CNC Router Vertical 4' x 8'

Hi,

I intend to design and build a vertical (or near vertical) 4' x 8' CNC router that is sufficiently beefy to mill aluminium or take deep and wide cuts in wood.

I currently have a 2' x 4' flat bed router but my workshop is too small to go 4' x 8' flat bed, hence the vertical design.

I am greatly inspired by this build by Dean (JazzCNC over on mycncuk forums) but I hope to make it a little more solid and also faster if I can.

[https://youtu.be/4STVbn_EMoc

So to get started I have chosen my spindle and drive motors.

Spindle: I intend to use a 3KW ATC water cooled spindle. Going ATC is hugely important for me. I have a 2.2Kw spindle now and I never use all it's power so I feel confident the 3KW will be enough even with a more ridged machine.

Drive motors: I intend to use 3 -phase closed loop Nema 34 steppers with either 9nM or 12nM torque. Dual on the moving gantry axis.

Attachment 460696

I did consider AC servos but advice to me was that their tuning can be very sensitive and easily get upset by backlash or resonance in indirect coupling like belts. I plan to use (admittedly short) belts to couple motors to ball screws and rotating ball nuts. The closed loop steppers are more plug and play and will easily provide the speed and acceleration I need while leaving more money for things like spindle pneumatics and vacuum table. Plus 3-phase motors are pretty smooth and quick.

I've started by calculating inertia and acceleration for the dual 9Nm drive for the moving gantry via rotating ball nuts. Each motor has a rotor inertia of 3Kg so 6Kg total (I am ignoring rotating ball nut inertia). I intend to use 2525 ball screws (25mm pitch) which means a gantry mass of 40Kg would provide an equal split of inertia between the drive system and the moving mass. I see no harm in pushing the moving mass inertia slightly above the drive system inertia, so I think this gives me a target total moving gantry weight inc. Z axis of about 60Kg-70Kg. The spindle is 14Kg so I want to aim at a gantry beam mass of about 30Kg-40Kg.

I have made some static stress simulations for 1.5m gantry beams to see which shapes and materials give the least deflection.

Out of the ones below it seems that ali box section tube with additional wall thickness added by ali plate epoxy bonded to the main beam gives the best results. Damping as well as stiffness is also desirable so I am keen on the constrained layer of epoxy adding damping. Ali tube with epoxy granite fill was my favourite before trying that.

Attachment 460698
Attachment 460700
Attachment 460702
Attachment 460704
Attachment 460710
Attachment 460708

So now I have my gantry beam design I will start designing the X and Z axis.

I quite like the idea of using HGH35 rails with 2525 ball screw for the X axis at least because they are large enough to naturally give clearance for the ball nut. Is there any downside to using larger rails like this rather than say HGH25, other than cost? More friction?

I also want to ask, will it be a bottleneck to use slimmer rails on the Z axis than the X axis? They will be shorter rails but that won't change how strong the block grabs the rail. So I think if I use HGH35 on the X axis I should use the same on the Z axis. Correct?

Hi Simon - each axis is different and there is no reason to use the same rail on all axis. Its convenient in terms of ordering and modeling however. Plus you can rotate cars from the z when they wear to other axis if they are the same size... As a rule use the biggest rails that meets the budget. 35mm is considerably stiffer then 25mm. Looking fwd to your work...Peter

ask the motor supplier for a torque/speed chart as well...

I should have replied. I do not have a specific stiffness goal, but just making comparison between practical options.

Anybody got some design ideas for a retractable dust shoe for ATC spindles?

Working on the Z-axis design now and I want to grab the spindle as near the cutter as possible for stiffness, but I'm not sure how to do a retractable dust shoe without having the cylindrical spindle body sticking out a fair bit.

I know most retractable dust shoes probably use pneumatics but I was thinking to use a nema17 stepper with threaded rod and a micro-controller to monitor the height of the material surface and dynamically adjust the dust shoe height. Good idea, bad idea?

Attachment 460992

you can create a 2nd lighter gauge Z axis for the dust shoe behind your gantry or you can just level and set it manually for each job

https://i.imgur.com/gd8w41j.png

Hi Everyone!

