Designing a mill using off-the-shelf parts- taking on a challenge

[Kaleb]

While looking on eBay, I have seen plenty of components like linear slides, ballscrews and trapezoidal (like ACME) threaded rod and nuts being sold for people building their own CNC routers. This made me think; might it be feasible to design and construct a small milling machine using some of these components? If so, I could design it with the features that we engine builders find useful in a mill, and also try to get around some of the problems associated with the commercial mini-mills.

So, what sort of mill should this be? Well, I have been thinking of designing what is essentially a scaled-down version of this:

which is known as a plano-mill because of its similarity to a metal planer, which is essentially a shaper with a moving table.

My plan is to use two Hiwin linear guide rails running in parallel as the slideways, with the table running on at least four of the matching carriages. Here's an eBay listing for a length of the rail for those who don't know what I'm talking about: http://www.ebay.com.au/itm/HIWIN-HGR30-600mm-LINEAR-MOTION-RAIL-GUIDE-SHAFT-CNC-MACHINE-SLIDE-BEARING-/291086767135?pt=LH_DefaultDomain_15&hash=item43c61f941f

I'm still not sure whether to use ballscrews, as they are costly, and might be somewhat difficult to use with a handwheel due to the fact that they advance 5 or 10mm per turn. They would however make climb milling easier since there's no backlash to worry about.

I intend to fabricate the bed and columns out of heavy steel plate and bar, which should hopefully make this thing nice and rigid for its size.

So, what are your thoughts? Could this be a worthwhile project, or is it just an expensive pipe dream that would never work out in the end?

 

[bazmak]

Start a thread,posting thoughts and ideas and maybe make it a joint effort or build
With lots of ideas.It would be most interesting to what ,if and how many come out of the other end.I for one would follow with interest. Complex frameworks
can be fabricated from cheap scrap offcuts of angle and channel etc instead of castings.Await with anticipation,regards barry

 

[Swifty]

Kaleb, I will be following along with great interest, I have always thought about building one with linear rails and bearings, something along these lines http://zealcnc.wordjot.co.nz. I would definately use ball screws though, just be sure that they are ground. There is a lot of information on the net about building your own. Here's how someone else did it http://www.5bears.com/cnc.htm.

Paul.

 

[Wizard69]

While looking on eBay, I have seen plenty of components like linear slides, ballscrews and trapezoidal (like ACME) threaded rod and nuts being sold for people building their own CNC routers. This made me think; might it be feasible to design and construct a small milling machine using some of these components? If so, I could design it with the features that we engine builders find useful in a mill, and also try to get around some of the problems associated with the commercial mini-mills.

There are plenty of machine tool parts online, to be worthwhile you will want to use more robust hardware than is seen in most low cost routers. Also you need to consider how you will machine everything

So, what sort of mill should this be? Well, I have been thinking of designing what is essentially a scaled-down version of this:

which is known as a plano-mill because of its similarity to a metal planer, which is essentially a shaper with a moving table.

These same machines can be had with rotary spindles and are often called planer mills when so configured. I was briefly exposed to one out of high school. Now that was a very long time ago, very late in the 1970's. The thing that impressed me was the electrical panel which was massive. The machine had four spindles and each one had drives with massive tubes. I had never seen such large tubes.

If I remember correctly the table was driven by rack and pinion. I don't believe there was an expectation of positional accuracy for the table.

One thing to be aware of is that it is amazingly easy to find all sorts of spindles for routers, it will be harder to find one suitable for machining metal and also easy to integrate.

My plan is to use two Hiwin linear guide rails running in parallel as the slideways, with the table running on at least four of the matching carriages. Here's an eBay listing for a length of the rail for those who don't know what I'm talking about: http://www.ebay.com.au/itm/HIWIN-HG...35?pt=LH_DefaultDomain_15&hash=item43c61f941f



I'm still not sure whether to use ballscrews, as they are costly, and might be somewhat difficult to use with a handwheel due to the fact that they advance 5 or 10mm per turn. They would however make climb milling easier since there's no backlash to worry about.

Use ball screws if you intend to CNC (actually a good idea). For manual climb milling you still run the risk of the axis getting away from you with ball screws.

