CNC – What’s the Problem?

[Nigel KParticipant @nigelk]

Does anyone know why there are there comparitively so few hobby CNC mill offerings out there from the likes of Warco, Axminster and Chester? Is there some regulatory reason perhaps? Lots of people build their own but ready-made options seem to be thin in the ground.

[Nigel KParticipant @nigelk]

[Tony Pratt 1Participant @tonypratt1]

Both Arc & Warco did CNC mills at one time I believe?

Tony

[Anonymous]

I suspect margins on the machines are low and the cost of support with users adding their own computers and CAM software is high. Let alone when the electronics play up.

Andrew

[JasonBModerator @jasonb]

As did Axminster, infact they may still do them but only to educational customers.

Cost would likely be one factor, with the Sieg KX-3 machines last seen selling here for about £8k there are very few hobby users who can justify the cost but plenty of tinkerers who could probably do it for less. Though the base converted machine won't have all the extra items that a purpose made one will.

There is also the question of back up and support, there are enough people buying standard mills with little knowledge of how to use them let alone CNC versions and the cost of having someone who can support customers questions and repairs if things go wrong would not be economical for the small volume of sales. This is one of the main reasons ARC stopped selling the Sieg CNC machines about 2 years ago.

Webeco spring to mind as selling CNC versions of their mills or the smaller Sherlines.

[Ian Johnson 1Participant @ianjohnson1]

Not many companies sell ready made CNC mills for all the reasons Jason stated, Axminster and Arc Euro pulled out of the CNC market at around the same time for various reasons including support difficulties, people making their own from manual mills and a small market, they were both selling essentially the same machines too.

For hobbyists I can only think of Amadeal in the UK who are still selling CNC mills, with their Syil X5, but I think the operating system is Mach3 which is no longer supported by New Fangled Solutions, they only support the later Mach4 system. Worth checking this out though, if buying one!

Another hobby CNC manufacturer is Tormach in the USA, but I don't think there is a UK distributor. So options are a bit limited in the UK, so it seems like the best option is to make your own by converting a manual mill into a CNC.

Second hand factory made CNC mills seem to be thin on the ground too, I can't remember when I last saw a Sieg KX1 like mine for sale.

IanJ

[Bob StevensonParticipant @bobstevenson13909]

The main "problem" is that very few people have the blend of skills and interests required to successfully set up and use CNC….even among those would be CNC'ers only a small proportion actually reach the fruition of a finished project.

In my clock club (EFHC) we have quite a few people who are interested in CNC and see it as a handy method of making the more tedious parts of clocks. I n the last few years this number has increased. However, out of our hundred odd members only one person has been truly successful in building his own system on which he can make any parts that he needs and it's no accident that he used CAD programs for the last 30 years of his career, is also a highly skilled machinist and also enjoys (and understands) electronics.

So basically you need to have three specific and dissimilar hobbies, and be very determined not to mention moneyed to succeed. This is a VERY small percentage of the population. There would be a big market for CNC if some of these disparate skills could be drawn together….for example; were there to be a faster and less demanding way to instruct the system with teh dimensions and shapes required without having to spend two years learning CAD then many more people would be attracted. As things stand it not only requires about two years to build the system (while learning CAD at th same time) but it typically takes another year or two to fight thru the many failures and dissappointments that automatically ensue in a very long teething process.

[Ady1Participant @ady1]

It's a bit like with cars

When cheaper self driving CNC versions become available will we all race off to buy one?

How much fun is it to push a button?

It's not about the destination, it's about the journey

bla bla

[Martin ConnellyParticipant @martinconnelly55370]

"How much fun is it to push a button?"

It's more fun than constantly cranking handles. Wasn't that used as punishment in prisons at some time in the past?

There are a lot of people who are part way to just pushing a button with DROs being a common addition so having a bolt circle option available. Then there are all the ways people have added motor drive to the lead screws so they can just push a button. An ELS is just short of CNC on a lathe.

You can use CNC machines in MDI mode (manual data input) where you use it just like a manual machine but instead of cranking handles you just instruct the machine to move to the required position at a given speed. Not all CNC is create a program, set everything up, press a button, remove finished part.

Martin C

[JasonBModerator @jasonb]

Well the cheap CNC routers are available and quite a few seem to be getting bought as a simple way into CNC.

