Project Abstract

      I will work with Dr. Oh this term to develop a Two Degree of Freedom (2 DOF) Laboratory Helicopter. This device would be used for demonstration and mechanical engineering laboratory classes. It would be used to study aerodynamics, fluid mechanics, nonlinear control, and more. It could also be used as a platform for more advanced experiments. The device already exists and is available from a supplier (Quanser), but the high cost is prohibitive for most environments. The goal of this project is to develop the highest quality version of the device which could be produced for a more reasonable $500-$600. We may share this plan with the academic community by publishing a paper in a scholarly journal. I will be concerned mainly with the mechanics of the device, and a basic control system. The control system will be further developed by a group with appropriate expertise.

Picture Source: www.quanser.com      

      The  helicopter itself is comprised of a central column, about which an arm can rotate. The placement of airfoils is identical to many real-world helicopters: one on the top front for lift, and one on the tail to compensate for the torque effect produced by the main airfoil. The  helicopter will be able to pitch and yaw by manipulating the two airfoils. This device presents several design challenges, including an electronics system with components that can rotate indefinitely. (See Figure 1). We are in no way reverse engineering Quanser's device, as we do not have one of their models. They have made a description, pictures, and movies of its operation publicly available though. See their version here.

Dr. Oh is the Project Advisor. We have developed this abstract together, and plan to adhere to the schedule below. I will consult him or his Graduate Students for any problems, and meet with him regularly to discuss progress and re-assess goals. I will spend approximately 6-8 hours per week on the project, not including a weekly program meeting and a weekly advisor meeting. Throughout, I will create and maintain a web-page logging the development of the project. My final deliverable will be a formal 20-minute presentation. Please note that the timeline and deliverables of the project have been modified with Dr. Oh's guidance throughout the course of the project.

Source: http://fargo.itp.tsoa.nyu.edu/~ms1436/sensors/optical.html

      What is an optical encoder? Basically, it's a spinning disk with holes in it. (It gets more complicated, don't worry.) One one side of the disk is a light source. On the other is a light detector directly across from the source. As the disk spins, the detector can see the light when there is a hole, and can't see it when there is a solid part. Thus, the signal will go on and off (digitally) as the disk turns. What good is a disk with holes in it though? To make it useful, a shaft is connected to it. That shaft can then be coupled to just about any rotating system. By recording how many times the signal goes on and off, we can tell how far the shaft has rotated. By recording the time of each "off" signal, we could find the angular velocity. More interesting things come from that point on. Note that some mechanism is also present to tell which direction it is turning in. This is most likely in the form of a second set of holes, purposely offset. Mechanical encoders are also available. They would do something like depress a switch as a gear ridge passed. They lend themselves to slow, simple movement though. Thus, an optical encoder with its capability of high speeds and accuracy lends itself to this application. Note that encoders only give a relative position, not an absolute one. It would have to be calibrated and "zeroed" to give a useful position. Slippage would be a major problem, because the software and hardware would then be out of synchronization. [My HTML editor didn't remind me to save any of the above "Optical Encoders" section so it was all lost and re-done...I must find a new HTML editor.]
      I experimented with a US Digital Optical Encoder and "Encoder Data Acquisition USB Device" called the "USB1-S." To experiment with it, I drew a circle and marked its degrees with drafting tools. Yes, the pointer that I attached really is a binder clip and a post-it note. Hey, it worked. (See below)

After calling the company to get the drivers for the device, I was able to connect it to my PC via USB. This is a convenient alternative to a DAQ (Data acquisition Card). Basically, it allowed the signals from the encoder to be sent to my computer for processing. The program that they included works on a basic level as is. It allows you to see a diagram and associated numbers as you turn the shaft for up to 4 encoders. The manual included a good deal of programming and electronics instructions, which are beyond my understanding. Some features are:

• logging what happens to a file, either continuously or only when something moves
• Triggering other programs when a certain event occurs (Such as Encoder #2 moving more than 90 degrees, or traveling 10000 rotations)
• Variable Resolution (More on that soon)
• Multiple Counting Modes

       The manual and full features of the hardware and software can be found here: http://www.usdigital.com/products/usb1/

