CAD & CAM content

I do intend to write articles on CNC printing for MEW but first I have to find what I want and buy it, probably next month.

3D printing will be used to make patterns for casting new model designs where I can design and prove the design before spending money an castings.

I started CNC programming when I was about 24. I am 68 now and spent over 30 years programming large and small CNC machines. I taught myself to do this but I could not have done it without Model Engineer to teach me how to machine metal.

I learnt the rest the hard way. My “formal” apprenticeship started in the school library at age 12 reading Model Engineer. I went on to do all sorts of machining, first 5 years mainly turning with some milling and then mainly milling.

CNC was a natural progression and I learnt several different programming languages including g code which is not as had as you may think. Everything I needed to know about g code was written on a sheet of A4 paper and I just cut and pasted it into a program and built it up piece by piece until it was completed but one piece of paper wpould not make up a good magazine article but a code library would.

Different functions would do different thing at different sizes. I used to write a lot of these functions in quick basic (a computer language) and run them to prove that it worked then transfer it to CNC using parameter programming.

This meant that to cut a hexagon on a bit of bar, all I had to do was insert the cutter diameter and the AF of the hexagon into the program and the machine wpould machine any hexagon within the capacity of the machine.

This sort of function library might be useful to programmers but for all I know, do CNC users use software programs to machine everything and no programming is involved? In which case, is there much need for CNC in MEW?

You tell me what you would like to learn in MEW and I will see what I can do. What CAD and/or CAM programs do forum members use nowadays?

Edited By JasonB on 05/04/2022 16:02:09

A lot of programs have wizards that will do simple things like hexagons and hole PCD as will any half decent CAM where it is just a case of clicking a few buttons to set your peramiters but where it really comes is is the fact it is no more work to produce a complex shape than it is a simple hexagon. CAM will do a simulation for you to watch a virtual cutter remove the waste from your virtual part and highlight any tool clashes and crashes.

Then there is full 3D milling, don't think I would really want to write 100,000 lines of cod6ethat some small parts I make need which is why I said the old articles are now dated.

For example how any functions would it take to produce this, there is draw ft angle on all the edges so every time the tool goes round with a 0.2mm stepover all the radius need to alter. CAM took me less than 5mins and includes various cutting strategies to best deal with the part from Adaptive clearing to ramp, scallop and pencil 3D finishing paths plus a simple drilled hole. Just short of 100K lines of code and 158mm long x 11mm thick. Cut from Apple and good enough to go from CNC to paint and then to be cast with air set sand in iron. Drawn with Alibre and CAM with F360

Edited By JasonB on 05/04/2022 12:24:39

Edited By JasonB on 05/04/2022 16:02:38

Hi Jason

I wrote a simple function to mill flanges like that. Did not take long. I used an old psion otganiser with maths pack.

Sounds like things have moved on.

I used to use Turbo Cad but never got as far as the 3D programming.

What would be the best set up for CAD CAM today that would be affordable to buy?

Edited By JasonB on 05/04/2022 16:02:59

Maybe best start a new thread rather then let this drift too much, also plenty of what CAD thread son here. You may also want to take a look through this thread which covers my learning curve with the CNC as there are examples of the CAD drawn parts and the CAM as well as machining.

I expect your function was just a 2D profile, now think what that needs to put draft angle around the edge, join them all together and fillet the edges

Edited By JasonB on 05/04/2022 13:20:43

Edited By JasonB on 05/04/2022 16:03:35

I don't use CNC on metal but find 3D printing very useful. When plastic parts are up to the job, 3D printing can manufacture them more or less directly from a 3D CAD model. Quite a lot of Model making could be done this way - parts for model railway layouts, Tram and Boat fittings, components, electronic enclosures, light mechanisms, and difficult shapes generally. Small parts of any complexity can be printed if they can be drawn. Not particularly suitable for making steam locomotives, but then blow torches, Myfords and rotary tables aren't ideal for making Quadcopters!

My interests lean towards design and experimental work rather than model making: more SMEE than LBSC. At the moment I'm exploring 3D printing as a design aid. Designing and developing a new metal mechanism consisting of several interrelated parts is difficult, and a lot of time is wasted making prototype parts in metal only to discover the shape or dimensions aren't quite right, or that the assembly interferes, or the design can be simplified by turning one part into two, or vice versa.

