DLG Molded Project
I have received quite a bit of feedback about doing a molded project here in the states from several different fliers. I want to start a thread to track the progress, but warn that it will not be as rapid as many of my build threads.
I probably won't reveal all of the details of what I'm doing because there may be a production run later on, but I do want to show how I'm going about trying to create all of the necessary tooling on a reasonable budget. Before you ask, I haven't defined yet the definition of "reasonable" but I'm trying to keep it under $1000 for tooling. I know you're thinking "I can get a prodution moldie for that, and probably a second one for the cost of the materials and time to get an airplane flying. You're right. You can. This thread and the build will be geared toward people who aren't familiar with the processes used, and maybe we'll shortcut a couple of things to keep production costs reasonable for people who work for a living. The goal is to get more people here in the US making moldies in their garages and basements.
Let me start by saying what tools I'm using. These are probably the cheapest tools that are capable of doing what you need them to do.
1. Airflow analysis software. I use XFLR5. It's reasonably easy to use compared to others out there, and other than some complex 3d calculations on spanwise flow and a few other nuggets, it's pretty capable. I would dare to say it's much more capable than I am, and probably more capable than most modelers.
2. 3D CAD software. I use Rhino 3D version 4. Rhino is extremely powerful and can be licensed to a single user for under $1000 with all features. That's cheap in the CAD world. If you're fortunate enough to be married to someone in the education field, they have a substantial discount for teachers, students, faculty, etc. That lands you a license for under $200.
3. A good machine shop or about $10k to plop down for a low end CNC machine. My machine shop has a Shopbot PRS Alpha which is a reasonably accurate 3 axis machine that is capable of doing wood, MDF, Renshape, and even aluminum if we're really patient. I have looked at purchasing or building a CNC for about three years. Finally I decided that I can have someone else sweat over it and worry about broken cutters, hold down, dust collection, space, material handling, feeds and speeds, etc. and pay them far less than that machine would cost me.
4. Friends who can help you with production work and tooling creation. You can't do this by yourself. There are large, fragile pieces at play here and a second or third set of hands is often necessary.
With this all being said, I will end the post with a teaser picture of the finished Rhino drawing of the model. Over the coming weeks I'll show you how I drew this, then disassembled it into production files for the CNC, how the molds are finished and created, then how real parts are created from the molds.
Enjoy the picture, and I'll try to post a little bit of what I've done every day for a few weeks.
You want to make a molded DLG. Where do you start? I won't go into the details of airfoil analysis because there's a bunch of information out here about that. It's kind of boring too. Tweak this, test, tune that, test, tweak this again, test again, etc. It's not terribly exciting, and it involves alot of compromise. Speaking of compromise, get used to the word. I've probably read about 30 books on full scale and model aircraft design and aerodynamics. Every one of them included the word compromise. It is, was, and always will be a part of the game.
So you have a design roughly picked out. How do you make that design become a reality? You have to get reasonably proficient at a 3D CAD program like Rhino. I like Rhino and picked it because you have one-stop shopping here. It's got some capabilities for drawing in 2D or 3D, as well as capabilities for rendering, light analysis, and even some toolpathing. It may not be perfect but it's what I chose. Beware of older versions. I use v4. The previous version had some really big issues with alot of the functions (boolean in particular) that you'll need to complete the molded project. With the current version loaded up, I'd suggest that you take the time to download the tutorial files and go through them COMPLETELY. I'm fairly good at 3D CAD work and I still learn something every time I go through those tutorials. If you're unfamiliar with CAD at all, they'll take a bit of time and may be frustrating. Trust me though, they're less frustrating than trying to draw up something without knowing what you're doing. As a test, I had a computer-savvy friend of mine try to draw a stick figure in Rhino without any training. He spun his wheels for over 30 minutes and couldn't do it. Back on topic...
So we need to think through a lot of things before we start this project. Here's a short list:
1. Do I have the patience for this?
2. Do I have the budget for this?
3. Will my wife divorce me and take my kids if I spend too much time on the computer?
Hopefully you have two yes's and a no. If you have any other combination, stop now and save your pennies for a Euro-moldie.
Seriously though... it is work. It's alot of work. It involves some re-work. Frustration will set in and you will question why you're doing it. Dedicate a certain amount of time a week and only allow that much time to it.
It's surprising how easy it is to design airplanes that can't be built. Play with any airfoil software and you'll soon find yourself optimizing some wierd looking thing that could never fly in reality. All through the process you have to keep asking yourself a couple of questions:
Can this be built?
How will I build it?
How much time will this take to build?
If you don't ask these questions, the time factor can snowball. My first fuselage was oval in shape and I was parting the molds on the "thin" side with a wing set into the fuse. I started looking (proudly I might add) at the dihedral angles that I'd wrestled with how to draw, and soon I realized that I had acute angles with sharp corners that couldn't be milled into the mold material. Crap! Either I find a different machine shop with a 5 axis machine (and triple my budget) or I rethink how I was shaping the molds.
