Advice on a DIY 30' X 3' X 3' CNC

[treegezer]

Hello everyone, first time posting. I am with a student group at ASU and we are looking to build a machine that could cut foam plugs for canoe hulls that would later be used as forms to make concrete canoes. I have been lurking for a while now and have come up with a basic game plan.
Build a gantry that would roll along the floor on an angle iron track pulled along by a steel cable pulley system. The gantry would then have a y and z axis carrying a router.
I am not familiar at all with CNC machines. So I am looking at using the 200oz/in steppers and controller kit from hobbycnc as hopefully that will reduce the complexity of the project. I plan on building the gantry out of ¼” plate and ½” box tube and using drawer slides for the movment.

Any other suggestions? Pointers? What software?

I know it wont have much accuracy but the way we do it now ( cut each section individually approx 200 sections with a hotwire and laminate them together) isn’t any better. I would be extremely happy with .25in accuracy. I plan on building this over the next three months so I am sure I will have more questions.
Thanks in advance,
Stuart

[mxtras]

Software? = Mach 3 - without a doubt. It's free for a fully working version, but is limited to 1000 lines of code and 25Khz step frequency in 'free' mode - $160 fixes those limitations if they are a problem. For many, these limits are not an issue.

I would look at the weight of the gantry and the rolling resistance of your guiding and possibly reconsider your 200OzIn motor selection - that seems a bit weak, but I could be off. I guess that depends on your accel and speed.

Scott

[wizard]

Originally Posted by treegezer Hello everyone, first time posting. I am with a student group at ASU and we are looking to build a machine that could cut foam plugs for canoe hulls that would later be used as forms to make concrete canoes.

There is just something about the concept of floating concrete vessels that bothers me. I know they work but it just doesn't seem right.

I have been lurking for a while now and have come up with a basic game plan. Build a gantry that would roll along the floor on an angle iron track pulled along by a steel cable pulley system. The gantry would then have a y and z axis carrying a router.

Not to be negative here but how do you expect the router to handle the profile of a canoe with just three asxis? The problem as I see it is the length of the tool and the diameter of the router, combined I don't think you will have the clearances to go deep enough. Granted I'm thinking about an conventional canoe and the 3D shapes they have.

I am not familiar at all with CNC machines. So I am looking at using the 200oz/in steppers and controller kit from hobbycnc as hopefully that will reduce the complexity of the project. I plan on building the gantry out of ¼” plate and ½” box tube and using drawer slides for the movment.

Somebody will have to do the math on this. That is figure out the wieght of the gantry and the required torgue to get the work done in a reasonable time. It is foolish to assume you can get good results by randomly selectiong components.

A packaged system would allow you to get on line quicker there is no doubt there. Just make sure it will do the job.

Any other suggestions? Pointers? What software?

Yes EMC. It is free which always helps and if you get the BDI install should be trouble free. Being open source there are no limitaitons to the software what so ever. Further due to continous development it has come a very long way.

I know it wont have much accuracy but the way we do it now ( cut each section individually approx 200 sections with a hotwire and laminate them together) isn’t any better. I would be extremely happy with .25in accuracy. I plan on building this over the next three months so I am sure I will have more questions. Thanks in advance, Stuart

Well you are doing concrete over foam so the accuracy thing is understood. For the Y & Z you shuld be able to get by with simple leadscrew drives. The cable drive I don't know about, if it was me I'd get a long length of light timing belt or chain.

Have you considered just getting smarter about your section laminations? I'm thinking why not CNC the sections. Yeah you still end up with alot of sections but your accuracy should be much better and the sections should blend to gether better. This would require a much smaller machine which you should be able to get on line quicker. Further if the sections aren't to thick you should be able to manage the 3D profiling with off the shelf hardware. I don't know how thick the foam is that you are using but if it is 2 to 3 inches this ought to be very doable low cost.

Personally I see the long machine taking you way to long to build and get working. Do a small router and you have a machine operating much faster. Another plus with a small machine would be the abiltiy ot keep the swarf contianed.

Thanks'
Dave

[Art Ransom]

Whers are you located?

[treegezer]

Remember we are making a plug not a core. The foam is a positive casting of the hull. So as long as the hull has no tangent line steeper then the angle between the bit and the collar of the router we should be alright. Am I correct in this assumption? If not could we some how tell the software that the 4th axis is controlling the pitch of the router and not a typical lathe like situation?

