Attempt at 5 Axis. Constructive criticism is welcome.

[dfaszer]

Hi all,
I am going to attempt to make a 5 axis router table. I have the bulk of the design done and am just missing a few of the finer details and feel that I'm close to the point where I can order some material. My goal is to have a 4' x 3' x 12" cutting area. The material for construction will be aluminum. I am planning on using stepper motors and acme screws for my power transfer. I need to double check my calcs for my acme screws, but so far I have 1"-10 for the X and Y axes and I think 3/4" for the Z axis. The estimated weight (minus the base) is about 150 lbs.

For those that have the software, or desire to look at it, here is a link to a step file for download (let me know if the link doesn't work): https://app.box.com/s/6b5fycwxox7mqudn05urpm0k5op533d9


The first picture is of the overall design. Currently, I don't have a base modeled in. I will either be purchasing a table from my work, or order one from online. The four uprights in the corners will be attached to the top of the table and will also house the acme drive screw and the motor for the X axis. The part of the base that has the grid in it will be made out of cement for the self leveling and rigidity. The grid will be for a vacuum system.


The second picture is of the gantry. I went with the split design like this in attempt to make it so that the forces that would be applied to it would be more uniform instead of having a large moment about where the next component interacts.This will ride on a piece of angle and roller bearings on both sides.



The next pictures is of the Y axis. The large hole in the middle is where the acme nut will go for the acme screw. I don't have this modeled in yet. I'm not really sure what to do about that. I know that brass is a self lubricating material and would work well for a nut, but it is also pretty expensive. I only plan on cutting wood, so a wet lube on a steel nut wouldn't work very well. Would a dry lube work well? What about a plastic nut? I plan on making my own acme nut.

In the second and third pictures, I have bearings on the top and bottom outsides. These will slide on angle pieces on the gantry. I have it designed so that the plates holding everything together have slotted holes for any alignment issues. The Z axis tube will also slide on bearings thru the center of this pieces as well.



The next set of pictures is the Z axis with A and B axis (I believe that's what the additional axes are called). There will be a steel rod connected to the motor thru a coupler which then rotates the parts. I have it so that router will be in line with the axis of rotation. This will help more in attempting to write the code for cutting. The large slot in the vertical tube is clearance for the cross acme screw. In the second picture, it's a little hard to see, but there are two small holes where the top motor is. The hole placement isn't finalized, but those will be for constant force springs (similar to a contractors tape measure) to help raise the Z axis up so there is less strain on the Z axis motor. (I have most of the components in the pictures set to transparent so that you can see the internal components).





My main concern with the design is how much flex that it will have. All of the box tubing (with the exception of the 4 corners) are 2" x 3" x 1/8" 6063-T5 aluminum. Most of the plates are 1/4" 6061-T6 aluminum, and the angle pieces are 1/8" 6063 or 6061 aluminum. I have tried doing some FEA thru SolidWorks, but I am new to that part of SolidWorks, so I am not sure how accurate it is. With just the Z, A, and B axes isolated, a 20 lbf on the tip of the router, it will deflect about .014". When I try to add the next component to the FEA study, it errors out. Does the .014" sound accurate? I have read that 20 lbf is on the upper end of the forces that I would see when cutting wood. I'm assuming that the .014" would be acceptable for a roughing pass and that when I do the finishing passes, the force and ultimately, the deflection, would be much less. My gut tells me that I need more bracing on the gantry to try to limit flex, but I'm not sure how much is needed.

Any feedback is welcome.

Thanks,

Similar Threads:

• Need Help!- Starting Machine #2, Looking For Criticism...
• Material Constructive Parameter
• 1st Router Design constructive criticisim please
• Newbie- first design, after constructive criticism
• New Machine Build- Just want some constructive criticism

[pippin88]

Your main problem is your rails. It looks like your X axis is just riding on top of angle? It needs more than gravity to stop it lifting.

What is your budget?
I wouldn't go near attempting 5 axis unless you can at least afford supported rails (though I strongly suggest linear profile bearings which are available for a lot cheaper these days - they are amazing).

[hanermo]

You basic premise is very good.

The excution is flawed, somewhat, due to;
Section sizes are too small.
Your error stack up.
You must use profiled linear rails - absolutely no way to make it work otherwise.

Mass Vs volume is wrong by 2 orders of magnitude.
In other words, its about 100x not stiff enough.

The main error is the same I have seen about 100 times.
You are instinctively thinking of strength, and not about rigidity.
This is common, as only metrology and machine tools need rigidity, and everything is in the world we done is about strength first.

