Member postings for Andrew Johnston

Forming the real spectacle plates was a mixed bag. One total bog up and one that looks ok. Here's the one that looks ok, after trimming the flanged rim to roughly the right depth on the mill:

I've also learnt some valuable lessons about flanging allowances that should make the remaining plates easier. I've run out of oxygen and acetylene so that'll have to be in the New Year.

Tomorrow I'll go up to the gliding club for a fettling session with angle grinder and power drill on the cylinders and supporting castings prior to machining.

Andrew

Out of idle curiosity I looked up the quoted losses for the VFD I bought to drive the motor on my Pultra lathe. The unit is by Yaskawa and is single phase in, 3-phase 200V out rated at 0.75kW in normal mode. The table gives a heatsink loss of 14.6W and an interior loss of 14.4W. The table quotes a total of 29.1W. That's a shade under 4%, which is pretty good.

Note to SoD: There are many companies in the far East that churn out commodity components in high volume and low price as they haven't had to do the development R&D. They're not companies that sell through distribution. I've been told to shove off because they weren't interested in volumes of 100k per year for a custom transformer, minimum was 100k/month.

Andrew

Looking good!

What material are you using? To some extent size errors are due to the material shrinking as it cools. For PLA I find that a good starting point is to scale by 1.005, ie, half a percent. Thin walls print best if they are a multiple of the nozzle size. My nozzle is 0.4mm so I chose 0.8mm, 1.2mm and so on. A raft or brim helps the first layer to adhere. It's usually controlled in the slicing software.

Andrew

Quite so, but that's not what Robert was saying. If one has a VFD capable of x watts output but one is only running at a maximum of x/2 power then the effective resolution of the signals such as output current is reduced by half. In practise that may not matter, but an axiom of signal processing is to start with the biggest signal, and best S/N ratio, one can.

It will also depend upon how the signals are processed. Any cheap processor based on mid-range ARM cores will have floating point hardware and probably sigma-delta ADCs. So loss of a bit or two isn't critical. But really cheap VFDs couldl use really cheap microcontrollers without floating point, at which point loss of resolution can be more important for algorithm stability.

I doubt that VFDs use discrete IGBTs, or MOSFETs, these days. Way too expensive to assemble. They will use 3-phase H-bridge modules with dies mounted directly on a substrate. You get better heat flow from die to heatsink that way too.

Fundamentally I agree with Clive. for any properly designed VFD there is no point in using a rating of twice the motor power. Simply use the next power rating up from the motor power.

Andrew

Finished my annual machine tool maintenance program. Good clean up and hoover round, chuck out the junk and scrap metal. Put things away if I want to keep them. Oil all zerks and check and fill levels for all gearboxes and spindles. Stone the tables on the milling machines. I've been given several sheets of neoprene rubber, so I've cut some mats to put in front of the main machine tools. We'll see how they get on with swarf, but they should keep my tootsies warner than standing on a concrete floor.

Just got the other side of the workshop (bench and hand tools) to sort out. I suspect there's lots of stuff that can be recycled, which should free up some space.

Also CNC machined radii on the remaining three corners of my fixture for hot bending the spectacle and front plates for the traction engine. Tomorrow I'll have a go at forming the real plates.

Andrew

I have a new method for assessing swarf temperature; it's called the oh beep test. If a bit of swarf gets on your hand and you say oh beep then the swarf is hot enough and you're making the cutter work for it's living.

The shell mills from Arc look good and are very reasonably priced. Following the positive review by Jason I intend to get a couple, and R8 and ISO40 arbors, so I can use them on all three milling machines.

Andrew

There's one problem with saying it's ok for experienced people who know what they are doing to use such techniques. How do they get experienced in the first place without taking risks? Personally I wouldn't use the technique, or at least I'd make the tool from thin gauge plate so there isn't a sharp edge.

It's strange that he winds the handle like crazy on the dividing head to engage the appropriate slot on the indexing plate. On my similar dividing head you can disengage the worm and simply rotate the chuck, much quicker.

Andrew

Paul: Thanks for the note on rivets; although I can't see myself going to the Caribbean just to buy some rivets! I have a casting for the chimney cap, but I plan to spin my own from copper sheet. Once I've got a chimney made I'll leave the top few rivets blank until I know whether they're going to be in the way or not.

Mark: Useful tips on using the rolls. Once I have a former made I'll put them into practise as it would be good to use the rolls for at least an initial bending.

This morning I refitted the mirror in the bathroom. Some idiot fitted it last night but 100mm too low.

This afternoon I experimented with hot forming some 3mm steel sheet. Here's the partial former, test piece and quick homebrew wooden mallet. Note to Jason: look away now. The theory behind a wooden mallet is that it doesn't mark the hot metal:

Having done an initial trial with a copper mallet I can say that the wooden mallet is much better. Oxy-acetylene was used as a heat source. A close up of the test piece:

And a nice sharp internal bend:

Next job is to add the second radius to the former and then form the spectacle plates for the traction engines. Then I can add two smaller radii to the former and make the front plates. Note that the fullsize engines have different radii on the spectacle and front plates. Gotta keep the rivet counters happy!