I intend to fabricate the bed and columns out of heavy steel plate and bar, which should hopefully make this thing nice and rigid for its size.

I'd seriously consider heavy steel tubing unless you have good resources for fabricating from plate. Done well the machine will end up extremely heavy, even so I'd consider epoxy granite infill. A machine suitable for machining steel and cast iron needs to be extremely still relative to what is required to machine wood or aluminum. Something like 10 times stiffer.

So, what are your thoughts?

Sounds like fun!!😄😄😄. However don't underestimate the effort involved. Well effort and money.

Could this be a worthwhile project, or is it just an expensive pipe dream that would never work out in the end?

It could be very worthwhile. The problem is how well it will turn out is directly related to the person doing the building. That includes the ability to design the machine, the willingness to see the project through and the ability to sustain the costs. If you look at all the different router build examples out there you will see machines that barely work and others demonstrating very nice workmanship. While there are fewer mill builds out there you basically have the same reality, your results depends upon your experience and what you are willing to put into the machine.

 

[bazmak]

This is the sort of project that interests me and given enough input from other members i may give it a go.Any other interested parties who may give it a go
should add to the pot of ideas.My own initial thoughts are as follows
If we start at the top of the tree with full CNC,rails linier brgs and ground ball screws/nuts to bottom of the tree with dovetails,gibs and leadscrews.My thoughts are somewhere in the middle.Non CNC, cheaper commercial rails and linier brgs
with rolled ball screws.Fabricated where possible from std steel sections and possibly cannibalised parts from cheap chinese tools.Ie cast iron T slot Tables from from old/broken machine tools or cross drill vices etc.Spindles from
old second hand mills etc.Any input on sourcing such parts most welcome

 

[mu38&Bg#]

I built my mill using linear slides I found on Ebay. I cast a tensioned concrete column. It's been making parts for 10 years. http://www.dieselrc.com/x/temp/other.html The linear slide components are Aerotech with 4mm pitch screws and direct drive servos. 2um resolution with 12x12x6", 305x305x150mm travel. The slide on Z is too light, and the base should be stronger as well. I've been planning to improve these areas for years, but it will probably never happen unless I decide to rebuild the entire machine with a larger x axis.

 

[Wizard69]

I built my mill using linear slides I found on Ebay. I cast a tensioned concrete column. It's been making parts for 10 years. http://www.dieselrc.com/x/temp/other.html The linear slide components are Aerotech with 4mm pitch screws and direct drive servos. 2um resolution with 12x12x6", 305x305x150mm travel. The slide on Z is too light, and the base should be stronger as well. I've been planning to improve these areas for years, but it will probably never happen unless I decide to rebuild the entire machine with a larger x axis.

Nice machine, I don't believe I've seen this one before. 12" of travel on the Y axis is pretty good.

I'd like to do a DIY CNC but have to always battle with demands for cash elsewhere. I think I need to find a low maintenance cabin in the woods with no property taxes. I'm not sure such a thing is possible anymore.

 

[Kaleb]

I've started doing some CAD for the bed. It is based on a length of 380 x 100 channel, with two 1 metre long guide rails bolted to the top. I haven't put the bolt holes in just yet. At the moment, I'm intending to use the 28mm wide HGR30 size rails.



To (hopefully) get adequate rigidity, this channel will be reinforced by welding several 6mm thick gussets to the inside running in both directions. The cutouts in the middle are there to allow better access for welding.



What are your thoughts so far? Does this look like it will be rigid enough for the job?

 

[Swifty]

What are your plans for machining the surfaces where the rails are bolted on? Both surfaces have to be perfectly true to each other, one rail is often keyed in, and the other is set parallel to this.

With all that welding, there is going to be a lot of distortion to cope with.

Paul.

 

[bazmak]

yes it will be rigid enough,but too much welding,lots of distortion.The top face will need milling flat.For the size of channel the top web is thin compared to the 2 flanges,skimming in the mill will reduce even further.Would suggest a
heavy thicker top plate reinforced with say two RSC or angles etc

 

[Swifty]

As Bazmak says, perhaps a heavy top plate with stiffening ribs underneath. You would have to get it stress relieved, and ground on top.

Paul.

