Here is a list of all the postings Andrew Johnston has made in our forums. Click on a thread name to jump to the thread.
|Thread: Free Electricity|
The original T&Cs for my amateur radio licence, back in the 1970s, banned end-fed wire antennas longer than, I think, 75 feet. That was because in the early days of radio people near the long wave transmitters were extracting enough energy to power a light bulb or two.
|Thread: Use of Colour on Drawings|
Of course the use of colour in assembly, as opposed to engineering, drawings is not new. This drawing was done in 1904:
|Thread: Gold Plating|
The only time I had parts gold plated I let the chemistry department at RAE Farnborough do it. I don't think the training centre would have been happy having cyanide compounds in the building.
|Thread: Use of Colour on Drawings|
1. My preference
2. Maybe, although a shaded isometric view doesn't add much - an exception would be an assembly or partial assembly where different shades would delineate the parts, and it would make sense if the shades reflected the materials
3. Definitely not, adds nowt and tends to obscure the engineering detail
|Thread: Zyto Owners Only. Other Riff Raff can keep out :P|
That's no problem, I wouldn't give a Zyto house, let alone workshop, room anyway.
|Thread: Part built Allchin 1.5 inch|
Ideally one doesn't run similar metals together, cast iron is one of the few exceptions. However, for a model which only runs once in a blue moon it probably doesn't matter.
Of course you could always make prototypical hollow cast iron pistons, although in such a small scale they probably wouldn't have any weight advantage over aluminium. My main objection to aluminium is the clearance needed to allow for differential expansion.
|Thread: VFD recommendations|
Hmmm, my Yaskawa J1000 inverter bought to drive my Pultra instrument lathe is 240VAC single phase in only, not running on two imputs of three phase. And it was made in Japan not China.
There's no reason why a single phase input VFD can't output 415VAC, it just needs more electronics, although offhand I can't find one advertised. The smaller VFDs just have a rectifier on the input to create the DC bus. Which is why the manuals state they don't meet the EU standards for harmonic distortion without an external filter. However it would be possible to have power factor correction on the input, which is simply a switch mode power supply with a control loop that keeps the input current draw in phase and proportional with the input voltage. The PFC is essentially a boost converter so it could easily be set up to produce a DC bus voltage high enough to produce 415VAC from a single phase input.
The output stage of a VFD looks like this, with IGBTs:
where U, V and W are motor connections. The half bridge drive signals are a bit more complicated than a square wave. Each half bridge produces a PWM signal swinging between 0V and the DC bus. The PWM changes on a cycle by cycle basis so that a filtered version of the PWM would produce a sine wave. In general IGBTs are prefered to MOSFETs in VFD applications.
A while back I wrote some notes on operating and driving IGBTs but deleted them in a huff after being told by a forum member that I didn't know anything about the fast Fourier transform. Which was a bit odd as part of my Ph.D was concerned with ways of computing it in hardware in real time with 100MHz bandwidth signals. If anyone is interested I can reinstate them.
To elaborate on the switchmode notes provided by SoD most of the small off the shelf DC-DC converters do indeed contain a transformer, as this gives the useful feature of an isolated output voltage. However, most small switchers do not use a transformer. There are a myraid of topologies but for low power, say <100W, there are four basic types.
1. Buck converter - converts to a lower voltage than the input, essentially a switch producing a PWM signal and followed by an inductor/capacitor low pass filter.
2. Boost converter - converts to a higher voltage than the input, essentially using the flyback voltage from an inductor to boost the voltage
3. SEPIC - can convert to a voltage both below and above the input - stands for single-ended primary-inductor converter. This uses two inductors in the switching. Oddly both inductors have the same voltage impressed on them, so the currents are the same, and they can be placed on a single core. But to be clear the dual winding is not a transformer
4. Flyback - which does use a sort of transformer although not in the conventional sense - it can produce just about any output voltage depending upon the turns ratio of the primary and secondary windings.
Frequencies are getting higher, many of the low power ICs can switch at low megahertz, so the inductors can be tiny. There are of course many more topologies especially as one gets to higher powers where transformers are driven by full bridges with phase control.