For example, if you bolted granite tombstones, or surface plates of 100 mm thickness, of 200 kg each, into each plane x,y,z, the design would then work.

As planned, you will get vibration.
This will make the whole thing not work.

[wizard]

Even for wood working 0.014" is a lot. For some thing's of course it doesn't matter but for most things needing more that 3 axis I would have to Imagine that it would be a problem. Further I suspect your FEA results are optimistic.

Before we go to far here you might want to tell us the intended usage. This has a signifcance bearing upon what might be good enough.

Given that I don't see this machine even coming close to giving you the sort of results that most people would want. The high Z working area implies a significant problem with torques on the machine. You need to remember that in order to get your cutting envelope you need considerable height for the 6 axis to allow for the head that comprises the extra two axis. How much actually reaction force you will have is unknown due to it being usage dependent. However I suspect that I could get more than 0.014" deflection out of that machine by pushing on the spindle with my pinky.

Oh by the way concrete does not self level. In fact it is hard to get concrete really flat.

[dfaszer]

I strongly suggest linear profile bearings which are available for a lot cheaper these days

I do currently have the X axis riding on top of angle. Do you by chance have an online source for the some line profile bearings? What do you consider cheap? $100 for 1', or $100 for 5'?

Section sizes are too small.
Your error stack up.

Could you expand on this?

You must use profiled linear rails - absolutely no way to make it work otherwise.

I understand that the linear bearings are superior, but I don't understand how it won't work otherwise (at least for the Y and Z axes). Could you explain?

For example, if you bolted granite tombstones, or surface plates of 100 mm thickness, of 200 kg each, into each plane x,y,z, the design would then work.

I am hoping this is a little bit of an exaggeration and that I don't actually need 200kg per axis? If I'm understanding your example tho, I need to add some plates to my machine to try to increase the rigidity of it? The main spot that I see would be the side of the gantry.

Before we go to far here you might want to tell us the intended usage

I don't really have an intended use for a 5 axis. I just like the idea of trying one since I have't really seen any DIY ones. I figured that if I couldn't get it to work, that it wouldn't be that hard from me to reduce it down to a 3 axis. I do have some gun stocks that I want to have made and a couple toys for the little one, but other than that, I haven't thought too much as to what I want to make with it.

In fact it is hard to get concrete really flat.

I knew that concrete was hard to level due to the aggregate in it, but I was under that impression that cement, which doesn't have any aggregate, was self leveling.


Thanks for your feedback.

[ger21]

I don't really have an intended use for a 5 axis. I just like the idea of trying one since I haven't really seen any DIY ones

There's a reason why you haven't seen any. Even a halfway decent one would be extremely complex, and quite expensive.

There are really only two reasons to build (or buy) a 5 axis.
1) You have work to do that can only be done on a 5 axis.
2) Just to say you have one.

Option 1 will make you money.
Option 2 will waste your money.

[alan_3301]

I would like to see some details on the 2 rotary axes, as far as construction, bearings, etc.
The rest of the machine looks ok for machining foam, but you will certainly get excessive deflection with wood or anything harder.

I would like to attempt some sort of 5 axis implementation on my machine eventually, for foam and wood, so I'd like to see how you approached it.
I can see now, that they look direct driven, which may not be good enough for your application.
you'll only have 1.8 degrees resolution.
I'd want to have 0.1 degrees resolution as a bare minimum. Ideally it would be 0.001 degrees per step, but this would require servos and gear reduction.

[louieatienza]

They make self-leveling cement used to smooth out concrete floors prior to laying down flooring. I can't say exactly how level it gets but some skill at floating it out close is necessary. You can get pretty close with regular cement if you screed it.

As for what you want to do, most 5axis diy CNCs have high rails, with the z dropping down, it's a little easier to make it more ridgid. I don't think relying on the screws to hold the gantry legs down will work well.

For gunstocks, you should be able to do it with 3axis and fixtures, or just add a rotary 4th axis.

The nomenclature for naming axes? A revolves around x, b revolves around y, and c revolves around z. Your design is known as an a-c multiaxis head.

[wizard]

I do currently have the X axis riding on top of angle. Do you by chance have an online source for the some line profile bearings? What do you consider cheap? $100 for 1', or $100 for 5'?

The problem here is designing a angle iron bearing system that will remain stiff enough over the length you have for the various axis. The Z axis for example will put tremendous force upon the Y axis saddle due to the long lever arm.

