iSixSigma

Help with Helicopter DOE

Six Sigma – iSixSigma Forums Old Forums General Help with Helicopter DOE

Viewing 26 posts - 1 through 26 (of 26 total)
  • Author
    Posts
  • #42262

    Jim P
    Participant

    Although I generally use a catapult for DOE class exercises, I would like to implement the “Helicopter DOE” described in the paper, “Teaching Experimental Design Techniques to Industrial Engineers”.  A pdf file of this paper can be downloaded on the internet at https://www.isixsigma.com/forum/showmessage.asp?messageID=10296
     
    The article provides a sketch of the dimensions of the paper used to make the helicopter but does not describe how to fold the paper, or where to place the paper clip(s).  I’d appreciate hearing a description of how to fold the paper and where to place the paper clip(s).  If you could provide any information on technique (height of drop, release method, measuring device, etc.), that would be a big help to me as well.
     
    If you post your response with your daytime phone number, I will call you after reading your post.
     
    Thanks.
     
    Jim 

    0
    #133362

    Jim H
    Participant

    OK – you know what a helicopter looks like? You  flold one of the blades one way (pick one), and the other blade folds down the opposite way. Ends up looking like a “T”.
    Put the paper clip(s) on the bottom.

    0
    #133363

    Jim P
    Participant

    Thanks, Jim H.  That is what I had guessed, but I wasn’t impressed with the performance.  From a height of 7 feet, it didn’t stay airborne for very long, and didn’t start twirling until it was almost halfway to the ground.  I held the helicopter with the wings down (like an upside down T) but didn’t notice any improvement in the onset of twirling when I turned it the opposite way (right side up).
    Anything else I need to know?  Do you use only one response variable (seconds) or do you use two (seconds and distance from target)?
    Jim P.

    0
    #133364

    BTDT
    Participant

    Jim P:Fold it like thisto make it look like thisCheers, BTDT

    0
    #133365

    Sheri
    Member

    Jim – I have a diagram that shows the places to cut and fold.  It’s not to scale but shows what you are looking for.  Please send me your email address and I will forward it to you.
    Regards,
    Sheri

    0
    #133366

    BTDT
    Participant

    Jim P:If you introduce two CTQs, you need two sets of specification limits. This is not usually done. The responce variable, “distance from target,” is quite skewed and non-normal. It will require a data transformation and a process capability calculation with only an upper limit.You will have two sets of factors and two prediction equations. The other issue is that the most accurate helicopters and those that drop like a brick, whereas the ones that stay aloft the longest time will drift all over the room.Students must be able to choose the optimal balance of the two CTQs by having a ‘cost’ function to normalize the opposing CTQs.Cheers, BTDT

    0
    #133371

    Jim P
    Participant

    Thanks, BTDT.  I understand now.  Could you provide the factors and levels used, and approximate seconds from X number of feet of altitude?
    Jim P.

    0
    #133372

    Jim P
    Participant

    Thanks, Sheri but someone already provided a sketch just prior to your post.  See the post from BTDT.
    However, if you have information on factors and levels, plus appoximate numbers of seconds at X number of feet, I would like to see that. It would help me plan for the classroom tests.
    If you can send an email, my address is [email protected]
     
     

    0
    #133376

    BTDT
    Participant

    Icarus:

    Icarus:The factors in the paper you have are frequently changed so
    you have seven factors. This makes it possible to use an eight run screening design
    with no replication as the first go. Based on the simple results, you will then
    be able to design a five factor, ResV, partial factorial with replication = 32
    runs. This is what we do. The body widths are changed by folding only, while
    the wing lengths are altered by cutting. The factor ‘taped wing’ refers to a
    small piece of tape applied at the junction of the two wings in an attempt to
    keep the wings open by making the wing fold a little stiffer.The benchmark drop time vary by room, the height of the
    person dropping and whether EHS will allow people to stand on chairs and
    tables. Make a quick helicopter out of copier paper, drop it from shoulder height
    and double your time.Cheers, BTDT

    Factor

    Low(-)

    High(+)

    Paper type

    Stiff card stock

    Copier paper

    Paper Clip

    No

    Yes

    Taped Body

    No

    Yes

    Taped wing

    No

    Yes

    Body Width

    1.42”

    2.00”