With some rough calculations a the 200 oz/in motors would require a 10:1 gear reduction to drive a 6 in spool to move a 50lbs gantry.
EMC looks nice since we run an ubuntu system in the shop.

We are using a 2in foam. However cutting the sections is only half the battle when laminating to produce the plugs. Indexing them and gluing them together is a long process. I cant find the actual numbers at the moment but we have done some calculations on laminating a large block using only full ½, and ¼ sheets of foam vs the cross sections we do now and we where able to save 3 or 4 sheets of foam due to waste along with less foam adhesive which gets expensive quick.

To clarify our design choices here is a little background. We compete in a southwest regional conference each year for the American Society of Civil Engineers where cost in material and labor is around 30% of the total scoring rubric. Currently all the teams produce and form hulls the same way. It takes about a day and a half for 3 guys to build a plug from raw sheets this includes cutting two sheets at a time as the hull is symmetrical end to end. So, the advantage of having to report little labor cost is large not to mention the reduction of actual costs.

As far as swarf goes when we hand build hulls we still have to run a planer. To contain the particles we enclose the area in plastic. Place a large fan on one end blowing into the enclosure and another on the other sucking it into a trash can and attach the exhaust of the planer to a shopvac. It works very well and could work with this setup too I think.

If it was for personal or multi purpose use I agree I would cut the sections on smaller table machine and laminate them. However unless there are other issues I believe this may be the optimal solution.

Originally Posted by Art Ransom Whers are you located?

I am in Tempe, AZ
More questions:
Are the hobby cnc drivers adequate?
What kind of bit to cut Expanded Polystyrene Insulation?
Anyone have a better source for large foam sheets or blocks? ( Buying insulation from Home Depot currently)

Thanks for the comment and keep them coming. I am learning a lot and may be tempted to build a small table for personal use.

[jeffs555]

So as long as the hull has no tangent line steeper then the angle between the bit and the collar of the router we should be alright.

Every canoe I have seen is close to vertical near the gunnels, so a fourth axis to tilt the router would probably be necessary.

[jeffs555]

I am not familiar with cutting foam, so not sure if this would work, but what about using a hot wire cutter and using it like a draw knife. The long axis could drag it along the side of the plug, and the other axes could tilt it and move it to cut the profile. It would take many passes, but the waste cut off would be solid sections, not the fine dust that a router would create.

Another option would be to cut out sections like you are doing now, but with a cnc machine they would be much more accurate. You could cut out locating holes, and stack the sections on a couple of 2x4's to make laminating easier and more accurate.

[andy_ck87028]

Originally Posted by jeffs555 Every canoe I have seen is close to vertical near the gunnels, so a fourth axis to tilt the router would probably be necessary.

Previously on this forum, I have seen suggestions to solve this problem simply by tilting the table. Assumedly the CAD software has to be able to generate G code based on a model that has been spun by the same number of degrees as the table tilt.

That would still be multi-pass but an easy solution none the less.

Andy

[pen25]

4 axis would be sweet to do think others have used 4 axis to build thiers. also rather then cable check out the plasma section and look what someone did with 40# chain. i also like the idea of hot wire to cut the plug. if it were a neg mold then it would have to be routed.

[wizard]

Originally Posted by treegezer Remember we are making a plug not a core. The foam is a positive casting of the hull.

To be honest you lost me a bit here. I was imagining a core around which the concrete was cast. You seem to be describing a molding technique wheere the concrete is cast in the mold????? I was thinking more along the lines of making a conventional canoe.

In the end it doesn't matter as the problem would be the same.

So as long as the hull has no tangent line steeper then the angle between the bit and the collar of the router we should be alright. Am I correct in this assumption?

Exactly! Granted I can't say I have a huge amount of back goround in canoe making but every one that I've every seen has some complex curves transitioning into vertical surfaces. So how would a verrtically mounted router handle that? With a real CNC spindlle and an extension you might cover the surfaces but even here a 3 axis machine is not really the best choice.

If not could we some how tell the software that the 4th axis is controlling the pitch of the router and not a typical lathe like situation?

This is certianly one way to do it. The problem is software. By that I mean finding software (low cost or free) that can handle the hardware configuration. EMC is flexible in this regard but I don't know if there is a kinematic solution available for that machine configuration. I'd suggest getting on involved with EMC and some of the low cost CNC software forums to see what the feedback is.

There is I believe a solution available for the fourth axis being rotary. That is nice but then that would be a very long axis and more machine complexity. The only good thing is that foam is nothing in the way ow weight or machine reaction forces.

