Member postings for Nealeb

As mentioned already, a lot depends on what you are holding and where. I have a lathe with 5C spindle nose and lever-operated closer (elderly British machine) which came with a set of metric collets. I have bought a set of imperial as well, and that works fine for a lot of workholding. Outside that range or for odd sizes, I have a Burnerd Multisize collet chuck, which is a bit like a giant ER collet holder but with its own unique collets (came with machine - would never have bought otherwise!) I like collets for various reasons on the lathe, not least being good for small work without chunky chuck jaws getting in the way. On my CNC mill, though, (no tool-changer, manual only) I use ER32 collets for tool holding. Specific imperial sizes to suit a small range of imperial cutter shanks plus full range of metric. Why imperial as well? As mentioned in another thread, while the metric set does cover a continuous range of sizes, close-fitting ER collets are a little easier to use. Manual mill uses ER32 collet chuck, plus selected R8 collets for most-used cutter shank sizes. ER32 are also used for work-holding in things like Stevenson blocks (useful little devices - see Arc Euro). CNC router uses ER20 for tool-holding - again, mostly metric but selected imperial.

ER32 set originally came from Warco; all other collets from Arc Euro - except metric 5C which came with lathe. Never measured any of them but never felt I needed to worry about them for my purposes.

So, horses for courses! If I did not have a native 5C lathe, I would probably be ER32/ER20 throughout.

Edited By Nealeb on 03/05/2023 07:08:13

I think the sums are pretty easy here.

Let's say that 20" of tilt give a 1mm movement of the bubble - about right? So we need to find the radius of the curve in which 20" of rotation is equivalent to 1mm at the circumference. Think of a giant toroidal tube if you like, of which the vial is the tiny bit that you can see. Clearly, rotate the toroidal tube and the bubble will move, giving the numbers we are working with. 20" is equivalent to 1/3', 1/(3*60) deg, 1/(3*60*360) revolutions. So the full circumference is 1mm times (3*60*360), which my calculator makes 64800mm

Radius equals circumference/2*pi, =10313mm.

So approx radius of vial is about 10m. Substantially less than the radius of the earth?

My first lathes were Myfords, all with "tool movement" calibrations. Made a lot of sense when the lathe plus vertical slide was also used for milling. Current lathe is an old-but-good Smart and Brown 1024 which is calibrated in diameter. However, the transition was significantly eased by the fact that it came with a DRO which, like many, has a radius/diameter button. It means that I work in diameter - take a cut, measure, set DRO to that diameter, and then machine down to required diameter. Yes, you still have to do the right thing with depth of cut, but the sums are easy enough (and simpler than the imperial DoC to metric equivalent for which I still have little feel). Can't remember the last time I read a feedscrew dial...

Do you have any CAD or draughting experience? Do you have access to anyone locally or otherwise who can give you help? Particularly if you are starting with 3D CAD, where some of the concepts and ways of approaching it are not necessarily intuitive, getting some of those initial ways of thinking under your belt would save a whole lot of grief and pain! For evidence, just look at some of the recent threads on the subject...

As it seems that you want to get started with 3D printing, you are going to need 3D CAD. Personally, I reckon that Fusion 360 has a lot going for it. Very good for 3D printing. I also use Solid Edge but decidedly over the top for 3D printing. There will be plenty of other suggestions; I've only mentioned things I use personally.

Without knowing the applicaiton, over-cutting into the corners is another approach. Means that you can insert a rectangular plug (or whatever) without needing to round its corners. Tidier with CNC as you get controlled cutting but I guess you could calculate coordinates to just take you into the corner. For this kind of thing, I aim at an arc in the corner that just touches the corner of the hypothetical rectangle, with generally whatever cutter I am using to rough out the rest. The resulting gaps either side of the corners can look quite neat - but it all depends on just what you are trying to achieve!

Sorry - can't remember off hand. I'll try to remember to check the label on the packet when I'm next in the workshop. I'm sure that there are recommendations online, though - it's commonly-enough used.

It does appear to put some kind of coating on steel but whatever it is, it seems to take paint well enough, and does not interfere with subsequent silver-soldering.

I use a citric acid-based pickle in my ultrasonic bath - the combination of warmth and vibration cleans up after silver soldering very effectively. I have two bottles of solution made up. One for steel and one for copper/brass.

Watching the "big press for Tesla" video, I noticed that he kept talking about "casting". To me, that means pouring or injecting something more-or-less fluid into a mould, letting it harden, then releasing. I appreciate that injection moulding in particular needs a lot of force to hold die components together during the injection process due to pressures involved. I had a trip recently around a local injection-moulding factory where they also make the moulds and was surprised at just how robust the moulds were for relatively small plastic crates.