Obviously I chose to reshape the molds, and even that took me down a path of more decisions: Do I use a four part mold instead of two and keep the shape the same? Do I use a two part mold and modify the shape? I think you can see where this is going.
I chose to stay with two part molds, and simply seam it on the sides instead of top and bottom. This will add an extra layer or so of cloth in the areas where it could do some good too, so it's not a bad thing, right? We'll see.
Some other things to consider as you're doing this project are as follows: The things that are easiest to draw are the least desirable. Stuff you want (smooth curves, accurate airfoils, sleek lines with fillets and chamfers) are the things that will take time to draw up. Take your time drawing it and you won't be sorry later.
The picture I shared above is a rendering of a drawing that I created that has all of the major surfaces. This picture has the hardware removed, and only has the smooth surfaces. I use this as a starting point for all future drawings. My current project has about 30 drawings right now. Some are overviews like the one above. Some have individual surfaces and the associate 3D mold files that I send to the machine shop. Some of those were edited after my machine guy called back and explained to me that there's no way he could accurately produce what I emailed him. My point here is to get the whole project planned in advance rather than drawing a wing first then realizing that you don't have a good way to attach it to the fuselage, etc. More thought up front leads to less swearing down the road.
I own Autocad and use it fairly often, but it would be nice if you could give a little info on using Rhino for lofting the wing. I have gone through the tutorials a few times and it is not the easiest program ever (although I can draw a stick figure with it PDQ). I am just used to ACAD.
Last edited by will_newton; Aug 19, 2009 at 02:25 PM.
I don't have a machine shop. My day job is in IT. I should have more clearly stated that "the machine shop I use" has a shopbot. I have taken a bunch of screenshots of Rhino along the way to share as well.
To answer the question quickly, you can either loft the wing (which is my preference) or do a two rail sweep. Both work, but lofting gives you a few more options to close the edges, etc. I have lots of detail to share, and will eventually share my cutting files and everything.
Great job on the stick figure!
More stuff to consider:
Airfoils are typically plotted in Profili or some other software with a zero thickness trailing edge. That doesn't "fly" in the real world. You need to modify (truncate) your airfoils for use in molding before you import them to a drawing. Even though I ALWAYS start by drawing my fuselage, I'm going to jump into the wing because I think you'll get more excited seeing a wing drawn and rendered with a minimum amount of work.
Here's what I do:
After tweaking and fiddling for seemingly weeks to get the right airfoil progressions, cambers, thicknesses, and washout progression, save all of your files in .dat format. Next, decide what TE thickness you want to build your molds to. I like .3mm personally. That works out to around .012 inches. This thickness gives you enough room for two layers of cloth before on either side before closing the mold, and more importantly it gives you room for a bit of splooge to hold the top and bottom together. Obviously if you leave .3mm depth on top and bottom, your trailing edge finished thickness will be in the neighborhood of .027-.030 since you won't be getting your molds to close perfectly. Trying for a thinner TE is really risky since you're dealing with super thin stuff now and the margin for error goes up a bunch, as does the likelihood of chipping or breakage.
Anyhow, back to how to mod the file. DAT files use coordinates in pairs like this...
0.99981 0.00034 *
0.99914 0.00039 *
0.99716 0.00054 *
0.99180 0.00092 *
0.98365 0.00152 *
0.97278 0.00231 *
0.97998 -0.00009 *
0.99382 -0.00024 *
0.99980 -0.00033 *
(BTW... this coord file already was "roughened" to about every third coordinate before I started working with it.)
Coordinate files typically start at or very near 1,0 and end up typically very near 1,0 with numbers positive being above the centerline and numbers negative being below. As always, the X axis coordinate is the first number and y being second. Coordinates at the start of the file are the trailing edge top surface. The middle of the file contains the leading edge top and bottom, and the end of the file contains the trailing edge bottom surfaces. Since we want to make sure that our thickness (y axis) is never less than .012" per mold side at the trailing edge, I tend to start in the beginning and end of the file at the 1,0 (trailing edge) and delete pairs of coordinates that have a similar x axis value and have a y axis sum of less than .024/chord. A pair of coordinates to me is one top and one bottom with a similar X axis coordinate. Not all .DAT files line up perfectly so you have to use your head a bit. Some have more coordinate density either on top or bottom. For example, with a 7" chord, delete pairs whose Y axis sum less than .0034. I have marked those in the raw .dat above with a * for clarity.
Why divide by the chord? Remember that your file doesn't have a concept of measurement or units, only relative numbers from zero to one. If your drawing is scaled in inches, when you import your DAT file, it will be a 1" chord. I do all of mine in centimeters, and my files import with a chord value of 1cm. If you just delete everything smaller than ".012" in the y axis of the file, you'll then scale your airfoil up and the trailing edge thickness will scale with it.
