[Infr1234]

Hi,
I'm interested in getting in to hobby CNC milling with the end goal of producing a CNC able to make small parts from aluminium. I've started this thread to get an idea of what sort of things I ought to do along the way.

My intention is to build something small and reasonably portable for cutting auminium parts. I'd hop to achieve a precision comparable with or slightly better than that achieved with typical consumer 3d printers, I am not after extraordinarily fine tolerances for ultra-precise parts. I'm interested in being able to mill aluminium simply because it can make parts so much stringer than 3d printed plastic, things like gears, shaft couplers, small mounting brackets and plates...

I have a lot of 3d printing experience but CNC is something I've mostly just seen discussed online. A lot of what I've seen online is geared towards big routers for wood cutting, or very heavy knee mill type machines cutting steel parts. I'm particularly interested in making a personal project of building my CNC machine, more satisfying than buying one, and also far more able to tailor it to my use cases.

I thought I'd start with a few questions, focused mostly on the mechanics, I've some thoughts about CAM software which I'm working through at present, better to work out a few more things there before making that the topic of discussion.

I've a list of questions below of things which have occured to me so fr:

1. Spindle head. I can see these available online as large chinese made motors with fitted collets and often coming with a drive electronics box to convert mains power to high voltage DC and give speed control. But I've also seen a lot of CNC machines built using something like a dremel or corded router tool. Which is the better option? Some of the videos I've seen seem to sugest that even for milling metal the power tool router type can be better as the bearings are lower down and more able to handle side loads. My use case is milling Aluminium.

2. Holding the workpiece. How is this usually done. Clamping a block of metal would seem fine for most of a part, but at the very end you've got a situation where the CNC is cutting away the very last pieces of metal which hold your machined part to the bigger block, at which point I would guess the part would come loose, collide with the milling bit in all manner of funny ways and wreck everything. I've seen examples of wood cuting CNCs using a spoilboard with self-drilling screws going up from beneath to anchor in to the part, so the part is held from below. This ofcourse leaves screw holes in the bottom of your part and isn't so practical for screw-driving in to aluminium.

3. Liquid profing for coolant? Is coolant essential from the very most basic CNC design upwards? Does that mean everything below the working area of my CNC needs water(or other coolant liquid)proofing? Does one effectively need to have a small tank structure around the bed to contain coolant and stop it flowing on to stepper motors lower down. The same goes for catching chips from the milling and ensuring they don't end up in the leadscrews/ballscrews?

4. General machine geometry, for cutting small 2.5D and 3D (no undercuts) parts out of aluminium what is the best form to use? 2 axis "bed slinging" with the spindle moving on z only? Spindle moving on x, y and z(assumedly a bad idea as stacking so many moving axes after one-another ruins stiffness and puts much bigger weight loads on the first one to carry)? Gantry router(surely for big workpiece requirements only?)? Workpiece moving on x or y, spindle moving on z and the other of the horizontal axes?... I'd be expecting to construct the CNC from aluminium extrusions, 3d printed brackets, and I can probably order laser cut metal plates from an online service too. All construction would be with machine screws, no welding. Depending on advice I could open some future threads in the T-slots forum about this.

5. Milling bits and behaviours. Is it the case one needs very different types of motion up milling and down milling? Dotoolpaths all have to account for whether the cut is up milling or down milling at any given point? As well as many bits being unable to do plunge cutting? This must really complicate toolpath planning if you have to swap bits every time one needs to start cutting a new pocket? Or does this only apply to larger bits, I'd be expecting to probably work mostly with 1mm and 2mm.

6. Electronics control board. I understand that milling is a much more user interactive process than 3d printing. A printer you just load a file from an SD card and it spends hours moving around to print it, but I understand a mill needs a lot more manual processes at points during operation. With a printer you can just run the same gcode again for a copy of the part, whereas I understand with milling you're often looking at very different gcode for the same part if milling it from a different sized initial starting block for example. How does this affect the choice of control board? My understanding is that the main open-source options are GRBL running on a microcontroller versus LinuxCNC which needs a full computer dedicated to it(single board computer like a Rasp Pi?), what fully are the advantages and disadvantages to each.

Thank you

P.S. Forgot to mention, when I say small I'm imagining I'd never do any single part > 10cm x 10cm x 5cm (z), although if it feasible i'd be interested in getting a working volume of 15cm x 15cm x 15cm (might open opportunities for making parts I'd never previously considered). Most of the parts are probably going to be sized within 5cm x 5cm x 3cm (z). I'd be after a precision somewhere similar to or slightly better than typical budget 3d printers, somewhere around 0.05 to 0.1 mm, tighter if it is easy to do so.