However, my impression was that the "Tesla" press and manufacturing process was more akin to forging, maybe in just one big shove rather than by multiple hammer blows, and there was no sight of the kinds of things you would expect to see in a foundry for handling hot metal. Are these "castings" actually one-piece forgings, made by squashing (to use a technical term...) a suitable blank, cold, into a set of dies?

My background does not include this kind of manufacturing, but I'm interested to know how the big boys do it!

I used to use TC but only ever in 2D. When I first bought a copy (cheap, found in a closing-down sale of a software place in the US when on a work trip!) I thought it was great. And it was - better quality drawings in less time than I could otherwise manage. I'm a scribble on the back of an envelope kind of model engineer otherwise. I upgraded a few times but never to a point where I found the 3D part usable. I worked, with difficulty, through the tutorials but never managed to do anything of my own. 2D was fine, though - until I started more ambitious drawings when I found the ability to go back and change things lacking. Anything significant to change and it was easier (for me) to scrap it and start again.

Then I found OnShape, which was a revelation in terms of 3D CAD. I did some drawing office training many, many, years ago so was at least aware of some of the principles but not enough to make unlearning too hard when picking up 3D. I've tried to help a few people get going with 3D CAD and I try to emphasise that you are not trying to create a set of engineering drawings, but to build a 3D model from which the software creates the drawings later. OnShape was OK, but at that time a bit limited, and along with a few other ME colleagues we all more-or-less at the same time found and started using F360. Much better! And the CAM was a bonus. 2-3 years ago I moved to Solid Edge as being more powerful in the CAD department, although I still export models to F360 for CAM. I found F360 to start getting a bit fragile when building more complex assemblies and SE seems to work better for me. I use 3D modelling rather than 2D drafting because I am building a loco to a published design and building the computer model gives me so much more insight into how the bits go together than just looking at the 2D drawings. And it's needed anyway to use CNC, so not wasted effort.

Anyway, as a model engineer. I moved away from TC because I could never get past "drawing in 2D only" and even then struggled to get the flexibility and ability to edit that I have found with the newer tools. 3D suits me much better, and I haven't had any particular problem adapting to the 3D modelling approach.

I'm another "design in SE, export .stp, import to F360 for CAM" user. The restrictions in the free F360 version were very frustrating, especially after moving to a CNC mill with a semi-automated tool change procedure (tool change is manual but uses tool-height sensor to automatically set tool height offset before continuing).

I use the F360 plug-in from here which replaces g1 pseudo-rapids with real g0 (misses a few but does a pretty good job) and also allows multi-tool gcode in one file with tool-change commands. I've been using it extensively for a few years and seems to do a good job. Makes up for the worst limitations of the free version's CAM (although still only 3-axis...)

I have an ABB inverter that is nominally three-phase but also capable of single-phase operation. The downside is that you seem to need to downrate it somewhat when running off one phase. These things work by, in crude terms, taking the input AC, rectifying it, and charging a big capacitor which gives a more-or-less smooth DC voltage. This provides feed to the inverter circuitry which generates the variable-frequency three-phase output. Running off three-phase with a bridge rectifier at the input, the capacitor is topped up at 6 times mains frequency (each half-cycle of each phase). However, with single-phase input, you only get one-third as many top-ups (hence comment above about increased rectifier current, by about three times). That means that between top-ups when the capacitor is being discharged into the inverter circuit, its voltage falls further than if fed by three-phase. If this dip between peaks exceeds some amount, my VFD trips with a "phase input error" - it rightly attributes the problem to a missing phase at the input. It works fine with my lathe except that it won't quite handle the load at the highest lathe speeds (it's a variable-speed lathe with expanding pulley arrangement for speed control).

Mind you, in my case it's aggravated by the fact that there is a crude voltage doubler on the input to go from 240V to 480V to feed the VFD which effectively halves the input frequency, so the downrating in your case might not be too bad.

Just for information, most of the Chinese inverters used in home-built CNC routers and the like with the ubiquitous 2.2KW spindle motors have exactly this kind of single-phase/three-phase input option and almost all of us using them run them very happily off single-phase.

That is the way I was taught, oh-so-many years ago in the Marconi DO School, as an electronic engineering apprentice. It also translates well into useful 2D drawings for use in the workshop when you are using a DRO-equipped mill, for example. Set up your DRO datum to match the drawing and away you go.

However, those trained this way are the worst possible students when it comes to 3D CAD modelling! The usual way to build up a 3D model is to start with a "sketch" - which is much more than a sketch and does contain accurate outlines and dimensions. Then you extrude that sketch into a 3D shape which can then be further modified. However, the idea is that the sketch represents "design intent" rather than the dimensions used in a traditional engineering drawing. You would mark two holes as being some distance apart and equi-spaced about the centre of the work, for example, rather than dimensioning each separately from the datum. Why? Well, later design changes - like making the piece a bit longer or shorter - will leave the holes the same distance apart and still equi-spaced about the centre. Or just change the hole centre spacing and they remain symmetric about the work centre. That's a simple example and the real world can get much more complicated but the same principle holds. The model expresses "design intent" and forget engineering drawing conventions. Those are handled later in the process. However, this approach so strongly contradicts the fundamental principles of traditional draughting that it takes longer to get those habits unlearnt than it does to get the new approach on board! And some never make the transition and never really get to grips with 3D CAD.

