Planning concrete horizontal mini-mill build, advice?

[j2345]

Hey all, I am planning a machine to work aluminum and maybe some light steel work, and would appreciate whatever feedback and advice you have.

The basic plan is a horizontal mini mill, constructed by pouring CSA concrete or similar on top of a 2’x3’ paving stone, with a horizontal flat planar surface to mount the plunge/Z-axis hardware, an X-axis table on top of that, and a vertical surface to mount the Y-axis motion components and attached spindle. Travels are tentatively 400x300x200mm, the main requirement being that I have a part 300mm long that I would like to be able to work on both sides of. Budget is a soft limit of $2000CAD, not including planned later 4th axis addition.

Right now I feel pretty set on the general frame format of the machine, and am mostly wondering about the specifics of what electronics, linear motion components, and controller to use. At first I had decided to use knockoff HGR15 linear rails and 1204 ballscrews with cheap NEMA17 steppers and Ramps control board, but after researching this may be inadequate? Moving up to larger motors would allow use of stiffer 16mm ballscrews and maybe higher pitch too for faster movement. There are also many pre-built linear stages available on ebay, which would greatly ease assembly and alignment at the cost of about +$200 per axis, tempting, but is it worth the money? Keeping in mind I don’t have access to much in the way of metrology and alignment would be bodged by some straight wire arrangement in combination with an indicator and a square.

For controls I am now thinking about mach3 since there are many packages with steppers, drivers, power supply, and USB board for running on any spare PC. I’m also considering Linuxcnc and have scrounged up a old computer with parallel port, but am not sure if it provides enough outputs for a 4 axis or if latency is low enough for proper operation.

Attached a picture of the basic arrangement. Apologies for the lazy unfinished CAD, while constructing the assembly I quickly realized I would be wasting a lot of time arranging components that would likely be changed.

Exanples of prebuilt linear actuators:
https://www.ebay.com/itm/203498392580
https://www.ebay.com/itm/113361541312
Cost is about 300USD vs 200CAD for equivalent length chinese HGR20 and 1605 or 1610 ballscrew. I think I will be going with a prebuilt stage for the X axis at least, since I need some sort of table surface anyways. (Planned to use a 40160 extrusion if building own)
https://www.aliexpress.com/item/1005002292740058.html

Spindle: https://www.aliexpress.com/item/32985816072.html
Typical 2.2kw 24krpm spindle, I don’t have high hopes for milling steel with this but it should hopefully produce a good few aluminum parts before dying.

Steppers: https://www.aliexpress.com/item/1005002200264402.html
About 300CAD for 3x 2Nm Nema 23s and associated hardware plus usb control card. This is just an example, I feel like there is probably better deals to be had. Would appreciate tips on where to find good price/performance motors.

Thanks for reading, and feel free to ask questions about anything I may have forgotten to mention. I look forward to hearing your feedback.

Similar Threads:

• Newbie- G0704 - build planning advice requested
• Newbie- Planning on 1st build and need (LOTS) of advice
• Newbie- My X2 Mill Machine Planning and Build
• Horizontal hone build - need advice
• Planning a mill, need spindle motor advice

[ardenum]

Check AXBB controller board AXBB-E ethernet motion controller and breakout board combo - CNCdrive - webshop. It's relatively cheap and well used all around. It works with UCCNC, Mach3 and Mach4

The more precise your CAD model is the less real world issues you'll have.

For an HMC you're gonna need an angle plate or a tombstone in case of the 4th axis. I've seen steel angle plates on alibaba in range of $150-200. The narrow axes might give you stiffness and vibration issues, try finding ones with width that complement your build volume.

You have a long way to go so good luck.

[awerby]

That doesn't look like any horizontal mill I've ever seen. It seems more like a vertical mill with a horizontal spindle. Why are you doing it this way? Horizontal mill tooling needs to be supported on both ends, and it goes much slower than that spindle turns. I don't see any advantage to doing this except perhaps that the chips will tend to fall out, rather than stay in a pocket. But supporting most workpieces will be a lot more difficult.

The idea of casting concrete onto a paving stone is interesting, but unlikely to work very well. It would make more sense to start with a granite surface plate and use epoxy/granite to build your tower, or column, or whatever that is on top of it. Why are you running the linear tails out into space like that? They don't have any rigidity on their own, if unsupported. Using pre-built linear actuators, though, is a good idea; they're often less expensive than the components they contain, and come already mounted square and flat.