Why did I have more coordinates deleted at the top of the file than the bottom? The example I used was already reduced in resolution, and for whatever reason it had more coordinates on the top surface than the bottom. Notice that I stopped deleting with a pair of coordinates that had X values around .95xx, or where the thickness (sum of y coordinates) exceeded .0034.
Understand all of that? I know it seems very complicated but it's really not... and you need to consider trailing edge thickness before you go any further. Drawing stuff with a zero thickness trailing edge is fun and pretty, but you can't manufacture it.
Another more clumsy way to do the same is to import the file as a .csv and it will appear as a point cloud in Rhino... Create a polyline or interpolated curve through those points, scale that polyline or curve to the correct chord, then trim off the trailing edge where it is less than .024" or .6cm. A word of warning about this process... you can't use the select all or group select the point cloud to create the polyline or curve. Your trailing edge won't "close" because that function basically plays connect the dots, and it assumes that the dots CLOSEST to each other should be connected. This usually results in wierd square wave patterns near the TE because the top and bottom coordinates are closer to each other at a given chord than the next forward or aft pair. Play with it and you'll see what I mean.
Importing the Airfoil...
Now that you've created your airfoil files... the easy part is next. Click import, select a file type of "csv" and then select space as the delimiting character. Browse to your airfoil file, and viola! You have a point cloud at 0,0,0 that looks like an airfoil. You may have to zoom in because the total chord will only be one unit of measure. If your drawing is set to MM scale, it will be a MM long, etc. Next, you need to create a curve or polyline through those points. Interpolated curves will give you a smoother surface where polylines will give you possibly a bit more accuracy. I do curves. Do an interpolated curve through points and select all of the points in the cloud. It's important that you select the points manually which is a pain, but it ensures that you have all directions throughout the airfoild correct. I always start at the TE bottom and work forward, around the LE and finish at the TE. Select a closed curve so the TE completes. If you don't, you'll have to add a polyline from the top to bottom surface at the TE and ensure the direction is consistent with the other line.
Once you have this smooth curved airfoil, drag the line (without the point cloud) to a "safe place" in your drawing since you'll probably use it over and over. You could save it off as an object as well so you can import it into other drawings later... this is what I do.
Now select the airfoil curve, click transform, scale 2D, then select the point at the very leading edge and type in the scale factor. If you are working in CM and your curve imported as 1cm long, you would type "18" to create an airfoil shape that is 18cm long.
Now you can also set the washout. Select that airfoil curve again, use the rotate function, select the center line of the airfoil from the LE (end snap and ortho help here) and type in the number of degrees you want to rotate it.
Now you have a closed polyline of the correct size and washout for a particular use in the aircraft. Repeat this process for all "breaks" in the plane.
I'll post more on the lofting process later. Is this the type of information that you guys want to see or do you want me to jump right into molding stuff? I prefer to add all the foundational detail.
Congratulation Tom, I do know what you are getting involved with and as it is not your line of work I have to take my hat off!
For the TE problem: your choice of .3 mm for thickness is a very good choice. May I suggest that you “extend” your TE by 10 to 15 mm and “scribe” your mold where the wing should be “trimmed” back? This is very easy to machine (it has to be positive in your mold to create a negative line in your molded product). This way you get a better definition, constant thickness and better control of your layup. It is easier to trim the top and bottom skins as well and will help for the positioning of these skins for the closing of the molds during final assembly/bonding of your wing.
This thread will actually be MUCH more useful with this type of info. Most threads like this are just a group of pics tiled "Here's a picture of my mold". Info on the CNC wing development process of molding DLG wings should include EVERY aspect you can stand to report whether good, bad, or ugly.
Think of yourself as teacher and we are the students, just like school-except we actually care what you're talking about!
I suspect you have a lot more experience with rhino than I have. My issue with the loft function is that it did not follow the exact curve that I wanted for the LE and TE, paticularly when the taper began to increase near the tips.
Can you let me know how you address this issue?
I ended up using catia for the wing surface model.
There are several different types of lofting, in particular "loose" and "tight" as well as a few others. The "tight" has always followed VERY closely to my shapes where "normal" and "loose" do their own thing a bit. If you want it exact to a LE and TE curve, you almost have to do a two rail sweep.
Those are the functions that I'll be covering first, and I should get them posted tonight.
Right now I have to go watch "Bolt" again with my son. After he hits the sack I can post some detail.
As promised .... screen shots.
|Category||Thread||Thread Starter||Forum||Replies||Last Post|
|Discussion||Small DLG scratchbuild project on the side||GreenAce92||Hand Launch||11||Jun 26, 2009 10:31 AM|
|Sold||Unknown DLG molded wing||Dbox||Aircraft - Sailplanes (FS/W)||3||Mar 24, 2009 09:46 PM|
|Micro DLG - Collective project||ClosetRCModeler||Hand Launch||101||Apr 27, 2005 04:18 PM|
|Blastr - DLG molded wing||Graphite||Product Announcements||0||Jan 15, 2004 09:40 AM|