CNC for Molded RC Sailplane Construction - RC Groups
Thread Tools
Sep 19, 2017, 02:19 PM
Callused Thumbs
rafterrc's Avatar
Discussion

CNC for Molded RC Sailplane Construction


Some of you may have seen or even followed this build log: No-Name 2.5m F5J/ALES Sailplane. I started with an attempt to do it with little more than hand tools. Somewhere down the path I concluded that getting an accurate bottom (convex/inside curve) wing surface was going to be difficult if not impossible. The upper (convex) surface came out pretty well.

I began a search for a 3D capable CNC router, and found one at a local Maker Space/middle school shop. It is a large X-Carve from inventables.com. It currently can be had for about $1500. You might ask if I'd rather spend the money on one of the offerings by the usual list and be money ahead? That thought crossed my mind, but I must confess a strong desire to acquire the skills to do it myself.

Over the summer, the X-Carve came home on loan. I was able to cut the molds for what has become known as the Obsession. A number of operational constraints ensued including a request that the waste board NOT be milled flat and parallel to XY. I built fixtures instead. The waste board varied ~3mm in heighth across it's surface. At the end of summer, the X-Carve was returned to school, and limited access for long milling operations. This constraint considerably reduces progress.

I began a search for CNC routing capability that would be in-reach of most modeling budgets. Besides, it makes little sense to build a $5,000+ CNC router that sits idly waiting for the next project. I am a rancher and chemical engineer that expects money to work for me. My partner in crime emailed me a link to an instructables.com article on a candidate. You can find more about it at MPCNC. It is the MPCNC - Mostly Printed CNC. Another thread on this forum has introduced it.

If you have seen some of my posts, then you will know that flexibility ranks high on my list of design criteria. The MPCNC has it in spades. Other candidates fail miserably. One of the neat aspects to the MPCNC is that you build its working dimensions to suit the application. For molded sailplane applications that would mean long and relatively narrow with a fairly small Z capacity (75mm is more than enough).

The MPCNC is NOT a production machine. The $5,000 CNC router might accomplish in 2 hours what the MPCNC will do in 10 or 12. As a hobbyist, time is not of the essence. Alfalfa grows slower. So the remaining criteria relates to accuracy. If you start with a Syner J foil .dat, then you want a Syner J foil reflected in the resulting molded wing. What you will see in the following is building an MPCNC and associated systems like electronics and software to accomplish just that.
Sign up now
to remove ads between posts
Sep 19, 2017, 03:09 PM
Leonard
mac44mag's Avatar
I'm the partner in crime that Jeff referred to in his intro post. We've worked on several projects together over the past couple of years, and have been dabbling in laser cutting and now CNC. I'm lucky enough to have access to a 50W laser cutter at the local library Maker Center, but their CNC machine is simply too small for what we're wanting to do.

Our passion is building unique, high performance RC sailplanes. As such, we need a machine with an absolute minimum of 24"x48" working area to accommodate 3 and 4 meter Open class sailplane wings, as well as the 3-piece 2 meter wings being flown in the emerging 2 meter F3-RES class. It is anticipated that the CNC will be applied to both built-up wings (i.e.: cutting individual parts) and carving both negative and positive molds for laid up components. We need a machine that will cut anything from 1/64" ply through aluminum plate, and it must do so with very high accuracy. Airfoil tolerances are critical at this level.

From Jeff's research, the only machine that we've found that appears to meet our budgetary requirements and offer the flexibility and scalability that we need is the MPCNC, specifically in my case, the Lowrider.

I will be highly interested in reading the comments and suggestions from fellow group members.

Leonard (mac44mag)
Last edited by mac44mag; Sep 19, 2017 at 03:12 PM. Reason: Vocabulary
Sep 19, 2017, 04:42 PM
Registered User
With 4 foot rails you will achieve one of your aims...flexibility..I've built a makerslide extended e-shapeoko machine with 1500mm long axis and it definitely needs anti-flex supports every 500mm or so on the long axis, and double cross beams on the short 750mm axis so as not to deflect excessively, and to help achieve the kind of accuracy I think your after.

Because of the design of the shapeoko derived machines support brackets can be included on the inner face of the makerslide profile with out fouling the carriages.

If I was to start again I think using the Openrails C section extrusion would be my preferred approach, but it doesn't easily adapt to belt drive with in the C section, which would protect the drive belts from the flying crap created when cutting. There are some extensions to the Oxcnc machine which look promising in terms of increasing rigidity and hence accuracy.

This machine looked to me a better cost performance balance while still being within the hobbist budget zone.