Way back when one designed a switch mode power supply you were lucky to get a basic control IC to which you needed to add a significant number of key components. With the advances due to mobile phones and laptops there are now hundreds of ICs available with some high quality simulation and design packages available for free from the manufacturers. There are also a range of nifty topologies such as buck-boost converters, boost converters that can disconnect the output and flyback designs that sense the impressed secondary voltage on the primary during the off time so you don't need to design an isolated feedback voltage circuit, yippee!
If that hasn't bored the pants off everyone nothing will! A final note; I always used to prototype switch mode converters because they never quite did what it said on the packet. But the newer ICs are very good, with good application software and notes so one can be confident of designing and going straight to PCB without getting egg on face.
|Thread: Milling feed/speed question.|
A mechanical speed adjuster allows the motor runs at its design speed, aka base speed. It's the speed at which the rate applied voltage is just sufficient to drive enough current through the windings to give the rated power. It also means that constant power is delivered to the spindle across all speed ranges unlike a VFD.
My CNC mill has two belt ranges within which the speed is varied with a VFD. I know what spindle rpm corresponds to base speed on the motor and I try and stay above that so the motor is always running at constant power. It's one reason I moved to running small cutters at high rpm and feedrates compared to the Bridgeport.
|Thread: What Did You Do Today 2019|
It'll be interesting to see how they perform. Also interesting to note that the pressure angle is 20 degrees, a lot of the far east cutters are 14.5 degrees. They seem to be missing depth of cut information, so some experimentation may be needed. One thing we can say for sure is that whoever was operating the engraving machine was imitating a newt.
Edited By Andrew Johnston on 13/01/2019 18:52:32
|Thread: Milling feed/speed question.|
Been there, done all that!
When I first started with my lathe I bought some 2" steel bar from a tool shop in Cambridge. Had all sorts of issues with finish; smooth for an inch and then suddenly tearing, and then back to smooth. After much experimentation I got nowhere. Starting thinking about taking up knitting instead, even bought an embroidery kit! Then I talked to a professional machine shop who were making some boxes I'd designed. They said we don't buy from that company; their steel is carp. I bought some more steel from a professional stockholder and suddenly I was getting good, and consistent, finishes.
It also took me a while to work out that the way to stop chatter on the Bridgeport was to up the feedrate, sometimes quite considerably.
You live and learn; just wish I was a bit quicker on the uptake.
It's a controversial view but I would expect a machine tool to be designed to cope with the cutting forces generated. Obviously a machine with 25hp would be heavier and more rigid than one with 1hp, but the same principles apply. It's also not just about the mass of metal, but having the metal in the right place.
All machine tools deflect so cutting parameters need to take that into account. It's not obvious but increasing feeds may well be beneficial. If the feed is small and the machine deflects a little, that deflection can be a significant part of the feed; equals chatter. With a higher feedrate the deflection is a smaller proportion so much less chance of chatter.
As well as the rigidity and power available it helps to have a basic understanding of the cutting process. Remember that neither the material nor cutter know whether they're on a small or large mill.
With smaller machines it's better to run small cutters fast rather than large cutters slow. On my CNC mill I run cutters much faster than on the Bridgeport, although both mills are the same horsepower. The rule of thumb of one cubic inch of steel per minute per horsepower works well for me. In theory the power needed per unit metal removal goes down as cutting speeds go up because the metal in the shear zone is hotter and hence softer.
That's enough pot stirring for now; time to go and fly the tug plane.
|Thread: Copper tube wall thickness & pressure withstood.|
The problem with relying on a formula taken out of context is that one doesn't know where it comes from, or what limitations apply. There are three sorts of stress to take into account; hoop, longditudinal and radial. Radial is only applicable is the tube is considered thick walled, ie, tube diameter is greater than 10 to 20 times the wall thickness. Longditudinal is applicable is the tube is under internal pressure, as I suspect a tube on a steam engine will be. The formulae, some calculators, can be found online.
The biggest problem is the value of tensile strength to use. It can vary rather more widely than suggested by SoD depending upon the exact material and it's state, eg, annealed or work hardened. And the copper losses strength at higher temperatures, even with steam at 100psi the loss needs to be accounted for. Unless one knows for certain what state the material is in take the lowest value for tensile strength.
|Thread: First attempt at threading on a bantam - all didn't go well|
Same as EN24 and, to some extent, EN3B. That's why you need to run fast with carbide inserts on these materials.
