[Siddvo]

Hi there,

I'm fairly new to the CNC world and currently in the process of constructing my own CNC Mill. I'm using a Box steel frame filled with epoxy granite or concrete; I haven't decided yet. Right now, my main focus is on designing the frame, specifically the X-axis. I'm drawing inspiration from the Grizzly G0759, but I'm building the frame from scratch and opting for linear rails instead of dovetails.

As I explore various DIY CNC mills online, I've noticed a common approach where the rails are fixed and not moving in the X-direction, but only the bearing blocks move. Similar to what Seabass Engineering does in his build: link. I hope it is clear what I'm trying to say; English is not my first language.

I'm grappling with the decision on the best design option. Initially, it seems that the first approach might offer more rigidity. The spindle always stays in the middle of the bearing blocks, ensuring maximum rigidity, and it could potentially allow for longer travel with the same length of rails since the bearing blocks don't need to be as far apart.

However, I can't help but wonder why this design choice isn't more widespread. Am I missing something crucial here? I'd greatly appreciate any advice or insights you might have on this matter.

Thanks.

[Bdrangdo]

I have just posted a practical working design for a VMC using standard compononents and linear rails on Grabcad. It was modelled in Solidworks but I also included step files that shoud work on any cad program. I am planning to build this myself. The frame is fabricated from mostly 10mm plate. It is free to download from Grabcad. Dave

[peteeng]

Hi Sid - No link, a diagram or screen shot of the design would be good. Peter

[Siddvo]

Sorry, I must have forgotten to paste the link: https://youtu.be/LAUtoVJ5m4k?t=556

That's how it's looking so far. Picture attached.

The two other pictures are the two designs for the X and Y axes. I hope it makes my point more clear."

[peteeng]

Hi Sid - The saddle is always a bit of a brain drain to design. Conventional arrangements come from the time when friction centers had to be aligned with drive lines and ways had to be machined and had slick slip. Now with rails and cars these two issues have dissolved. So you can have static rails or moving rails up to you. Your spindle outreach will need to have extra support. Its surprising how much flex is added by a large flange like you have designed. You need to have the outreach as close to the cars as possible or include webs that run onto the top of the cars if possible. Personally I'd stay away from concrete or EG filled steel structures they don't give you much more stiffness. Just use very thick hollow sections much better. Plus once you fill them with that sort of material you can't drill through them to add that bracket or box you will want to add next year. I see you are using plate for some parts. Plate is a very good way to go for the whole machine. Don't use SHS or RHS anywhere. This allows you to tailor the local thickness and do things that hollows are limited in doing. Edge bolting thru thick plate creates a very good connection... A fully bolted machine is a very good thing to do for this style of machine. Welding creates many problems. If you have access to a mill that can make good flat surfaces then your on the right track.

Personally I like the job to stay still and the machine to move. So look at moving column designs, this allows the machine to be better optimised as the table does not move...Here are two Mori machines for instance. I like the LHS arrangement... Thats a Mori M1 I like it because the X axis mechanics are up out of the muck. You will need to shield your ballscrew and motor etc if you put it in the bottom of the machine....Peter

[peteeng]

Hi Sid - You say its a "Mill" what do you intend to machine with it? You have modeled a high speed spindle on the machine which implies its more of a router than a mill. If you intend to mill steel then the spindle is the wrong type. If you intend to mill aluminium then thats OK as long as you get a spindle with very good bearings. Since you are in design mode consider other configurations then the bridgeport or single column type. If you are doing metal work you need a very rigid machine and the single column type exists because of historical limitations (200 years ago) that do not exist now and they are not particularly rigid. There are high rail designs, moving column designs, portals etc that maybe better for your job... So tell us what you want to do and there maybe a better way... Peter

[joeavaerage]

Hi,
my mill, I designed and built over about three years, and has been in service for 2.5 years now. It looks very similar to the middle of the three pics.
That is the linear rails are fixed to the table and 'upside down' and the bearing cars are facing upwards and mounted on the saddle.

My previous mini-mill was like the last pic you posted. I always had problems trying to keep coolant and swarf out of the works. I decided therefor the the
'upside down' approach would go a long way to solving that issue. The second consideration was a peteeng describes the saddle if very hard to design and build such that the X and Y
axes are exactly perpindicular to each other. The arrangement I chose was to have a saddle on 'top' of the Y axis and another saddle 'underneath' the X axis. The two saddles could then be bolted and even
adjusted if required. In short this approach was a vast simplification and improvement over my first design mini-mill years earlier.

The last feature, ie the simplification of design of the saddles has proven to be a winning combination for me. The first pic is the X axis in normal mode, ie with a sheet metal skirt in place to
defeat coolant etc as best as possible. The next two are with the skirt removed. You can clearly see the bearing cars facing 'upwards' and the rails 'downwards' You can also see the topmost of the
two saddles, which are no more than two 275mm x 275mm x 20mm steel plates ground flat and smooth in a surface grinder....so simple.