While I'm not the first to dismiss cobbled together linear bearings, as they can work well for some uses, in this application I believe you are getting good advice to look at a different bearing solution.

Could you expand on this?

You need stiffness in your beams. The beams just look a bit small for a machine with the working envelope described and the relatively high placement of the Y axis gantry above the X. That being said at least you are using tubing for the Y axis beam supports.

I understand that the linear bearings are superior, but I don't understand how it won't work otherwise (at least for the Y and Z axes). Could you explain?

Well here is my perspective.
1. The bracketry that is shown in the picture appears to be rather light. You need a structurally sound way to support the bearings.
2. Angle iron is far from precision steel and varies widely in hardness. So adjusting the bearings to achieve a consistant contact across the length of the beams will be a problem much less getting everything square and parallel. Even if you get the machine assembled fairly accurately, a big if, the angle iron will wear in, The problem here is that even minor clearances will result in big errors at the end of the Z axis. A common approach on some machines is to spring load the bearings but you have a very long lever arm here and that would mean needing extremely high preloads on the springs.
3. Tolerance stack up means that every bearing in the machine contributes to error at the tool, ever twist in the frame does also. Your linear bearing approach would be the weakest link in your machine.
4. It is questionable as to the beams being stiff enough for the projected usage.

I am hoping this is a little bit of an exaggeration and that I don't actually need 200kg per axis? If I'm understanding your example tho, I need to add some plates to my machine to try to increase the rigidity of it? The main spot that I see would be the side of the gantry.

What you need depends upon what you want to achieve. 0.014" is pretty bad in my mind even for wood working especially any wood working needing a 5 axis machine. However perspective counts here. If you are cutting plywood for boat building 0.014" might be perfectly acceptable.

In the case of this machine it isn't just a matter of adding gussets here and there. Bearing mounting, the types of bearing and their placement all play a roll in how stiff the final machine will be.

I don't really have an intended use for a 5 axis. I just like the idea of trying one since I have't really seen any DIY ones. I figured that if I couldn't get it to work, that it wouldn't be that hard from me to reduce it down to a 3 axis.

I'm not going to dismiss the value in a challenge but if this is your first build I'd strongly suggest going the other way. That is build a robust three axis machine that can be upgraded to a 5 axis in the future.

I do have some gun stocks that I want to have made and a couple toys for the little one, but other than that, I haven't thought too much as to what I want to make with it.

A four axis machine would certainly do the gun stocks well.

I knew that concrete was hard to level due to the aggregate in it, but I was under that impression that cement, which doesn't have any aggregate, was self leveling.

Well their might be some out there that self levels but I've never seen concrete level itself nor have I've seen motar do so. More importantly concrete is crap without aggregate or other reinforcing materials.

Thanks for your feedback.

I'm most bothered by your bearing solution, the rush to five axis and member stiffness.

[dfaszer]

Ok, well it didn't take you guys long to convince me to get rid of the 5 axis idea. I'll try for 4 axis then with the standard 3 axis and the 4th rotary.

That being said, should I go with the same type of concept as I have but with improved bearing rails? louieatienza stated that 5 axis machines have high rails to help with rigidity. Would it help for a 3 axis as well?

I found this video in another post. Would this type of rail be sufficient? I was thinking that if I went with the high rails, that I could run an angle piece the length of the X axis rail and have the bearing clamped down on the angle.

Extended Linear Carriage with ABEC 7 Bearings - CRP102-00 | CNCRouterParts


As far as the building materials/extrusions, it will probably all be aluminum. My work has a program that lets employees buy "standard" company stock at a much discounted price than if were were to buy it online or at a local metals place. Currently the 2x3x.13 box tubing is the cheapest my work offers. The next cheapest is 2x4x.25 box tubing, but that is double the price (other sizes are available too, but the price keeps going up). I don't mind splurging on the larger size if it is required, but I'd prefer to be able to stick to the smaller 2x3 box tube if possible so that I can splurge elsewhere.

[victorofga]

the 12 in cutting area let you do nothing really of axis work...

lets make a smple calculation..

any spindle you can use at least 8 in long... your spindle with collar or bracket will takes up at least another 4 in

so 12 in area sufficient for 1/16 sized stuff approximately..

the Z axis has to carry in vertical movement the Z slider, the router two motor and a bounch of mechanoic..

guess what, the Z motor might be need larger than X or Y axis..

tooltip will be from rail from gantry at least 20 in...

basically the head what configuring everything.. the drives in the head need to be sufficient strong to hold the tool , if possible with minimal backlash..
your option probably will ending at a wormgear..

steel tubing, much more suitable for this project.. smaller size more rigidity
and price way cheaper than your discounted aluminum..

you not really find 5 axis hobby machines for two reason..