My process typically starts with a rough sketch - back of an envelope, or squared paper. I often develop complex shape ideas with a 2D drawing package (QCAD) because old-school drawing board technique is the bees-knees for capturing dimensions and layout accurately. Or, I go straight to a 3D CAD package and create the part in three dimensions by adding and removing a series of shapes, each shape being defined by a simple 2D sketch. Thinking in 3D is quite different from 2D, and it can take a lot of effort to unlearn 2D technique. But once grasped 3D, the technique is extremely powerful.

Both 3D-CAD packages I use can draw 2D projections which I can cut metal from manually in time honoured way. Or they can export a parts geometry as an STL file. 'Standard Triangle Language' is a computer generated description of a 3D object's outer surfaces. Not normally altered or written by humans, though it can be.

To do CAM, STL is usually translated into G-code by another program. G-code isn't really meant for humans either, but it can be edited manually if required.

For 3D Printing I use Cura, which is a 'slicer' producing the g-code telling my printer where to accurately squirt tiny blobs of molten plastic as necessary to create a solid representation of the 3D model. This is an additive CNC process done reasonably well with a desktop 3D printer costing a few hundred pounds.

Although slow, 3D printing is ideal for prototyping because it makes most parts faster than I can! 3D savings require the problem to be above a certain level of complexity - designing groups of closely related parts where my brain can't hold all the information, and everything has to be in the right place. Designs can be debugged and developed using automatically made plastic parts, not requiring me to plan a metal cutting workflow, or to apply my dubious metalworking skills, until I know the prototype is tickety-boo. Not true CAM because the final metal parts are still made by me with manual lathe and mill etc.

I don't do subtractive CAM, where a CNC machine creates shapes by cutting metal off, partly because I rarely need the facility, but mostly because I haven't caught the bug yet!

I use two different 3D-CAD packages.

FreeCAD is genuinely free (no charges), but is a slightly buggy development project, where certain features are missing, and many others aren't useful to Mechanical Engineers. Most useful for developing single parts because it can't do joints yet.

Fusion360 is a well-developed commercial product focussed on Mechanical Engineering. It does multipart assemblies and joints, intuitively I think. All the parts of an engine can be modelled, assembled together, and animated as a working engine. For prototyping, a good assembly function reduces the need for 3D-printing. Much to like in F360. Unfortunately, the full function version is no longer free to hobbyists: the free version is restricted in various ways to encourage users to buy the full version. Being a cloud product makes some users nervous about having their licences cancelled, or their proprietary designs being visible to AutoDesk, or hacked.

My main point is the value to Model Engineers of 3D-Printing from 3D-CAD, I think more people would gain from it than CNC milling, which because the latter is isn't the best way of knocking out a few parts. Cutting metal with CNC pays off if the shape requires a curved cutting path and when between 10 and several thousand parts are wanted. Often quicker to make a few parts manually than to set up a CNC machine. As most Model Engineers only make parts in small numbers, I suspect CNC will only actively appeal to a smallish subset of the community.

Could be wrong. Again!

Dave

Edited By JasonB on 05/04/2022 16:04:21

The CNC can do simple parts quite quickly but as I still like to do manual machining like many others I tend to keep it for items that are where it really comes into its own such as those curved in 2 or 3 dimensions.

Most of what I do it the CNC are one offs and the setup time is not that long, often just putting a block of metal into the vice and locating it just as you would any part on a manual machine. Take something like that stick of flanges I posted earlier, would have needed setting up 5 times on the rotary table to do the basic outlines and then either another 10 times to do the smaller curves or a lot of filing with buttons. Plus you can always be making something on the manual machines while the CNC is doing it's own thing.

Funny enough When asked if I could make the pattern I did say why bother as the average builder likely to me making the engine could make the flanges from bar stock on the rotary table so castings are really superflous

I know of a builder on another forum that has done away with most of his manual machines, just keeping a larger lathe as non CNC. He produces every part on his CNC mill and CNC lathe using wizards for simple parts and somewhat outdated software for much of the other which needs a lot more input on his part to produce the parts. These are not simple engines but multi cylinder aero and racing car engine in miniature.

Edited By JasonB on 05/04/2022 16:16:28

What are people using to control their CNC machines? I am about 10 years out of date.

Jason’s thread from 3 years ago to date on KX3 still seems to be MACH3.

Fraid so!

I make onesies and twosies on my CNC mill as well as larger quantities of parts, like the spokes for my engines. I also do a lot of 2D parts, as well as 3D and 4th axis parts. Some parts could be made on manual machines; others like these bevel gears, cannot:

Many of my parts are designed in 3D CAD, irrespective of machining method, so the only extra needed for CNC is the CAM program. The time for that is often offset by simpler workholding and set up.

