[flanajb]

Hi All,

I am in the process of designing my first cnc build. Plan is to build a machine that can route 8'*4' panels.

My first plan was to build the sub frame using extruded aluminium, but I decided against this as I just don't like aluminium framed cnc machines and I think the resale value would be poor as they look diy.

I decided that a steel frame was the way to go, but the issue I have is that once I have welded the frame up, it will need to be heat treated to remove the stress. This is not an issue, as my main concern is how I can install the linear guide rails.

I have seen people bed the rails on epoxy, but that seems like a lot of work and pretty hit and miss. Just wonder whether it is possible to get a company to surface mill grooves in the x-axis rails for the linear rails. This would ensure the rails are parallel and in line to the required 0.001" tolerance I want to achieve.

Any ideas would be most welcomed ?


Thanks

[DMF_TomB]

basic machinery alignment rules

1) bolt it down so you can use a precision level and avoid leaning on it as your weight can distort readings.

2) shims are common where it cannot be machined and equipment can be grouted with cement or epoxy for better support between shims

3) temperature expansion can seriously affect rail straightness. if a rail is warmer than whats supporting it the warm rail will expand and the extra length will cause rail to bow or snake. sometimes rails are slotted or in alignment grooves so expansion can occur but rails stay straight

4) rail parallelism is done with a tram gage, basically a rod with a inside micrometer or a rod with a dial indicator to see if distance between rails is the same. often a magnetic base and dial indicator can have magnetic base replaced with a block or angle and the rods, clamps and indicator can then be adjusted for different distances

5) visit
TKT Products
they make tools for lumber sawmills. basically setting up equipment to run of rails or tracks and aligning feed rollers. short of using optical alignment surveying tools for smaller machines they have some of the best designed equipment i have seen.

[Mountaincraft]

My 'plan' (so far) when I rebuild my table, is to buy the steel piece by piece and have them cut it to exact lengths as I buy it.. drill holes at home... When I have all pieces, save the money up to have 'them' weld it together using their welding fixtures and superior welding equipment and skills.. Then put it on their big mill and surface mill both rails... Then have fun getting the thing home and unloaded...

I will then cut several 'alignment jigs' on my existing machine before I tear it apart out of laminated layers of high quality plywood or MDF... These jigs will be used to hold my existing 80/20 aluminum rails in place as I bolt the existing 80/20 rails to the new milled base rails... From this point I should just be able to bolt everything on top, taking time to adjust and align things as I go.. being confident that it's all flat and parallel..

At this time the plan is to use 3/8" walled steel rectangular tube frame rails, with all the supporting structure being out of 1/4" and 3/16" steel tube..

I want to make it as rigid and heavy as possible.. with replaceable slide in steel framed material beds designed for various purposes.. sheet goods, 3D work, indexer work, plasma, etc...

[flanajb]

have 'them' weld it together using their welding fixtures and superior welding equipment and skills.. Then put it on their big mill and surface mill both rails...

We think alike my friend. That was what I was thinking, but was unsure whether there are companies out there who can mill the surface x axis rails for you once the whole frame has been welded together ?

Have you obtained any quotes on how much that would cost to have done ?


Thanks

[Mountaincraft]

No, but I've had other work done by them.. So based on that, I'm guessing that their 'labor' would be around $600-800... I know they can do the milling, as I've seen their mill... It's huge! Like aerospace huge...

[flanajb]

No, but I've had other work done by them.. So based on that, I'm guessing that their 'labor' would be around $600-800... I know they can do the milling, as I've seen their mill... It's huge! Like aerospace huge...

Does the company have a web site ?

I live in the UK, but it would be handy as it will enable me to ask if companies here in England can perform the same service.

