I am looking at the MPCNC for the subject application. I would like to run both wing molds and accompanying fuselage molds in one pass without moving the stock. Working area needs to be on the order of 300-400mm x 2000mm. The critical dimension relates to airfoil reproduction in the finished mold (Z).
In the past, I have used parallel finishing steps to accomplish <.2mm across 800mm wing section lengths.
Can I reasonably expect to hold tolerances like that with a current version of MPCNC? If so, then what rails should I use along with the DW 660? I suppose that I could “cut” the wings into sections to achieve a shorter max length of the Y, if it is not capable at the 2000mm working length…
Yes, I am equipped to build a support table to accomplish this.
I think if you want to keep it real nice you might want to do stainless rails and build a long rectangle with a set of supports on the long side. The short cross member will support the whole machine on the long rails so the one long cross member shouldn’t be an issue.
I can’t honestly say what tolerances to expect but <.2mm across 800mm is up there depending on your base material that is really hard to hold on most any machine.
I have been getting up the CNC learning curve with an X-Carve at a local MakerSpace, and using MDF. The next step in mold quality is usually something like Corian, but in general, the advantage is usually more parts from one mold. On the other hand, removing a big part of one face of MDF brings huge movement because of stress relief. Corian is more stable.
I did pick the short dimension expecting to get a payoff in support of the long. I suspected that 304 would be the answer.
Is there a pic of the supports? Is an .stl available on the thingiverse page?
I agree that achieving those kinds of tolerances is difficult with any machine, but I can use the money they cost for other things. Convex mold surfaces are difficult all by themselves.
Kind of a different issue, has anybody used a Super PID controller for the DW 660? Getting the right feeds and speeds would go a long way to dealing with the machine harmonics.
There are a few supports on TV, anything will work really you only need one per side and it isn’t doing too much.
I was asking the superpid guys last night for a bulk buy but it looks to be out of the question, they did tell me next month is a 15% off deal. I don’t know of anyone that actually has one but I do think it can solve a lot of issues. The machine is pretty good in my opinion so I have been exploring all other low hanging fruit before committing to another full re-design.
I have no idea, I am more worried about “spindle” run out on the dewalt (or any inexpensive tool) and all the other things that can easily be .2mm off.
Accuracy and precision has been coming up a lot lately. When I built this that wasn’t exactly my goal. I needed a CNC I could afford. Then I have been making improvements ever since. Now we have software that is catching up to us and all the sudden accuracy and precision is being questioned. I can’t honestly say I have ever or would actually know how to give an accurate number as an expectation. Way to many variables, I mean most people print there own parts on a cheap import knock off printer that has never even been checked for a square axis and they mostly seem to work. I just know most of us without really putting a ton of effort into it get within .2mm on our regular cuts in most any material.
Are you needed no more surface variance that .2mm or no more than .2mm off over the length of 800mm? You are very far into the realm of most people can’t actually measure that, even hard on a constantly variable surface like a wing. I would think you would actually need a cnc probe to actually test that right? I do not own the tools to make either of those measurements so I can promise anything I have not tested.
Accuracy perspective: Years ago before current modern sailplane airfoils and computer modeling of wings, a group of people were wind tunnel testing an airfoil. Somebody stuck a piece of Scotch tape on the upper surface, and drag dropped 30%.
Practical perspective: The hinge line for flaps and ailerons must be straight. I stuck a straight-edge (one of those fancy ones) on the hinge line, and tried to stick a feeler gauge under it and no go.
If I am pushing the envelope with the MPCNC, then I am probably pushing the envelope with manufacturers of internationally competitive almost ready to fly sailplanes. I know of nobody that sells them who guarantees or even lists any degree of conformance with the airfoils they supposedly used. The real “accuracy” test is whether or not it wins contests. In all likelihood, the airfoils used are those designed with a computer modeling program, scaled to length of the chord, and then machined with a CNC router (all sorts used but generally higher end). CAn you say expensive? What sells it is the designer of the airfoils (no guarantee the buyer actually got that airfoil series), the weight, and any track record of contest wins. Or for some buyers, so and so uses it. It’s just like few golfers benefit from a $1000 driver.