The 3D CAD packages that I have used do allow entry of fractions, though - and even calculations in the dimensions used in sketches. Useful when you are building a 3D model from old 2D engineering drawings.

I wonder how much difference the newest generation of hybrid buses will make? A bit odd the first time I rode on one; silent take-off from the stop with the engine only starting once we were rolling. I assume that this is to avoid the black smoke effect pulling away from standstill but I guess it might also make a bit of difference to engine life by avoiding so much operation in the carbon-forming regime?

A lot of my projects end up like that - with a big bang...

Couple of thoughts.

It is not unknown for a later version of a product to write its output in an updated format that earlier versions of the software cannot read. There is also a difference between Community Edition and the Solid Edge 2D version and it seems possible that this is creating a problem. However, it does seem reasonable that as you are using something that might be considered an "industry standard" format, this kind of incompatibility should not happen! If it were a proprietary format (like a .par file) that might be understandable.

The other thought is that SE specifically prevents files written by CE from being read by the commercial version. "Industry standard" argument would apply, you might have thought, as above so this would only apply to proprietary file formats. Maybe! And again, would this apply between CE and the 2D version? Dunno!

I have never used the 2D version - just CE, so no experience with this problem. I have frequently written .stp files from CE and read them in F360. Not sure what that means about the wider range of formats but seems to suggest that SE in principle writes standard format files OK.

'From time to time I get the chance to earn a few brownie points by printing paper patterns or templates or some such for my wife. I was recently asked to produce a couple of circles of little dots with precise circle dimensions and doing this in SE was straightforward. I then print using the standard print dialogue but use the Microsoft "print to PDF" printer. I then have a file I can pass to my wife, although she generally asks me to print it anyway! Good for another brownie point. I went through this process for my concentric circles and, printing from Adobe Reader at 100% scale to an HP desktop laser printer, it came out to exact dimensions as measured with the beaten-up old plastic ruler on my desk. If not, it would have been a trivial job to calculate and set the print scale factor when printing from Reader.

As mentioned, the printer scaling can be applied at so many unseen points in the print work flow that it would be difficult to point the finger at any one of them; better to have a repeatable process that can be simply tweaked to calibrate. Given the alignment errors I always get when printing the patterns scaled up and tiled across multiple pages, there is sufficient mechanical distortion going on between paper and printer that better than a millimetre or so accuracy across the page is probably about as good as I would expect with my setup.

Edited By Nealeb on 23/03/2023 10:39:37

F360 can also produce stl files for printing but it is not necessarily obvious exactly how! It does work, though - I tend to use F360 for simpler 3D printer models and SE for more complex engineering assemblies.

I would also vote for Solid Edge in your situation. Fusion 360 is, I would argue, a little easier to learn and use (not least because it has a somewhat simpler user interface) but it gets upset if it is cut off from the internet. You can download models to work on locally (I do this if I need to work on my laptop with no network connection occasionally) but it's a bit of a faff. SE is happy to work on an isolated machine and all models are stored locally. Both are essentially commercial products but are free for "home and hobby" users.

I'm sure that there will be lots of votes for something like FreeCAD as well, but I haven't used that for a few years and can only suggest things I actually know.

I still think that the whole "metric/imperial stuck in the dark ages" argument is a bit of a red herring. It's easier with a DRO and its metric/imperial switch, but even decimal imperial dimensions aren't that bad with a calculator if you are using metric graduated handwheels or vice versa. Well, apart from the Myford 125 division leadscrew handwheel anyway - a pain for dimensions greater than one revolution.

What really fouls it up for me is the use of fractional dimensions on the still-used drawings from our usual suppliers. Pretty much anything I use in the workshop, including working copies of the original drawings, will end up with scribbled decimal equivalents because, curiously, none of my machines are calibrated in fractions...

Like many others here, as a pragmatic engineer (is that a tautology?) I use whichever is easier at the time. My loco model drawings are imperial, covered in fractions. Sometimes I use a scan of the original in the workshop, sometimes a drawing from the 3D CAD model which is in mm. If the former, it probably has decimal equivalents scribbled on it because my DROs are metric/imperial agnostic but they don't do fractions! The only other problem area is screwcutting. My imperial Myford made cutting imperial TPI threads easy. However, cutting metric "pitch" threads on a metric lathe often means swapping the TDI drive gear, and you need three of those to cover the usual range of threads on my machine.