View Full Version : BUILD LOG: Steel Framed CNC router

02-10-2011, 07:55 PM
I am currently in my second year at collage and I am building a CNCrouter for my corse work. I have uploaded some design pictures and wondered what people thought of my design. It was all designed on Google sketch up. I am now quite a way into the build and will upload some other pictures shortly

Many thanks


03-10-2011, 05:52 PM
Here are some more pictures of the build. The main of the frame was mig welded together and I have used an old saw bench inverted as the base, this way I can use the T slots to hold things down. I have already made the bearing blocks. They are 31x41x50mm and were drilled out to 23mm first and then put onto the faceplate on the lathe for boring and reamed using an expanding reamer to 32mm. Then they were honed exactly to size for a tolerance fit. Unfortunately I am not using supported rails but that is what I had available and I am working on a tight budget so maybe an upgrade at a later date.

The router will be belt driven because of the faster speeds and also because it is a cheap option. This will be the second build I have undertaken as the first attempt was a MDF construction which was fine for woodand plastic but I wanted something to do aluminium. The idea behind the design was to reduce the moving weight as much as possible but to keep whole thing stiff. This way there is less mass to move around when cutting and this also means that faster speeds can be achieved. The red frame is all steel and this is a material that I can weld easy and also is cheap but I have then used aluminium for the Z axis to save weight.

I have now got to finish the z axis and assemble the beltb rackets and stepper motors. One of my teachers at collage has access to a steel laser cutter and he will be cutting out some mounting brackets for me for accuracy and to save time.

I wondered what people thought of my design. Do you think I will be able to cut aluminium??

47014700469946984697469646954694469346924691469046 89

07-11-2011, 07:44 PM
The build is now finished well at least until I change something!

The project has also been entered for the National Engineering Competition so a big thankyou to all the people from this site as I have got many design ideas from you people

I have some pictures of the finial design...



Kenneth Paine
01-07-2012, 10:03 AM
Hi Joe,

I just stumbled across your thread by accident whilst Googling "budget cnc router" and I cannot believe nobody has commented yet.

I like your design very much and it has provided me with a lot of inspiration. The build and finish quality is also very good and a lot of thought has gone in to keep everything neat, tidy and purposeful.

I have yet to build my router but I have been developing its design in my head for a while now and my concept coincidentally shares a lot with your design:

- Off the shelf budget materials
- Belt drive
- Diy spindle using brushless outrunner
- Ability to tackle aluminium

My router will be used mainly to cut balsa and ply for my first hobby which is RC aircraft, but I have recently bought a mini lathe and a mini mill and my interest in metalword has mushroomed as well, so my router design must now also be able to tackle aluminium and eventuually handle a 4th axis module. How well has your design coped with aluminium?

The only deviation I would make from your design, apart from dimensions, is not to use precision ground rails. This is purely to keep within the budget/diy ethos of my project. I will probably adopt a linear gude system as used by the Momus router (Google it).

Thanks for sharing your work. I am impressed and inspired.

Kind regards,


01-07-2012, 05:08 PM
Hi Kenneth

The first thing that I would mention is that I have had some problems with the designing of the brushless spindle, The main one being cooling the bearings and the motor. The outrunner motors are designed to have constant air flow over them and a computer fan just won't cut it, however the motor seems to still work fine when hot and always provides enough power. The other problem I had was that the motor has a 8mm shaft and I wanted a ER16 collet chuck so in the end I made a complete new shaft that would fit the 20mm bearings and then the 8mm motor, Quite a hard job as the shaft has many different diameters and they all have to be exactly on centre.

I have since found this design which I am hoping to replicate which is far better than mine : http://dieselrc.com/projects/cncspindle/

The next thing i would say is yes mine does cope with aluminum but very small amounts, the unsupported rails bend and cause serious chatter, but it is nothing to do with the belts they are fine so only go for the supported type of rail, The momus type look good but quite hard to set up maybe 16mm supported rail from china is a possibility?

I have a few more pictures just before it went to the show looking all shiny :)




01-07-2012, 05:14 PM
Hi Joe,

I just stumbled across your thread by accident whilst Googling "budget cnc router" and I cannot believe nobody has commented yet.

Thanks for sharing your work. I am impressed and inspired.

Kind regards,


Kenneth Sorry but I've just got to say this and please don't think I'm just being negative, the only reason is to save you some pain and disappointment.!!

This design has some serious issues and use's just about every design you want to avoid for a good strong accurate CNC machine.!! . . . . I'll list them in bad news order.

Gantry driven from just one side with narrow bearing separation.
Flimsy Gantry using Unsupported round rails for strength.
Unsupported round rails on all axis.
Bad Z axis design and placement creating long lever affect promoting resonance.
Belt drive with no reduction direct onto motor giving poor resolution and causing premature wear of motor bearings.

Not lots of issue's but they are the worst things you could possibly build into a CNC machine.!! Let me explain some of them better.!