Which raises an interesting point what do most of the hobby CNC users want/need? I would say a good proportion are not the usual "model engineering" type who may want to machine steel and iron but come into the machining as a need to support other hobbies where non ferrous metals, composites and wood are more likely to be what they want to machine. In these cases their needs are likely to be better met by a gantry type machine rather than one based on a traditional mill. These are often fitted with high speed spindles some with ATC ability which again will be far better suited to these uses than a slow mill head with unwanted quill and noisy two stage high/low gearing.

As to how much fun pushing a button is that's debatable but I do find it very satisfying to see a part that I have drawn up and processed in CAM beginning to emerge from a block of metal. And lets face it you can always leave the CNC running on it's own while you get upto whatever fun you like with your free hands.

[Ex contributorParticipant @mgnbuk]

Hobby CNC machines are very cheap compared to industrial equipment, but I would suggest that most hobby users will require far more support than an industrial user who would have more of an idea how to use what they have purchased . Support costs money – the hobby user has spent (for a hobby) a relatively large amount of money and expects support, but the margins on the hobby machine are most likely insufficient to provide the level of support that (generalising here !) an untrained and inexperienced hobby user may expect. Add in a whole additional layer of equipment for the supplier to have to learn how to support & stock (relatively expensive) spare parts for, there doesn't seem to be much of a business case to support offering this type of machine into the hobby market IMO. As a derivation of the old joke woudl have it : How do you make a small fortune selling hobby CNC machine tools ? Start wih a large one ….

There seems to be a misconception that a deep knowledge of CAD is a pre-requiste to operate a CNC machine. While it does depend upon what you wish to achieve, a great deal of industrial use is simply programmed at the machine "long hand". At my current employment in a machine shop that specialises in producing graphite components, we have 7 CNC milling machines / machining centres & 6 CNC lathes – and no CAD / CAM facilities. The operators program everything at the machines. It should be said that we don't do 3D machining on the mills & some of the setups can get challenging, but for the most part CAD or CAM would not help with those anyway. With industrial CNCs we do have full tool offset capabilities that may be be missing from the hobby controls & may complicate some operations if not available.

Nigel B.

Edited By mgnbuk on 01/04/2021 08:35:13

[JasonBModerator @jasonb]

Nigel, how do your operators know what the part looks like so they know what too input. I would have thought you get a CAD drawing of the part? Suppose a hobbiest could use a pencil drawing to enter code from if not into CAD.

Interestingly one thing that had put me off going down the CNC route was all the talk of G-code and having to learn a whole new language. I reality I've not had to write a single line of code yet produce some quite complex often 3D parts. All done with free CAM.

[HOWARDTParticipant @howardt]

One of the problems is the cost of the hardware to turn a machine into a cnc. At the basic you have three servos and ball screws, that can be a chunk of money for a cheap machine probably more than doubling the basic price. Then the question is how far do you go to make the machine better for just a little more money. So then you move from dovetail slides to linear ways and so it goes and you end up getting away from the sub £1000 machine to something over £5000 and your market has disappeared. Making you own either by modifying a standard machine or building from scratch is OK if you have the skills in design and manufacture, but then you need a machine to make the parts. In addition you have the software to drive the machine, many choices and needs research with costs from freeware upwards.

[Nigel KParticipant @nigelk]

I am extremely grateful for all these interesting replies – thank you all.

Listening to everyone I get the impression that the support issue is the most likely reason why CNC is so seldom added to hobbyist machines. Sadly model engineering is seen as an old man's game but perhaps this will change when younger people, more familiar with IT take over the reins.

I use a 3D printer a lot and use FreeCAD as the design tool. Yes, there are lots of others, and FreeCAD can be a real nuisance for various reasons, but there are aspects of that system which particularly suit me – the fact that it is free being one of them! How I would love to go back to using SolidWorks if only there was a cheaper 'lite' version without all the amazing but unnecessary and hugely expensive features you get with that system.

I am interested by CNC because of its capacity to do repetitive tasks. If I want 200 tiny but equally spaced holed in a piece of metal then CNC will do that without getting distracted! It also offers the chance to learn something new. A 3D printer is an absolutely fantastic resource and I use it every day. For example, I bought an SDR radio from eBay which arrived yesterday. I was frustrated because it won't sit at an angle so that I can see the screen. It took me perhaps 5 minutes to draw up a little stand and about 35 minutes to print it. Fantastic! I guess I'm hoping to have the same creative experience with a CNC mill although I'm under no illusions about the increased difficulties involved – not to mention the weight!