[ardenum2]

1. You want a router spindle at the least, no dremels of any kind, those are toys. For a production machine must hate an ATC(automatic toolchanger), you wouldn't want to sit by the machine and change the tools manually every 5-30 seconds, a few examples:

direct tool holder https://www.aliexpress.com/item/1005002474457550.html

iso20 toolholder https://www.aliexpress.com/item/1005004798512348.html

bt30 toolholder https://www.aliexpress.com/item/1005004206523750.html

belt driven bt30 https://www.aliexpress.com/item/1005005131530025.html. You'd need a stiffer/heavier frame and run it with a spindle or a servo motor.

3. You can machine aluminium dry or with mist coolant, you can configure the amount of mist so it doesn't drip down and/or air blast.

4. Double column raising gantry is the stiffest you'd be able to do.(first picture). Personally I'm more in favor of a moving column, not as stiff but stiff enough that offers more versatility(second picture) is a willemin macodel, you can change your front-end mounting hardware any way you like it. You can start off as a 3-axis and easily turn it into 5 axis, because axes are separated from each other, there's enough room for everything.

Having said that here's a few more frame ideas 3D machine tools models

5. I hear these guys are good and cheap for end mills www.winstarcutting.com


Aim for C3 ballscrews and C3 supports and you'll be fine precision wise.

C3 ballscrews https://www.aliexpress.com/item/1005004778503006.html

C3 support https://www.aliexpress.com/store/gro...515670103.html

[joeavaerage]

Hi,
milling aluminum takes quite a bit of doing. It requires a rigid machine, way less (approx 10 fold) than doing steel but a lot more than wood or plastics.

May I suggest you look at one of the many drill-mills around, this is an example:

https://rongfu.com/mill-drill-machin...-machine-rf45/

This is not a CNC machine, but many have been converted to CNC. I think it a good example of the rigidity required, it will do a pretty good job in aluminum and even with care steel.
If you go to build a machine of your own you'll need to match or even exceed this machine for rigidity. As you can imagine, that is no easy task. Do not be lulled into thinking
'I'll take it easy, I don't need such a tough/rigid/expensive machine' because you'll be very disappointed if your machine is not rigid.

All the other questions fall into line....after you decide on the rigidity of the machine required to do your work.

For instance the spindle...you can use the low speed high torque spindle that comes with the mill pictured or you could use an asynchronous 15000-24000rpm spindle and VFD. Either will do a good
job in aluminum. The low speed high torque spindle is the only answer for steel, high speed spindles don't have the torque to do steel. I have two spindles, an 800W 24000rpm spindle for soft metals,
plastics and engraving, and a 6.1Nm (cont) 3500rpm servo motor powered spindle for doing steel and stainless.

Clamping parts is a very broad subject and is very much the skill of milling metal parts. The number of parts that have been destroyed because the part has shifted or pulled out of the vice!!!

I personally use flood cooling when cutting metals and plastics. In both cases its more about flushing chips out of the cutzone to prevent them from being re-cut....that's when they weld themselves to
the tool and wreck everything. Others use mist or air blast...I use flood cooling. My new mill is totally contained and still I get leaks! If you use flood cooling you must consider how to contain it.

Up and Down milling is more a concern with manual machines with backlash. Even small amounts of backlash can cause real drama when Up milling. Most CNC machines have zero or at least
vanishing small amounts of lash and you can mill either way with not much in it. Finishing passes where you hoping for the best finish you might chose one strategy over another. Materials like stainless
that work harden I prefer to Up mill.

I use Mach4Hobby ($200)and an Ethernet SmoothStepper ($195) and my own design breakout board (approx $200) and have done for many years. GRBL can be made to run mills but its original
intent was the 3D printer market. I think there are better choices for a genuine CNC mill, including Mach4Hooby, UCCNC, LinuxCNC and others. Having a dedicated PC to run a machine is the norm.
My machine has its very own PC, its not hooked to the internet EVER, nor even a network. It runs NOTHING other than the software required to run and tune my machine....nothing else.