[j2345]

Check AXBB controller board AXBB-E ethernet motion controller and breakout board combo - CNCdrive - webshop. It's relatively cheap and well used all around. It works with UCCNC, Mach3 and Mach4

The more precise your CAD model is the less real world issues you'll have.

For an HMC you're gonna need an angle plate or a tombstone in case of the 4th axis. I've seen steel angle plates on alibaba in range of $150-200. The narrow axes might give you stiffness and vibration issues, try finding ones with width that complement your build volume.

You have a long way to go so good luck.

I've looked at the UCCNC offerings, my main concern is that there seems to be no support for screw mapping or closed loop control with linear scales, which I think would help with cheap ballscrew accuracy? A Mesa 7i96 is about the same price or slightly cheaper and seems to offer more flexibility combined with Linuxcnc, at the cost of being more difficult to setup. For the rotary axis, it will be one of those with the rotary axis parallel to the x-axis, with headstock and tailstock mounted to the table, not the typical rotary table HMC arrangement. The parts I want to make are vaguely tubular as well as long and slim and can be supported at both ends that way. I understand I will need some sort of angle plate to mount a vise facing the spindle. I hope to cast a right-angle block out of concrete, once the machine is running it should be relatively simple to face two sides and get them reasonably flat and square.

That doesn't look like any horizontal mill I've ever seen. It seems more like a vertical mill with a horizontal spindle. Why are you doing it this way? Horizontal mill tooling needs to be supported on both ends, and it goes much slower than that spindle turns. I don't see any advantage to doing this except perhaps that the chips will tend to fall out, rather than stay in a pocket. But supporting most workpieces will be a lot more difficult.

The idea of casting concrete onto a paving stone is interesting, but unlikely to work very well. It would make more sense to start with a granite surface plate and use epoxy/granite to build your tower, or column, or whatever that is on top of it. Why are you running the linear tails out into space like that? They don't have any rigidity on their own, if unsupported. Using pre-built linear actuators, though, is a good idea; they're often less expensive than the components they contain, and come already mounted square and flat.

Indeed, this is not a typical horizontal mill with the doubly supported arbor, I plan to use endmills exclusively. The horizontal arrangement is just for better chip evacuation and hopefully rigidity and ease of construction, and parts mounted on a rotary axis should have the same access whether from the top or side. I had considered starting with a granite surface plate, but they are quite expensive here and tough to drill into for threaded anchors. Plus, I would still have to build a flat and square vertical surface for the Y axis. I was hoping to use some scrap granite countertop or other cheap flat surface to cast the concrete surfaces against, so they would hopefully be flat without needing grinding or other post machining. The rails are just too long because of my laziness with CAD and downloading premade models, sorry. Of course when buying I would get appropriately sized rails, except maybe an inch or 2 of stickout on the z-axis for extra tool changing clearance.

For building the thing, I planned to first cast the horizontal flat base, and either try to cast the vertical surface in place, or cast the column separately and then bolt it to the base. I guess the second option will be much easier to square. If the prebuilt linear acuators are worth the cost then I will get to work making a updated CAD model.

Thanks for the feedback everyone.

[j2345]

Hello, I've updated my CAD model so hopefully you can get a better idea of the plan. I've incorporated the prebuilt linear stages for X and Y, Z remains using HGR20 and 16mm ball screw. I also added some chamfers for chip clearance, hopefully the rigidity won't be too affected. The first attachment shows the Z motion seperate from the saddle so you can see the layout, the 2nd picture shows it assembled. The cars and ballnut are spaced so that I can screw the saddle plate to the Z axis components from the top, then remove it from the rails and screw to the X axis actuator from the bottom side before final assembly.

My main concern at this point is wondering if it's even worth it? Just the linear guides are costing about $1000 in total, about the price of a small mini bench mill. Then 500 for the spindle, leaving only 500 in the budget for concrete, electronics, table, stand, and whatever else. The main advantage of this build would be that it has the linear rails and ballscrews already, but if the rigidity and performance isn't likely to be significantly better, then it would seem to make much more sense to just convert a cheap minimill like many others have done. I would appreciate hearing everyone's thoughts on how you would expect it to compare. Thanks for reading.