    Body Length

    3.00”

    4.75”

    Wing Length

    3.00”

    4.75”

     

    0
    #133378

    BTDT
    Participant

    Jim P:I knew ‘Icarus’ was another thread, but the nom-de-plume seemed far more appropriate to your post. Don’t fly too close to the sun.Cheers, BTDT

    0
    #133379

    Jim H
    Participant

    One of the factors not usually meantioned is the weight of the paper stock – the stiffer the paper, the more repeatable/predictable flight characteristics you will get (to a point, obviously). Also, if you want a fast spin, you can make one of the factors the angle of the blades – instead of bending them so they are flat across, bend them like a ceiling fan blade.  Just like a ceiling fan blade, the steeper the angle, the more air will move.
    I would embed a picture, but I am not a posting stud like BTDT.

    0
    #133380

    Jim P
    Participant

    Thanks, Jim H.  Yes, I figured I’d use two different weights of paper.
    I’m not a posting stud like BTHT either.
     

    0
    #133382

    BTDT
    Participant

    Jim P and H:I don’t know whether I want to be labelled as some kind of stud with all these other high fliers around.Cheers, BTDT

    0
    #133383

    Jim P
    Participant

    Good point on the transformation needed for the distance from target, BTDT.  I have other examples I can use for optimization exercises instead of one like this that pits seconds against distance.
     

    0
    #133384

    Jim P
    Participant

    Good stuff, BTDT.  I’m not sure how I will run this yet, whether as an 8 run with, or w/o replicates, or a 12-run without replicates. It will depend on where it is in the course, near the middle, or near the end. I’m still trying to determine the optimum order of introducing the various concepts.
    This is the first time I will be running a DOE course with everything created from scratch, and using Minitab software for the first time in the DOE class.  Previously I used video tapes from Greg Montgomery to cover about 50% of the course, and different software for the exercises.  With the current course I’m developing, it’s like starting all over again, so I’m still modifying the order of the material to where I think it will “flow” the best in class for students.
    Have a good weekend.

    0
    #133385

    K. Lo
    Participant

    everyone knows it’s not the size of the wings or the height of the drop; it’s all about the color of the paper.
    We used both Statapult and helicopters in my sessions, part of the challenge we had in the first trial was that we were not given instructions how to cut them out or fold them, it was just paper with lines on it. We found that everyone cut them out and folded them properly from the start, so that wasn’t a factor in our class.  I don’t recall using paperclips, but isn’t that something you’d want to include in a factorial design? can’t you predetermine a “hi, med, or low” placement and use those in your DOE?

    0
    #133386

    Jim P
    Participant

    The COLOR of the paper is significant???  Nah, I don’t think so.  You wre just kidding, right?  The WEIGHT of the paper, yes.
    The paper clip can be one of the factors as described in the paper in my original post above.  You can use none, one or two if you want a 3-level design.
     
     

    0
    #133387

    Jered Horn
    Participant

    This may go without saying, but…
    I can understand wanting to know all these details as the instructor, but figuring all this stuff out, as a student, is an important part of the lesson.  Don’t tell them everything you know.  Let them figure it out on their own.  Maybe they’ll teach you something.
    And, yeah, knowing where to cut/fold shouldn’t be one of the things they have to figure out.  I can send a better template that’s been “poke-yoked” to anybody that’s interested (email at [email protected]) —- PLEASE let’s not create another LONG thread of requests with people’s (vultures) email addresses.

    0
    #133389

    BTDT
    Participant

    Jim P:

    Jim P:Just a quick point about the colour of the paper.There are a huge number of BBs who have run the helicopter
    experiment and are used to seeing yellow card stock and white copier (20 lb)
    paper.In addition, because so many learned DOE with 2-level
    experiments, the paper clip was either on or off. Mixed two-level and
    three-level designs are a bit advanced for most training classes.I have a couple of stories about teaching DOE because it is
    Friday afternoon.We once had a class with a PhD aeronautical engineer from
    Sikorsky. He was adamant that he knew how to make the best helicopter. I
    convinced him to make a statement and then prove his claim with the data. At
    the end of the day, he proudly stood in front of the class and proved that his
    intuition was, by far, the worst thing to consider when making a paper
    helicopter. The lesson was not lost on the class and we all applauded.We did another class in France and one of the engineers who
    had already had two heart attacks was not going to do any such menial work as
    make paper helicopters. I happened to have Montgomery’s DOE book with me and
    asked him to examine the set of data on quality of Pinot Noir wines given the
    DOE set up with fermentation temperature, type of oak barrel, etc. Being French
    and admitting to enjoying the ‘occasional’ glass of wine, he jumped into it.
    His report out was about the most important factors in making a quality Pinot
    Noir. This was coloured by his own opinions about what was really
    important and what they really should have considered. He did prove that the
    gauge R&R on tasting had enough scatter that very little could be learned.
    Of course, the gauge R&R would have been much better if he had been
    on the panel.Cheers, BTDT

