down. Power utilization ranges between 5 and 25 watts; assuming speed was only 0.5 mph (which is conservative -- it may be closer to 1 mph) a strouhal number of about 0.3 was achieved (fish achieve strouhal numbers between 0.2 and 0.4). More to come in terms of measuring bollard pull and free swimming speed, as well as fixing attitude and trim. Toward the end of the video, the camera is down in the water with the fishBOAT, which is an interesting view.

First water for Phish v2. Two modes of ESC control are demonstrated. In the first mode (in the first section, underwater), the motor and inertial mass oscillate back and forth, changing direction. The speed of the inertial mass passes 0. In the second mode, the motor and inertial mass go in only one direction. The inertial mass is cyclically braked and accelerated, producing torque reaction, but without changing rotational direction. Due to the quirks of this unsensored ESC (it does not have Hall sensors), the second mode produces a better rhythm; crossing 0 in the first mode is hard for the unsensored ESC/motor combination. Maintaining speed in one direction provides back EMF and allows the ESC to maintain a better pattern. However, a sensored ESC will produce  a better rhythm in both modes.

Next is adjusting depth.
Below is one of my early flukes for v2 and the interior of the hull. Looks good!

Dry land training with v2! This is still using the ESC from v1. In development is a new ESC. The new ESC will be smoother.

Bollard pull: about 3 grams for about 23 watts of power (2.73 Amps, average across the power cycle, times 8.4 volts is 23 watts). 
Bollard pull would be a bit better in deeper water, but still not bad. This is TOTAL power draw by the ESC, Arduino, the measuring system (which together draw about 8 watts), and the motor (drawing 15 watts).

The fishBOAT swimming, shot from below. The new fluke is working well. The fishBOAT is going about 0.28 mph. The motor is loud. It is right on top of the camera, and it has a primitive motor. V2 has a much nicer motor.

First "free swim" for the fishBOAT! For its first free run, Phish-1 towed its barge reliably and with low power utilization.  
Phish-1 didn't have steering (and I didn't want to go swimming), so I had to keep it close and there are a lot of things to improve, but the concept works!

The fishBOAT connected to a "barge" via a harness. It can operate without the harness, but it is helpful for this prototype.
Next is waterproofing the electronics and letting it "swim free" in the ocean. 

The fishBOAT generates serious thrust!

The fishBOAT swims!  This was basically the first real test, so there are a TON of issues and LOTS of things to improve.  But, wow, it works! 

YouTube Video

One of the biggest issues in this test is that the fishBOAT is in a tub of water and the craft is interacting with its tether, rather than being free-floating.  The craft is pointing up, in fact, because interaction between the craft's thrust and the tether is pushing the nose up. A couple of times I gave the craft slack on its tether and I could see it swim rapidly across the tub.  There are a lot of significant variables which I have yet to optimize, such as how far out the fluke projects from the body (and how much deflection the fluke undergoes) and the frequency and the amplitude of the motor oscillations.  It is commonly understood that fixed propeller tests do not do as well as dynamic tests.  There is more work yet to do before I can set the fishBOAT free, but this clearly shows that the basic idea works.

This is video of the fishBOAT with a rubber skin.  Getting ready to go in the water!

This is a video of the fishBOAT operating with a first version of the improved ESC from Helldyne, Inc., of Bainbridge Island, WA.  It works great!  You can listen to the motor and tell that I am not driving it very hard.  The Helldyne ESC allows me to adjust the frequency and amplitude of the motor oscillations.  When you hold the craft in your hands, it feels like a fish which is trying to swim!

Soon, the fishBOAT will be swimming in the water!

This was in early August, 2014.  This was a rough test, but the hull is now assembled, the motor can reverse, and the tail and harness are attached.  The length of the tail is adjustable.  The Electronic Speed Controller ("ESC") is still fairly unresponsive in this video, so I'm only driving it one direction and braking in the other.  This is also in the air, without the dampening effect of water.  Even with these significant limitations, this still shows that ample power is available with only modest use of electricity. This craft is going to work!  Helldyne, Inc. of Bainbridge Island is working on an improved ESC which will produce a nice, sinusoidal oscillation of the craft.

2014-08-02 fishBOAT test

Here's a close view of the fishBOAT:

fishBOAT close view

The video, below, was from the first test of the torque reaction engine in the fishBOAT hull.  See here for a summary of the torque reaction engine This test only shows the motor going in one direction (counter-clockwise), which only drives the boat in a clockwise rotation. This first test clearly demonstrates that significant force is communicated to the tail through rotation of the inertial mass by the electric motor.  

First test of the torque reaction engine in the fishBOAT

This picture shows the motor mount on the bottom hull section and the motor.  The motor has a lead shell which Ballard Machine Works fabricated for me. The wires from the motor exit the hull through the three wire-sheaths (the red, yellow, and black straws poking up around the motor mount).  A harness will gather the wires and bring them back to a "barge."  The barge will hold batteries and, later, a solar panel.  I am also making a tail or fluke.  The top of the motor has a bearing which I am mounting into the top hull section.  

I designed the tow craft prototype in Sketchup and intersected the 3D model with 1/16" layers.  I pulled out the intersection lines to a PDF and a friend cut them with a laser cutter on 1/16" wood layers.  I stacked the wood layers and bonded them together.  I put a layer of fiberglass on the exterior of the hull in the area of the center of displacement.  I made it in two halves, so that I could put the motor in it.  It turned out very well!  I used this technique to build the much larger fishBOOT.

This prototype is approximately 12" long.  It is a reasonable size for an inexpensive data acquisition drone.  Because the torque reaction engine is entirely sealed, because it has only one moving part, and because it is very efficient, it will be able to operate for a very long time (days or weeks) with minimal supervision.