[pippin88]

Why a horizontal? They are good for batches of parts using a tombstone. But generally a vertical machine is easier for most uses

[ardenum]

Hello, I've updated my CAD model so hopefully you can get a better idea of the plan. I've incorporated the prebuilt linear stages for X and Y, Z remains using HGR20 and 16mm ball screw. I also added some chamfers for chip clearance, hopefully the rigidity won't be too affected. The first attachment shows the Z motion seperate from the saddle so you can see the layout, the 2nd picture shows it assembled. The cars and ballnut are spaced so that I can screw the saddle plate to the Z axis components from the top, then remove it from the rails and screw to the X axis actuator from the bottom side before final assembly.

My main concern at this point is wondering if it's even worth it? Just the linear guides are costing about $1000 in total, about the price of a small mini bench mill. Then 500 for the spindle, leaving only 500 in the budget for concrete, electronics, table, stand, and whatever else. The main advantage of this build would be that it has the linear rails and ballscrews already, but if the rigidity and performance isn't likely to be significantly better, then it would seem to make much more sense to just convert a cheap minimill like many others have done. I would appreciate hearing everyone's thoughts on how you would expect it to compare. Thanks for reading.

looking good, next step would be making semi transparent blocks for stuff like travel area, working area, collision boundaries etc. to see if things line up properly. Make sure to also include a block that represents the longest tool you'll be using. I think this will make your Y axis much shorter then you think. I believe the general rule of thumb was max. tool length = 5*tool diameter, longer then that the tool needs to be of the dampened kind.

also way covers, here's a pic of way cover sizing by Krius. This has been most helpful for me.

https://i.imgur.com/zmcDPrj.png

[j2345]

Why a horizontal? They are good for batches of parts using a tombstone. But generally a vertical machine is easier for most uses

I believe a horizontal layout is better for my purposes. I don't plan on working any large flat parts that could be laid down on the table of a vertical machine, my planned parts are to be held in the rotary axis for side working, or held in setup blocks on the table for working on the ends. The parts being about 8" long, a vertical mill would require standing them up for working on the ends which would eat a big chunk of Z, while they could be laid flat on the table in the horizontal configuration. Additionally, the horizontal should be more rigid with a shorter and more direct load path, as well as easier to construct as there's no need to build and attach a large overhanging head to create the clearance for Y-axis table movement. It seems to me that main disadvantage to the horizontal configuration would just be the larger footprint, but you're the second person to mention this so I may be missing something important?

looking good, next step would be making semi transparent blocks for stuff like travel area, working area, collision boundaries etc. to see if things line up properly. Make sure to also include a block that represents the longest tool you'll be using. I think this will make your Y axis much shorter then you think. I believe the general rule of thumb was max. tool length = 5*tool diameter, longer then that the tool needs to be of the dampened kind.

also way covers, here's a pic of way cover sizing by Krius. This has been most helpful for me.

https://i.imgur.com/zmcDPrj.png

I'm not quite sure if this is how it's done, the images attached show a green block for working area and red for travel clearances. The working area is shown based on a 6cm stickout tool. Thanks for the way cover info, annoyingly it looks like they will eat at least 4-6cm of travel on the near ends, unless I can figure out how to create clearance for them without affecting stiffness somehow.

[ardenum]

It seems to me that main disadvantage to the horizontal configuration would just be the larger footprint, but you're the second person to mention this so I may be missing something important?

I'm not quite sure if this is how it's done, the images attached show a green block for working area and red for travel clearances. The working area is shown based on a 6cm stickout tool. Thanks for the way cover info, annoyingly it looks like they will eat at least 4-6cm of travel on the near ends, unless I can figure out how to create clearance for them without affecting stiffness somehow.

Actually a horizontal setup lets you utilize the space much better then a regular VMC in that its more narrow but can be longer, for me that's my preferred footprint, instead of the usual square. I agree with the stiffness point, personally I don't like the type of Z axis you see in routers. It's hard to make it stiff unless you want to pay extra for more machined parts or can fabricate if from steel yourself.

Good attempt as there's really no guide on this anywhere, here's mine: https://i.imgur.com/ZSK3uuS.png

red block is a boundary box for spindle collisions, the max volume the spindle occupies
blue is a boundary box for the thickest and longest tool I'll use.
yellow is work part that's held on the rotary

Have to say your frame is quite nice because you can change the spindle anytime and you can use an off the shelf spindle even something like this (https://ctbservo.com/product/motoriz...011-2500-12000) if you ever have the need for it.

also you have space inside the footprint, on the right side of the spindle for an ATC. With the moving table you can make enough clearance to mount a lot of tools and a toolchanger arm.

