[Student_Eng]

Hi.

I am an Engineering student. I need to build a large CNC metal drilling-machine, 6m (240”) long and 2m (80”) wide and high. The machine will have only small forces because it does not need to mill, it only has to be able to drill small locating holes. The accuracy has to be within 0.25mm (0.001”) . We have a very small budget (2500$ US I’m from another country so don’t know how cost compare) so linear guides will be too expletive and need to use a DIY guide method. We thought about using home made V rollers but I’m not sure if the accuracy will be achievable. Anny ideas or other threads or literature that I can check out will be appreciated.

Because of the size the surface will not be machined where the guides need to be fixed. Any advice on minimizing the possibility of miss alignment will also be appreciated.

The machine will only have rails on the one side where it will be driven by a servo motor with encoder. The long axis motion will be done with rack and pinion and the other two axis with screw and nut. The frame will be made from square tubing. Does the over all design sound feasible? This is my first time working with CNC so any advice will be appreciated.

[kdoney]

I would suggest a cantilever system which saves you a lot of money on the base and uses 1 linear guide (required for .001) per axis. Even small linear guides that are well mounted will be more precise than any diy guide system. If not, think about going vertical. Gravity is usually in a straight line. I think you must rethink your linear guide decision. Holding .001 is harder than it sounds. Justify part of the cost as axes stability mass. Remember that a separate workpiece table only has to be normally stable in the Z axis but the tool platform has to be extremely stable in x and y. The only way you are going to be able to stay within .001 with your budget is a slow moving Kevlar belt driven motion system.

[kdoney]

Oh, and one other thing. Looking for this tolerance, you will have to map the actual final movement to the expected movement to the controller. This will allow the controller to adjust it's commands to move the machine to the actual requested location. Borrow a glass scale, like a good DRO, gather the data on precision of each move on each axis throughout their movement. Then tediously input this data into a screw mapping file. Mach3 has this,.

[Student_Eng]

Thanks.

We plan on using a cantilever system. How will one attach the linear guides accurately if the surface of the base can not be machined? I could not find linear guides long enough for this application, is it a problem to join shorter guides?

Will the belt drive not experience whip over the long distance? Or is this not a big problem.

[LGSM3]

I would level my base as best as possible then build a shallow form at the perimeter, pour epoxy to get as flat as possible where your bearings will be.

Other than that, I might contact a company like Bell Everman, relay your goal and hope that they might have some spare servo belt parts lying around that they might help out an ambitious student with. Never hurts to ask


Sent from my iPhone using Tapatalk

[kdoney]

If, as you say, "The machine will have only small forces", then the base of the guide will support the working force and the only requirement will be to mount the guide straight and level. If the forces are small enough, or the guides large enough, only the guides will be needed. The two guides should be properly connected at the fulcrum point and the least precision declared for either of the guides will be the expected precision of the machine.

A Kevlar belt, well attached and supported throughout its length, at what I said was a "slow moving" pace, should not "whip" and should provide the precision required. I will say again here that for the money, you have to map the deviation of the motion machine. Kevlar should repeat these findings long enough to prove the system and be of some service.

I see no other motive force working. The cost of the guides should take almost all of your budget and every other means of motion are just too expensive.

I have to say that I have never worked on a machine with these dimensions. I would think that the long side (6 m) should be the one supported by the base. 3 meter guides are available from IGUS and a carefully engineered hand-off procedure could be developed to pass the other axis from one section of the guide to the other.

[leMaitre777]

Hi Guys im with Student_Eng on the same project. Comments on this design are wellcome... bring it to me true.
Rack and pinion m=2,stepper motors, aluminium extrusions.
Im busy designing the long axis with a timming belt aswell.

[leMaitre777]

I have 12meters of I beam soo using them will cost me nothing. I tried to create a setup in which you first align the Beam with the 4 holes at the bottom base plate and then the extrusion by means of the cut groove which allows for alignment.

[vegipete]

Have you tested how stiff that extrusion+rail is over a 1.5m span? I suspect you'll find it deflects far too much to be useful.

[leMaitre777]

Hi yeah I also suspected that the deflection will be too much and yes it was ALOT (17mm , 0.67in resultant) even with an extrusion double the size on the previous figures I posted. (That was 45x90 used a 45x180 in the simulation) .
I then tried using two 45x90 extrusions, got an deflection of 0.28mm, 0.00113 inwards as displayed in the attached figure noting the definition of the axis. This is on the edge and definitely not wanted. (Used a load of 300N (Safe value for the z-axis assembly) on the edge with an uniform load accross the span of 280N for the weight of the members.

I’m thinking of maybe installing a steel member in between the extrusions for the added effect of the steels resistance to deflection. (A upright flat bar between extrusions, or a C-Channel as it serve a cable track as well) I’m just concerned about how I will mount a steel member to the extrusion without influencing the extrusions natural straight form.

[vegipete]

I was thinking of the stiffness of the long horizontal rail between the vertical I-beam posts.

Could you cut the I-beam into 2 6m lengths to support the horizontal rails? It might take a lot of fiddling to get the rails straight but it could be done with screws and shims.