Craig

[peteeng]

Hi Craig - Do you mill steel with that spindle? Peter

[Siddvo]

I intend to primarily cut aluminum but also steel. The spindle is just a model I already had downloaded; it's merely a placeholder, and I haven't decided on a specific spindle yet.

What are the limitations? Is a gantry-style mill more rigid? Why are most commercial CNC mills still single column?

[Siddvo]

Hi joe, yes, that's what I was planning to do as well. How is it with maintenance? Are you using oil or grease on the rails?

[joeavaerage]

Hi peteeng,
yes I do sometimes. Its 800W 24000rpm aircooled, so its a very VERY poor solution for steel.
If I use a 3mm tool at 12000 rpm -15000rpm (113m/min- 141m/min) with plenty of coolant, its slow but it does work. I use this little spindle daily and have done for years and for business. I CANNOT afford to blow it up...
so I treat the occasional steel job with great caution....if I'm of the opinion it puts my little spindle at risk I refuse to do the job.

As you may know I made another spindle based on a 1.8kW 3500rpm servo, 6.1Nm cont and 18Nm peak. Its great for steel and stainless. Unfortunately I wrecked earlier in the year
by allowing coolant to get into the optical encoder. I have bought a very rough, but hopefully still functional replacement servo. It's in customs as of this morning so I should have it by the
the end of the week. Came from California.

Its a bit of a pain swapping the spindle over, but what can you do? At least short of spending ten of thousands on a spindle. High speed low torque spindles are great for many things but hopeless in steel.

Craig

[joeavaerage]

Hi,
a squirt of grease every few months is adequate, at least with good seals on the bearing cars.

Why are most commercial CNC mills still single column?

Several reasons, not all historical, but unlike peteeng single column machines can be made as rigid as any other......but it is true you have to design and analyse well to achieve a stiff design.
One other reason is that most machines are still made in cast iron....and cast iron is a sure sign that God does love humans after all, and single column machines tend to be cast in two pieces,
the upright column and the base. They are essentially long and narrow which is convenient for casting.

In my case I had three axis beds cast in iron, and then precision machined, the cost nearly killed me, be the result is very satisfying. Thereafter I was to bolt the cast iron beds to a frame.
I would ideally cast the frame but the cost would be enormous so I made a frame out of 32mm plasma cut steel. I reasoned that the cast iron axis beds were going to be as good as I could possibly afford to make
them and they would form the basis of my machine for the next twenty years. The frame may change, but the axis beds will not.

To date that strategy has served me well. The philosophy that resulted in this strategy was that if I design well I can achieve good results (accuracy, rigidity) without requiring hugely expensive parts
or materials and equipment. In those areas where there is no choice but to use the best materials and techniques to get a good result, is do so to the best of your ability, I won't say irrespective of cost,
but at whatever cost necessary to get the result.

Craig

[peteeng]

Hi Sid - As Craig infers you will need another solution if you intend to cut steel often and reliably. I wouldn't say "most" commercial machines are still single column. If the company is casting machine parts and they work they have little reason to change. Sure you can make a single column as stiff as you like... sort of. But now's the time to explore other solutions that over time maybe better for you. Building a machine takes time and effort so you may as well start at the "right" place. You shall consume lots of time sorting out a steel spindle... Peter

[joeavaerage]

Hi,
peteeng is right, milling steel 'separates the wheat from the chaff'

To cut wood and plastics might require rigidity R. To cut brass and aluminum requires 5R. To cut steel requires 20R minimum.
If you are designing to cut steel then the cost will be something like five-fold over cutting aluminum.

Amongst the costs are the spindle. To cut steel requires low speeds but high torque.......yet much of your machining will be done with high speeds and low torque.
You CAN buy a spindle that will do both but they are huge, heavy and expensive, say 10kW and $5000 to $10000.

The only solution I could come up with on a hobbyists budget was to have two spindles. Typically the spindle (or spindles) is the single biggest expense of any build,
you need to think about that first. The rest of the machine will fall into line with the decisions you make regarding the spindle.

Another important choice is size. Most newcomers, myself included, want big......but that is a mistake. Big costs a fortune. If you make a small machine you can make it exceptionally stiff, but
to get that same stiffness in a large machine requires HUGE parts weighing untold and costing even more. My first mini-mill had travels of 180mm x 180mm x 180mm, and I used it for nearly nine years.
My current machine has travels of 350mm x 350mm x 350mm.

I would estimate that to double the size of the machine will increase the cost tenfold AND retain the same stiffness. Be aware the extreme cost penalty that comes with increasing size.
Being very clear eyed about what goal you want to achieve is an essential skill with CNCing. So many machines start great but because of budget all sorts of bad compromises had to be
taken later.

