[tommyt]

I am considering building a 3 axis CNC router table for home use to use as a mold making machine. I would be willing to spend up to about $3500. If a table like the one I'll describe is no where near my budget please let me know. Also after reading my entire post what would it cost to have molds cut if I outsourced? Who would I go to?

I build and fly radio controlled 1/4 to 1/3 scale sailplanes and there a couple of 1/3 scale sailplanes that are not on the market that I would like to scratch build. I have been building R/C models since 1980.

My requirements are a table that could route/mill a 120" X 24" X 9" half mold down to .002" out of high density foam, MFD or hardwood or .0005" or less if I decide to eventually go with aluminum stock for the wing molds. I would prefer that the wing molds be female molds.

I know that this is a very large/long table. I am mostly interested in accuracy and not speed. I could care less if it took days or even a week to route a fuselage mold.

Each sailplane would need about 8 molds. L & R fuselage half, left upper and lower wing half, right upper and lower wing half and upper and lower elevator half. Each half molds would be joined after receiving a layup with fiberglass, carbon fiber, kevlar and resin and allowed to cure. Once cured the halves would be separated and a finished fuselage or wing would be pulled from the molds.

I know this would be an ambitious project but I am sure I have at least most of the skills and tools necessary for this project. I am somewhat of a "Jack of all trades master of none". I am very computer savvy having assembled 6 desktops in the past and although I have no formal electronics background I used to build complicated Heathkit electronic projects as a teen with my dad. I also have extensive cabinet making skills and my own shop tools. (table saw, band saw, router table...) and I have a neighbor who works in a regular machine shop that would be willing to do some side work for me.

Regards,
Tom

[jcc3inc]

Sir,

If your .002 and .0005" are absolute accuracy, the $3500 will not do it. If the .002 and .0005" are resolution, that is you can move your cutter in increments of less than .002 or .0005", then you might have a chance. Even so, you will need good components to build such a machine and $3500 might be tight. Milling metal requires a much more rugged machine than doing the foam.

Regards,
Jack C.

[solodex2151]

That is a really long table...As far as commercial options, the only thing even in your price range in the small 4'x4' shopbot. That will cut aluminum and with it's open sides you might be able to make it work.

That 120" length is going to kill you. Are you sure you need that much? Can you build a couple of smaller molds instead of big one?

As far as the accuracy goes, you will pay dearly for it, especially the 1/2 mil you want for the aluminum. Standard machine shop work is generally 1 mil. Also, I don't see the point of all the accuracy requirements given your working materials. MDF and Hardwood will change in size by 1/32" anyway so don't even think you can hold it to those tolerances.

So, suppose we use a polymer for the molds and the expansion problem is gone. Let's start with the 120" axis. That length demands rack and pinion. Belts won't hold the accuracy and leadscrews/ballscrews need to be extraordinary thick (and expensive) for those lengths. A quality linear bearing system is a must as well. Skate bearings won't work for those accuracies. The rest should be handled fairly easily with ballscrews. You will need to invest in quality motors and controllers. Also, spindle runout on routers tends to be about 1/2 mil to a mil, so you will need a professional spindle. Cost start adding up here.

One more question: Are you starting a business with this? If so, you need to think long and hard about buying one. As a business owner, unless you are making the CNC to sell to the world, you need to buy a cnc because you get support, a product that works, and a solution that is ready NOW. Building a machine that you rely on to produce goods that are the majority of your income is a recipe for disaster.

Cheers,

Nate

[1000IPM]

That's a tall order for a $3500.00 budget. I put together a 2' x 4' table w/ 6" of Z travel - constructed from extruded aluminum for about that price range w/ electronics and software to run it. The problem you are going to encounter is the length of the table you require. The only way that I can see this working is by a rack and pinion system. The CNC router we have in the shop is 4' x 10' but cost in the $50,000 class when purchased from CAM Tech in Canada. Your materials in HD foam will cost about $1200.00. If I had a project this big - seeing as you will fiberglassing your shells from CNC forms - I would section the forms to fit on a smaller machine and make them in multiple pieces. Your tolerances that you are looking to achieve are also not a reasonable expectation from MDF or foam. Good luck.

[drmosh]

solodex speaks the truth...

first off, you won't build a highly accurate, 120" router for anywhere close to $3500. the cheapest way to go would be to buy a 5x10 router, but even then it's going to be $12k by the time it gets to your door. i would strongly suggest trying to engineer a solution that could accomodate 4ft sections. you will pay a major price penalty for jumping to 10ft lengths.

secondly, and even more importantly, the point about building a business is dead on. you ignore those words at your own peril. if you're already making these products then you know if a cnc will be money well spent. if it is, then buy a commercial unit (with support!!) that you've seen with YOUR OWN TWO EYES cut the way you need it to. then, finance the cost out past where it pays for itself (plus interest) and there you go - great business investment. just about every manufacturing company makes those decisions every day. but once you start relying on a machine to make you money, you better have a number to call when it goes south.

[harryn]

Your project represents an interesting challenge, buy a mill / router or buy / build molds, and how much to get started.

