View Full Version : DIY CNC machine for steel?

25-02-2016, 12:18 AM

I'm new to CNC and am currently planning my first DIY CNC build. I have many questions as you'd imagine but i'd really appreciate any help you can offer!

I plan to use a 2.2kW chinese spindle, which i believe has the power to cut steel (does it?):

http://www.ebay.co.uk/itm/2-2KW-WATER-COOLED-SPINDLE-MOTOR-2-2KW-VFD-DE-4-BEARING-MILL-GRIND-ENGRAVING/170746857178?_trksid=p2047675.c100005.m1851&_trkparms=aid%3D222007%26algo%3DSIC.MBE%26ao%3D1%2 6asc%3D20140122125356%26meid%3D2da02f6d0bc04ec7ba2 f4bf375b509e4%26pid%3D100005%26rk%3D1%26rkt%3D6%26 mehot%3Dag%26sd%3D181806514041

I get the impression that with the right kit, it's not TOO difficult to cut aluminium parts on a DIY machine. By the right kit, i mean anti-backlash nuts for leadscrews and rails/ bearing carriages for linear guidance. From my research these are the best options for good overall stiffness on a budget.

I've seen videos of people cutting STEEL on these machines, which i'd love to do. I get the impression that the main design criteria for cutting harder materials will be the stiffness of the rig, however i'm not sure about how to calculate the required stiffness as I can't find a reliable way of finding the maximum force on the cutter. I have access to FEA software, so if i knew the forces involved, i could get a good idea of what i'd need of the frame. I won't be cutting steel all too often, and the spindle is quiet, so I don't mind if I have to feed extremely slow or take super shallow cuts.

I don't want to just over engineer the hell out of it, although it's a tempting option, as I'd like to keep the price down.

I've attached a picture of my design as it stands, in case you're interested. Let me know if you've any ideas to easily increase the stiffness. The plates are 1/2 inch aluminium, the ally profiles are 40x80. This gives a 300x300 cutting area 100mm z travel.


How can I make my DIY CNC build cut mild steel? (even if only super slowly)

What are the main design considerations for cutting metals?

Are there any ways to improve my design without killing the budget?

Any and all help greatly appriciated!



Boyan Silyavski
25-02-2016, 04:09 AM
Hi Luke,

I see you have been impressed by other peoples videos where they cut steel on their routers. So lets cut the crap- the answer is:

If you want to properly cut metal- buy or make a proper metal cutting machine.

Detailed, i write this down for the benefit of others in the future that drop here with similar questions:

-with the machine you have drawn would be next to impossible. Forget aluminum, all need be steel, all needs to be braced in all directions, the thicker and the heavier-the better. You could use aluminum, but must be reinforced with steel.

-The spindle.
There is cutting, there is what we call scratching. This spindleis NOT made for cutting steel. Means bearings will fail at one moment. These bearings have 1 year life if machine working 8h a day. Thats what Chinese manufacturer states. And that routing wood. Many people, including myself have used them much more and on materials like aluminum.

The ideal spindle for a small machine to cut metal is the BT30 (http://www.aliexpress.com/wholesale?initiative_id=SB_20160225044522&site=glo&shipCountry=es&g=y&SearchText=BT30+spindle&page=2), coupled with pneumatic release for tool change and belt driven by servo motor. Thats about 2000$ from China. If you weld and have a machine shop you coul dput together a frame. They /China/ sell a very nice VMC frame (http://www.aliexpress.com/item/Milling009-cast-iron-frame-BT30-spindle-metal-vertical-Milling-CNC-machine-bed-metal-engraving-3-axis/2037578440.html?spm=2114.01010208.3.307.Y3ZTZr&ws_ab_test=searchweb201556_2,searchweb201644_2_505 _506_503_504_10020_502_10001_10002_10017_10010_100 05_10011_10006_10003_10021_10004_10013_10022_10009 _10008_10018_10019,searchweb201560_3,searchweb1451 318400_-1,searchweb1451318411_6448&btsid=1533f5d7-1b33-46ce-b161-f1c3e41983ee) for that setup but is expensive. Thats one of the reasons you will not see many DIY mills in forums. Cause its cheaper and better to buy and retrofit old mill.

-the power needed
One thing worth noting is that yes, if you have very sturdy setup, use small bits, have rigid table, proper cooling and calculate properly speeds and feeds you could machine mild steel. I dont see how thats to happen if you dont use some software like HSMAdvisor (http://hsmadvisor.com/)or Gwizard (http://www.cnccookbook.com/CCGWizardv1.html?utm_expid=2217679-153.8v_dbZAmTwylWIIDySxbfw.1&utm_referrer=https%3A%2F%2Fwww.google.es%2F)to finely calculate what exactly you are doing. From there you will see the forces involved and the power needed to do that.

I myself use HSMAdvisor, inside there is an profile of 3kw 18k rpm spindle, use that to see what happens. I highly recommend getting the 3kw versus 2.2kw for metal.
Have in mind that most spindles use all ceramic bearings or partly. There are dedicated high speed spindles for metal that cost more and they say with better bearings. I have no info if thats reality or Chinese tricks.