I got a bit of time over the Easter weekend to work on this design.

I created a dust extraction shoe design which I will 3D print.

I also worked on the frame. I'd appreciate feedback on this in case I have missed any axis of flex that need bracing or you just have a better suggestion?

Considering the size (1.8m tall and 2.8m long) and the heavy 100 x 100mm 5mm wall tubing, I would never move it again once it's welded together so I am planning to bolt sections together and use bracket plates (5mm steel). Do you think M16 bolts will be big enough?

Attachment 461564
Attachment 461560
Attachment 461562

FEA simply to highlight modes of flex, not for an absolute displacement figure.
(I'll bolt it to the floor)

Attachment 461566

Hi Sash - The diagonal bracing is not needed at the back. This plane is outside the loadpath for machining so is not efficient use of material. Better to put the weight into the machine bed vs the frame. Design it in two pieces A) The machine bed and B) the frame. In this way its easier to break down for transport and its easier to true the machine bed vs the entire frame which will be massively stiff... The frame can be quite light but the bed has to be stiff.. Peter

There are two considerations in the bed 1) plunging stiffness and 2) XY stiffness. If your only doing profile panel work then XY is important. If your doing moulds and lots of heavy plunging then you will need more cross members to create a stiff bed....Plunging takes a lot of load!

The dust extraction looks fancy but will actually be very poor.

Will you be using high pressure low volume (vacuum cleaner) or high volume low pressure (dust extractor)?

You will never remove all the fine dust (that actually affects health) with a vacuum cleaner.

Quote:

---

Originally Posted by pippin88

The dust extraction looks fancy but will actually be very poor.

Will you be using high pressure low volume (vacuum cleaner) or high volume low pressure (dust extractor)?

You will never remove all the fine dust (that actually affects health) with a vacuum cleaner.

---

Can you expand on why you think it won't work well?

To be clear, it will have a brush attachment and move down to touch the work piece. It's not going to sit up in the air like that.

I use a Yorkleen WV2000 - https://www.yorkleen.co.uk/product/wv2000/

Airflow Rate 125 l/sec (+/-5%)
Filters 3 stage, up to 0.5 micron
Suction 2500mm WG (+/-5%)

Quote:

---

Originally Posted by peteeng

Hi Sash - The diagonal bracing is not needed at the back. This plane is outside the loadpath for machining so is not efficient use of material. Better to put the weight into the machine bed vs the frame. Design it in two pieces A) The machine bed and B) the frame. In this way its easier to break down for transport and its easier to true the machine bed vs the entire frame which will be massively stiff... The frame can be quite light but the bed has to be stiff.. Peter

There are two considerations in the bed 1) plunging stiffness and 2) XY stiffness. If your only doing profile panel work then XY is important. If your doing moulds and lots of heavy plunging then you will need more cross members to create a stiff bed....Plunging takes a lot of load!

---

Hi Etee, :p

Do you mean the bed or frame doesn't need cross-bracing? I think it must, because there will always be un-even loads in the frame and especially there will be a resonance loops that goes around the outside of the bed making it trapezoid.

If you simply mean it's best to cross-brace the bed rather than the frame, I did start that way but it actually produced 35% stiffer results in FEA and used less material to brace the back of the frame than the bed.

Despite that I think maybe it is better to cross-brace the bed like you said because that will also provide a better supported cutting surface and secondly I can use (or sell) the router as a horizontal system if I ever want to.

Attachment 461624

Quote:

---

Originally Posted by ssashton

Can you expand on why you think it won't work well?

To be clear, it will have a brush attachment and move down to touch the work piece. It's not going to sit up in the air like that.

I use a Yorkleen WV2000 - https://www.yorkleen.co.uk/product/wv2000/

Airflow Rate125 l/sec (+/-5%)
Filters3 stage, up to 0.5 micron
Suction2500mm WG (+/-5%)

---

I recommend you go read up on dust extraction.

Bill Pentz is a good source. Dust Collection Research - Home
Woodworkforums.com have a lot of info on dust extraction also (Look for BobL).

You have a HPLV system.