Craig

[peteeng]

Hi 1234- starting from the end:

1) for a mill 150x150x150mm a conventional format is best. So look at small mills for a muse. Maybe a fixed gantry or bridge design is to be considered. So you have a moving table, an axis across the bridge and a Z. Try not to use 2 screws for the Y. On such a small machine this is not needed
2) from the top - use a 24000rpm water cooled spindle. A good quality one with good bearings. Does not need to be big 800W or 1500W is a good size. You need one with small runout as you are using tiny tools and these need to run true
3) Your table will need space for vices. Top hat machining is common if the part has a two parallel areas. You machine the part then flip the billet over and mill off the "hat"
4) your machine will be wet so it will need to be made from materials that can be wet all the time. Plus you will need an enclosure to contain the swarf and coolant
5) Plastic brackets will not be stiff enough for this venture. It will be a 100% metal machine, Stay away from extrusions. They are expensive for their weight, not very stiff for their size and a poor solution. They are convenient and that's about it.
6) If your going to use 1 and 2mm bits then you need a fast spindle as discussed before. You will learn all about milling when you get there. You don't need special plunging tools but tooling is a big learning curve
7) You will need to learn about CAD and CAM systems. Have a look at Fusion 360, freecad and others
8) You will have to forward your thinking way past 3D printing to effectively mill aluminium. Early days but the forum will sort you out. Peter

[ardenum2]

Forgot to mention but if you can accommodate the weight of a granite surface plate, you could build your machine on top of it, these can be heavy but it would save you a lot of trouble and help you align all the other components at the same time, usually come with a dedicated steel table too. a 800x500x100mm plate would weight about 150kg but give you enough space to build on top of it. Around $500 for a grade 0.

re. 1 and 2mm endmills, I missed this part, like Pete mentioned, you need high speed for this so scratch that "belt drive bt30" that I mentioned above.

[joeavaerage]

Hi,
if you think you can get genuine C3 and C5 ballscrews for the prices in the links above I think your dreaming.

You know the saying...'if something sounds too good to be true then its likely false'. I wouldn't trust that site as far as I can throw them.
I paid more, much more, for BNFN double nut C5's from THK second hand than this crew want for C3's new. I call BS!!!

Any fool can write what looks like a C3 certificate.....and I bet that's what they've done, the Chinese have a long reputation for that sort of stuff.

If you want C3's or C5's, both highly desirable stay with the known manufacturers THK, NSK, Hiwin etc

Craig

[ardenum2]

Hi,
if you think you can get genuine C3 and C5 ballscrews for the prices in the links above I think your dreaming.

You know the saying...'if something sounds too good to be true then its likely false'. I wouldn't trust that site as far as I can throw them.
I paid more, much more, for BNFN double nut C5's from THK second hand than this crew want for C3's new. I call BS!!!

Any fool can write what looks like a C3 certificate.....and I bet that's what they've done, the Chinese have a long reputation for that sort of stuff.

If you want C3's or C5's, both highly desirable stay with the known manufacturers THK, NSK, Hiwin etc

Craig

I don't really care, I'm not willing to pay more than that, I'll get what I'll get. Paying a few grand for 'brand name' for each C3 ballscrew makes no sense because you can get linear motors for those prices and the choice is obvious.

[joeavaerage]

Hi,
that site is claiming 1um per 300mm. That's three times better C0 and C1 grades....and you think that claim is legit?

Craig

[pippin88]

There are some cheap C3/C5 that may be genuine - TBI and DTK - based on online presence / websites etc

[joeavaerage]

Hi,
yes I have seen those sites too, and like you find them more believable. They are asking around $380USD for a 32mm dimeter 5mm pitch double nut C3. That price is still incredibly cheap
by comparison to THK/NSK/Koruda/Bosch Rexroth etc.......but could be genuine.

I'm sure New Zealand is no different than other countries whom have imported steel from China of grade xxx but when it got here and tested it was found to be s***t. Do you trust
these guys?

There are plenty of Chinese products and manufacturers whom do have my trust and their products match their claims.....but there are many more which don't.

If I'm paying over $1000USD on ballscrews, and that is exactly what I paid about two years ago now for three THK ballscrews second hand from Korea, then I want to be 100%
the screws match the specs. Despite being secondhand they turned up with the original test certs and matching serial numbers? I'm sort of wondering if they had ever been used?
Anyway I got what I paid for and they are superb.

Craig

[Josh_Carr]

There's no real standard for cast iron parts out there, like there is with extruded aluminum and 6030 rails - bummer!
Would be nice if there was, but usually you want to build a big machine for rigid cutting of metals.
You could buy a mill cross table, but these seem fairly light, and with CNC you want something heavy.
I went with a 500kg VMC and will try to convert it myself with closed steppers.
I'm kinda stuck on the 4th axis design, so many options out there (AverageJoe's looks like one of the nicest!)
There are lots of other options with DIY. The granite slab is a great idea, I see a lot of designs that end up trying to add weight with granite filling, anyway.
For what OP wants to do (aluminium only) if you built something out of the heavy walled 6030.
I suspect the vibration would be a non issue, and you wouldn't have to worry about hunting for cast iron on craigslist, or granite casting, surface grinding, linearity, and all that jazz.