Craig

[peteeng]

Hi Sid - Its worth investigating the steel spindle first. Its likely that you will need a pulley system that will take up space so the current placeholder won't work. Attached are images of a plywood machine I built that flies through plastics and timber and aluminium. But I expect it would struggle on steel. Its an 800W water cooled 24000rpm spindle. Its been sold to a fellow that will make fibreglass surfboard fins with it. So get familiar with steel parameters pretty quick so you sort the spindle first... This machine is called Frankie and its sort of a prototype for an aluminium plate build. So all the ply parts would become 16mm al plate. It would be remarkably stiff as the ply is very very stiff. 17mm ply is equivalent to 10mm aluminium. But the ply machine can't be a wet machine ie flood coolant which you will need also..... Peter

[joeavaerage]

Hi,
another idea that might be useful is considering the upgrade path. Budget may preclude you from doing exactly what you want, but maybe you can work up to it.
For example the spindle. On the basis of cost a high speed low torque spindle is very compelling, and you'll do swags of work with it and they are very well priced.
Its not good in steel, so maybe down the track you get yourself a low speed high torque spindle. If you have thought it through, like leaving the available room and
having the extra Z axis thrust to accommodate a heavier spindle then swapping at some later date is a breeze.

One area that I would caution is not amenable to upgrade are the linear motion parts, the ballscrews particularly and to a lesser extent the linear rails and cars.
The reason is that if you decide on a lowly C7 ballscrew, with the idea that you'll upgrade to C5 or C3 later then you are restricted to getting the exact C3 ballscrew
to physically match the existing C7. That will cost a fortune. The linear rails are somewhat easier to replace but the same general principle applies. For this reason I would suggest you get
the very VERY best ballscrews and rails/cars you can afford right from the outset with a view that they NOT be upgraded later, but last the life of the machine.
This is reflected in my design where I bought the best I could get (ballscrews and linear rails) and mounted them in precision machined cast iron beds......they are not really suitable for
upgrading....they are what they are, and will remain so until being scrapped twenty years from now.

Steppers are a reasonable candidate for upgrading to servos. You can simply swap a 23 size stepper with a 23 size servo, or a 60mm stepper with a 60mm servo.

Things like servos, spindles, coolant pumps and filters, adding a fourth of fifth axis can, with a little bit of forethought at the design stage can be swapped out easily and cheaply.
Linear motion parts not so. That leaves the frame. If the frame is integral to the linear motion parts, that can be a problem.

For instance you may have noticed in one of the earlier pics of my machine that the top of the column (two L shaped pieces of 32mm plasma cut steel) extends 300mm above the top of the
Z axis. This is so when I put my trunnion fifth axis (recently completed, a year long endeavour costing thousands) on the table I can unbolt the Z axis bed and 'shift it upwards' and so regain the
Z travel lost by piling the fifth axis on the table, I did that deliberately....because I knew that I wanted a fifth axis at some point so I made the frame just that bit higher to accommodate it when I did.

Craig

[Siddvo]

Peter, Your machine looks really cool! Is the frame entirely made of plywood? The problem i have with gantry-style machines is the minimal Z-travel, but yours seems to have a good amount, along with a massive frame. Plywood is not an option, as you mentioned it's not compatible with flood coolant. If I have to make the frame out of steel or aluminum, it will likely be very heavy and probably expensive due to its size.

I was hoping I could use a high-speed spindle for aluminum, and at least have enough torque at low RPM for small end mills in steel. Vfd controlled spindles with encoder feedback ar supposed to have torque even at low rpm. I thought that would be enough. The Placeholder spindle is 24k rpm 2.2 KW


Craig, I am planning to ad a 4 axis to somwere down the road. not a big one so i need the z travel. currently in the model there is about 300mm of z travel possible while if the spindle is at the most bottom position the collet nut is about 10mm from the table. i hoped thet would be enough.

Im using hgh20 rails and rolled sfu16 ballscrews(I already had them). Cant i just upgrade the same model just a higher class?

I have 4 nema 23 steppers, which i hope have have enough "umpf" to move the Axis, i plan on using for now but at some point i will probably upgrade to servos for faster rapids but thats just not in the budget. For now i want to get the frame right because as you sad its not easy to upgrade. The z is probably going to be to heavy but i hope i will be able to get away with it by asisting the stepper with some counterweights.

[joeavaerage]

Hi,

I was hoping I could use a high-speed spindle for aluminum, and at least have enough torque at low RPM for small end mills in steel.

Your hope will be cheated. The slowest you dare run even a water cooled 24000rpm spindle is 6000rpm. It will have its rated torque of 0.87Nm, but not more.
That suggests the biggest tool you could use in steel is 6mm, as it results in a surface speed of 113m/min, which is about the max for uncoated carbide in
mild steel. Personal experience suggests that 0.9Nm is either inadequate or marginally adequate.....you'll be down to featherlight cuts. Any more than that
and the spindle will stall and you snap the tool....YET AGAIN!!!!. Ask me how I know!