Ian Johnson mentions Amadeal. I enquired there about the Syil machines but it seems that they are no longer the dealer for that manufacturer so I am nervous about buying from them. But I do like the Syil X5. When you look at pictures of it on the web it looks like the sort of colossal beast that you'd find in a factory. However, inside the large cabinet is in fact a smallish, bench-top-mountable unit that might work. It is definitely a contender.

Finally, in answer to Martin Connelly's comment about the use of handles and MDI (Manual Data Input) mode, I would like to investigate this. I like the idea of having rotary encoders wired up in a way that I could use them to work the machine by hand if required. Since I am 'into' electronics this might not be such a difficult thing to rig up.

I will continue to give this some thought and watch out for the interesting comments…

[Ex contributorParticipant @mgnbuk]

Nigel, how do your operators know what the part looks like

Paper drawings as part of the works order. These may have been sent to us as a CAD prepared drawing or as a hand drawn sketch. Quite often the CAD prepared drawings are the most difficult to follow, have basic dimensions missing or are dimensioned in bizzare ways. Some simple parts are just a written description i.e a cylinder x O/D by y long or a plate x by y by z thick.

Most of our milling machines have Heidenhain conversational controls, which don't use G code (they have the option to do so if required). The lathes are 50:50 conversational Heidenhain and conventional Fanuc & my current work project is to retrofit a conversational Fagor control to a Gildemeister lathe (cheaper than a Heidenhain – this particular lathe doesn't warrant the extra cost). The conversational turning controls use cycles similar to Mach 3 "Wizards" to build up a program, though they can also run conventional G code. & can convert a cycle generated program into G code.

For Nigel K – I understand that FreeCad also has a CAM "workbench" to prepare toolpaths from the part you have designed. I have not explored this feature yet – it could be a useful tool for when I get my Triac back together, as I intend to use it initially with a GRBL controller which does not support tool radius compensation.

Nigel B.

[Nick WheelerParticipant @nickwheeler]

Posted by Ady1 on 01/04/2021 06:53:25:

It's a bit like with cars

When cheaper self driving CNC versions become available will we all race off to buy one?

How much fun is it to push a button?

It's not about the destination, it's about the journey

bla bla

Standing in front of a machine cranking handles to scrape material off a block isn't fun. It's work. Tedious and time consuming work.

Brushing the removed material away from the part and removing it from the machine so it can be put to the purpose it was made for is the fun bit.

Spending time in the workshop isn't my hobby. I do it because the real hobbies require making/modifying/fabricating stuff.

I'd like to reduce the amount of time I spend doing this, and the best way would be to buy a Bridgeport sized mill. But I don't have enough space for the benchtop machines, so automating my X2 mill to do the boring work while I do something else is sensible. That something else could be assembling parts, designing the next ones, or making a coffee and learning the leads of Cambridge Minor.

Edited By Nicholas Wheeler 1 on 01/04/2021 10:38:04

[Brian HParticipant @brianh50089]

Fascinating subject this and one that I am increasingly becoming more interested in, especially with the lowering of prices.

I've been having a play with the hobbyist version of Fusion360 which also has a cnc machining capability which can be used as a simulator, I don't have any actual cnc machinery, yet.

Brian

[Nigel KParticipant @nigelk]

I love the 'Nigel, how do your operators know what the part looks like' comment. This can be so true – not so much in my case because I'm the one who will be doing the designing but…

When I was working in historic motorsport I would sometimes be handed a bundle of drawings for an old racing car and asked to make it. With precious few photos of the prototype his was like trying to figure out how to assemble a 3D jigsaw. Often I didn't know what I was looking at, so yes, it certainly helps if the maker knows what they are trying to make. We used to ask permission to visit motoring museums and use the exhibits to try to figure out what was what.

One of my fabricator friends told me the story that he was handed a drawing for a sheet of metal that required hundreds of holes to be made in a sheet of metal to fantastic precision. He did what he was being asked to do and reamed out each hole carefully to the required diameter. Once he had made it he found out that it was a part of a ventilation system and that the holes were simply to allow the air to pass through. The draftsman had incorrectly left the tolerances from a previous job on the drawing which my friend had unwittingly followed. An expensive air duct!