First the gantry driven from one side.? This should be obvious but it still amazes me I see it used or considered.!! Basicly driving from one side the twisting forces on the gantry increase greatly with every MM the Y axis moves away from the driven side creating an ever increasing lever. This results in the gantry racking and deflecting the tool massively, when cutting some hard materials like Aluminium it's very likely to cause sticking and stalling with even moderate depth cuts.?

Unsupported rails don't offer any strength regards combating twist etc and the gantry mainly uses them to do just that, they also act like spring boards. Just watch the bars when gymnasts use the parallel bars to see the same affect.? Even short length unsupported bars can suffer from resonance which will show at the cutter.

The Z axis besides using unsupported rails hangs too low down from the gantry basicly creating a long lever, which again will flex and resonate the tool when cutting any moderately hard material. Really it defeats the point of high sides, the goal should really be to minimise the distance from Y/Z axis mounting point to the cutting tool tip.
This can be down by either raising the material to the tool or lifting the bed plus matarial to the tool but in this case the low hanging Z axis restricts this.?

Belt drive works good and his very accurate but in this case it's badly implimented.?
It's direct drive with no reduction ratio means the pitch or distance traveled for 1 revolution is huge, this lowers the resolution greatly but it does allow high speeds which really arnt needed 95% of time.
The direct drive onto motor shaft stress and over heats the bearings and kills motors.!
To give you some idea the pulleys look around 20 Teeth so will be in the 30mm diameter region. Using Pi this gives a circumference of 94mm. So your distance traveled for 1 rev = 94mm So assuming the usual 200 steps per/rev and half stepping microsteps(400) the distance traveled for 1 step would be 0.24mm which is really not a good resolution.! . . BUT it gets worse.?
Stepper motors have a usable RPM limit of around 1000rpm with a 94mm pitch this gives 1000rpm*94mm=94,000mm/min (94mtr/min)which is great if you want to go from A to B very fast but with average cutting feeds being in the 2-2.5mtr/min region and for hard materials like Aluminium or some hard woods being more like 1000mm/mtr then this puts the motor in an rpm region it just can't work at and give smooth performance resulting in cogging and jerks.
IE: using average cutting feeds 2500mm/min the motor will need to spin at 2500/94=26RPM at 1000mm/mtr = 1000/94=10.6 RPM

So has you can see these are really low RPMs and just about useless for detailed or sub 1000mm/min feed rates which is common in every day cutting.

To use belt drive affectively and keep the motors working at RPMs they like best it needs minimum 3:1 ratio, the other thing being belt drive doesn't actually workout cheaper than ballscrews by the time you have have bought pulleys,belts,bearings and made plates etc and it's certainly much more work.!

Like I say I'm not being deliberetly negative just trying to save you some disappointment thru my experience of building belt drive machines.!! Thou the other issues are far worse IMO and will show in the work produced.!!

Hope this helps.!

01-07-2012, 05:47 PM

Yes all the points you have mentioned are correct apart from the problems about the gantry being driven from one side. The bearings are separated by 200mm and then have a large I shaped bar between them which reduces the racking effect, yes this is not ideal but it is more than adequate for little jobs in balsa and ply. The resolution is around 0.04mm and I am using 12 tooth pulleys which again are not going to product the best ever CNC machine but it is cheap and will work.

But thank you for those points, I am planning to change to supported rails and ball screws, all bought but just need to find the time to do it!

What is your opinion on the brushless spindles, you seem to know far more about quality CNC machines than myself



Kenneth Paine
01-07-2012, 05:51 PM

Thanks for your reply. I notice that you have added an electronics box and water cooling since your first posts. Very nice.

I was considering using a heli motor which has forced air cooling built in. The down side is that their shaft diameter limits the size of ER collets to ER11 (from memory) without rework. However,this is not too much of a problem for I don't need a heavy duty machine.

If I achieve the quality of finish with mine that you did with yours I will be very pleased.

Funnily enough I also found today the spindle you wish to replicate. It is impressive, but way beyond my skills or available time. Good luck and keep us posted.

Kind regards,


Kenneth Paine
01-07-2012, 06:12 PM
Hi Jazz,

Thanks for your advice with which I agree entirely. Surprisingly, I had considered all of what you said and would have done as you said anyway for I have been researching the subject for a couple of years. However, it would have been rich of me to come up with these points having not built my machine yet.

Design shortcomings aside, I still think Joe has done an excellent construction and finish job, and I look forward to seeing his MKII version if he gets round to it incorporating improvements to the shortcomings of his MKI.

One great idea you have given me Jazz is that of lifting the work up to the to the spindle to minimize Z axis flex. You see, I want a minimum net Z axis clearance of 150mm for when I get round to carving fuselage plugs out of MDF for 1/3 scale gliders, yet my machine needs to be relatively light and mobile for I have a very small workshop. Thanks to your idea I will now design in "mezzanine" sacrificial boards for metal work and anything that does not require the full Z axis depth. When I want to use a fourth axis or carve thick plugs out of MDF I can remove the mezzanine sacrificial board to enable the full Z axis.

