-
Fixed gantry or column for mill rigidity?
Hello all,
I'm looking to build a new CNC/manual mill. I have a ~1,000 pound (450 kg) table that has been ground flat, with a surface area of 42" x 36" x 1" thick (106 x 91 x 2.5cm).
The build criteria is simple: be as rigid as possible, within limits.
To that end, I'm leaning towards a fixed bridge design using heavy wall square tubing, 35 - 45mm guide rails, and as big a diameter ballscrew as I can practically fit under the axes.
Here's a scratch design I've been noodling with while contemplating ideas. This is not a final design and I'm not really looking for feedback on it, just to serve as a conversation starter:
https://lh3.googleusercontent.com/pw...9=w490-h297-no
Soooo... if you were building a mill from scratch that can hog through steel and not break a sweat (or end mills), what would you design? Is a fixed gantry a better design (within its compromises) than column? The commercial machines I've seen on YouTube easily cutting steel have all been column types... hmmm.
Thanks!
JR
PS: good to be back. Looks like I've been bitten by the CNC bug again ;-)
-
Re: Fixed gantry or column for mill rigidity?
A fixed gantry (or "bridge") design like you've drawn is more rigid than a moving gantry, and it can be made very massive, since it doesn't have to move. The "column" type mills are typically small machines; larger ones are usually massive castings in a "C" configuration. But mills like the one you've drawn are becoming more common in industry. They are easier to build if you don't have an iron foundry, and work well if attention is given to rigidity in all parts and assemblies.
You say this is a "CNC/manual" mill; I'd suggest you choose one or the other. Ball screws are used for CNC, they're not used for manual mills because they tend to back-drive - you'll need a way to lock down any axes you aren't using for a particular move. Is this machine supposed to have dual spindles and another axis parallel to Z? If so, you'll need a way to drive the two separately, this gets complicated. Also, you'll need longer rails for X, and a footprint about twice as long as you've indicated if the tool is supposed to reach the whole table area.
.
-
Re: Fixed gantry or column for mill rigidity?
Much appreciate the feedback, Andrew.
I set out to build a moving gantry many moons ago to cut PCBs. It ended up being 52" x 60" (132x152cm) and mostly routed wood and aluminum parts for itself. I have cut a number of mild steel projects on it but need to go very slowly with that.
Good to know about fixed bridge gantries. Looking at the different types, they seem to be the most rigid. Compared to a column, it's also likely easier to build in a garage.
The sketch above is a moment in brainstorming time. Nothing's set, other than perhaps the table/base dimensions. The columns are 6" (152mm) wide, 1/2" (12mm) wall. The rails and trucks are 45mm, "winged" or wide type. Roughly, X and Y have 20" (508mm) travel. Z will have to use a smaller rail guide.
I imported a drawing of my home-built ER25 spindle and it looks like a toy next to the 45mm trucks. Have a laugh:
https://lh3.googleusercontent.com/pw...E=w434-h313-no
For reference, that spindle is 2.75" (70mm) diameter. I was so proud of my spindle...ha! Most of my current tooling is R8 since I run a Bridgeport, but it looks I'll eventually have to buy or make something like a BT30 spindle for this new machine.
Yes, sorry, when I say "manual" I mean manually controlled motors. I have a DIY pendant I use in the wood router that let's me operate it manually, somewhat. That machine runs steppers and an old laptop that can't really keep up with faster manual inputs. I already have the motors for the new one: NEMA-52, 15NM, 2.3kW AC servos - though I think Z will use something smaller.
Thanks,
JR
-
Re: Fixed gantry or column for mill rigidity?
Quote:
I have a ~1,000 pound (450 kg) table that has been ground flat, with a surface area of 42" x 36" x 1" thick (106 x 91 x 2.5cm)
Well that's a handy thing to have on hand !
You're probably already aware of the WADE'O router but here's a link anyway:
https://www.wadeodesign.com/design-details.html
It's not designed to be a mill but maybe there's some useful info there.
Bridge mills are great but I'm not sure that it's an easier DIY project than a C frame mill. WADE'O made it look easy but he cheated by starting off with all the good stuff ready made :)
-
Re: Fixed gantry or column for mill rigidity?