Thanks

[DMF_TomB]

many steel frames sag from their own weight. some can be machined out if supported in milling machine similar to way machine part will be used. if part is clamped down tight to table of milling machine it can distort as soon as milling machine clamps are loosened. it can need shims even on milling machine.
.
many parts warp or distort from welding and as material is removed from machining. some parts are heat treated to stress relieve part so it warps less from machining.
.
as a field machinist i often has to compensate for parts sagging from their own weight by measuring and adding shims. also how a part is bolted to a concrete floor can influence frame twist. basically a 4 legged chair with a short leg can be pulled down tight to floor with mounting bolts or adjusted with shims or jack screws.
.
a precision level that reads 0.005"/ft or 0.3mm or 0.4mm/m per line increment is an extremely useful tool for rough alignment.

[flanajb]

many steel frames sag from their own weight. some can be machined out if supported in milling machine similar to way machine part will be used. if part is clamped down tight to table of milling machine it can distort as soon as milling machine clamps are loosened. it can need shims even on milling machine.
.
many parts warp or distort from welding and as material is removed from machining. some parts are heat treated to stress relieve part so it warps less from machining.
.
as a field machinist i often has to compensate for parts sagging from their own weight by measuring and adding shims. also how a part is bolted to a concrete floor can influence frame twist. basically a 4 legged chair with a short leg can be pulled down tight to floor with mounting bolts or adjusted with shims or jack screws.
.
a precision level that reads 0.005"/ft or 0.3mm or 0.4mm/m per line increment is an extremely useful tool for rough alignment.

There seems like a lot more to this than I thought. I was hoping having 3m x rails that were 150mm * 50mm * 6mm section will not sag if supported at the ends and in the middle by the sub frame would have very little distortion.

[kuhnkejer]

You should look at carveone build he started with a wood base then welded up a 12 foot base. His forum is around 250 pages but its a great read. He starts making the steel table at around page 90. His table is working great now and even looks good. I will making my next Cnc like his. Plus the design is cheaper then rails

[flanajb]

You should look at carveone build he started with a wood base then welded up a 12 foot base. His forum is around 250 pages but its a great read. He starts making the steel table at around page 90. His table is working great now and even looks good. I will making my next Cnc like his. Plus the design is cheaper then rails

Thanks. Do you have a link to the web site ?

[ger21]

I think that it would be very expensive to have the base machined to .001" over the entire 8-10ft that you'd need.

This should be required reading for anyone building a large machine. Lots of good information here.
MadVac CNC - home made 4'x8' cnc precision gantry router

[technodream]

thanks for your insight!

me and my build partner are having similar thoughts and deliberations about our cnc build.

we used a welded steel table as a frame, the rails and r&p will be mounted on two rectangular steel tubes i will call "rail holders" (dark green in picture) which in turn will be mounted on the frame using M6 bolts.

the proposed calibration procedure is -

1. rail straightness - mount each linear rail on a rectangular tube, bolt and shim (using 0.1mm spring steel sheet) against a 50x10x2000mm aluminium ruler which hardly sags under own weight. after reading this post, also considering grouting with polyester.

2. rail-rail parallelism - mount each rail holder using only two bolts, one on each end. instead of a tram gauge, i suggested we connect the two rails using the linear bearings and a stiff 80x40mm 80/20 profile. then push the thing back and forth, and shim the rail holder where appropriate until minimal resistance is felt.

3. rail-work table parallelism - hook up a dial gauge to the above contraption, with sensor pointing downwards, and move back and forth on X axis, adjusting work table as neccesary. change Y position, repeat.

would a level really be helpful in this procedure, and where?

cutaway view shows rail holder in yellow, rack in red, table in gray (should be hollow)

note that our plan is to mill plastics and mdf at 0.1mm resolutions, and perhaps also aluminium.

thanks again for your insight

[PaulRowntree]

t

would a level really be helpful in this procedure, and where?

In theory it is not required, but having a simple way of checking your machine to intermediate accuracy makes the build and alignment process a lot easier, especially if the table legs have adjustable feet.