In any event good luck!!!

regards JH
Sep 19, 2017, 10:00 PM
Callused Thumbs
rafterrc's Avatar
JH,

I am well aware of the piping flex issue. It appears that 1" OD 304: stainless tubing should accomplish the job along with a center support. The shorter dimension will help the longer one, too. Absolute rigidity brings its own downsides, too, chiefly, weight.

I am of the opinion that even the $5K varieties require some artistry to get them to produce accurate wing molds.
Sep 20, 2017, 05:20 AM
Registered User
I'd be interested to know how you get on

. I've used this bit of software (https://www.autoleveller.co.uk/) on a slightly warped/uneven bed and it worked well . Whether or not it could compensate for rail sag is a question I don't know the answer to. It won't compensate for lateral displacements due to the rails moving though

regards JH
Sep 21, 2017, 04:57 AM
Registered User

Welcome to the club ...


I don't have the low rider variant ( AFAIK the project uses 23.5 mm conduit, in Europe it's more common the 25 mm) but I've build (and sold) a quite big MPCNC (120x120 cm). For my needs it was a bit of an overkill and I settled for mini-MPCNC (based on 16 mm conduit).
Just a few advices:
- the CNC is not extremely rigid, don't expect from a 500 USD CNC the same level of performance as from a 5000 USD CNC
- there are several sources of errors and you can do some tricks to minimize each of them (see details in the following lines)
- try to "square" and "plane" as accurate as possible. "Square" is to have the same diagonal between legs in order to minimize skew (but don't worry too much about it, an error of 1-2 mm is achievable and is barely noticeable for a wing mold). "Plane" is to have the 2 pairs of XY rails as parallel as possible (have the legs at the same height). Again, don't sweat about it, 1-2 mm is not a big deal, plus you can use the autolevel feature in Marlin firmware (mesh autolevel) to make the system "read" (and use) the real geometry of the system.
- try to maintain a certain tension in the timing belts...they should "sound" like a guitar string. Not too tight though... Use some tensioning springs mounted on the belt and/or adjustable belt tensioner system
- as I said, don't worry too much about XY accuracy (if you think about it I don't care if in a mold the whole wing is displaced in one direction with 1-2 mm as long as the Z height is OK). Z accuracy is your real target here, and it has two sources:
- compression deformation on XY rails: easy to fix, just put some constant height supports in the middle of the long axis (or go crazy and do it every 10 inches or so).
- something specific to MPCNC (but not on low-rider I guess) is torsion deformation on cross rails: don't go overboard with a long Z stroke (5-10 cm should be enough) and limit the weight of the spindle (for mini-MPCNC I use a 50 g brushless motor). Still good advice for low-rider: try to make the Z axis mechanism as tight as possible (but not too tight to get bindings). Definitely use a backlash mechanism (spring between 2 nuts)
- take you time and mill it slowly (feed rate and DOC) ... you job is to do things right from the first try and not to loose time with re-runs
Sep 23, 2017, 12:45 AM
I can fix that ...
scottsdalejohn's Avatar
This looks like a great winter project! Since the lowrider (appears) to use the top work surface as the X axis do you think you can keep the wheels/work surface clean enough so as to not impact the accuracy of the Z axis? I'm guessing that the design was intended to "through-cut" or used for projects where the accuracy of the Z axis was not that important, unlike your intended use in mold making. I like the idea of a 3-D printer being able to make the majority of the parts, but wonder if a design along the lines of the machine Jeff was able to borrow for the summer might be better starting point for your intended purpose.

Thanks to both Jeff and Leonard to getting this project going!
John
Sep 23, 2017, 12:56 PM
Callused Thumbs
rafterrc's Avatar
Let me see if I can shed some light on this journey.

We know the starting point is that we have a pile of stock (MDF, Renshape, Corian, DIY Corian, Aluminum, etc.). The intermediate destination for this project is a mold from which a wing will be made (the real destination). Within some tolerances, the mold should be a close facsimile of the .dat files from which it is derived. Therefore, we will need to have some understanding as to whether we have arrived at the destination.

One requirement is that the machine must double as measuring device with a probe.

The Vehicle for the Journey


Alternatives:
  • Experienced operator with a CNC Router. Cost: material+operator+machine time every time you want something milled, but accurate.
  • CNC Router known to have accomplished the job. Cost: many machines in this category are expensive on a modeling budget (I don't care if you've got the money for a $75K Legacy machine. )
  • Buy/build one that has a high probability of getting close. (joescnc.com, for example)
  • Buy/build one that might get the job done with some modifications.

I am aiming at the last one with an all up cost of some fraction of the cost of an off-the-shelf competitive FXJ sailplane. I would like to get close to the cost of CNC Router plus airframe production materials at approximately the current cost of a 4M FXJ molded sailplane. That way one could build the next of any particular design at half the cost an off-the-shelf one.