[John HaineParticipant @johnhaine32865]

Nigel, worth checking out these people:

**LINK**

Converted manual mill with ballscrews and motor mounts – add steppers or servos as you choose. They also supply these and electronics kits, but you can build your own off eBay for less if you have the knowledge. I have built two sets of electronics, one for my lathe and the other for mill, they probably cost less that 200 and 300 respectively.

I use Mach 3 which works OK but paying $175 for a bit of obsolete unsupported software these days is not really a good idea. Mach 4 is supported but more expensive, there are also others like UCCNC which have good reputations. Don't plan on using a parallel port, these days a motion controller is the only way to go.

GRBL is free but is only a motion controller and there doesn't seem to be anything around that replicates things like wizards etc though there are a load of user-interfaces that can drive it. It also doesn't have radius compensation and the main branch doesn't support backlash compensation because it requires too much code for the Arduino Uno which is the target machine. There's probably a space for a "conversational" program to drive GRBL.

Handwheels are initially attractive, I went to the trouble of buying a couple of encoders to fit on the lathe, but actually I don't miss them and haven't bothered. The standard Mach 3 / parallel port configuration doesn't have enough I/Os to support two, and you can get equivalent functionality for just one axis with a "ShuttlePro" USB device, or actually now I use an Xbox wireless controller that has joysticks.

[Nigel KParticipant @nigelk]

Thanks John, very interesting.

Regarding CNC4YOU, I did contact them and they said they'd email me with details but never got back to me. But yes, they are certainly worth looking into.

Interesting what you say about hand wheel and encoders – a useful tip.

If you look at the professional cabinets they have controllers which do the same job as PCs. Did you look to see whether it might be worth buying one of these instead of using a PC? Because I don't need to go fast I suspect I could get away without using servos BUT I am worried about PCs' predisposition for carrying out housekeeping when they feel like it which I imagine could cause problems with steppers. I know that the PC-based CNC community get very good at running 'clean' versions of Windows but I'm wondering whether I should simply bite the bullet and get a 'proper' controller.

[SillyOldDufferModerator @sillyoldduffer]

Posted by Nigel K on 01/04/2021 09:39:16:

…

Listening to everyone I get the impression that the support issue is the most likely reason why CNC is so seldom added to hobbyist machines. Sadly model engineering is seen as an old man's game but perhaps this will change when younger people, more familiar with IT take over the reins.

…

I am interested by CNC because of its capacity to do repetitive tasks. If I want 200 tiny but equally spaced holed in a piece of metal then CNC will do that without getting distract…

I will continue to give this some thought and watch out for the interesting comments…

I'd agree with that, whilst noting the factors against hobby CNC often combine to put people off:

• The group who imagine that CNC is just button pressing don't buy in because they value manual skills. And a member of team Simples who buys CNC could be a support nightmare! The customer who has completely underestimated how much learning he has to do and is constantly on the phone demanding basic tutorials or claiming he's been mis-sold. None of us are good at Reading the Flipping Manual, and several volumes of new-speak may be too much.
• CNC shines at repetitive work, and I suspect most model engineers don't have much call for that. Most of the stuff I make doesn't require several copies, and it's as quick to make simple one-offs manually as it is to set up CNC. If more than half a dozen are needed, CNC is the answer.
• Apart from repetition work I enjoy manual machining!
• I rarely make the complex curved items CNC makes easy because it can move the cutter simultaneously in 3 dimensions. (Not in metal, though I often 3D print peculiar shapes.)
• The learning curve is enormous for anyone who isn't computer literate, or thinks in 2D, or doesn't have the time or inclination to take on complex software packages packed with mysterious features, and no obvious starting point. Retired chaps may have the time and money needed to learn the ropes, but frankly I don't have the energy to do everything I'm interested in.
• CNC is expensive in a hobby notorious for careful spending!

Add all these together, and hobby CNC moves into the same class as many other industrial processes Model Engineers are unlikely to do at home such as stamping, rolling, precision grinding, or cutting with laser, water jet or plasma arcs. The list of metal working methods we don't use is longer than those we do.

At the moment metalworking CNC isn't for me, even though I see many advantages, have the necessary computer skills, and could cough up the cash.

As always the need for a tool depends on the job in hand: my guess is a substantial minority of Model Engineers would make good use of CNC, and the percentage is going to rise in the future. A not to intimidating starting point is 3D printing in plastic, which I can confirm is a jolly good thing! It's a relatively simple form of additive CNC, with a burgeoning consumer market, where much of the complexity has been hidden by the software, and there's massive support available on the web.