#1 designing one is not about 2 more axis.. designing is about a concept.. a whole different concept..

# 2 to making a compact head... you have to be very creative... what you drew up... don't looking a workable concept.. I mean that wont be rigid at all..

there are two working concept out..
check them and you might get ideas of..


one is a gantry type like you want ot make
another is rotary table type..

both design working and they rigid sufficiently..

https://www.kickstarter.com/projects...is-multi-fabri

Pocket NC

probably there more coming..
however while they are looking simple... it doesn't mean you can make one simple....

[pippin88]

Linear profile bearings are available for about 150 per meter with two bearing cars (Hiwin is usually the cheapest brand, cheaper on Aliexpress than ebay) I used BST Automation on aliexpress and had a good experience.

[daniellyall]

if you wont a rock sold machine you need a good size budget, low cost uselessly means a good size anchor it can be done but you will regret it if it does not work like expected and you will turn around and build another one that works.
there are lots of good and bad designs online find one you like the look of post it here and go from there and yes you will get told in no uncertain terms whats good and bad most of them no there stuff so listen

[dfaszer]

I've looked at some more 3 axis designs and I've decided that I will try a go with taller rails. I'll post back when I have something drawn up.

One thing that I noticed on pretty much all of the machines I looked at is that for the Y axis, there is one bar spanning across the gantry, and that the Z axis is then cantilevered off of that. Is there an advantage to having it this way as opposed to what I have originally drawn up on my gantry. On mine, I had 2 bars spanning the gantry and the Y axis is centered between the two. I feel that with my design, it would be more secure as the Z axis isn't cantilevered.

and price way cheaper than your discounted aluminum..

With my work discount, the 2x3x.13, 6063-T5 aluminum box tubing is $0.28 /inch my price.
The 2x4x.25, 6061-T6 aluminum box tubing is $0.68 /inch my price.
I've done a little online searching (not much), but I haven't been able to find a site that can beat these prices for either aluminum or steel.
I realize that the steel is much better for this application, but I'll try to design for the aluminum tubing purely because of the price. Once I get it redesigned, I'll to do a cost comparison and weigh the pros and cons of a steel frame and aluminum frame.

Thanks for the feedback. It has been very helpful.

[Ultranoob]

Hi!
Im in a 5 axis proyect too, still with tthe design and the correct components.
Want to know 2 things:

1st: What kind of controller board are you going to use? I was thinking in smoothieboard. Maybe you can get one free if you use it in the proyect. Smoothieboard - Smoothie Project

2nd: Why your structure is not cubical? In my research the 99% of profesional 5 axis cnc has cubical structure. I think you will gain more "stability". ( but im not an expert im a noob in this forum) 5axismaker | 5axis cnc machine 5-axis CNC Router etc.

So good luck with your proyect, i will be reading this post to help in mine too.

[wizard]

I've looked at some more 3 axis designs and I've decided that I will try a go with taller rails. I'll post back when I have something drawn up.

One thing that I noticed on pretty much all of the machines I looked at is that for the Y axis, there is one bar spanning across the gantry, and that the Z axis is then cantilevered off of that. Is there an advantage to having it this way as opposed to what I have originally drawn up on my gantry. On mine, I had 2 bars spanning the gantry and the Y axis is centered between the two. I feel that with my design, it would be more secure as the Z axis isn't cantilevered.

The bar or bars in your case are part of the gantry just so we don't get confused here.

I suspect simplicity plays a huge role in these designs. The problem with the two rail approach is that they almost certainly need to be welded together heat treated and machined afterwards. The problem is the more stiffness you seek out the more precision that is needed in the linear bearings. As such you would only benefit from this sort of structure if you went to profile rails. Even then your weldment would need to be machined precisely to keep the rails on the same plane and absolutely parallel.

Contrast that with a single beam that may require no welding or minimal welding. You can often get by without machining by epoxy leveling. Of course you can do epoxy leveling on a dual beam stucture but I wouldn't suggest that it would be easy. Further the top of your beam has to be absolutely parallel to the surfaces that the linear bearings attach too.

In a nut shel I see it as dramatically more difficult especially if you have a limited shop to work with.