With some rough calculations a the 200 oz/in motors would require a 10:1 gear reduction to drive a 6 in spool to move a 50lbs gantry. EMC looks nice since we run an ubuntu system in the shop.

Yep there is even a page floating about the internet with scrips and stuff to do the install. I'm not a fan of ubuntu so can't comment on it. There is a fair amount of interest in ubuntu and EMC so that is good.

Please understand that there are two versions of (maybe three) of EMC. One called EMC1 or the original stable version, one called EMC2 which is fairly along in development, and the BDI install which is EMC1 that is maintianed as a seperate project.

We are using a 2in foam. However cutting the sections is only half the battle when laminating to produce the plugs. Indexing them and gluing them together is a long process. I cant find the actual numbers at the moment but we have done some calculations on laminating a large block using only full ½, and ¼ sheets of foam vs the cross sections we do now and we where able to save 3 or 4 sheets of foam due to waste along with less foam adhesive which gets expensive quick.

I understand all of the above but do realize that a bit of design effort could help with putting all the pices together. You could route keying slots and sequence numbers into the panels for example.

Even better if you have a machine that can handle full sheets of foam you could turn the panels out length wise. This should drastically reduce the problem with mate up of the parts. You still end up with a modest router compared to doing the canoe in one shot.

Taking the above idea a bit farther you could make a rotary fourth axis economically on this size of machine and do 3D sections of the canoe lengthwise. An 8 foot axis is far more doable than a 30 foot one.

To clarify our design choices here is a little background. We compete in a southwest regional conference each year for the American Society of Civil Engineers where cost in material and labor is around 30% of the total scoring rubric. Currently all the teams produce and form hulls the same way. It takes about a day and a half for 3 guys to build a plug from raw sheets this includes cutting two sheets at a time as the hull is symmetrical end to end. So, the advantage of having to report little labor cost is large not to mention the reduction of actual costs. As far as swarf goes when we hand build hulls we still have to run a planer. To contain the particles we enclose the area in plastic. Place a large fan on one end blowing into the enclosure and another on the other sucking it into a trash can and attach the exhaust of the planer to a shopvac. It works very well and could work with this setup too I think. If it was for personal or multi purpose use I agree I would cut the sections on smaller table machine and laminate them. However unless there are other issues I believe this may be the optimal solution.

Interesting competition! The idea that the canoes are symmetrical leads me back to thinking about a smaller machine to do half the canoe.

I am in Tempe, AZ More questions: Are the hobby cnc drivers adequate?

I don't realy know but with a 10 to 1 reduction I'd be concerned aobut the top speed you could realize out of your servo/stepper. If you are being scored on saving labor the machine will need to operate reasonably fast.

Look up Gecko drives as these are about the best you are going to get for stepper control in a reasonable price range. They also have servo controls.

The problem is that you will want to move reasonably fast through the foam to prevent melting and burning. So you may need a larger drive to keep you speeds up.

What kind of bit to cut Expanded Polystyrene Insulation?

I don't know for sure which would be best but a couple fo things to consider. One is that HSS might be better than cabide if it is extremely sharp. Two you don't want to have the helix on the mill pulling the foam up. Look into Onsurud, hope I spelt that right.

Anyone have a better source for large foam sheets or blocks? ( Buying insulation from Home Depot currently)

It always pays to look into the big yelow catalog - McMaster-Carr and then the big blue catalog - MSC. You should have these catalogs on hand when attempting this porject anyways, though I don't see them as big foam suppliers. It may pay to check the yellow pages for local plastics and contruction suppliers and of course boating suppliers. Try looking up this stuff on the net.

Thanks for the comment and keep them coming. I am learning a lot and may be tempted to build a small table for personal use.

Addictive isn't it!

[pen25]

the 4th axis could be scalable. so instead of only doing 3foot plugs you could scale to the size of the boat. lazer guides are cheap now days and would allow the scaling to be sone pretty acurately. that would allow you to use hotwire outside. if you look at things like a cnc paint gantry. since its foam you wouldnt need to exactly have to have something heavy duty and would allow you to mount a router as well. i would use a trim router as its lighter..

[vacpress]

there is an inexpensive(free?) way to generate 4 and 5 axis toolpaths... it is a lite version of 3D Studio Max and some add on.... I cannot remember the names, sorry. I do know there is a thread or a few here about it, and searching google for 'rainea cnc' may find someone who uses it, and i think he may be the wreiter of the plug in also...