On the left is a picture of the main program screen. Note the time stamp on the left, which is continuously counting in microseconds. The hardware can accept up to 4 encoder inputs, so 4 are displayed on the screen. Notice on the right that values have been changed for all 4. To do this I simply plugged the one encoder that I had into each port separately. The software could do 4 simultaneously though. 
      I then began to do some quantitative experimentation. I first noticed that the dial on the screen moved in the opposite direction to my pointer setup. It would match if my setup were upside-down. I tried to find a way to reverse it, but did not. I'm sure with some clever programming it would not be a problem. The basic interface for each encoder allows you to name the encoder, see the position, adjust the resolution, change the counting mode, and reset the position to zero. The resolution didn't do what I thought it would at first. I expected the software to see 1 full rotation every time I did one on the encoder. I then expected the resolution to break up that rotation into smaller pieces as I increased the resolution. After checking here, I now know that the resolution actually changes the number of pulses per revolution. That's how many times the encoder says that it hit a solid point per revolution. The disk inside must be very precise, because the resolution goes up to several thousand. I played with the resolution until I could get the software to match the rotation of the encoder. It works best at 980 as far as I can tell, so I imagine that the number of openings on the disk is around 980. The counting mode (1x, 2x, 4x) seems to just multiply the movement of the encoder by that number. For example, after I calibrated the resolution to 980, I changed the counting mode. At 2x, the software turned 180 degrees when I turned 90. At 4x, it made a complete circle. I also noticed that every time I moved the encoder, a green LED marked "activity" lit up on the box next to the encoder input.
      This system offers a great deal of possibilities. With some programming, it could be very useful for my application and many others. The USB-1 device costs $300 though, and is rather large. It's also overkill for what I want to do, so I know we will have to use some other device to interface with the computer. This will probably be in LabVIEW, which is also provided by US Digital. We'll eventually use a DAQ of some sort to simplify the operation and lower the cost significantly.

[Setting up the website and creating this section took about 7 hours rather than the 3 that I had scheduled for it. Hopefully it'll go smoother from here on.]

                                 uea-inc.com  (FLASH Animation)
                                 sliprings.com

Note that there are also alternatives to the traditional slip ring, such as Rotating Electrical Controllers (RECs). For now, I will focus on slip rings. They are a tried and true technology, will serve our purpose, have many suppliers, and that's what Quanser used. On another note, it appears that building a slipring of sufficient quality from scratch in this time would be impossible. It would have to come in the form of a stackable kit or something similar. So far, I have been unable to find any such thing. It looks like we ll have to buy it, hopefully at a reasonable price. If the prices look too high though, I'll certainly try to learn how to fabricate one. Now to make it practical for this project, we need to know how it will be used in the helicopter. First we need a layout, so we'll start with the one from Quanser. (Top of Page) My design drawing is soon to come, even if it's hand-drawn. For now, I have assembled a parts list:


Parts List

    
From the above design, it seems we will need five channels in the slipring:

Slip Ring Channels

Potential Slip Ring Suppliers (Not a final list of course)

http://www.rotarysystems.com/
http://www.mercotac.com/ (alternative to slipring)
http://www.uea-inc.com/slip/index.html
http://www.polysci.com/SlipRingcatalogsummary.htm
http://www.smallparts.com  (website down when I tried it)
http://www.grainger.com
http://www.mcmaster.com/ (NONE?)
http://www.robotcombat.com/links-suppliers.html

More to come soon.

            

1. Slip Ring - The largest expense of the project will be the slip ring at $100-$125. I am planning to go with an Electro-Tek H-Series slip ring, unless FabriCast  can offer a better deal, perhaps on a more robust slip ring. They’re on PST, so I’ll call today (Friday) when they’re open.

Estimated Cost: $100-$125

2. Supplies for the base and various hardware – I plan to build the body of the helicopter out of balsa, per your suggestion Dr. Oh. The base could probably be made of something denser, hopefully a machinable plastic, metal, or harder wood. The vertical supports could be of several materials, including metal rods or wooden dowels. All of these things I expect I could find at a hobby store, unless your group has a supplier for such things. Bill, please let me know how I should go about searching for / purchasing these items.

Estimated Cost: $50

3. Motors/Airfoils – Still remain to be discussed specifically. After some rough calculations which I will check with someone soon, it appears that the main airfoil assembly will have to produce 7.6 N of thrust. Does that value seem at least in the right order of magnitude? I am uncertain of the power requirement for the tail assembly. The Quanser airfoils had diameters of 280 mm and 203 mm for the pitch and yaw, respectively. Again, I’ll heck this with someone in the lab.

Estimated Cost: $20 (for both motors) + $40 (for both airfoils)

4. Electronics - Still remain to be discussed specifically. We’ll need two optical encoders, a power supply, a board, wires, the op-amps (as you described Dr. Oh), and probably some other small things. Will we be using this with other lab equipment, or should it be stand-alone? (Especially for the power supply)

Estimated Cost: $120 (for both Encoders) + $40 (The rest)

                                    Total Estimate: $450

2-22-05: It's Alive!