Hi, and thanks for the post and link. I had watched his video on YouTube while searching for fixed gantries. Very nice build! It was going great until he said he built it "to cut aluminum" :rolleyes: Hopefully he joins this conversation, if he's still around. Maybe I'm just dreaming that I'll be able to cut steel, but I'll try.
Yes, mounting the square tubing does worry me. I'm also sort of cheating by starting with a Blanchard ground table that will be my datum reference for the rest. Grinding the square tubing flat will be a challenge, however.
One thing I'd like perspective on is how I mounted the rail blocks upside-down:
https://lh3.googleusercontent.com/pw...t=w552-h305-no
I'm not sure why more machines aren't built this way; I must be missing something. My rationale is that the Z and X axes will *always* be exerting pressure along the X plane range. It really doesn't matter where Y moves to, Z is always on the same X plane. If that is so, why bolt the supporting trucks to the moving table and having them be away from where Z and X are exerting pressure?
In other words, if the trucks are mounted to the table and you're cutting at the edge of your Y travel, the trucks will not be under Z. But if the trucks are fixed to the base and along the X vector, there will always be maximum support from the rails and trucks regardless of where Y moves to. Sure, heavy loads from parts on one edge of the table might tilt it slightly (like Bridgeports do) but that's a lesser compromise compared to not having support under Z, IMHO. There's also the added advantage of not having the rails exposed to chips as they're always tucked under the table. I'm missing something, aren't I?
Regards,
JR
-
1 Attachment(s)
Re: Fixed gantry or column for mill rigidity?
Seeing as no one else has replied I'll toss in 2 cents.
The blocks you're looking at have huge capacity. The question may not be will they handle the load (?) but rather what is the lifespan of the components used in this orientation (?)
It's not hard to imagine where the stress points will be on those blocks when the table is at the extremes of travel. These manufacturers have engineers to answer technical questions.
I don't how responsive they are to questions from DIY builders but it's worth a shot ?
I built a router using this concept but with linear bearings that are a joke. Deep groove bearings were added to the outside corners of the moving table (not shown in picture) to prevent the tilt.
The original plan was to have the screw move with the table in rotating nut style but I may never get around to that modification. The router is lag bolted to a hefty torsion box.
Attachment 443492
-
Re: Fixed gantry or column for mill rigidity?
Hi JR - Its usual to mount the table rails on the bed and have the cars on the table. Probably easier to assemble. But your arrangement will work, this will keep the load in the cars "kern" ie centre. Seems you have oversized everything so it will be a right beastie. Bridge, gantry or double column mills are common in very large machines. There is a width at which the column cantilever becomes too wide for stability so then the "bridge configuration" takes over. But it scales up and down very well. To answer your original Q there is no doubt that a fixed bridge config is one of the stiffest configs. Its downside is the larger footprint required for the sliding table but all designs are a bunch of compromises that produces a "best" design. Peter
The next discussion is if you are going to weld this structure together are you going to stress relieve it? This is associated with what accuracy are you aiming at for the machine?
-
Re: Fixed gantry or column for mill rigidity?
Hi, and thanks both for your insightful replies.
Cycle, the idea behind mounting the trucks fixed to the bed is so they are always under the cutting forces. With the trucks moving, there is a rail there under the spindle but the truck could be elsewhere, not supporting the work. As I see it, the compromise of possibly flexing when the table is at the end of travel is a lesser evil than cutting with little support underneath. At that point, the table will really flex.
Peter, the trucks I'm looking to get are the "wide" variety that can be specified with over or under mounting. True, it's easier to mount the rails down and then slide the Y table over it.
I have been reading a really insightful paper on the topic, the "Principles of Rapid Machine design" by Eberhard Bamberg, PhD. <spoiler alert> It essentially concludes a variation of a fixed bridge machine to be the most rigid from the group evaluated, which also included a column mill.
While the analysis proposed requirements include a 5-axis design - a far more complicated design that skews the conclusions as it takes this into account - it opened my eyes to two important factors: the square/rectangular beam I had chosen is in itself a compromise (it twists vs a round tube), and the Z axis in a bridge design makes it much weaker than a column design. This is contradictory to the paper's conclusion; bridge is better.
