View Full Version : [Project] Mola Project

10-30-2009, 06:12 PM
Hello everyone,

My name is Charles Whitaker and I'm currently working on my 11th grade engineering project. For this I decided to create a Oceanic Sunfish (Mola mola) type blimp:


Here's a concept on how the final project will (eventually) look like:


My reasonings is to create a more efficient and flexible blimp system by utilizing the unique body shape and movement of the Oceanic Sunfish:

YouTube - Mola mola The ocean sun Fish

I have done expensive research by reviewing various papers on the subject. The most prominent is this paper:


which have set the perimeters which I want the final product to met. I want the bot to be able to travel .75 meters a second (the average form the study) with the fins completing a full cycle after 2 seconds (about .5 hertz a second). I have also decided, on part of my limited knowledge of programming (which I'm hoping to improve on), to make this robot an RC vehicle until I have more experience programming.

Anyway here are the parts that I'm planning to use for the body:

Arduino Pro Mini Microcontroller - 5V/16MHz
Dynalloy Flexinol Wire 0.010" HT (probably not that size, odds are it'll be with a smaller diameter)
Chloroprene Weather Balloons (For the prototype)
Wooden Poles (For a support structure in the body)

For the custom controller I'll be using:

SFE Concave Button - Black
SFE 2.4GHz Small Duck Antenna RP-SMA
SFE Reverse Polarized SMA Connector - Vertical Mount

I also thinking of using a On Shine 433MHz High Sensitivity Transmitter/Receiver Pair to connect the Controller with the Mola:


and powering them both with their own SFE Polymer Lithium Ion Batteries - 3.7V 860mAh:


The microcontroller will be programmed with an SFE FTDI USB to Serial Basic Breakout Board (3.3V):


Anyway the project will be starting in the next few weeks (possibly next week) so I'll be posting my progress on the project.

I also opened an thread for help on Robotshop's forum. Several people there helped me choose some of the parts for this project:


This project was inspired by the Oceanic Sunfish of course and Festo's animal inspired robots:

YouTube - Festo AirPenguin
YouTube - Festo AirJelly
YouTube - Festo Air_ray

10-30-2009, 08:48 PM

Cool idea, and it's nice to see someone actually do their research. A few thoughts though on your part choices.

First off, I'd seriously stay away from those 433Mhz modules, they can be a real pain to get working -- something like an XBEE radio pair costs a bit more, but it's nice to know it's going to always work. What are you putting in the controller though to send commands out? I'm assuming a small micro, but I don't see any mention.

Second, I notice on that other thread people saying you *could* use the Arduino to directly power the flexinol -- that's a bad idea if the current is >5-10mA. First off, it builds up heat on the Arduino, and could pop an output. Second, the contraction rate of the flexinol is dependent on the current, if you can't instantly source enough current your bot will move slowly. You don't need a full h-bridge as they suggest, a simple signal transistor will work (unless a signle line is driving a couple feet of that flexinol, in which case you might need a power transistor).

Lastly, it looks like you plan to power your 5V arduino off a 3.7V battery, but I don't see a mention of a boost regulator to do that.... if you're set on that, no problem, but if you haven't looked into your power circuit, you need to check it out (I'm not sure of your level of experiance here).


10-30-2009, 10:58 PM
Thanks, I have several other projects that I have a lot of research done on that I'm hoping to get started later in the year. A few specialized version of this Mola bot are also included but I need to do more research on them. A little sneak peak of the Masturus bot and Ranzania bot's inspiration:

Sharptail Mola Masturus lanceolatus

Slender Sunfish Ranzania laevis

XBEE Radio Pair, you mean this:


I didn't know if the controller needed a micro or not. Perhaps another Arduino Pro Mini Microcontroller, but I don't know where to get any pins for it?

Okay how do I control the signal transistor with the RC transmitter? Do I connect the RC transmitter to the micro controller or to the transistor? Sorry if this sounds stupid, but I'm new at this. Do I even need the microcontroller for the Mola's body? Wait I do, I forgot to mention the Dimension Engineering Buffered 3g Tripple Axis Accelerometer:


The power supply ... I think made a mistake with the micro controller that I posted. I meant to post the 3.3 volt verison:


But you're right I do need to review my power setup. I'll new on robotics (the construction and circuitry at least) so any critique on the system is welcomed.