All molds get post-sanded. The real desire is after coating with auto paint basecoat and clear coat, does the resulting wing exhibit a mirror finish? It gives an illusion of low drag. Never mind some will put turbulators (fancy Scotch tape) on the wing.
It would be cool to win a big contest with a plane made from a CNC router that most any modeler could buy. I suppose that’s my real objective.
Hopefully, you can now see why I am pushing. Apologies in advance.
I am not saying it isn’t possible, I do think it is. I just don’t want to make promises that are out of my control. I can’t even make a test cut to say if I am getting anything near that as I don’t have a granite block to measurements off of. If you plan on touching it with sandpaper after milling, I see no reason that this will not do what you want, a good clear coat is probably more that 0.2mm thick anyway.
I was more trying to convey you are in uncharted territory and those are some big boy numbers. I recently had someone ask for 0.002" (o.05mm) tolerances like that was a usual expectation, to that I answered with a swift, “no way”.
All kinds of people are cutting slots and drawing squares and comparing things with a ruler or cheap calipers, what you asked for is beyond that, but most pretty easily get with in 0.2mm, so with some decent CAM you will be fine. But be prepared to make a lot of test cuts and spend a lot of time truly dialing things in to get into the better tolerances. There is a lot to learn on the CAM side of things.
It appears as many want information about tolerances and capability. Sad thing is is that yes, there are far too many variables. Build size, tightness of screws, the printer that printed them, it goes on forever! I remember me and dui were going to face off in a milling contest of preset gcode on a 3040 chinese cnc and my MPCNC. Never know what happened to that, guess time got in the way. If I had another machine to test, then I would probably do a write up on my findings. I got a lot of MPCNC performance related thoughts in my head all the time, considering I use mine for at least 2 hours a day.
My school’s robotics club is apparently getting a cnc router of some type, name not given. Can’t wait to get my hands on it. Nervous though cause nobody is gonna have a clue on how to use it and we all know how bad things can get lol. If it is anything sub $1500 I think it is worth my time doing a short comparison.
I have spent more than a little time studying and navigating (metaphors help) complexity. Complex circumstances exist when the outcome of a cause cannot be predicted. Simple exists when anybody can do it. Complicated circumstances are when experts help. In complex circumstances, experts do not help and probably hinder progress. “Too many variables” sounds like complex circumstances are present.
Ryan, thanks for the honesty. It’s all too rare today. I will be a customer.
One strategy for navigating complexity is the safe/cheap to fail experiment. One cheap to fail experiment is a standard coupon. For my application that would be a short section (50mm) of airfoil milled with the same gcode near the edge of the work volume and then in the middle. Conduit and 304L at different diameters.
Then repeat for various lengths of Y rail. Start with conduit. This same coupon approach, then gets applied for different bits, different spindles, different builders, different setups, different CAM, etc. In fact, get the whole community to run one coupon of your choice - protein molecule rotation style. Just have them/us run it, “measure” it, record specific details of their machine, and send in the results.
That is kind of the problem, I can’t give a number. I don’t like to even make videos anymore. If I do or say something and someone prints there own parts and buys all there own stuff uses a dremel and a drywall bit and can’t match my settings I get called a liar. So it is best if I just don’t give hard numbers. The community has specs all over the place. Some people are getting perfect measurements, some aren’t. I would love to do some sort of standard test object so people can check the machine, could be fun.
I think with stainless rails and some time spent getting the CAM right you can do amazing things.
Kevin is my new favorite example. He got frustrated at first but listened and asked for help and tried things and turns out, I do believe, he has the fastest machine to date. With specs he is happy with and chewing through aluminum now. Correct me if I’m wrong Kevin but you are getting within 0.2mm on aluminum or better on some fairly large parts?
Convex mold surfaces are difficult all by themselves.