Here's a quick depiction of what I see as a weak point in bridge designs, with my apologies for the crude drawing:
https://lh3.googleusercontent.com/pw...k=w474-h325-no
Note how the spindle is lowered. The closer it is to cutting, the weaker its support becomes, just when it needs it the most.
But a column Z is fully supported throughout and, in fact, it is *best* supported when cutting the work:
https://lh3.googleusercontent.com/pw...j=w429-h349-no
The column is a weaker support structure when looked side by side a bridge, but not so much when the machine is cutting the work, IMHO. Another issue with bridge design is that it puts the weak cross of Y and Z up on the bridge, the weakest point of the machine. Contrast this with a column where the cross axes are supported by the machine base.
Peter is right, this is all a series on compromises vs function and requirements. I'll probably change my mind a few more times and battle my OCD before settling on a compromised pick.
Regards,
JR
-
2 Attachment(s)
Re: Fixed gantry or column for mill rigidity?
Hi JR - I have analysed many gantry and machine configs and each has + and -'s... One thing that can be done is to not have the top rail on the gantry face but move it to the top and back of it. This achieves two things 1) It makes the force on the rails smaller as their distance is further apart so the local deflection is less 2) It makes the centre of action of the moment closer to the shear center of the beam which means it reduces the lozenging of the section. But this then means the saddle is L shaped but my modelling shows its a bit stiffer. This also gives the opportunity of putting the drive on top of the gantry vs on the front if this is of use.... so many combinations.
https://en.wiktionary.org/wiki/shear_centre
https://www.quora.com/What-is-Shear-center
bambachs thesis is a must for any engineer in this field.
To extend your logic on the column config a rising Z bridge solves some problems. Effectively its a double column machine. I intend to do this next machine...
Cheers Peter
-
Re: Fixed gantry or column for mill rigidity?
JR, the problem with a C frame (what you are referring to as a column mill) is the lever length needed to get decent Y travel.
If you don't need much Y travel and want lots of Z travel, a C frame can be a good choice.
If you want lots of Y, a bridge (dual column) may be a better choice as the spindle is much closer in as the table can travel under the bridge.
-
Re: Fixed gantry or column for mill rigidity?
https://www.youtube.com/watch?v=f8nTNY0FO6o
Here's a large scale mill with moving everything. The rising gantry is called axis W. Peter
-
Re: Fixed gantry or column for mill rigidity?
Quote:
Originally Posted by
peteeng
Hi JR - I have analysed many gantry and machine configs and each has + and -'s... One thing that can be done is to not have the top rail on the gantry face but move it to the top and back of it. This achieves two things 1) It makes the force on the rails smaller as their distance is further apart so the local deflection is less 2) It makes the centre of action of the moment closer to the shear center of the beam which means it reduces the lozenging of the section. But this then means the saddle is L shaped but my modelling shows its a bit stiffer. This also gives the opportunity of putting the drive on top of the gantry vs on the front if this is of use.... so many combinations.
Indeed. I came to the same conclusion when I built my first CNC router back in 2004. The execution was entirely lame, as I had neither the tools or the knowledge, but something was telling me that second rail on top was better than coplanar to the first. Since this pic, I've moved the second rail further back and added a 1/2" (12.5mm) steel plate across Y:
https://lh3.googleusercontent.com/pw...q=w377-h282-no
On that last design with the blue machine, I was loving the fact that the entire Y axis dropped down on Z. That does solve many of the issues with stiffness and makes for a nice hybrid of both types of machines. But of course, the double ballscrew and belt add their own issues.
Regards,
JR
-
Re: Fixed gantry or column for mill rigidity?
Quote:
Originally Posted by
pippin88
JR, the problem with a C frame (what you are referring to as a column mill) is the lever length needed to get decent Y travel.
If you don't need much Y travel and want lots of Z travel, a C frame can be a good choice.
If you want lots of Y, a bridge (dual column) may be a better choice as the spindle is much closer in as the table can travel under the bridge.