Interestingly, I actually went through somewhat of a similar experience (with a flashlight hobby). It might seem different, but it is just as crazy and can be just as addictive.

I originally chose to outsource all machining work, as I had no machining tools other than some hand tools. The bottom line, is that while it makes a lot of sense to outsource production to a "real" production shop, what kills you, and them, is prototype costs. Coding in an intricate design takes time, and the shop also needs the part to run fast as their machine / machinist cost real money.

It is also just dang near impossible to design something right the first time, at least for me, and that included having others involved that did some of this work professionally and had 3 D imaging software.

Skipping over your question about whether it can be done for $ 3,500 or not, I will throw in my 2 cents that I wish I had first built (or purchased) a router / mill BEFORE I started throwing money at machine shops to build prototypes and short production runs. I spent way more than $ 3,500 (at least double) on flashlight protos. Sounds crazy doesn't it.

I am not saying that the machine shops were unreasonable, I just underestimated what it really takes to do the project.

You can gain some more information related to mold build costs by posting in the RFQ section (request for quote). If you want to keep your specific plane a secret, pick something in the same complexity range and ask for a quote based on that. There are a number of high quality shops around that monitor the RFQ forum. I would suggest that you try to use someone local if at all possible. There is nothing like the helping hand of experienced machinist and die makers.

BTW, there are some decent used, manual, inexpensive large mills around in the used market (not ebay). It might make sense to try to retrofit a stepper motor / Mach III kit to one which is large enough for your project. That might be a viable approach to obtaining the desired accuracy. I will say that it takes a very patient person to consistently make parts that are less than +/- 0.005 inches, but it is routinely done. ( just not by me) Getting to 1/10th of that takes real knowledge, real pain, and some real world experience.

So, bottom line, if you are serious about being in a business like that, think not just about this particular plane, but the next 10 planes you will build up. If the bite is too large, maybe there is a shop looking for work that will do a joint venture with you. 50 % of something is better than 100% of nothing.

[harryn]

Just curious, have you considered having the planes machined from a foam blank, covering it with a carbon fiber / resin mix, and then just dissolving away the foam ? I know it is not quite the same, but maybe it can work.

[CarveOne]

There is a commercial machinable solid plastic material called RenShape 440 that is available in blocks large enough to make the fuselage and wing moulds being discussed. 6" thick and 10' long is not a problem if you have the ability to buy it. It is not cheap. It can readily be cut to size in a common bandsaw with wood cutting blades. It can readily be epoxied together to make thicker, wider, or longer blocks which means that scrap pieces can be re-used. Machining marks sands off easily with 80 grit sandpaper, and 180 grit will finish it up enough that it will wax out to a shiny finish for epoxy/fiberglass/carbon fiber/Kevlar lay-ups. This material is normally used for small production runs. For long term high production use, machined and polished aluminum is better. Search for Renshape 440 on the web for information

However, getting 0.0005" accuracy will not happen with this solid material without an extraordinary amount of sanding and measuring effort, and it will not maintain that kind of accuracy with normal use.

This material is usually cut on a cnc router such as ShopBot and others. I work for a small company that uses moulds made of this material but we get the machined moulds from another company and do the fiberglass lay-up and assembly work.

I'm a long time R/C modeler who doesn't think that the stated accuracy requirements is reasonable for scale R/C gliders. Is there another reason for such a tight requirement?

CarveOne

[yham]

Tom,

check www.MechMate.com web site. You can download plans for free. This machine is 4'x8' and plans are written so that way that you may change dimensiones of table as you like.

Milosh

[solodex2151]

Tom,

check www.MechMate.com web site. You can download plans for free. This machine is 4'x8' and plans are written so that way that you may change dimensiones of table as you like.

Milosh

I am not sure that the Mech Mates give the desired accuracy. Anyway, I am pretty sure that scaling the Mech Mate to the 10' axis length will probably mean that it is out of his budget range. Worth a look though...

I also agree with CarveOne on the point that the accuracy you're looking for is not appropriate for RC planes, unless you need to build it to mil-specs. If you can give some more insight to why your requirements are in place, we can help you better.

Nate

[CarveOne]

A few more points to consider:

1. Assume that you have two mold halves that meet the stated requirements, lay up a fiberglass/epoxy "skin" in each half, add any internal structure such a bulkheads (formers), join the molds with an acceptable method then, then when fully cured, you pop the fiberglass object out of the molds. Due to shrinkage during curing the object will be smaller than the stated requirements.

2. The wax layers, and the mold release material layer will also change the dimensions of the final object.

3. Even with structure added inside the object, there will be some small amount "oil canning" between the bulkheads.

There are many other factors to account for in order to arrive at an object that meets +/-0.0005" tolerance in all points on the surface of the object.

This is why I bring up the point that the specs seemed unrealistic to me. I don't intend to imply that it is impossible, just that it will be impossibly hard to accomplish with a cnc router table and the materials that are available on the open market. It really doesn't matter what the intended object is, glider or otherwise.

The company that makes the molds for us doesn't outsource jobs and I don't know the costs of making them. I would think that large size and higher complexity makes the costs go up exponentially.

CarveOne