• Welcome to The Forum for Gauge 3 Model Trains.


NOTE : The self-registration facility is disabled.

G3 Society Members : Must Log In to view some limited access Boards.
Use of Personal Data and Cookies http://g3forum.org.uk/data.html

Main Menu

CNC machining

Started by Nick, Apr 02 2021 22:51

« previous - next »

0 Members and 1 Guest are viewing this topic.


In another forum https://www.model-engineer.co.uk/forums/postings.asp?th=172061 Ian implied that I hadn't been very specific in my concerns about the rigidity of a low-cost CNC machine he was thinking about, and I accept that is a valid criticism. To make amends I am starting a new thread to separate the machines from the software.

I will say right at the beginning that my experience is with industrial, not hobby, CNC, but the laws of physics don't change and neither does the technology. Anyone coming at it from the hobbyist angle is more than welcome to chip in.

Rigidity, then. If you take a sheet of steel or aluminium and hit it, it rings. That is because it is vibrating, and the amplitude of the vibration can be large and it lasts a long time before it all damps out. Do the same thing with a block of steel or cast iron, and it rings but the amplitude is smaller and it damps out more quickly. Repeat with a block of mineral like concrete and the result is a "thunk". Any vibration is damped out very quickly indeed.

When milling, every time a cutting edge strikes the workpiece, that is a little hammer blow which causes the tool and the workpiece to vibrate, and the vibrations are transmitted through the structure of the machine. If the structure is metal they are transmitted very well and if the structure includes relatively thin sheets, they vibrate and stay vibrating for a long time after the initial impact. That vibration is transmitted back to the tool which tends to bounce on and off the workpiece, and that is what we call chatter. If the structure is cast iron the vibration energy is lower and dissipates sooner, and if it is concrete it dissipates almost immediately.

The lack of rigidity is not inherent in a gantry machine, it is inherent in the material it is made from and the way it is made. I've seen a gantry machine being used to machine the cylinder head of a marine diesel engine as big as several houses put together. It was mostly made of epoxy-bonded concrete, and yes, it was rigid enough for the job.

The low cost 6040-type machines (let's call them routers, because that is what they really are) that I have seen are made mainly from sheet steel or aluminium or aluminium extrusions. For cutting wood, wood products, waxes and plastics they are adequate because the cutting forces are sufficiently low in soft materials. Possibly aluminium, there are lots of videos of aluminium being hacked about in various ways, but one never knows exactly what sort of aluminium it is. Brass? I'm not so sure. Steel? Take a deep breath.

Then there is also the question of spindle speed. Routers come with high spindle speeds, and industrial CNC machines operate with spindle speeds much higher than we use for manual milling. Why? Because the controlling variable for cutting is the speed of the tool relative to the workpiece. Operating at a high spindle speed means you can run high feed rates and cutting depths and remove metal more quickly. Industrial CNC machines, software and operating procedures are all about minimising the time on the machine, because that controls the cost of the parts being made. Shaving off a few seconds of machining time can make the difference between commercial success and failure.

But high cutting rates mean large cutting forces, and so the problem of rigidity of the machine returns (actually, it never went away). Industrial machines are made massive (and expensive, but the cost is recouped in reduced machining time). As hobbyists we don't have the same requirements and we can take as long as we like over a job, so we can choose lower feed rates and cuts. But that in combination with a high spindle speed means the cutting process is far from optimum. With a low rate of cutting in combination with a high surface speed, the tool tends to skate over the surface so that the surface gets scored and abraded rather than cut cleanly, and the cutter quickly loses its edge. Any machinist knows blunt tools and good surface finish are incompatible.

Unfortunately, running at a lower spindle speed which is quite possible with a variable frequency drive, is rarely the answer because of current limits. Too much current will overheat the motor (even with cooling) or the drive electronics. The quoted power will be at the rated speed. At lower speed, the power available will be a lot less than that. No, the answer is we need motors that are rated for much lower speeds, in fact, the sort of speeds used in manual machines.

In fact, there is nothing magic about CNC machines. To the machine, the only difference is that a motor rather than somebody's arm drives the table. Industrial machines have gone a different route because of commercial requirements, but to us hobbyists the desirable requirements and design features are much the same as for manual machines. As heavy and rigid as realistically possible, with a spindle speed range that is compatible with the cuts and feeds that we learn to use on a manual mill.