    0
    #133391

    Horticulture
    Participant

    Folks,
    If you want to mess with the students heads even more, act like a real customer and along with giving them the measurable outputs of flight time and distance from target, also give a fuzzy y – “spins nicely”
    Its another measurable that the DOE can be analyzed against and reinforces the idea that it’s good to have many outputs for each DOE run.
    You can then slide right into working with the Minitab optimizer function.

     

    0
    #133392

    Tim F
    Member

    I had an interesting experience when doing a helicopter experiment. Besides many of the factors here, I added paper color, figuring it would be a good way to illustrate a factor that had no effect. Weel, the color didn’t have any main effect, but I got a significant interaction between color and weight of the paper (cardstock vs typping paper). I repeated the expreiments and found the same interaction.Well, I got ahold of a sensitive scale and weighed the paper. Turns out that the blue card stock was lighter than the white cardstock, but the blue typing paper was heavier than the white typing paper! Not a lot in either case, but enough to be measureable. There was the interaction I had observed in the data!
    Tim F

    0
    #133435

    Jim P
    Participant

    BTDT,
    I enjoyed the story about the helicpopter PhD.  It might have been a travesty for your course had he been RIGHT.  Heaven forbid.  I’m glad that worked out for you.
    I never taught DOE in France, but did in Jamaica and Suriname, South America.
    I may not do any mixed level designs, but am considering a 3-level one, and strongly considering a Central Composite Design as well.  I’ll decide once I know what can reasonably fit into 4-days of training.
    Jim
     

    0
    #133437

    Jim P.
    Participant

    Tim,
    Great observation.  I’ll definitely keep that in mind – the fact that colored paper might be heavier than white paper. And if it is, I wouldn’t be surprised to observe some type of effect, either main effect or interaction.  Thanks for the heads-up on this.
    Jim

    0
    #133438

    Jim P
    Participant

    BTDT,
    I forgot to mention …I have a paper on the capapult from Quality Engineering magazine, circa 1990.  In it the author describes the efforts by a group of engineers to model the catapult, using engineering principles only.  They did a nice job – AFTER 4 DAYS OF WORK. The group using DOE did a good job too in JUST A FEW HOURS.
    If you want, I can get the literature reference. I need to warn you however, it is a really tough read due to all the mathematics used by the engineer group. Perhaps the only real value in this article is showing it to the “doubting Thomas’s”.
    Jim
     

    0
    #133468

    Kyzmit
    Participant

    I’ve performed this experiment twice, and I’ve learned that it’s certainly best to give it plenty of room to fall (second or third floor balcony, at a minimum) to be sure that the worst flying and fastest dropping helicopters have plenty of time to find their groove (some do, some don’t) and minimize error due to reaction time of the poor guy with the stopwatch.
    Also, get creative with the factors and levels (or let your students decide on them… 5 minutes of brainstorming can lead to some surprising suggestions that are really interesting to test).  Modifications to the design of the helicopter, such as rounding out or pointing the wings, additional folds on the wings or at the base of the wings, making the base or wings double thick, punching holes in the base, etc. can lead to some interesting flying patterns that break up the monotony of watching paper… spin in the air… again… and again… and again.

    0
    #133502

    Jim P.
    Participant

    Kyzmit,
    Good advice.  I plan to do this from the second floor anyway.  A height of 6 feet just doesn’t allow much time for flight.  As for letting their factors run wild, I may have one team using standard factos and levels, and have two teams do their own thing.
    Jim

    0
Viewing 26 posts - 1 through 26 (of 26 total)

The forum ‘General’ is closed to new topics and replies.