Yes, I agree. C-frames do limit Y compared to a bridge, and this was another factor leaning me towards a fixed bridge design. As for Z, I guess I would be looking at Z height rather than total axis travel. In the years I've been using my current CNC router, I've never needed much of Z travel. Of course, if I had it, I'm sure I'd be saying I couldn't live without it. I think if I account for a vise, maybe a future 4th axis, holders and tools, the actual travel of Z shouldn't be more than a few inches.
Regards,
JR
-
Re: Fixed gantry or column for mill rigidity?
Quote:
Originally Posted by
peteeng
Here's a large scale mill with moving everything. The rising gantry is called axis W. Peter
That is just amazing. Super complex to DIY, however, and I suppose if you threw enough mass at a machine it could probably be made out of marshmallows :)
Cheers,
JR
-
Re: Fixed gantry or column for mill rigidity?
Quote:
The commercial machines I've seen on YouTube easily cutting steel have all been column types... hmmm.
Back in another lifetime I ran a Mazak VQC 20/50B bridge type mill. Pound-for-Pound it was a rigid machine. Unfortunately I can't find photos of the bridge with the swarf protection removed.
Quote:
Cycle, the idea behind mounting the trucks fixed to the bed is so they are always under the cutting forces
The same as the router drawing I posted earlier, the tool is either over one of the middle bearings or somewhere in between the pair of middle bearings. Why 3 bearings/axis ? I had 6 blocks and didn't know what to do with 2 spares, no other reason. Like the Mazak my router is decidedly rectangular.
-
Re: Fixed gantry or column for mill rigidity?
Hi JR - I'd use independent motors vs the belt and pulleys. Motors are cheap. Then I'd use UCCNC and an AX-BB controller so you have 6 axis to play with. I've started drawing up the mill already...Peter
-
Re: Fixed gantry or column for mill rigidity?
Hi JR - Its a bit tricky figuring out where square bearings should be. This is because they can support a moment. This means the reaction can be in any direction (its an indeterminate problem). If the rails are the round type then the reaction is easy to predict visually. The sq bearing closest to the load will be the dominant load bearer (ie you could remove the secondary bearing and it will still work, just not as good). The next one will be secondary. Need to do FEA to figure out exact reaction trajectories and how the load is shared then try to get these through the shear centre.
https://www.youtube.com/watch?v=ZlUX5OkP8vU
DMS use a flat front face but semicircular rear shape for their gantries...Peter
-
Re: Fixed gantry or column for mill rigidity?
Hi and thanks much for your continued help.
Cycle, I found a few images of that machine in various states of partial undress. I think I can picture the rest in my head. Very nice machine, especially for the built year. I'm actually surprised at how solid square way machines are compared to modern guide rails ones.
Peter, yes, controllers, etc is next. When I built my router years ago, it was all about parallel ports and Mach2. Now it's USB or Ethernet and everyone seems to be wanting to move away from Mach.
A bit off-topic but, I'm not familiar with recent controllers. I thought I would be using Mach3 or 4 on my new build and something like a "SmoothStepper" device over Ethernet. I'm not clear where the line is drawn between the computer NC software and the external device. Do these devices interpret G-code outright or is something like Mach under control, ramping paths, setting PWM, etc?
I remember how upgrading Mach versions actually improved how smooth and quiet my steppers ran. Hopefully the external devices got that right as well. I'm a little apprehensive on relinquishing that level of sophisticated control out of the PC but that's probably because I have no experience with the external controllers. So AX-BB and UCCNC is better than SmoothStepper and Mach4?
Welp, I better do something with this now. Last night I bought this thing:
https://lh3.googleusercontent.com/pw...w=w400-h300-no
It's an old CAT40 out of a retired Shizuoka B-5V. Somehow "biting more than I can chew" keeps sounding in my head ;) I wanted something that can take a 1" (25mm) tool shank, as I have a few of those. There also seems to be a huge variety of CAT40 holders, at least here in US, and cheaper than some smaller tapers. A new BT40, belt-driven spindle would have been better but cost and shipping is a bit high coming from China. Maybe I'll do that later on.
Regards,
JR
-
1 Attachment(s)
Re: Fixed gantry or column for mill rigidity?