 

[Wizard69]

I've started doing some CAD for the bed. It is based on a length of 380 x 100 channel, with two 1 metre long guide rails bolted to the top. I haven't put the bolt holes in just yet. At the moment, I'm intending to use the 28mm wide HGR30 size rails.

I'm not sure I agree with the others here, if you intend to mount those 600 mm rails that is a long unsupported span. I do agree that you have a lot of welding ahead of you, that is a good reason to suggest box sections with internal gussets. Yes you still have welding to be done but not to the extent of your channel approach. Of course if you have the channel already you might consider welding up a box section.

To (hopefully) get adequate rigidity, this channel will be reinforced by welding several 6mm thick gussets to the inside running in both directions. The cutouts in the middle are there to allow better access for welding.

Like has been said, lots of welding and lots of distortion. You can overcome that of course but don't underestimate the extra work.

What are your thoughts so far?

Well you should certainly stiffen up the bed with the structure outlined. The only question I would have is will it be good enough for the span and everything built around it.

Does this look like it will be rigid enough for the job?

It certainly would be an improvement. The question of good enough really depends upon your expectations and what the final design is.

 

[mu38&Bg#]

I followed this build http://www.5bears.com/cnc.htm before I built my machine. Lack of access to any machines, let alone large enough to mill a welded frame (base and column), lead me down the path I took.

Greg

 

[hanermo3]

I scratch built a mill like this.
Its been finished 3 times, and the 4th will be re-finished within 2-3 weeks as a full-on VMC.

Steel build.
== 2000 kg.

Portal mill like the pic with moving table.
Table is 1600 mm wide.
VMC is 2400 mm wide, about 1800 mm tall.
1600 mm max work area (x) x y based on bed length, currently 1200 x 1200 mm.

Spindle was/is a Bp M-head.
Now I have a proper spindle, an ISO30 spindle with auto toolchanger capacity.
To be installed.

I started with small ballscrews, 0.750".
No good.

Now, with 32 mm ballscrews, 5 mm rise.
AC brushless servos, with (new) HTD8/30 mm belts at 1:3 - 1:3.75 ratio.

Servos are 400W AC brushless, 60 V DC, 5000 count, 3000 rpm, 1.3 Nm cont, 3.9 Nm peak (3 secs).
Peak force about 1500 kgf on x and y axis. Continuous force is 1/3 of that, == 500 kgf.
0.2 micron theoretical step size.
X axis linear rails: Hiwin 35 mm.

Bridge will go to new 35 mm linear rails (mass ==700 kg).
Vertical height of rails 1 m, with overconstrained movement.
ATM vertical (z-axis) is on the Bp quill only.
Vertical ballscrew is 25 mm thick atm.

Current structure uses tool steel, 20 mm thick (F1 calibrado in Spain).
Bridge beams, on their sides, are 2400x200x20 mm == 100 kg each.
2400 mm long linear rails are == 25 kg each.
Carriages are mounted to 50x100x700 mm tool steel blocks, that carry the x-axis crosswise bridge and the spindle.
Crosswise bridge is a box beam, 240x200 mm, 700 mm long, 20 mm thick, F1 tool steel.

I am using various step pulse generators in hw to get the needed high pulse rates.
ATM a pokeys-cnc on the VMC, and CSMIO-IPS on the lathe (similar, more accurate, more powerful).

The 4 iterations of the mill, plus 2 lathes, plus T&C grinder, have taken about 15.000 work hours, and 30k€ cash.
I also have == 2000 hour industrial experience on "real" industrial equipment, and factory training.

ATM standing on the free span, spindle head, == 80 kgf force, bends the spindle head down == 0.06 mm.
I expect to reduce this to == 0.02 mm shortly, as not all elements are yet rigidly fixed.

The table is too thin (60 mm) and will go to == 100-120 mm thick.
ATM table == 200 kg mass.

I hope/expect to be able to fully use a 2-3 kW servo as spindle drive.
Anything over 1 kw will be a success.
Making it half the size in x, would make it 8 times more rigid, and a 10-15 kW spindle would work fine.

You can ask questions, and I will be happy to help.
I can only do pics, once I have refinished it to commercial specs.
There are commercial considerations - sorry.

(Full sheetmetal, enclosure, flood coolent, chip shields, flex conduit for signals etc).