Kind regards,


01-07-2012, 07:51 PM
Thank you... I came 3rd in the end which was highly commended but I was highly recommended by Rolls Royce and a few other big companies. Also I have a 2 page spread in the DATA magazine about my project, not bad for a 6th form student with a small workshop and a 50 year old Myford :thumsup: . But this will be very useful for the future as I plan to become a precision engineer in the long term

Yes there are many parts that are far from perfect that I want to change but just need a bit of time and money, then I should have a far better machine. I am just doing a few small jobs at the moment to try and raise some more funds :apathy:


01-07-2012, 07:57 PM
Hi Joe and Kenneth,

Joe please dont think I was calling your work and 100% agree with Kenneth you have made a marvalous job and I'm sure it's been very enjoyable and great learning exercise.

I understand the single sided gantry does what you need but my mentioning this was to show that in the main it's not a good idea or design and the potential for problems and error is large. . . . Like you have probably found, has does just about everyone soon after building a CNC machine your ambitions to do more and cut other materials rise quite quickly and this design would very soon become a very limiting aspect.!! . . . You certainly couldn't cut aluminium or hard woods with it and be accurate.!

I don't have any issues with belt drive and it's very accurate but it would have been far better to have a single motor using a ratio driving a shaft connected to belts driving the gantry from both sides. It's a win - win higher torque better resolution.!!
Regards them being cheap then I've built a few machines using them and to be honest found by the time you have bought bearings, shaft etc (To do ratio setup), whether that be housing type or machiing your own housings and cheap bearings it just doesn't workout much cheaper than ballscrews from china and the work involved is much less with ballscrews and although belt is good it just doesn't match ballscrews.!

Regards brushless Spindles then my experiences are mainly of the chinese spindles and find these are very good and fully recommended them.! . . .Other than these then I can't comment really has I don't have any experience of any other brushless spindle.

01-07-2012, 09:15 PM
JAZZ Thank you for your advice...

My plan in to try and raise some funds with my current machine (I have some jobs cutting foam patens for castings) and then move onto a more robust reliable machine that will happily cut aluminum.

In your experience do you think that a budget of 1000 is realistic for roughly a 800mm x 500mm x 200mm machine?? I understand this is a hard question as it all depends on the quality off the parts you buy, but for arguments sake say somewhere in the middle.

With regards to the spindle I have seen the water cooled type and I am considering buying one, do you think that something like this would be ok for cutting aluminum? http://www.ebay.co.uk/itm/WATER-COOLED-SPINDLE-MOTOR-2-2KW-INVERTER-VFD-2-2KW-VARIABLE-FREQUENCY-DRIVE-l9-/140661241892?pt=UK_BOI_Industrial_Automation_Contr ol_ET&hash=item20c0103824


02-07-2012, 12:28 AM
In your experience do you think that a budget of 1000 is realistic for roughly a 800mm x 500mm x 200mm machine?? I understand this is a hard question as it all depends on the quality off the parts you buy, but for arguments sake say somewhere in the middle.

Ermmm without seeing design or intended components it's hard to say but 1500 would be a more realistic figure but with some carefull buying/scrounging/stealing.:pirate: for a steel frame based machine using supported round rail and chinese ballscrews then it's just about do-able but you'll have absolutely no room for error in the design.!!

With regards to the spindle I have seen the water cooled type and I am considering buying one, do you think that something like this would be ok for cutting aluminum? http://www.ebay.co.uk/itm/WATER-COOLED-SPINDLE-MOTOR-2-2KW-INVERTER-VFD-2-2KW-VARIABLE-FREQUENCY-DRIVE-l9-/140661241892?pt=UK_BOI_Industrial_Automation_Contr ol_ET&hash=item20c0103824

Yes thats exactly the same type of spindle I use and recommend. For the money they simply can't be beat.
They eat aluminium for breakfast and will happly cut it for days none stop with just a simple cooling setup(sub Pump in 25ltr drum), I've had mine cut non stop for 34hrs and still only be mildly above warm and no where near hot.
They are very quite compared to a router and power wise they blind even the highest powered routers which can't really hold a candle to them. They are flexible regards tool size with collets from 1mm to 13mm with far superior runout to routers. If you look on the forum you'll see pics of the aztec calender I cut and here is a vid showing the detail and thats just using a 90deg spot drill.!! . . . Far finer detail is possible with proper engraving bits.!!. . http://www.youtube.com/watch?v=wjDvimSdH8M&feature=plcp

They do have a one draw back and that is the minimum 5000rpm speed, go below this and the torque drops away very quick and not really usable so they don't lend them selfs to drilling using large drill bits.!! . . . I get round this by simply pocketing large holes.!

22-06-2013, 11:53 PM
Hey everyone, my name is Marco, I live in Sao Paulo Santo Andre, I like a lot about cnc time ago and built a lot like the one that posted this here.
Do not know much to write and read in English, Google translator is doing this to me.

cnc do Marcao - detalhes.avi - YouTube (http://www.youtube.com/watch?v=_veA6GqMBQw)

23-06-2013, 10:45 PM
Hi Joe, congratulations on coming 3rd. I am sure the commendation was very well deserved. I think that what you have achieved as a sixth former is brilliant. I bet you next effort will be really super. Good luck with your future career plans. G.