[rcheli1]

You might want to try the tight wire method detailed here

[Mountaincraft]

Does the company have a web site ?

I live in the UK, but it would be handy as it will enable me to ask if companies here in England can perform the same service.

Thanks

I don't know.. But they're called Gerlingers, in Redding California...

[Mountaincraft]

You know, once you build your table and it has all it's parts on it, and everything is tightened down, typically, most people install a sacrifice board and then use the machine to mill itself... If there's little play in everything, then the surface job should follow any imperfections, so that when a material is secured 'flat' to the sacrifice board, you can get some pretty tight tolerances, as your machine calibrated itself to account fro any imperfections..

I don't know for sure about this, but wouldn't most materials expand and shrink with temp changes to exceed the .001 tolerance your after? I would that the CNC table would.. and the larger the size of the table or material, the more the total expansion and contraction, no? Or are you going to have a climate controlled environment?

When I have mine milled, I'm not after any specific 'tolerance' on the table itself so much as making the thing move as smooth as possible... I just want it as parallel as is reasonable... as every 'error' gets added to the next error during assembly.. So the more critical I am in the early stages, the more forgiving later stages will be.. I'm also very interested in mass and solid construction so that I can push it at higher speeds...

[flanajb]

You know, once you build your table and it has all it's parts on it, and everything is tightened down, typically, most people install a sacrifice board and then use the machine to mill itself... If there's little play in everything, then the surface job should follow any imperfections, so that when a material is secured 'flat' to the sacrifice board, you can get some pretty tight tolerances, as your machine calibrated itself to account fro any imperfections..

I don't know for sure about this, but wouldn't most materials expand and shrink with temp changes to exceed the .001 tolerance your after? I would that the CNC table would.. and the larger the size of the table or material, the more the total expansion and contraction, no? Or are you going to have a climate controlled environment?

When I have mine milled, I'm not after any specific 'tolerance' on the table itself so much as making the thing move as smooth as possible... I just want it as parallel as is reasonable... as every 'error' gets added to the next error during assembly.. So the more critical I am in the early stages, the more forgiving later stages will be.. I'm also very interested in mass and solid construction so that I can push it at higher speeds...

I agree with what you are saying, but my main concern is that the x axis rails need to be dead level, otherwise, if you were doing a groove and they were not, then you would get variation in groove depth over the length. Even if you mill the bed, this will not solve the problem as the piece of material you are machining will need to flex down into the groove on the bed.

I have found a company in Dorset, UK who have a 3000 * 1700mm cnc bed for milling. So I am going to make my bed out of 150mm * 50mm * 6mm steel box section for the x axis rails and 75mm * 50mm for the bed rails and then get the company to mill the bed and more than likely get them to install the linear guide rails for me.

[ger21]

What you probably want to do is weld bar stock (6-10mm?) on to the box sections and have them mill that for the rails. Have them mill a shoulder to align the rail to.

[flanajb]

What you probably want to do is weld bar stock (6-10mm?) on to the box sections and have them mill that for the rails. Have them mill a shoulder to align the rail to.

Hi Gerry, That is a good plan and it will mean I can get the rails aligned that way.

Thanks

[Mountaincraft]

I agree with what you are saying, but my main concern is that the x axis rails need to be dead level, otherwise, if you were doing a groove and they were not, then you would get variation in groove depth over the length. Even if you mill the bed, this will not solve the problem as the piece of material you are machining will need to flex down into the groove on the bed.

The 'key' there is to make sure your material is held down 'securely' to the milled surface, so there is no flex.. IOW, it follows the contour of the sacrifice board closely..

Even if you get this table 'perfect' you're still going to have to come up with ways to pull the whole material flat to it's work surface, otherwise the material itself will introduce imperfections and out of tolerance situations...

I don't know much about what you plan to do with it, or what materials you'll be using, but you might want to also consider a quality vacuum holding system for it... otherwise all this effort at perfection might be in vain...