-the machine design
So yes, if you go that way, its better to have the gantry fixed. Do your research and see build section in forum. There are discussed machines exactly made for what you say.

-me personally
i would buy me a small mill or retrofit big one. I in fact am just finishing a very sturdy build that could definitely do that what you want, but dont plan to f__k my spindle cutting steel

Hope that helps

PS. below a snip from HSMAdvisor, As you see spindle power is not the limiting factor. it would be the stiffness. sliders are adjusted for standard stiff machine/ mill/

25-02-2016, 09:25 AM
You can also opt for building in epoxy concrete (often called epoxy granite).
And yes, you can use a chinese spindle but the bearings have to be ok.

A guy close to me built an epoxy concrete machine with a chinese spindle (I think 2200 watts but up to 18000 rpm) and can easily mill steel, to about 200s of a mm.
He used old servos but you can get simmilar results with closed loop steppers.

Have a look over here in a dutch thread (maybe use google translate):

Here it is (http://cnczone.nl/viewtopic.php?f=8&t=8410&hilit=epoxy&start=406)
Here it is in action, steel 12000rpm, 4500mm/min feed, 15mm doc, 0.5mm, at long life settings:


He says he can also do 20000rpm en 9500mm/min but that is no good for tool life.

26-02-2016, 06:52 PM
Thank you both for your responses.

After further research, I agree that the easiest way forward for me is to buy a small, cheap manual mill and convert it to a cnc machine.

I found this on Ebay:


And here's a conversion:


Couple of questions:

Firstly, does this look like a rig that could cut steel to any degree of success?

Secondly, I know that manual mill can have quite larger backlashs. Will this be an issue in a CNC conversion and how can it be overcome?

Many thanks,

Luke Blades

26-02-2016, 08:40 PM
Here is picture of a mill, it's a Bridgeport - the most common mill ever made i think, I have one just like it. It's 1100kg of solid cast iron and has a proper 2Hp motor on it.

It is capable of milling steel, but will only take relatively light cuts due in most part to it's versatile nature - the head can swivel, rotate and tilt, these are great features but make it weaker when it comes to real milling.

Steel needs lower speeds and more power, yes it can be done with high speeds but you can expect short cutter life and you will only be scratching away the metal.


That being said, I am building a gantry style CNC machine and fully hope to be able to do fine milling in aluminium but nothing tougher, i am looking at cutters of around 2 - 5mm diameter.

I really would look closely at what you need to achieve but steel milling is going to be a tough nut to crack.

Conversions can be done, plenty of Bridgeports have been CNC'd - ball screws are a must and a good machine as a foundation is a big help as lot of them have had a hard life.

26-02-2016, 09:53 PM
Don't touch that Proxxon if you want to mill steel. That Bridgeport would easily mill steel and cast iron, with quite heavy cuts, that's why they were probably the most common mills ever made. I have a Myford VMB which is a lot smaller than the BP put I can take say a 1 or 2 mm cut with a 12mm cutter in mild steel with ease. There are lots of smallish mills that will do similar, look for example at the SX3 from Sieg. This is available in slightly different forms from several suppliers such as Arceurotrade and Axminster. There are also similar machines from the likes of Warco and Chester. These can be converted to CNC, and especially the SX3. It is best to use ballscrews to replace the normal ones to get rid of backlash. I

Sent from my iPad using Tapatalk

26-02-2016, 11:46 PM
I'd love to get myself a bridgeport or a SX3 but i have neither the money nor the space.

I'm looking for something <700 which will leave me some room for tooling etc.

Thing like the:
Warco WM12

You guys have any idea as to the best way to go for a mini-mill that can be converted to CNC and cut small steel parts?


Boyan Silyavski
27-02-2016, 06:08 AM
I have seen nice converted mini mills go for 1000-1300. So meanwhile save some money. Your best bet would be to buy already converted machine from somebody who wants to upgrade.

But for 700, i dont see how that would happen, se better be real from the beginning.

What pieces are you exactly trying to achieve? i mean, what would be the typical job on the machine?

27-02-2016, 07:19 AM
I got a Novamill without its electronics off eBay for 610, built the electronics for less than 200 using 2m542 drivers, BoB and PSU from China. The Novamill is about X1 size but a different league for quality.

Sent from my iPad using Tapatalk

27-02-2016, 08:15 AM
The SX2 from Arc looks quite nice. It has a better table than the standard X2, basically the same as they supply on their version of the X1. You can also get it with an R8 spindle, R8 collets are economical, grip very well, and don't incur the overhang of an ER type collet chuck. And they don't get stuck! I think quite a few people have done CNC conversions of this class of mill including fitting ball screws. Another point about CNC is that since you don't have to stand around twiddling feed screws, the computer can take multiple smaller cuts so machine rigidity isn't maybe quite so important.