Hi JR - There are many many controllers to choose from. From simple to complex. from zero $$$ to 1000's $$$. Mach 3 should be avoided as its unsupported and superseded. Although it works and there is a huge fan base. Humans have lots on inertia in learning new things unless put on the spot. Mach 3 was built around it controlling the computer clock internally. Seemed a good idea at the time when computers were simpler. Computers use internal clocks to time actions and internal activities. These clocks change there timeing for different things. A CAM clock cannot change time so M3 freezes various clocks so it keeps true time. I'm not an M3 user thats just how I read the guff. I nearly was an M3 or M4 user but due to a chain of events I ended up with UCCNC and I'm happy there.
Modern software uses the computer for the application but an external computer for timing and motion control eg smoothstepper and UC100, UC300 and others.In this way they are seperate and no need to fiddle with computer internals. So do some research on controllers I'm sure someone in the forum will have a summary of whats out there.
My cheap machines I use "blackbox" https://www.makerstore.com.au/product/elec-ob-bbmcs/ an all in one solution with free software. My more sophisticated machines I use UCCNC-UC100. About to play with a AX-BB as I need 5 axis for the rising Z design. I have used a 5 axis BOB on a machine so I could use UCCNC-UC100 and the chinese BOB like the photo attached as I'm familiar with that. Peter
-
Re: Fixed gantry or column for mill rigidity?
Quote:
Hopefully the external devices got that right as well. I'm a little apprehensive on relinquishing that level of sophisticated control out of the PC
Linuxcnc uses only 'dumb' hardware or stated another way the PC remains in control. 'Dumb' sounds a little insulting to some of the excellent hardware available but I think even the Mesa Electronics guy himself refers to it that way. Other than Linuxcnc everything is moving in the direction peteeng described (or at least everything I'm aware of).
-
1 Attachment(s)
Re: Fixed gantry or column for mill rigidity?
Quote:
Originally Posted by
JRoque
Hello all,
I'm looking to build a new CNC/manual mill. I have a ~1,000 pound (450 kg) table that has been ground flat, with a surface area of 42" x 36" x 1" thick (106 x 91 x 2.5cm).
The build criteria is simple: be as rigid as possible, within limits.
To that end, I'm leaning towards a fixed bridge design using heavy wall square tubing, 35 - 45mm guide rails, and as big a diameter ballscrew as I can practically fit under the axes.
Here's a scratch design I've been noodling with while contemplating ideas. This is not a final design and I'm not really looking for feedback on it, just to serve as a conversation starter:
https://lh3.googleusercontent.com/pw...9=w490-h297-no
Soooo... if you were building a mill from scratch that can hog through steel and not break a sweat (or end mills), what would you design? Is a fixed gantry a better design (within its compromises) than column? The commercial machines I've seen on YouTube easily cutting steel have all been column types... hmmm.
Thanks!
JR
PS: good to be back. Looks like I've been bitten by the CNC bug again ;-)
For a table that width size you will have to have a bridge mill design.....a single column will not support the over hang for the Z axis to reach over the full width of the table you plan have but a double column or bridge will enable it to do so......but you could make the Y axis (under the bridge) the short size and the X axis (across the bridge) the longer size as the bridge can cover the long size easier as it's more supported.
Column mills normally have a long narrow table because of that reason.
BTW, the length of the mill base between the columns (Y axis) is usually twice the width of the table (X axis) to enable the full length of the table Y axis to move under the bridge. and to prevent too much overhang by the spindle (Z axis) from the face of the bridge
Bridge mills also have less Z axis travel due to the height restrictions of the columns which can be massive to make them less liable to back and forth movement due to cutter torque forces.
Attached is a pic of my favorite bridge or gantry type mill design.
Ian.
-
Re: Fixed gantry or column for mill rigidity?
Hi all,
Thanks for the clarification on the external controllers. If those don't interpret G-code directly, they probably nearly do. It'd be great to understand the Mach (for example) protocol. If Mach4 no longer generates the timing pulses, it would have to tell the external controller how many steps and in what direction it needs to move, and at what speed each movement is made. Plus it probably sends other things like curve radii. Anyway, I'm mostly speculating here; need to read more on this.