[pippin88]

For the specific setups you mention, a horizontal spindle makes lots of sense.

[j2345]

Okay, how’s this. Green representing working area with stubby 2cm tool and red representing spindle collision boundaries with the table at max plunge depth, and stock in yellow. I don’t have much experience but I think mastercam and the like allow you to input a simplified model of your machine and work holding, and will route tool paths to avoid collisions? Obviously I can’t achieve the full work volume without risking possible collisions with rotary, table, etc.

I’m glad to see some support for the horizontal layout, it seems to be less popular compared to VMC type builds but I’m convinced it does indeed make sense if your work holding allows it. ATC would be great but is not in the budget for now, if I find more funds later I would replace the spindle with one of those BT30 all in ones. Carousel would likely be located above spindle mounted to top of column, if there is vertical clearance above the machine.

Now onto electronics, playing around with the calculator at https://www.orientalmotor.com/motor-...ew-sizing.html I get <2Nm torque required even with the 10mm pitch screws, not sure how valid this is because I don’t know about accelerations or typical machining forces.
For steppers I am tentatively looking at https://www.aliexpress.com/item/4001034467781.html
2.2Nm with 1.3mH inductance. Kind of expensive for a stepper but it claims much less inductance than comparable Nema 23s, most of them seem to have >3mH inductance for that torque range. Makes you wonder if those claims are any good. I would appreciate everyone’s thoughts on motor sizing, I really have no idea about what’s adequate here. Also, if anyone has leads on cheap steppers and drives available in Canada that would be great. Thanks all for your feedback so far.

[ardenum]

Okay, how’s this. Green representing working area with stubby 2cm tool and red representing spindle collision boundaries with the table at max plunge depth, and stock in yellow. I don’t have much experience but I think mastercam and the like allow you to input a simplified model of your machine and work holding, and will route tool paths to avoid collisions? Obviously I can’t achieve the full work volume without risking possible collisions with rotary, table, etc.

Definitely, the best thing to do is import to cam and do some virtual milling, these blocks are basically just doodling that I use to make sure the workpiece is centered in the spindle work area, because setting up a postprocessor takes some time(a lot). In my file these blocks are parametric so if I change something on the frame they update accordingly so it's much faster to spot problems, at least for me.

Carousel would likely be located above spindle mounted to top of column, if there is vertical clearance above the machine.

Sounds good. There should be space for that because you can move the work part to the front when tool changing.

here's a hurco that's very similar https://db.hurco.de//files/products/...X80i-frame.jpg
if you google horizontal boring center, you'll find a lot of frame references for your build.

[mactec54]

Okay, how’s this. Green representing working area with stubby 2cm tool and red representing spindle collision boundaries with the table at max plunge depth, and stock in yellow. I don’t have much experience but I think mastercam and the like allow you to input a simplified model of your machine and work holding, and will route tool paths to avoid collisions? Obviously I can’t achieve the full work volume without risking possible collisions with rotary, table, etc.

I’m glad to see some support for the horizontal layout, it seems to be less popular compared to VMC type builds but I’m convinced it does indeed make sense if your work holding allows it. ATC would be great but is not in the budget for now, if I find more funds later I would replace the spindle with one of those BT30 all in ones. Carousel would likely be located above spindle mounted to top of column, if there is vertical clearance above the machine.

Now onto electronics, playing around with the calculator at https://www.orientalmotor.com/motor-...ew-sizing.html I get <2Nm torque required even with the 10mm pitch screws, not sure how valid this is because I don’t know about accelerations or typical machining forces.
For steppers I am tentatively looking at https://www.aliexpress.com/item/4001034467781.html
2.2Nm with 1.3mH inductance. Kind of expensive for a stepper but it claims much less inductance than comparable Nema 23s, most of them seem to have >3mH inductance for that torque range. Makes you wonder if those claims are any good. I would appreciate everyone’s thoughts on motor sizing, I really have no idea about what’s adequate here. Also, if anyone has leads on cheap steppers and drives available in Canada that would be great. Thanks all for your feedback so far.