There was a thread (we Brits love a pun) recently on screwcutting where the finish improved no end when the OP ditched the cheap insert and bought a professional one. I'd also agree with Speedy; if EN16 is anything like EN24 you'll need a high surface speed to get any sort of decent finish with insert tooling. I'd start with EN1A.
|Thread: Bronze for bushings|
The bronze will be fine for the application. Many bronzes have a tendency to grab drills; I use slow helix drills to start and then bore. I rarely ream; the last time was for my traction engine water pumps. Same as drilling, the material seems to close up slightly in that the reamer is very tight to turn by hand after reaming.. To get the water pumps rams to fit I had to grind them 3 tenths undersize. So either my secondhand reamer is fudged or the material closed up slightly.
I wouldn't bother with the reaming, just drill and bore. Measuring small bores accurately isn't simple. I usually make up a plug (or use a piece of the same silver steel as the crankshaft) and assess the fit once near to final size.
|Thread: Hole cutter|
If you're stalling a 3/4hp motor with a 4 thou DOC then you need to sort that out first. And changing the lathe is not the answer, it should be capable of much more. All talk of cutter types and methods is a complete waste of time if you don't sort out the stalling issue, as you'll neve be able to use them.
|Thread: Learning CAD with Alibre Atom3D|
Thanks for that David. I'm interested in skew bevel gears, just out of idle curiosity. I'm aware of the work by Kozo and others modelling the gears for logging locomotives. But their pinions have teeth that have considerable lean on them, not at all like the full size gears. I want to have a go at making a set that look like the real thing. I must have a login to the Alibre forum; simply (?) a case of finding out what it is.
I started modelling a set of skew gears. A passable version of the gear is fairly simple using a loft to a point on a circle rather than to the centre of the gear:
But I miserably failed to model a pinion. Skew gears are a complete fudge. As the pinion rotates the gear teeth not only rotate with the pinion but they must also slide along it. So that means the DP must be changing. So the gears can't mesh along a line, but must be at a point. I gather the original gears were rather noisy and prone to wear. On the Alibre forum post it was surmised that the full size originals were mucked about with in wood until they meshed, and then were cast from the wooden patterns. I'd go along with that.
Thought I'd better split the previous post into sections in case I got dumped for rambling on for too long!
Although the model creation is involved I needed it to accurately represent the gear as I would be making real gears from the model. The gears above were created as 1DP; to get the final 16DP gears it's a simple scale by 1/16. All the little errors created in the original modelling usefully get reduced by the scale factor. After some faffing with the CAM program the bevel gear can be machined on a CNC mill, with a 1mm ballnose cutter:
And the gears fitted on the governor:
The meshing is smooth. Some while back at a SCMTEG meeting a member, who is also on the forum, was observed having a detailed look at the gears and turning them to check meshing. He didn't comment to me, so either they were dreadful, or he couldn't find any imperfection to point out.
This is how I modelled a bevel gear, although be aware that I use Alibre Expert. First I calculate all the parameters. Then I set up the construction planes and axes to get a plane that represents the outer face of a tooth. Then I draw a cone that represents the root angle cone. The gear is initially drawn as 1DP.
Then a sketch representing the involute function of a single tooth is placed on the plane representing the face angle of the outer edge and adjusted to line up with the root cone. The involute sketch was imported as a DXF gear, the unwanted parts deleted and the DXF exploded to give a sketch consisting of straight lines. All in all a bit of faff:
The sketch is then lofted to a single point at the origin to form one tooth:
There seems to be bug in Alibre here; the loft won't work to a point even though the manual says it will.. I needed to draw a 2 thou square, 7 thou above the origin, and then the loft worked. Note that the outer edge of the tooth isn't a good fit. That's because the sketch was drawn on a plane rather than projected onto the cone that represents the outer surface of the gear. Next is a simple circular pattern for 16 teeth, although it takes a while to compute:
Finally a nunber of simple revolve cuts tidy up the model so we have a proper gear:
An assembly can then be created to visually check meshing:
Want the latest issue of Model Engineer or Model Engineers' Workshop? Use our magazine locator links to find your nearest stockist!
You can contact us by phone, mail or email about the magazines including becoming a contributor, submitting reader's letters or making queries about articles. You can also get in touch about this website, advertising or other general issues.
Click THIS LINK for full contact details.
For subscription issues please see THIS LINK.