I'm still trying to decide on the machine design. One day fixed bridge wins, others vertical column wins. Right now, I'm going through what/ifs with vertical columns. It's hard to figure out how all would go together without having critical things like the spindle and its motor. The spindle cartridge (pictured previously) arrives in the next day or two. That will give me a chance to take measurements, figure out how I'm going to drive it, and build the support box/head for it. Of course, I can't finalize that until I have a spindle motor....
I have been obsessing the past few days over the spindle motor. What I want and what I can get at a reasonable cost don't seem to intersect. I want a servo motor that can do about 8,000 RPM at the top. There are a million different high speed, VFD driven spindles everywhere. But I want to rigid tap, thread mill, maybe even automatic tool changing one day. Sure, there are tricks to get an induction machine to do these things but a servo has all that it needs natively.
What I can get that meet my requirements is a 3.7kW servo package, like this:
https://lh3.googleusercontent.com/pw...pk0YmN=s400-no
The darn thing is way too large for the small machine I'm planning. It's also 100lbs (46kg)! I can get a 2.2kW that's slightly lighter but it costs the same to ship and only about $80 cheaper than the 3.7kW. The Chinese are so imprecise with their quotes and leave so much details out. I wish they posted the complete specs of whatever they're selling, post the manual online so you can read it *before* buying and have a good explanation why a smaller servo package costs the same to ship as a larger one.
Back to the Alibabas and their express train of confusion.
Regards,
JR
-
Re: Fixed gantry or column for mill rigidity?
Hi JR - There are no "curves" in G code only point to point commands. They maybe very close together so the path looks like a curve. The g code generator has settings which define how curves are broken down into small straight lines. The motion controller looks ahead and keeps the tool moving according to the controller settings of tolerance, speed and accelerations. You'll figure it eventually. Peter
-
Re: Fixed gantry or column for mill rigidity?
A question or two......do you really want to build one for the experience or for economy.......and do you want to use it for production for saleable items or mainly hobby use?
You can go to great lengths for the ultimate and then find that you could buy the same machine on the open market for less and get it tomorrow, so the burning question is how much do you want to spend?
Ian.
-
Re: Fixed gantry or column for mill rigidity?
Hi all,
Peter, maybe I've been confused all of this time but, something like this: https://www.haascnc.com/service/code...value=G02.html. Now, that the external controller can't interpret that is certainly a possibility. I do understand that in the end, it's all small lines. In fact, from where I stand, there are no circles in the universe, only short segments of straight lines.
Hey Ian, no production here. I'm looking to replace a fairly large CNC router I built many years ago and a worn out Bridgeport that's more of a drill than a mill these days. I don't want a 10,000 pounds (4500kg) VMC but neither an aluminum extrusion frame machine from eBay. I do want the most rigidity for the money I can get, hence this thread. Like most of us on the DIY CNC pill, I will probably end up making parts for the machine itself in an endless loop :)
PS: between the time I started typing this post and when I actually submitted it, I bought a spindle motor! https://www.alibaba.com/product-deta...791875146.html It's a smaller 2.2kW instead of the 3.7kW I had in my shopping cart before. It's 30kg instead of the 46kg of the larger unit.
I picked this package for several reasons. First, the factory representative, Donald, was absolutely fantastic to work with. Remember my previous post where I complained about typical Chinese not being precise or forthcoming? Not Donald Chen. First thing he did was send me to his factory website to download the manual and drive software: Spindle catalogs-Hangzhou Bergerda Automation Technology Co., Ltd.. Nice! Then he stayed online while I asked a million and a half questions about the unit. The listed motor is for 380V but they are going to wind me a special version for 220V. Ha! How awesome is that?
The unit itself looks great, with stuff you would expect from a model servo drive, but not necessarily in a spindle. It has encoder output which I can feed back to the controller NC software to read precise spindle position and speed. Tapping is easy if you know where the tool is. ATC is also a breeze when you can command a position and lock the motor there. The drive also has step and direction, optically isolated inputs and a rigid tapping mode built in.
Regards,
JR