Horizontal mills are good, your design should be ok there are other machines that are similar in design

Here is a 5 axis Horizontal https://pocketnc.com/ a little smaller than your build

I built added a Horizontal spindle to my Bridgeport CNC Mill some snips of my completed build has rigid tapping

[j2345]

I've added more detail to the column assembly; I will try to cast it with 3 threaded inserts in the base of the column, and use bolts as jacks to adjust it to sit square. Then there is a thru hole I will run an allthread down and hold the whole thing in compression to the base. Lastly, another thru hole from the back face to inject grout into the interface between column and base.

Definitely, the best thing to do is import to cam and do some virtual milling, these blocks are basically just doodling that I use to make sure the workpiece is centered in the spindle work area, because setting up a postprocessor takes some time(a lot). In my file these blocks are parametric so if I change something on the frame they update accordingly so it's much faster to spot problems, at least for me.
here's a hurco that's very similar https://db.hurco.de//files/products/...X80i-frame.jpg
if you google horizontal boring center, you'll find a lot of frame references for your build.

I'm not familiar with post edits, I think i will leave that for after the machine can move. Yes, the design is certainly inspired by those borers, I like how you can achieve a big Z just by lengthening the base. I do wonder if my 250x220mm column will have rigidity appropriate for milling type operations.

Horizontal mills are good, your design should be ok there are other machines that are similar in design

Here is a 5 axis Horizontal https://pocketnc.com/ a little smaller than your build

I built added a Horizontal spindle to my Bridgeport CNC Mill some snips of my completed build has rigid tapping

Good to hear more support for horizontals. Very nice looking spindle, I would be interested in seeing some workholding setups you use with that.

In other news, I have ordered the linear actuators, total cost 530USD for a 300 and 400mm stroke. Now looking at aliexpress HGR20 and ballscrew sets, is it a good idea to match the 10mm pitch of the actuators or should I save 40$ and go with a 1605 ballscrew set? Will likely be using those 2.2nm steppers at 48v. Of course, perhaps the cheap generic rail sellers will be sending worse quality product than the more expensive listing. I have no idea if I need fast plunge speed, but I have a small diameter, high L/D drilling op that might require some fast feeding or retraction.

[ardenum]

I do wonder if my 250x220mm column will have rigidity appropriate for milling type operations.

For your combination of the steppers and spindle the column is plenty strong.

[j2345]

Haha, I'll take it. Any thoughts on 5mm vs 10mm pitch screws?

[mactec54]

Good to hear more support for horizontals. Very nice looking spindle, I would be interested in seeing some workholding setups you use with that.

Here is a good video of different work holding you can use vices angle plates anything really to get the job done and sometimes not that much difference then a vertical mill, except you can have better chip control

[ardenum]

what about the 4th axis rotary, are you making that yourself?

[j2345]

Here is a good video of different work holding you can use vices angle plates anything really to get the job done and sometimes not that much difference then a vertical mill, except you can have better chip control

Thanks for the info, it seems like various ways of clamping the top surface to the table are pretty simple.

what about the 4th axis rotary, are you making that yourself?

Yes, current plan is to cast an angle out of concrete, face and bore after the machine is running 3 axis, then bond in an aluminum sleeve to hold a harmonic reducer. Currently lurking ebay searching for a good and cheap one, I really want a thru-spindle bore on the reducer so I can tighten an expanding arbor from the rear of the head stock instead of having to go through the part. I have not done any CAD on this yet but it should be pretty simple once I find a suitable reducer. Hopefully a cheap NEMA17 I have laying around will be fine for this at 50:1 reduction.

[ardenum]

Thanks for the info, it seems like various ways of clamping the top surface to the table are pretty simple.


Yes, current plan is to cast an angle out of concrete, face and bore after the machine is running 3 axis, then bond in an aluminum sleeve to hold a harmonic reducer. Currently lurking ebay searching for a good and cheap one, I really want a thru-spindle bore on the reducer so I can tighten an expanding arbor from the rear of the head stock instead of having to go through the part. I have not done any CAD on this yet but it should be pretty simple once I find a suitable reducer. Hopefully a cheap NEMA17 I have laying around will be fine for this at 50:1 reduction.

So you will put the stepper not inline with the reducer but I'm guessing run it with a belt from the side or top? Make sure the reducer is big enough for your application, you wanna pay attention to torsional stiffness. You might need a break, but don't take my word for it.

There are geared reducers where the hollow shaft isn't obscured but they are quite expensive starting at $600.