Dave

[John HaineParticipant @johnhaine32865]

Nigel, if you search the archives here there has been quite a lot of discussion about the packaged controllers – initiated especially by the late John Stevenson, rightly highly respected here, who fitted quite a few of them in upgrading older Denford CNC mills. He rated them highly. There are a few China based suppliers.

The PC's housekeeping issues are avoided by using a separate motion controller that does all the time-critical stuff, with USB or (better) Ethernet link to the PC. Having said that I have never seen a problem with Mach 3 / Win XP which just takes over the PC at a lower level than DOS.

[EmgeeParticipant @emgee]

Nigel, a bit more of an insight:

I use a 1985 Emco 5cnc lathe and an Emco F1 cnc mill, both are programmed by manual G&M code entry, either at the machine control pad or via an RS232 link to a computer.
On the lathe a single line of code will face the end of a bar, another line will turn any amount off the bar at the chosen depth per pass and selected cutting feedrate.
Threading either metric or inch is also a single line entry for threads from 0.25 to 2mm.
To part the work off also only needs 1 line of code.
There are 2 methods of programming, Absolute or Incremental that is used below.
In the code below the bar is 16mm diameter extending 25mm from the chuck jaws.
The RH cutting tool is positioned at 16mm diameter and the end of the bar.
No decimal points are used so 800 = 8.00mm and 50 = 0.50mm
After each cycle the tool returns to the cycle start position.

Line 01 faces off .50mm for 8mm depth at 100 mm/min feedrate.
Cutting point of tool now 0.50mm from end of bar.
Line 02 turns 6mm off (12mm off diam) for 21.50mm length at 100mm/min
Line 03 Toolchange to threading tool, X&Z dimensions are the tool offsets
Line 04 Threading cycle .50mm depth of thread for 19.50mm at 1.0mm pitch

The 21.5 and 19.5mm dimensions are because the cycle starts 0,5mm from the end of bar.

It is only a 2 minute job to enter such a program at the machine control, takes less time than a computer takes to start and load the CAD.CAM program never mind about doing a sketch, converting to a model then completing the Manufacturing set-up then finally using a Postprocessor to output the machine code.
This is only true for simple turning and milling operations, for more complicated work and 3D milling CAD/CAM wins hands down, unfortunately such programs cannot be used by my old Emco machines without retrofitting, I have been messing with a lathe conversion and have CAD/CAM facility on that using Mach3.

Video showing sample G code entry here: **LINK**

Emgee

[John HinkleyParticipant @johnhinkley26699]

I have been having a "play" with some CNC software in the last couple of weeks and even as a complete novice, with resppect to cnc, can produce simulated toolpaths for a simple 3D drawing and subsequently the gcode to make the part, The software flow goes like this:

I designed part in Alibre 3D – loaded stl file into Meshcam V8 and did the necessary to generate the toolpaths – pressed the save gcode button.

This is the 3D part:

And the resultant toolpath looks like this:

This is for a roughing pass and doesn't machine the top surface. It took a number of experimental tries but the gcode is only a mouse click away and further toolpath generation will produce a finishing cut which will do all the final machining. The two holes where the big and little end bores go are intended to be for mounting to a machining fixture so that the blank can be flipped and a pocket machining run will carve out the other relief and recess.

I found that the main problem I had was learning all the new (to me) terminology.

I realise that this doesn't add much to the original query, but does illustrate that, with the right tools, control of a machine is within the grasp of most people.

John

 

Edited By John Hinkley on 01/04/2021 11:46:26

[JasonBModerator @jasonb]

A lot will depend on what you want to produce on the CNC, basic square/rectangular/round shapes and patterns of holes are not too bad to do by typing out the code but as said once things get complex CAM is your friend and will quite easily produce many lines of code once you start to get the hang of things.

Take a recently machined flywheel 45,000 lines of code to do one side and the other that had some extra features was just short of 84,000lines, not something I would want to do manually. Each side took maybe 10mins with CAM before I had a USB in my hand and was hot tailing it down to the workshop.

I think the Linux software has quite a user friendly screen for basic functions though not tried it myself.

May not have been noticed in my earlier post but Webeco are available here from Pro Machine

[Author]

Posts