The system of producing the mold includes CAM (free); the board that interfaces between your computer, gcode files, and the CNC; machine structure; stepper motors; and spindle/router. I have developed some facility with Fusion 360 CAM (free to hobbyists). The board I have selected is the Smoothieboard unless somebody convinces me differently. Rationale for the selection will be addressed in a future post along with links. I plan on a spindle since routers require something like the Super PID to get feeds and speeds right. After that, the list of credible potential CNC machines is short.

The MPCNC fits within the budget. JH's flex concerns are valid. For a build long enough to do most fuselages and wing sections of 400mm X 2000mm, the OX calculator estimates ~$1700 for the OX. With 1" 304L stainless rails (minimize flex), I'm looking at ~$750 all up without support table (likely MDF torsion box). I figure to go full length of 2254mm. If flex is too great I can always shorten them. 1" 304L tubing is $4.75/ft. 20mmx80mm extruded aluminum if higher.

The outstanding question which the probe will answer: Will the MPCNC get me there? Not even the designer knows for sure.
Sep 23, 2017, 03:31 PM
Registered User
Hi Rafterrc,

I ill follow your thread with a lot of interest.

I must give credit to Ender67 for opening my eyes on the potential of MPCNC. Following this path and the one on various Hotwire CNC design, I was striked that all those machines, including 3D printers could have, at hobbyist level, the same DNA: Threaded rods, stepper motors, ball bearings, printed parts.

I now have a 3D printer at home. Not sure I'll build a MPCNC, but always interesting to learn :-)

By the way, I believe that finding the right stock material is not easy for one-off builds taken into account the cost parameter. MDF looks less than perfect.

JMF
Sep 23, 2017, 04:02 PM
Registered User
Ward Hagaman's Avatar
I'll throw in my two cent's worth. I love CNC, I did it as a career, and I have a CNC mill, router and hot wire in my garage.

If you have a limited budget and your main goal is to mold FXJ sailplanes, and not CNC as a hobby in itself then I suggest that you build a CNC hot wire and splash your wings and tails from bagged positives. You can get more accurate end results than a cheap CNC router could ever get you since it will be a non contact system with no cutting force. Yes, you will need to make epoxy molds, but an epoxy mold is good for dozens of pulls.

Ward
Sep 23, 2017, 04:46 PM
Callused Thumbs
rafterrc's Avatar
Quote:
Originally Posted by JMF11
Hi Rafterrc,

I ill follow your thread with a lot of interest.

I must give credit to Ender67 for opening my eyes on the potential of MPCNC. Following this path and the one on various Hotwire CNC design, I was striked that all those machines, including 3D printers could have, at hobbyist level, the same DNA: Threaded rods, stepper motors, ball bearings, printed parts.

I now have a 3D printer at home. Not sure I'll build a MPCNC, but always interesting to learn :-)

By the way, I believe that finding the right stock material is not easy for one-off builds taken into account the cost parameter. MDF looks less than perfect.

JMF
With complete honesty, I will say that I hate MDF ... except for a fairly narrow range of applications. Yes, MDF is less than perfect to say the least in this application. But knowing it's behavior allows for methods to counteract them. It is somewhat stable until you take off one side; as you do in the mold making process. I suspect that HDF could be an alternative. One maker of MDF has a product for 3D application. But nobody within 250 miles sells either one.

I plan to use a coupon approach for setting up the MPCNC after making a probed determination that the work surface is flat and XY axes parallel to the surface or seeing if auto-leveling works. The coupon would be a suitably long section of an airfoil machined near y=0, at y=1000mm, and in between Y axis supports. When I get that far, you will see pics here. Then I'd use the coupon to test different CAM finishing approaches.

Jeff
Sep 23, 2017, 05:03 PM
Callused Thumbs
rafterrc's Avatar
Quote:
Originally Posted by Ward Hagaman
I'll throw in my two cent's worth. I love CNC, I did it as a career, and I have a CNC mill, router and hot wire in my garage.

If you have a limited budget and your main goal is to mold FXJ sailplanes, and not CNC as a hobby in itself then I suggest that you build a CNC hot wire and splash your wings and tails from bagged positives. You can get more accurate end results than a cheap CNC router could ever get you since it will be a non contact system with no cutting force. Yes, you will need to make epoxy molds, but an epoxy mold is good for dozens of pulls.

Ward
Tell me more about the process and the foam materials used. Unless the foam is something like Foamular 250, then it is out of the question given my very rural location.