It's been a long post and a slightly technical one, but I hope I have explained why I decided that a low-cost router-type machine really was not something I wanted. If your interest is making coach bodies, wagon bodies or model buildings it might be just the thing, but for locomotive parts and fittings, I think not. So what is the answer? Maybe that's something to come back to another day. It's late and I've been sitting at the computer too long.



I agree with Nick's comments, it is not appreciated the forces generated by the cutting tool i.e pushback the only way to overcome these forces is machine rigidity.

There are many mid-range milling machines on the market with round columns, it is difficult to believe but these columns flex under load, thus only light cuts are possible for a good finish.

There are few small CNC mills available that fit the rigged criteria
New Seig / Syil X3 £5-7K.    Tormach  £10K imported
Used Denford Triac, These machines are the best value for money can be found for around £2k and converted to modern electronics £1K.

Buying a CNC mill is a big investment but if you compare with a new G3 loco, around the same cost, then you can build as many locos you want for a fraction of new price.
I bought my Syil X3 around 10 years ago half today's price. To date I have built  12 G1 and 10 G3 locos, many coaches and wagons. Value for money?



My remarks elsewhere where not intended (in any way) to be critical of your views Nick.

I was really musing on the considerable leap (in both cost and complexity) of "CNC" over 3D Printing. One of the things about the ME Forum is that it covers a very wide range of interests - from people who make clocks, to those who build 6" (half scale) traction engines - or who restore motorbikes. So their machining 'needs' vary considerably. Some consider the Myford S7 as little better than an "expensive toy" whereas others feel they are overkill for most of their work. 

I do understand the need for rigidity when machining harder materials but also know that this can be mitigated to some extent by the depths of cut required and the associated feeds & speeds etc. So small parts in softer materials (non-ferrous) do not impose the loads that larger parts in harder parts will require. 

I made the point that CNC for Hobbyists makes sense where a) complex geometries are involved (even on single parts) or b) where many identical parts are required (that are either too expensive or unavailable commercially). Whilst understanding that the low cost 'router' gantry-type machines are far from ideal for serious metal removal - they are being used very successfully by clockmakers to machine (and drill) very small parts precisely.

So I do understand your reservations about these machines but I was really considering whether the cost/benefit ratio (where it could be usefully deployed) was a sufficiently good reason to invest in one.


Nothing's ever Easy - At least the first time around.


To Mike's list I would add a couple more options. I have two manual mills: a Sherline and a Chinese mill. The Sherline is small and light but it is very well designed and made, and I have surprised myself on occasion what it is capable of. Sherline sell a turnkey CNC conversion, but it works out at nearly £5k. They actually sell a complete CNC mill for not very much more than that so they probably don't expect many conversions.

The web is awash with Chinese mill conversions (some of which involve attacking the saddle with an angle grinder – horrors!) but being sensible about it and using good quality components, I figured the cost is £2.5-3k. If one were doing it from scratch, add the cost of the mill to that.

So whichever way you look at it, it's not cheap. I don't imagine doing enough commercial work with it to offset that, but neither do I envisage turning out enough locos to rival Mike (if only ...). So if I did it, it would be because I decided it would be fun to do.

And speaking of fun, I did sketch out my ideal machine, which is basically a highly scaled down heavy industrial machine. That means a gantry configuration made largely of epoxy granite. The working area is 300 (x) by 200 (y) by 100 (z). With a gantry configuration it is difficult to achieve a large z-axis travel. It fits within a 500 mm cube, more or less, and weighs 80-100 kg. The cost is mostly in the linear slides, ballscrews and electronics, so about the same as a Chinese mill conversion. If I decide I want a change from model railways I might just do it, but I will get no. 563 finished first.

Ian, please don't apologise, it is always good to have someone else react to what I write. I wasn't aware that clockmakers were using the low cost machines but I'm sure it makes sense for them. I decided early on that I would not consider it for G3, but maybe it would work for some of the smaller gauge work that I do. If you do take the plunge I'm sure we will hear about it.