[peteeng]

Hi Josh - By std I imagine you mean its geometric shape not its alloy. There are 100's of standard CI alloys. Cast Iron bar stock is common. If you do a search for fixture plates you will find companies that machine billet to size. Some fixture plate companies have std sizes and hole patterns. Welding tables are often cast iron and quite flat as well... They would form the base of a good machine. Peter

Buy Dura-Bar® Cast Iron Cut to Size - Price and Order Online (onlinemetals.com)

[ardenum2]

Hi Josh - By std I imagine you mean its geometric shape not its alloy. There are 100's of standard CI alloys. Cast Iron bar stock is common. If you do a search for fixture plates you will find companies that machine billet to size. Some fixture plate companies have std sizes and hole patterns. Welding tables are often cast iron and quite flat as well... They would form the base of a good machine. Peter

Buy Dura-Bar® Cast Iron Cut to Size - Price and Order Online (onlinemetals.com)

I'd love if we could pinpoint a factory in china or india that's willing to do quality one off iron/steel castings. Would open the door for a lot of quality machines for the diy.

[joeavaerage]

Hi,
the local foundry here can do up to 65kg of cast iron or steel in one pour, but can get stuff poured in Timaru, 150km distant, up to 1 tonne.
Works out just under $10NZD/kg including getting the pattern made. That is about $6.50USD/kg.

They did my mill beds for me, they were 115kg each, and I had three of them done for $3000NZD including the pattern, molding, freight to and from Timaru to be poured, the pouring it self
and fettling. Seemed expensive at the time......but they are the foundation of my machine and they are superb, stable, accurate and RIGID!

I've used 750W Delta servos and they work out at $1000NZD each....so the cast iron beds cost the same as the three servos. When you put it into that context then cast iron
is well priced. The fact that I have since bought another two servos for fourth and fifth axis means that I've well and truly spent more on servos than cast iron.

Casting is not cheap but it sure is good, great balance of stiffness and damping and made to your exact size and shape requirements.

Craig

[ardenum2]

three cast iron parts for 3000NZD is very cheap for a one off, you are very lucky to have this service so close by, I hope you'll go all out on your next machine and do something more inspiring than a regular C frame.

[joeavaerage]

Hi,

I hope you'll go all out on your next machine and do something more inspiring than a regular C frame.

Why? There is a very good reason that C frame type mills have been so popular over literally centuries, they are straight forward to design and build
and can be as rigid as any other design.

Craig

[ardenum2]

That's exactly why, popularity doesn't mean it's the most efficient design, it's just a copycut behaviour and low risk taking, pretty sure its the management deciding what the company builds based on what other company made instead of the engineer choosing the most efficient structure. Don't get me started on the saddle, not a stiffness issue but if you want to make bigger parts, the saddle can get ridiculously wide, it can look like the circus is back in town. You can't move that much mass efficiently either.

Take a look at Modig for example, their line of machines, you can clearly see that it was the engineers decision through and through. They completely refreshed their lineup.

I just think if you have this great service available to you and like you did, you decided to do this properly, with having it stress relieved and machined on top of that, you could have gone all out and made history.

[joeavaerage]

Hi,
There are many different designs out there, if one were superior to others it would dominate.....and no such design exists.

My design was predicated on three cast iron axis beds. They need to be super rigid and accurate...and they are. The best quality linear motion components I could get (affordably)
dictated that the length of the beds be 700mm and the travel be 350mm. By mounting these to a C frame I could at will or need shift those beds on the frame. The beds are bolted to the
frame and shimmed . For instance as and when I complete my fourth and fifth axis and when mounted atop the bed I would lose significant Z axis travel. I deliberately made the C frame higher
so that I could shift the Z axis bed to re-gain the travel that the height of the fourth and fifth axis take away.

I would ideally cast the C frame, but budget precluded me from doing so and so was made of steel. If however, I do have the money so to do I can replace the frame alone and yet retain the axis beds. Thus the C frame
is replaceable.

it's just a copycut behaviour and low risk taking,

That much I agree with, I have copied previous successful designs. I have tried to build the most rigid and accurate machine I can afford while retaining flexibility for future modifications.
The rest of your comments are plain disparaging.

on your next machine and do something more inspiring than a regular C frame.

I save the 'inspiring stuff' for the parts I design and make.....my machine is simple and plain but it does the job easily as well as any other design.

Craig

[ardenum2]

The rest of your comments are plain disparaging.

absolutely, I have no respect for companies parroting what other companies did instead of doing the actual work of innovation.

[joeavaerage]

Hi,
well that certainly explains your temperament.

In this instance I was never intending to innovate but rather make the best job using known designs, materials and technologies.

Craig