I am guessing that you would make a plug from the foam, cover it with say fiberglass so that it could be sanded to say 8000 grit, for example, then make the female mold with perhaps DIY corian. The tip section of a SynerJ 30 with short chord would be next to impossible, <== a guess.

Is a spline smoothed LE and TE achievable?
Sep 23, 2017, 05:53 PM
Registered User
Just some crazy idea I got in the last 2 weeks. I bought some sheets of perspex (plexiglass/ PMMA/ acrylic glass) in order to build the cases for my 3D resin printers and at the end I had to find a good use to 13 sheets of 100 x 64 x 0.6 cm clear plexiglass. Of course I did a quick calculation and it was enough for a 3.5- 4 m SynerJ wing molds.
First I considered laminating the sheets, milling, polishing... the whole process. Too much work, dubious results. Then I got it: why not laser cut it?
Small change to the program I created for drawing ribs for balsa cutting: compute ribs 6 mm apart, draw each rib in a 32x4 cm box, in fact draw them side-by-side for the left and right wing, cut the line in front of LE and in the back of TE (with some round or square trenches for resin), cut also 4 circles / holes of 8 or 10 mm diameter for positioning (also 2 holes in the rib), put these 3 parts (upper, rib and lower) on some rods, light sanding and polishing with acetone and voila ... first quality molds (plus the wing profile to be used as a plug). Not extremely cheap (6 mm acrylic is not a job for my 2.5w laser and it costs 1.2-1.5 USD / m to cut 6 mm acrylic) ... but definitely cheaper & faster & better results compared to MDF or Corian.
It doesn't have to be perspex, you can use 6 or 8 mm MDF also.
No need to worry too much about how accurate it is, you can count on a 0.25 mm up to 0.3 mm kerf (or whatever it is...it doesn't matter as long as you know the value), compensate it in drawing the rib (piece of cake, I do this stuff when I compensate for balsa or composite D-box) and the surface of the cut is quite smooth, I can post some pictures taken with the USB microscope. Definitely it saves a lot of time in finishing the mold, just a very light sanding and some smoothing with acetone (for perspex) or light epoxy (for MDF).
BTW, for a better variant of low rider search thingiverse for CNC Artisan3.
Anybody willing to try this new approach to molds / plugs? Just give me a PM with details...

Later edit: just a mockup of a set of lasercut molds... the first from a set of 300+. Cost wise (at least for me) it's better to cut the 100x64 cm sheets in smaller 20x30 cm pieces and to buy a 40W CO2 laser to cut these pieces at home ...this way it's cheaper and you have also the laser (right now is around 350 EUR on ebay).

Later later edit: BTW, one of the guys in the Romanian rc model forum wanted to build a 1 m scaled-down Synergy2 wing and asked for my help for milling the foam. I explained to him that it's quite hard and at the time I suggested a different approach: use my program for balsa ribs to laser cut ribs in 6 mm depron sheets (or XPS...used for floor insulation) as 6 mm apart ribs and position them on carbon rods (this is a job I can do with my 2 w and 2.5 w lasers...). (He abandoned the idea and bought a Whippit). OK, not pesrpex, but 6 mm foam floor insulation is cheap and we can get plugs (and from them DIY corian molds) and cores (with internal carbon tube spars). Anybody interested?
Last edited by ender67; Sep 23, 2017 at 06:36 PM. Reason: adding some image
Sep 24, 2017, 05:47 AM
Callused Thumbs
rafterrc's Avatar
Quote:
Originally Posted by ender67
Anybody interested?
Brilliant! A month or so ago, I thought about something like this, but discarded it. Pictures are definitely in order.

Do you think that the mold is strong enough for compression molding?

I assume this was the wing starting with SynerJ 100 at the root. How did you incorporate the polyhedral breaks?
Sep 24, 2017, 06:34 AM
Ok that's high enough
FabFlight's Avatar
This is how I fabricated my last joiner boxes.


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Question CNC tolerances for building molds Tim Harbour Hand Launch 18 Jul 24, 2017 06:35 PM
Mini-HowTo CNC Solid Molded Construction- Part Three - Cutting the top surface of the foam tom43004 Composites Fabrication 18 Oct 05, 2012 12:29 PM
Mini-HowTo CNC Solid Molded Construction- Part One - Creating the mold tom43004 Composites Fabrication 27 Jun 17, 2012 09:59 AM
Mini-HowTo CNC Solid Molded Construction- Part Two - Creating the vacuum hold down tool tom43004 Composites Fabrication 8 Apr 23, 2012 11:21 AM
Question Best mold construction method for Painting in the mold ScorpionRacing Composites Fabrication 3 Jan 14, 2011 03:28 PM