Your vacuum has a stated flow of about 264cfm. That will be with no restrictions and is probably way overrated.

You need about 1000cfm at the dust source (tool) to collect fine dust. Fine dust is generally not visible when suspended in air, and behaves somewhat like a gas rather than a solid. You need to remove the air.

If you just want to collect chips (the large bits that make a mess but do not affect health), then most attempts work ok.

Air is subject to friction. You have a lot of curves, some of which appear cosmetic.

Round smooth tubes are much better than other shapes (area vs perimeter - less friction).



(If feasible high pressure high volume systems existed, they would be great but unfortunately it takes energy to generate flow and to generate suction. Doing both takes a LOT of power)

Hi Pippin,

Thanks for the link I will have a good look there.

Initial reading suggests to me the recommended 1000CFM is based on a free space calculation, like maybe a hood over a sanding station. If the tool area is tightly enclosed by a shoe with brush that touches the work piece, then the story will be very different. Air speed inside the confided area will increase significantly.

I will consider putting a chamber around the entire machine.

Quote:

---

Originally Posted by ssashton

Hi Pippin,

Thanks for the link I will have a good look there.

Initial reading suggests to me the recommended 1000CFM is based on a free space calculation, like maybe a hood over a sanding station. If the tool area is tightly enclosed by a shoe with brush that touches the work piece, then the story will be very different. Air speed inside the confided area will increase significantly.

I will consider putting a chamber around the entire machine.

---

You have to have flow in to replace the air sucked out.

If you could make an airtight seal with the material and then have an inlet equal in size to the vac inlet then you could catch most. But this is never going to happen.

Lots of tools have poor flow in and so the dust extractor ends up starved for flow.

A vac system (HPLV) will not drop flow as much with restrictions but simply does not have the flow volume to catch fine dust.

Whether this all matters depends partly on exposure. If you are exposed to wood dust for only a couple of hours every week or two it may not be a big deal. If you spend hours most days in that environment you have much higher risk.

The cheapest simplest thing to do is to increase ventilation of the whole space. Wide open doors / windows, cross ventilation. Easy in mild climates, not so good in extreme / cold climates

In terms of your proposed design:
What is the inlet size for your vac system? 50mm? You want that size, in smooth tubing/ducting right to the cutting area if you can. (Too big tubing will result in chips dropping out of airstream and building up). Smaller ducting will reduce flow.


I am not an expert in dust collection at all. Just sharing some things I have picked up from reading.

Hi Pippin,

Lets calculate the air speed around the tool.

(EDIT: Oops made a mistake - updated now!)

The website says we need 3000FPM (feet per minute) to keep all debris suspended in ducting.

I have 4" main ducting and a 3" drop to the tool. So we can max the 3" duct CFM at 190CFM, but lets be conservative and say 150CFM at the mouth of the dust shoe.

The area enclosed by the dust shoe and brush is 0.3ft²

150CFM / 0.3ft² = 500FPM

So you are right that it will not remove all debris and dust. However since it is enclosed on all sides other than the brush where there is positive air flow towards the extraction point I don't think fine dust will manage to escape.

Hi Pete,

This seems to work very well. Using 50 x 100 rather than 100 x 100 for bracing with a more dense pattern. Should be cheaper, nice and stiff and also supports the bed surface more evenly.

Attachment 461742

What would people use for a primary bed surface?

I'll use vacuum hold down most of the time so there will be a secondary MDF surface on top.

My initial thought is 1" birch plywood. Good and stiff and stays pretty flat. It's a little inconvenient that the machine is larger than a 8' x 4' sheet which means it would need to be in sections! I was wondering about stone or concrete slabs like big floor tiles?

Hi Sash - Yes thats good but I would construct the bracing with end tabs and bolt them in. By the time you weld that design out it will warp and your bed won't be flat. Also by tabbing you will be able to adjust the flatness of the bed.... Or 30% or 50% weld don't 100% weld. Peter

Hi Pete,

Yeah it is indeed my intention to have L brackets that fix each X section to the main rails to allow slight adjustment. I'm no expert welder, lucky if I can get a bead without holes in it.