I started off a conversion of an X1 with this table, in the end the Novamill came up so I sold the X1 to a friend in manual version. It was clear that one could fit ballscrews, but you need to design the bits carefully so you can make them on the manual mill then just fit them whilst it's in bits. Also be prepared to at least fettle and possibly replace the gib strips, I made new ones from brass.

Sent from my iPad using Tapatalk

27-02-2016, 08:20 AM
I have a feeling your budget is way too low, even for decent work in aluminium, the likes of the 3040 on eBay will do very light cuts on ally, not steel, but will push your budget and fail after a short time. I had thought of getting a 6040 from eBay but after research I decided to build one better, the budget went out of the window, from 400 to over 1500 now but I am far more confident of results and life expectancy.

The Proxxon is a mini drill really, don't class it as a mill, for watchmaking maybe yes but anything bigger no. If looking at other smaller mills to convert, try not to look at those with swivel heads - these are very weak (i had one).

As said, what is your target project, if we know the subject we could guide a bit more :)

27-02-2016, 09:14 AM
Loads of model engineers are using machines like the X1 and X2 and the WM series for doing good work in CI and MS. What you don't want for steel is a gantry style router, a standard vertical mill format is much better, but of course you pay in reduced work area. I agree about swivel heads, they are not worth it, there are very few jobs where they can be used and the reduction in rigidity for everything else is a pain. What you can do with the X Series mills is fill the column and base with epoxy/granite chips which increases the mass a lot. The Arc SX2P doesn't have any swivels.

Sent from my iPad using Tapatalk

27-02-2016, 01:37 PM
My budget ~700 is for the machine prior to cnc conversion.

I'd like to be able to make aluminium parts and make some small steel cuts. The steel objects might be clamps for the mill itself or small hobby items (dice, keys little projects) I wouldn't cut deep or fast or with a large cutter in steel!

I am very tempted by the SX2. Is this likely to do the job?


27-02-2016, 02:18 PM
I would think it would be fine, Arc have a good reputation and Ketan the owner is well respected. If you can, go along to their showroom and have a look at the machine and talk to him about what you want to do. There's a very capable engineer called John Stevenson who works with Arc and has a lot of input to the machines they sell, they have a number of improvements compared to the generic X series. As I said above, if you went for this I'd get one with R8 spindle, the collets are cheap, hold well, don't get stuck, and with the cutter in a collet direct in the spindle you get more headroom and better rigidity.

Sent from my iPad using Tapatalk

27-02-2016, 02:27 PM
X2 looks ok, there is even a conversion kit for it...


27-02-2016, 04:49 PM
Indeed, but in the USA, out of stock, designed for the standard X2 rather than Arc's improved version, and using trapezoidal rather than ball screws.

27-02-2016, 10:06 PM
The siegs from Acer etc are quite simple to convert from experience spent one day on my X1 to get it up and running from scratch, now tidying and tuning everything. The X1 is only 250w but have already cut myself some steel T nuts with a carbide bit and some patience. Cost of my unit 580 for mill and 80 BOB and motors although had a few 3d printer parts laying around shop around and have a look if you can.

28-02-2016, 05:47 PM
Thanks for the support guys.

Obviously i'll need to upgrade the drive screws for the 3 axes.

Ballscrews are a popular opinion but i came across these:

http://www.ebay.co.uk/itm/LOW-COST-395-MM-CNC-STAINLESS-STEEL-M8-LEAD-SCREW-AND-DELRIN-NUT-3-sets-/151191033498?hash=item2333b0169a:g:uGkAAMXQVT9SsE~ 4

Super cheap in comparison with ballscrews and claim to have a very similar effect.

Does anyone have any experience with anti backlash nuts. Is it worth the risk or are ballscrews THAT much better?

Who are a good supplier of well priced ballscrews in the UK?



28-02-2016, 06:43 PM
Interesting, but personally I wouldn't touch them for a mill conversion. I'm using something similar (although home-built) on my current CNC router. They've been working reasonably well, but only because the router is an MDF structure that can't manage anything like reasonable speeds or cuts. For a mill - no way! There's a reason why people say "use ballscrews"..

28-02-2016, 09:01 PM
To get ballscrews, look to China if you are price conscious, guy called BST Automation on Aliexpress site, will have all you need, is very helpful and prices, well, I just bought all the motion - linear rails, carriages, ballscrews, nuts, brackets etc for my mini CNC build for the price of a couple of screws alone from a UK supplier.

I usually shop UK but I could not have done this build if I had.

28-02-2016, 10:41 PM
The Sx2p table has anti backlash nuts which are made of cast iron with a longitudinal slit, and an adjusting screw to take up the backlash. The threads are trapezoidal IIRC, and there is a common technique to make delrin nuts by heat moulding the material around the screw, so you could perhaps do your own. The advantages of ball screws are low backlash, near zero if you use double nuts, and low friction, I'd recommend them if possible. A good strategy would be to try the standard nuts and adjust the backlash as small as possible, try delrin if that isn't satisfactory, and ultimately if you need to, upgrade to ballscrews.

Sent from my iPad using Tapatalk