With my work discount, the 2x3x.13, 6063-T5 aluminum box tubing is $0.28 /inch my price.
The 2x4x.25, 6061-T6 aluminum box tubing is $0.68 /inch my price.
I've done a little online searching (not much), but I haven't been able to find a site that can beat these prices for either aluminum or steel.

That does seem good price wise. I'm just not sure the stuff is good enough cross sectional wise.

I realize that the steel is much better for this application, but I'll try to design for the aluminum tubing purely because of the price. Once I get it redesigned, I'll to do a cost comparison and weigh the pros and cons of a steel frame and aluminum frame.

It isn't impossible to mix the two! You can run into expansion problems if the area temperature isn't stable.

Thanks for the feedback. It has been very helpful.

There is a lot to be said for building a machine out of materials you can get cheap! In that regard you might consider making a 4x4 beam out of the 2x4.

Also aluminum does not respond well to welding so you need to focus on a bolted together or riveted structure. Further you might want to supplement those fasteners with adhesives. When doing all of this consider don't forget corner bracing, gussets and the like.

With these materials you could certainly do a light weight machine.

[dfaszer]

So I found a design for a 3 axis machine that I like. The build post is here: BUILD LOG: 1.7*0.74*0.4m Mill / Router building... - Page 12

I threw together a quick model. It is still very rough, mostly just a concept and all of the bearings and acme screws are just floating in air for now.
Iso

Front

Right



I've decided to go with steel for the building material. One aspect that I like about the build in the link above is that he has the ability to lower his working table if he needed to. I think I will try to replicate that aspect, so I might look at building the table platforms out of aluminum so that it is lighter to move. For the Z axis, I was thinking of just using aluminum plate. I still plan on going with a bolt together design.

I also found some linear bearings to use off ebay. I'm not sure if the 12mm diameter will be enough. Anyone have any thoughts on that? Also, do they seem like a good price for what they are?
For the X Axis: Linear Bearing Slide SBR12 1500mm 2RAILS 4 Blocks for CNC | eBay
For the Y Axis: 12mm Linear Slide Guide Shaft SBR12 1000mm 2 Rail 4SBR12UU Bearing Block CNC Set | eBay
For the Z Axis: 1P SBR13 500mm Linear Shaft Supports 2P SBR13UU Bearing Block Slide for CNC | eBay

Thanks,

(Since I'm no longer going to attempt a 5 axis, is there a way to change the title of this thread so that people don't come here looking for 5 axis?)

[ger21]

What do you want the tile changed to?

And I wouldn't go with those SBR 12.
12mm is too small. I'd recommend 20mm

[dfaszer]

I got a little bit farther with my modeling. I haven't had much time to work on it lately.

I updated the linear bearings to accurately reflect what the purchased item should be. I also added angle supports and a few gusset plates, still have a few more plates to add tho.

Here's a couple updated pictures:



In the first picture, where the large plates are, I originally had strips like in the second picture, but by the time I put all the individual plates on, there was barely any tube showing, so i figured it would be easier to just have a simple plate.

So far, I have 2"x3"x.125" steel tube for most of the construction. .25" steel plate for the gussets. The members holding the cutting surface will be 2"x3"x.125" aluminum. The main tube of the gantry is 2"x4"x.25" steel with the Z axis being .5" aluminum plate.

A couple questions that I have:
How does the design look so far?
For the main steel tubes, will 1/8" wall thickness be sufficient? Will .5" aluminum for the Z be thick enough?
For the gussets and angles, is it overkill? It seems overkill to me.
For the aluminum box that I made of plate around the gantry. I have easy access to some 4"x8"x.25" aluminum box. Would it be better to replace the box made out of plate with the box tube? In order for me to make it work, I would need to reduce the gantry tube to a 3"3" tube and add a couple small shims between the bearing blocks and the inside of the 4"x8" tube.. However, I think the 4"x8" box tube would be much more sturdy than the one that I could make out of plate.
Lastly, I have it set up so that I have 4 feet of travel in the X direction and 3 feet of travel in the Y direction. In your experience, would it be better to switch the two just for being able to place a whole sheet of wood on the machine? Would there be any disadvantages of this? The only real problem that I see with this is being able to get the part of the sheet hanging off to be flat with the table surface.

What do you want the tile changed to?

It doesn't really matter to me. Maybe replace the "Attempt at 5 axis" with Bolt together design?


Thanks,

[daniellyall]

found something that might put you back on the 5 axis
5axismaker | 5axis cnc machine