Andrew

MACH3 is still common on many hobby machines. Linux has a reasonable following too. Those converting older machines with new drivers or buying small routers will be using the software that they come with like Centroid Acorn. Pilotpath for the Tormach owners.

Any half decent CAM will be able to spit out the code to suit just about any control system

Edited By JasonB on 05/04/2022 17:08:02

Hi Andrew

Bevels like this used to be machined on horizontal mills and dividing heads.

Those are proper bevels not parallel depth ones as not many have a proper gear shaper in their workshop to make them

They're true bevel gears where the tooth profile changes uniformly along the face width. Neither of the mathods using a dividing head and horizontal mill achieves that. I've never machined parallel depth bevel gears but have machined bevel gears on a horizontal mill using bevel gear involute cutters. The special bevel gear cutters no longer seem to be available commercially.

Andrew

Beaten to it.

Edited By Andrew Johnston on 05/04/2022 17:24:53

Fascinating, I bow to your superior understanding and experience.

I still imagine there's a minimum part where CNC isn't worth the effort, but now I don't know what it is! Last thing I made on my lathe was this ⌀2.5mm Brass Pin with Top Hat:

As only the diameter and shank length are critical, I made this in a couple of minutes by bunging a brass rod in the lathe.  I only measured the diameter accurately. Gut feel is CAD and CNC are over the top but maybe not - 3D modelling the pin and producing g-code is trivial too.

OMG, perhaps I need CNC after all...

Dave

Edited By SillyOldDuffer on 05/04/2022 17:53:09

That is where a wizard would come into it's own. Simply enter nominal stock dia, diameter and length of the bit and set protrusion of stock. Machine will know from tool table where the end of t6eh cutter is so no measuring and no need to produce a CAD drawing then CAM.

Have a look towards the bottom of this page for a shot of the Mach3 Turn Wizards that will do basic cuts

Myself I'd still do it on the manual at that size but if it were say 50mm head dia and 29mm shank x 100mm long you could leave the machine to it as it took many passes.

Edited By JasonB on 05/04/2022 18:21:16

CNC turning using wizards can take the drudgery out of some machining operations. I always found boring to be particularly boring. Just removing metal to reduce diameters can be a lot of repetitive handle turning as well. If you are someone who is quite happy doing these sort of things then there is probably no point in you considering CNC but if you are getting to the point where these sort of things are becoming harder due to age CNC may be a good option enable you to keep working. You don't need to go full CAD/CAM to make use of a CNC enabled machine. You can still draw something and work off the drawing as you would if you were manually machining.

Martin C

I don't have a CNC lathe, so I'd have made the pin on a manual lathe, or the repetition lathe, depending on quantity. I chose my words carefully. I am making all of the studs and bolts, and most of the nuts, for my engines rather than buying them. They are mostly made on the repetition lathe - a sort of handraulic CNC. But I don't model the parts in CAD of any sort; just a quick paper sketch with key dimensions calculated.

Andrew

One of you CNC nerds should get the 'BEST CNC" private number plate currently on Ebay.

I have a manual lathe (Taig), a manual mill (SX1LP) and a 3D printer. I have never learned technical drawing. Things I want to make, I usually design using CAD (MOI). If they need to be metal, I print 2D sketches from the 3D models and manually produce the components on the lathe and mill.

If plastic will suffice, I print direct from the CAD model.

In my experience one can do most of what is needed using CAD/CAM and the controller wizards. Some explanation of how to use these probably would be useful as an intro for newcomers especially if not too biased to one particular solution.

Where I think a more "basic" approach is useful to extend these is when tackling jobs not easily done using CAD/CAM. Examples from my own experience are making mathematically defined profiles such as cams (an example being an equation of time cam) or cycloidal gears as most CAD systems might stretch to involutes but not cycloidal. Even in these cases CAM may be used to actually convert the shape to g-code.

Though I have a manual lathe (a Unimat 3) it doesn't get used, most of my turning is done on the CNC'd Myford using the wizards and the MDI (a way to input single line g-code statements). Looking at Dave's example this could be programmed using the wizard and cut probably in less time than it takes to do manually.

Although PathPilot (used with my Tormach) has a wide range of wizards, I've never used them. At some point I'll try the thread milling ones, out of interest.

I'm not sure why mathematical profiles are a problem? My CAD system will import spreadsheet numbers to as much resolution as one wants. I use it for involute profiles. I always design my gears as 1DP and then scale down as the last operation before exporting to CAM. So errors can be made arbitrarily small.

Andrew