I think also that you will need to triangulate to cantilevered beam. Try stack two extrusions with rails top and bottom and a C-shaped carriage running back and forth. Then the other side can hold another extrusion or whatever diagonally down to the lower long axis rail.

[hub]

You could make up/down stiffer like this:



And sideways is much more flimsy... Maybe something like this as an idea:

[Student_Eng]

Another concept we are considering

Attachment 281726Attachment 281728Attachment 281732

[victorofga]

you have to forget all alu extrusion.. they will eat budget up itself..

you can use steel squaretubes,, and some bearings are rolling on..

I think simplest would be looking for a sponsor..
if it were a later used equipment, someone will pay a lot you.. if it were a fictionally made something, then look for another project what can be simpler realized..


the size you want to build , no matter how little force, a 2 meter movement with a bridge cost less than a cantilever..

the 80 in height as you described, require at least 25-30 inches height of the bridge, that can support 80 in movement..
also preventing flexing the bridge need a 30 in length ...

unless you find everything at a recycling place free, I don't think you can put together this..for 6 meter workarea you need at least 6.5 meter of track..

you might try to imagine a bridge crane, and mounted onto the ""Z""...


your design you showing here.. I don't think will work..
you try to put the Z axis on a 80 in long arm end, and it will hang down 80 in?

probably a 2 year old infant can move the end about 2-3 in... at least.. versus you need 0.25 mm ....
for drives, your only solution is something like bycyclechain....

[bhurts]

Is this a project you chose to undertake or one assigned by the teacher? I hate being critical or saying that something is unachievable but in this case it is. If you had a tolerance of 1mm instead I would still say you could not do this for your budget. I'm not even sure you could build a structure that size that the materials won't move and contract way more then your tolerances allow. This is just being realistic. Your stated tolerance in the first post is 0.25mm is this a misprint. Did you mean .025mm which is .001 inch. The tolerance you wrote would be .01 inch not .001. I am not saying something like this can't be built or that your not capable of designing it. Only that you can't do it for your budget or without access to a world class large scale machining facility and a highly controlled environment with almost no temperature variation. The idea of using some kind of scale to map your movements would still not get you to those tolerances because they only take into consideration linear movements on each axis. That is not real accuracy in 3d space and the equipment to map those movements in real space is unreasonable to achieve for this type of project. Also drilling is not "only small forces" by any means. If you are only drill very tiny holes maybe but anything more requires lots of torque which then has to imparts those forces to some degree to the structure around it.

I'm sorry I have taken a severe point of view but in this case it is needed to impart some realism into this discussion.
I suspect if this is a teacher assigned project you are meant to do the best you can with what you have. That will end with you failing to hold to the guidelines you have set for this project. It will also teach you to be realistic in your expectations and work as a team to overcome obstacles. In the end that may be more beneficial then succeeding.

Just my opinionated opinion.

Ben

[kawazuki]

I gonna be ultra critical here but if you presented that to me - in your final year as an engineering student - i'd resign as you hadn't seemingly understood any basic engineering principles!
You have designed a device that would have deflection and flexation in just about every plane of operation! You have a series of linked boxes with no diagonal bracing to prevent
deformation, an unbraced projecting tool arm - neither laterally or vertically, forget sub millimeter tolerance - I think you'd be lucky to achieve sub 5mm precision..............need I go on?

[victorofga]

bhurts exactly

the drilling force relatively small, but...

drawing up the force vectors, youll find every gram force on at the dril, will result hundred of kg force on the other end of the arms

if he wants to build a light structure, then he has to look for bridges, or high voltage electric poles.. why and how they are braced..
or wooden trusses..

it is very common that folks learn any cadprogram, and they got the idea they can design

just like you buy a phonendoscope and try to acting like a doctor :-)

I just hoping teacher gives him some more insight..

if he can design some reliable stuff, he can find sponsor..

[leMaitre777]

Hey so after a many depressed nights after y`all broke the truth like that which I`m so glad you did (This should have been our first post but we were stupid). victorofga is in a fact a much nicer person than my study leader for the project haha believe it or not. So attached is the new current design.

Linear System
This design is far more superior in rigidity as you would note and makes use of linear rails for the y and x axis. Suppoted shaft rails will be used along the z - axis. The reason for the supported shafts are the degree of freedom that comes with mounting the rail to structural steel. With some agonizing struggles I'm hoping to get these rails allinged with the aid of shims (vertically) and laser alignment (horizontally).

Motors/Drive
The motors used for the long z-axis will be 7.7 Nm NEMA 34 motors used in conjunction with 5:1 timing belt reduction. A question I have here is whether the it will be needed to spring load the pinion to the rack. My guess is it will be best to do so but how I still need to figure out. It seems that all the solutions I can think of are not really bulletproof ideas. Any suggestions would be greatly appreciated.

Describing then whole design will take me ages to type out. Pictures speak a thousand words. Feel free to comment on the design and ask questions give it to me real.

[alan_3301]

Spring loaded, yes, but 5:1 reduction on 7Nm motors?
Do you really need 30Nm?? You frame wont take it.
What are you cutting anyways?

[leMaitre777]

Done wrong - Removed