11-03-2009, 07:50 PM
Anyway I created the circuit layout, but I have some questions.

http://photos-d.ak.fbcdn.net/hphotos-ak-snc3/hs107.snc3/15459_1101053861561_1683129241_209193_2497326_n.jp g

Okay so here is the RC thingy. I choose the same micro controller (link in previous post) that I'll use for the body as well as the same power supply. They'll also share an XBEE transmitter (link in last post). However I'm wondering if the transimitter is in the right spot? It can't connect to the analog pins because the buttons and joystick take up all of those. Also does the motherboard have a regulator that'll bring the 3.7 volt battery down to 3.3?

Body Circuit:

Okay so here we have the circuit that'll be placed on the body. lnxfergy, you said that I don't have to do an H bridge and recommended another approach, is this what you meant? Sorry if the question is stupid, but I'm in a rush to get this stuff done so that I can order the parts by the end of the week. If this right, should the muscle wires be placed on the analog pins or on digital pins? Right now I have them on analog.

Again sorry if these questions are stupid.

Thank you.

11-03-2009, 11:39 PM
What's that thingie on the bottom that has six lines going to it?

We need pin names, mspaint drawings included need pin names. :)

Muscle wires wouldn't connect directly to an MCU in any way. Waaay tooo much current draw.

An XBee Transmitter will go RX on the XBee to TX on the micro, TX on the XBee to RX on the micro. The XBee has a VCC that goes to power (3v) and Common GND

11-03-2009, 11:57 PM
The arduino mini does have a 3.3 volt regulator on it, if the was the mother board you were referring to. The maximum voltage is around 12 volts as far as I can remember. I think fergs was saying that you could use a transistor to run your muscle wire. If your not sure how a transistor works start here (http://en.wikipedia.org/wiki/Transistor). The pins coming out of the mini aren't going to provide enough power to actuate the muscle wire by themselves.

11-03-2009, 11:58 PM
How are you powering this project?

11-04-2009, 12:02 AM
>> Also does the motherboard have a regulator that'll bring the 3.7 volt battery down to 3.3?

It would be incredibly rare to find a regulator that only had four tenths of a volt of dropout.

Regulators, like diodes and similar components, drop the voltage going across them. So the maximum output is not some 1:1 correlatation. If I had a regulator on a 3.7v battery I'd typically be looking for 3v output from a low dropout regulator.

11-04-2009, 10:02 AM
Adrenalynn Wrote:
What's that thingie on the bottom that has six lines going to it?

That's the XBEE transmitter.

We need pin names, mspaint drawings included need pin names.

Sorry about that, I'll go back and fix it.

11-04-2009, 11:15 AM
The XBEE is a one wire device. One wire for receive and/or one wire for transmit. Without pin names on both devices, I'm not sure what you were trying to do, but it's lookin' kinda random. :)

It also wants power and ground, but yeah, there's only going to be two wires going from it to the micro.

11-04-2009, 11:48 PM
RE: Regulator on Arduino Mini
As Adrenalynn said, it's tough to bring 3.7V down to 3.3V (especially because in this case, a 3.7V LiPO will actually run from 4.2V at fully charged to ~3.2V when you should shut it off to avoid blowing it up). Hence my previous comment about a boost regulator (to take the 3.2~4.2V and boost it to something > 3.3V, the minimum an Arduino willl handle).

I actually wasn't recommmending switching to a 3.3V system, from the 5V. I think you'll actually find that it's easier to get a 3.7V Lipo range -> 5V boost regulator (Pololu sells several models for <= $10). 5V also tends to be easier to interface to other stuff (except in this case the XBEE), so if you plan to add other sensors, etc, later on, you may be interested in staying with a 5V micro, and using a boost regulator. The other added benefit of a 5V arduino is that it's more "standard" than a 3.3V setup. There's a reason they call them "Arduino Pro" modules, they are a bit harder to use (some libraries may not function as expected at 8Mhz vs the standard 16Mhz).

Note, you might take a look at my Basic Power Supplies Tutorial (http://forums.trossenrobotics.com/tutorials/how-to-diy-128/power-supply-basics-3258/).

RE: Schematic Creation
You should take a look at a real schematic program like EAGLE. Even if you aren't doing board layout, you can put in your schematics in an easy to view way that cuts down on errors. And although you'll spend sometime learning to use the new program, there are so many pre-drawn items that you won't have to draw your XBEE or your Arduino, they are already in there.

RE: Driving the muscle wire
A circle on a line, labeled transistor doesn't tell me a lot, but yeah, generally, the idea is to use a signal transistor to drive the larger load with your low-power I/O pin on the Arduino. You'll want to switch those transistors with an PWM (analog output, in Arduino terms) pin, if you want to be able to vary how much current goes to a load (not sure this actually does anything in muscle wire, since the current is so low, but it may allow a limited control on the speed of contraction)

RE: Other Parts
One thing I don't see on your parts list, is an FTDI Cable or In-System Programmer, which will be neccessary if using an Arduino Pro module.