As the Work Log above may have shown, I have not updated the website in some time. I have certainly been busy on this project though. A brief summary of what has happened since the last post:

1. Met with Mike Joyce to explore the basic theory, and get an idea of which motors and airfoils to use.
2. Made several purchases from McMaster-Carr with the help of Bill Green to get the basic mechanical materials.
3. Continued to work on choosing a slip ring, and got suggestion from Todd and Dr. Oh to explore using a telephone cord untangler (See one here: Phone Untangler). I have since purchased some untanglers and phone cord to test and develop for this application. The biggest barrier is that phone cords have 4 circuits and we need 6. Also, its current carrying capability may be an issue for running the motors. I'll hopefully look into this later.
4. Continued to refine design as cleverness and/or problems arose. Discussed machining with Keith Sevcik. Dr. Oh made contact with Mark Schriber at the Drexel Machine Shop to request permission for me to use the equipment.
5. Began making several trips to a local hardware store to get loose-end items like correctly sized washers, and angle brackets that aren't strong enough to hold up a deck (the ones from McMaster were) Also did some machining at the Machine Shop using the bandsaw and drill press. Mark and the shop staff are great guys. In out lab, cut steel angle brackets to the correct length with the masonry blade on the cop saw. Many sparks...good, safe fun.
6. Continued putting the pieces together, refining the design, and solving problems as they arose. Got the mechanical system up and running, and made another McMaster order. Also ordered the chosen slipring, after devising a way to connect it to the optical encoder without putting any load on the slipring.
7. Got an evaluation copy of Adobe GoLive, which I am now using to modify the webpage. I like it very much...it only supports my desire to redo this entire website in super-awesomeness. (How much of this I'll have time for remains to be seen.) An education license of Adobe Creative Suite could be a great investment. Also received 10 interviews for Co-op, and am currently working them into my schedule. Exciting in more ways than one.
8. McMaster order and Slip Ring arrive. Put it together and....It's ALIVE as the following videos and images will show:

First, a brief explanation of the parts involved (more detail to follow at a later time):

• Bases and uprights are made of HDPE, cut to size
• Supports between bases are threaded studs, with matching press-fit inserts. This allows for solid support while allowing the system to be disassembled for repair, modification, or storage.
• The weight of everything above the upper base rests on a simple turntable (bearings between two specially designed plates) so that the weight is not on the slip ring, as was with Quanser's design.
• The uprights are also made of HDPE, and fastened to the angle brackets with press-fit inserts (of a different size than for the supports)
• The helicopter body is made of balsa. It will have a nice balsa sheet fuselage for aesthetic purposes soon
• The arm on which the body rotates is an alminum rod, which will be coupled with the encoder soon.
• The motors were pulled from a blimp toy. Please note that they are not permanently mounted as of yet, so that I could easily modify their positions. Also note that the bolts taped to the arm served to manipulate the balance of the beam. Somethign similar but more elegant may be used in the final design.
• The apparatus shown in the second picture above is the optical encoder for yaw. It's not connected yet, as I require an english hex key set to work with the couplings.

Below are videos of the helicopter in action. You will need the Divx Codec (click here to get it) to view them. Note that as I had one power supply at my disposal, I was only able to run one motor at a time. The function of each motor is thus showcased seperately in the videos.

Here is a link to my final presentation: Final Presentation

The above presentation was given to Faculty and peers as the main deliverable of the MEM 399 class. After speaking with Dr. Oh, we determined that it would not be in the best interest of anyone involved for me to cotinue on the project independantly due to time constraints and other commitments. Instead, we expect that the completion of this project would make an excellent Senior Design Project, or could be continued by another honors student. We discussed the project in terms of Technology Readiness Levels (TRLs). You can find a link to the NASA document explaining TRL's here: Technology Readiness Levels. Dr. Oh assessed my current progress at TRL 4. He expects his Senior Design students to reach TRL 8. As such, please find two sections below giving a rough layout of how to get there.

The basics of the relevant TRL levels:
Click here for full NASA document. Both quotes below copied verbatim from the document linked at left.

TRL 4 (Current): "Component and/or breadboard validation in laboratory environment"

TRL 8 (Senior Design Goal): "Actual system completed and "flight qualified" through test and demonstration (ground or space)"

If I Could Have Done It All Over Again (with more time)...