I am planning to ad a 4 axis to somwere down the road. not a big one so i need the z travel. currently in the model there is about 300mm of z travel possible

If the centre line of the fourth axis is 100mm above the bed then your travel is reduced by 100mm, so only 200mm left. Can you not arrange to shift the Z axis up a bit when you get around to
a fourth axis?.

rolled sfu16 ballscrews(I already had them). Cant i just upgrade the same model just a higher class?

In most cases ground C5 and C3 ballscrews have different physical dimensions to C7's and its not possible to do a straight swap. There are of course Chinese suppliers selling
(claimed) C3 screws that are identical to C7's and do not cost that much more.....lying, cheating, BS b********ds! You could perhaps get a C5 or C3 made for you to match
the existing C7....but it will cost thousands....its just not practical to do it. If you want C5 or C3 type accuracy then do it from the start and make your machine to accommodate
the ballscrews you can get, not the other way around.

I was lucky and bought three 32mm BNFN double nut C5's by THK, perhaps the most prestigous manufacturer, near new second hand. They had six FK25 support bearings included.
Including shipping to New Zealand I paid $1000USD for them. If I had bought new I might have expected $3000USD each or $9000USD total! This purchase made my machine build possible.
The machine dimensions are made to get the best from these ballscrews, as they are the heart of the machine. I could not complete the design UNTIL I had them, so when I say 'I made my machine
to suit the ballscrews'; that is the literal truth.

I have 4 nema 23 steppers, which i hope have have enough "umpf" to move the Axis,

If the steppers are low inductance and you use high voltage drivers, say 80VDC or even more, then its entirely probable that stepper will be all you need. My first mini-mill had
Vexta steppers matched to low lash planetary gearboxes, and they were superb. OK they were a bit slow (1500mm/min) rapids but they had huge thrust and acceleration.
Given that it was a small machine, modest or even slow rapids are not a deal breaker. Having high reserve thrust and high acceleration is the gold standard.

My new mill has 750W servos, and they are scary fast, I seldom use more than 1/3 of their 25m/min speed. They eat any stepper ever made....but they cost, I had the budget at the time,
and it was always my intention to use servos so I did it. Despite the cost I will not be going back to steppers any time soon.

Craig

[peteeng]

Hi Sid -
1) at 6000rpm your 24k spindle will stall even with a small drill . Milling will be unreliable.
2) Plan your rotary in a forward or side pit. They take up considerable Z height and if not planned well, your machine will not be able to be upgraded easily. I suggest you do it now. Most people I deal with that say they want a rotary never implement it or it takes years to implement and by then they move onto a different machine.
3) The high rail design with walls has proven to me to be very stiff, even in MDF. So its a std approach for me now. Z=350mm but to have a long Z means you need a really stiff Z axis and saddle. It took me 3 goes to get this right once I went over 250mm. I make moulds and things that need a lot of Z
4) You will see in the Scoot images that scoot is steel. Mainly 3mm steel. Next one will be 2mm. I tend to prototype in plywood and make production machines in steel or aluminium. The ply bed and walls in Scoot can be replaced with plate so then it can be flood cooled. Next one, always a next one...
5) Where your heading is a heavy machine no matter what. Don't worry about weight up front, get the machine basics correct and future proofed (at least on paper) then review and cut weight... around version 20 or 30 you start getting to the real deal....
6) also with a 2.2kW spindle make sure you have the correct amperage available for the inverter. If you do not have the correct amperage, you will not get near 2.2kW! But to actually use 2kW is a big tool with a big cut...
Peter

[joeavaerage]

Hi,

Most people I deal with that say they want a rotary never implement it or it takes years to implement and by then they move onto a different machine.

Guilty as charged. I always wanted a fourth axis for my mini-mill, even bought a great big 'you beautie' servo reducer as the gearbox....but never got around to it.
In truth the gearbox was overlarge for the little mill. It sat around for nearly eight years before I dusted it off and made a fourth axis for my new machine. Its a near perfect
size (and rigidity) for my new machine. Even more recently I took the chuck off and built a trunnion table and a fifth axis, already posted the pic.

Took a very long time between the 'I want.....' stage and the actual spending the money and building it to the 'I now have.......' stage. A very significant part of the motivation for
doing it at all was I had set myself a challenge......so there is a personal satisfaction reason to complete it. I'm very glad that it is done because now I can progress to other matters,
repair my 'steel' spindle, a new cast iron headstock, a new electrical cabinet, work on my 3kW servo driven ATC spindle, new coolant pump, filters and resevoir....maybe through spindle coolant.

Craig