11-17-2009, 09:00 PM
Sorry for the delay in replying. I've been very busy lately with NaNoWriter. Anyway here's an updated part list from all the recommendations I've been provided and the research I've done:

(2) Arduino Pro Mini 5V/16MHz
(2) SFE Polymer Lithium Ion Batteries - 3.7V 860mAh
(2) Pololu 5V Boost Regulator NCP1402
(1) 5.5Ft Weather Balloon
(1) FTDI TTL -232R-3V3 USB - TTL Level Serial Converter Cable
(1) Dynalloy Flexinol Wire 0.005"
(4) SFE Concave Button - Black
(2) XBee 1mW Communication Module
(1) Phidgets Mini Joystick Sensor
(1) Right Angle Pins
(6) PNP Bipolar Silcon Transistor
(2) Dimension Engineering 3.3V 1A Switching Voltage Regulator
Wooden Poles 1/8"

I've downloaded the Eagle software and I'll start of the layout tomorrow at school.

01-20-2010, 08:45 PM
Wow, I'm so sorry for leaving this thread dead in the water. Um there have been a few delays in this project because I've been busy lately. Don't worry this project is not dead I'm still working on it so expect future updates.

01-25-2010, 04:42 PM
http://hphotos-snc3.fbcdn.net/hs190.snc3/19772_1147244056287_1683129241_297669_7375656_n.jp g
Does this look right? It brought up some questions that really need to be answered if you'd be so kind. Um ... well Adrenalynn said that the XBEE was a one wired device however the Eagle software kept telling me that I needed to connect the resets and that I had to have DIN and DOUT connected so I'm a bit confused. Also it said that I needed to connect the 5 V and a 3 V pins to a power supply which also has me messed up. Am I using the wrong parts? I tried looking for specific parts in lbr form so I had to improvise.

01-27-2010, 01:04 PM
A few ideas for you:

1. I would use two SN754410 H-Bridge instead of the discrete transistors. I have found them to work better and there are fewer wires.

2. If you don't need any intelligent behavior on the receiver end of the system, you can just use the 8 DIOs on the XBee and two SN754410 and you will not need an Arduino.

3. If you go to a kite or model airplane store you can get thin graphite rods that will be lighter and stronger then your wooden ones. The model airplane store guys will also have plenty of advice on how to build lightweight fins if you talk to them.

4. Connecting muscle wire is a PITA because you can't solder to it. The model airplane store will have some very small aluminum tubing you can cut in to short pieces and use to crimp the muscle wire to your copper wire. Also keep in mind that muscle wire is somewhat brittle compared to copper, and easily broken, so design your system so it is easy to replace a broken strand.

01-27-2010, 03:40 PM
I can't tell because of the scale of the drawing but it looks like the ground for the XBEE is separate from the other grounds. You should always tie all grounds together (unless you need an electrically isolated circuit, which you don't)

01-27-2010, 07:55 PM
Thank you darrellt and jes1510,

I'm considering trying both the H Bridge and Transistors and decide for myself which works better. I've had some people tell me the H Bridge would be better and others say the Transistors would be.

In the future I would like to have a controller that can tell me a bit more but for the prototyping phase I only need to control the bot.

I had considered Graphite rods but the size I need are beyond my budget. I could connect smaller rods together but that would create weak spots in the sides. I have a good idea already on how to make the wings light weight.

Thanks for the tip with the muscle wire.

I'll keep that in mind for the circuit's revision. Thanks.

01-28-2010, 01:01 PM
Just as a side note: "One Wire/Two Wire" refers to the data communication, not the glue. And Reset (RST) is not required in normal operation.

01-28-2010, 04:01 PM
So you need to build something really cheep and really lightweight? When I was a kid we would build hot air balloons out of a dry-cleaning bag and McDonald's straws for spars. A cotton ball soaked in alcohol was the burner. MickyD's has larger diameter straws made out of thicker plastic then standard, very strong. Well, strong for a straw anyway.

01-30-2010, 01:43 PM
Hello all,

Here is a revision of my previous circuit design:


and here is a revision of my Mola body circuit:


For the body circuit, I should note that I couldn't find any representation that could be used for a muscle wire so I left that area blank.

How do these designs look? Are the viable? I'll be sending out for the hardware and software next week so that I can begin testing.