• I would first do a thorough search of the literature to. Two main purposes would be to see if anyone has already done this, and if not, to see what is available to help. Relevant writing on lab devices, sliprings, etc. would be helpful. I did do a brief search to make sure I wasn't doing a redundant project, but didn't read much. Always good to know how your work is fitting into the scientific and engineering communities.
• I would more seriously consider building a custom slipring from scratch. This would of course be more work, especially if this project is published with recommended instructions for other academic institutions to use. The advantage would be a solid platform capable of yawing indefinitely, with an encoder mounted perfectly to it. It would also reduce friction, and ensure predictable sustainable performance. This would of course take a good deal of extra engineering to develop. The disadvantages versus using an off-the-shelf slipring would be: 1) Much more time involved 2) More difficult to reproduce 3) Possibly higher cost. I would look deeper into off-the-shelf sliprings as well, and perhaps commission a custom-made one from a company. It may be as simple for them as modifying a few drawings and making one up that does exactly what we need it to. Also, I noticed some sliprings with encoders built into them. None were the correct size for this application, but it may besomething that could be custom ordered. I also still have the Radio Shack phone Untangler, and never really looked into using that. It only has 4 conductors to be used, but someone may yet figure out a way to use it.
• I would solve the basic dynamics equations in full. This involves some complex moments of intertia, so it may be easier to determine that value experimentally. We got as far as a few pages of calculations, and then decided to let it go. It will definitely be necesssary for writing the code though. In the world of engineering, these calculations would be relatively simple, though perhaps tedious.
• Related tot he above point, I would figure out the exact force output and equations for the motor and airfoil system. This will definitely be required for the programming later. It would also help in choosing an appropriate motor and airfoil for the job.
• I would use the machine shop for almost all fabrication. Often I would use the chop saw in the lab, or try to measure things very carefully with a huge T-square. Needless to say, I think my fabrication could have been a lot more precise.
• I would order reverse-threaded (right hand / left hand) studs to use form McMaster Carr. For a few bucks this could be done even now. As it stands, I screw them fully into the bottom plate and then "unscrew" them into the top plate. They should be reverse threaded. I believe it was Mike that aptly said though, that right hand / left hand threads hardly ever work like you want them to. His idea then was to simply use a few nuts and washers to solidly connect the studs to the top plate. I agree that this would be a good approach to it.
• I would make the airfoils look nicer. These ones were pulled right off of a blimp toy. They look okay, but nowhere near as cool as Quanser's.
• I would MUCH more carefully do the balancing of the helicopter. Right now, there are some washers taped to the back. It works, but its precision ranks up there with trying to hang a picture frame without a level.
• I would work on the friction problems that the device now has. For some reason, when in pitch, the two side supports seem to move closer and further away, almost as if the support bar were acting as a screw. Tried to sand it to fix it. Weeeeird though. The turntable also seems to have some "sticky" spots which silicone grease didn't make much better. What can I expect from a simple ball bearing turntable? This issue may be solved by the second bullet above though ( custom slipring)
• I would draw the helicopter body outline in CAD, probably using a real-world helicopter as a rough template. It looks ok now, but I'm sure some CAD magic would make it look much better.
• I would create and implement the electronics required for the system. I did not have the time or knowledge for this step, but Dr. Oh assures me that for someone with the right knowledge, it would be simple.
• I would create and implement software to run the system. It would have a nice GUI, and would present a challenging, fun, and informative lab for the students using it. A mathematical model in MATLAB or LabView wouldn't hurt either. That way, the students could play with a "virtual copter" and then see it in the real world.
• I would write a manual and lab to go along with the system.
• I would test the system and refine it to get it in its best working condition.
• I would see if the engineering education community were interested in this design and if so, write a paper for it.

To bring this project to TRL 8:

Please find below a rough outline of the steps and time required to bring the system to TRL 8 as an engineering design team would. Of course more may be learned throughout the re-design process that changes time requirments, or adds or removes entire steps. All hours are given in "man hours" (male or female of course) that could be split amongst multiple people. Expect that the more people are involved, a small percentage more total time would be required because of communication and the such. I imagine that for a team of 4, a 5% increase in the time required for this reason would be reasonable. I have erred on the longer side of the time estimates to reduce the chance of surprises. Please take this only as an estimate though, and change what you see fit. A 10% buffer has been added for the unknown unknowns that are bound to happen.

The final estimate for a 4 person team assuming 3 terms at 10 weeks each comes out to be 3.4 - 7.6 hours per week per person devolted to the project. (I have included one hour for a weekly advisor meeting in the weekly esitimate). I have had to make an educated guess on some of the steps, so unforunately the estimates may be way off. I also have not included any senior design deliverables such as papers or a website, or "tutorials" as Dr. Oh likes his students to create. Depending on how these estimates shape up, this could certainly be a viable senior design project. It would be suitable for a smalle team as well (perhaps three students) as the time estimate now stands.

Good luck. To anyone who continues this project: please feel free to contact me regarding the 2DOF Helicopter project. I'd be happy to try to recall anything from my experience with it, and give any direction or advice I could.