Andrew

I don't bother reducing the feedrate for finishing cuts. When boring in particular increasing the feedrate at small DOC can help to eliminate chatter. I rarely use a feedrate of less than 4 thou/rev. Finish versus depth of cut is a complex subject. It depends on a number of factors such as material and speeds and feeds. For instance EN3B is poor with inserts and small DOC, but for austenitic stainless steels it makes little difference.

Andrew

Eloquently put. If I ever finish my traction engines I fancy making a clock, and darn right I'd be using CAD, CAM and CNC as well as hand methods.

Andrew

There are those that talk about it and those that do it.

CAD:

CAM:

CNC:

Real gears:

I thought that the involute tooth form gave primarily a rolling contact, with a little sliding. It would seem strange to use a tooth form for power transmission that has inherent losses due to friction over another tooth form.

Andrew

The boring bar shouldn't be deflecting that much. What insert are you usng and where did it come from? Also what feedrate is being used?

Even if the bar is deflecting it isn't a problem. As you get near to final size take a couple of spring cuts without moving the cross slide before measuring the bore.

Andrew

There are two obvious solutions, although I don't think one can automatically say one methed is faster and/or more accurate than another. It will depend upon the equipment available. One, drill and bore in the lathe. How quick will depend upon the largest drill available. My largest drill is 1-7/8", which still leaves a fair amount to be removed with a boring bar. Two, drill a series of thru holes, which slightly overlap, on a circle. Of course that's a lot easier with a DRO that has a bolt circle function. Once drilled the centre can be tapped out leaving a few tens of thou to be cleaned up with a boring head. With method two you get get keep the core in the offcuts bin as well.

Andrew

I made a set of buttons some years ago when I got my CNC mill. Very useful; but I'd make a set of three. That way you can use two for alignment and the third, in a different T-slot, as a stop. A set of hex blocks with an eccentric screw are good for holding work in conjunction with the buttons, while leaving the top free, like these three:

I bought the hex blocks and screws, but made the T-nuts.

I also made a set of buttons, but much taller than standard, which can be useful for larger work. I cylindrically ground my tall buttons, but that's a bit OTT. The tall buttons can just be seen in this picture:

Andrew

Only the purchaser can decide if the instrument is worth the asking price; there's no absolute scale of value for money.

While it might be a fine instrument it isn't really correct for aligning work on a lathe. Here are two instruments:

The instrument on the left is a dial gauge. It will accurately indicate the movement of the plunger. In the case of the one shown over 3mm by 0.002mm increments. However the plunger needs to be perpendicular to the work. They are normally used in a holder on a surface plate for comparison purposes. The instrument on the right is a dial test indicator (DTI), and is what I use for measuring runout and setting up the 4-jaw chuck on the lathe. Total movement is small, less than 1mm, but the probe can be moved to any angle, relative to the body, that is convenient. While the DTI doesn't give absolute readings it's much more flexible when it comes to setting up work on the lathe.

Andrew

I store MT1, 2, 3 and 4 drills, and sleeves, on a piece of MDF above my lathe. I just drilled parallel holes so that each taper goes through about 2/3 to 3/4 of the length. The same would work for reamers, although I prefer to keep my reamers separately packed in drawers. Completely OTT to make tapered holes.

Andrew

Jason: I found some beech on Ebay for about £40, but it was spalted. Needed to look that up! Seems a shame to use wood though as once the job is done the wood is redundant. At least with steel the swarf can be recycled and the former cut up and used elsewhere on the engines. My new kitchen worktops were not real wood - way out of my price bracket!

Paul: I've looked at your TT build, thanks. There seems to be very little about chimneys on TT, and many people seem to get them rolled outside. Being a paid up member of the awkward squad that means I want to do it myself instead. I've tried something similar to your trick with the round bar. I've now got a chimney that looks roughly right. But I made a mistake with the pattern - didn't allow for the thickness of the metal. So it needs a bit taken off the one of the straight mating edges. I've been wondering about the rivets and whether I can close them from the outside. Where did you get the flat head rivets? I've got pictures of the chimney on two full size engines. One has no rivets, which I assume means it was cast. On the other there are two columns of rivets, so an internal strap is used. That's what I plan to do to avoid the need to form a joggle.

Diogenes: Interesting idea. Since I already have the sheet I think I might do the reverse, ie, roll a "cylinder" from sheet and then pull it round a former as Paul did to get the cone shape.

Nowt is going to happen until the New Year in terms of ordering material. So I'll sulk think about it over the break while I have a go at hot forming the spectacle and front plates.

Andrew

There was a video online showing a 4 roll in action; basically a pinch roll but with free rolls in front and behind the pinch pair. But that doesn't help me.

I'd considered wood, but I've no idea what to buy or where to get it. I just tried an online shop; for 100mm by 100mm by 500mm long of American Ash I was quoted £13852.

Think I'll stick to steel!

Andrew

Not sure, but two washing up bowls three quarters full. say 15kg? I don't think refiners have an easy way to separate metals. I used to recycle at a proper scrap yard. Steel was just dumped and measured by weighing the vehicle on the way in and out. Copper alloys and aluminium were weighed separately as they're way more valuable. Last time I went 70kg of steel was worth £6 whereas 4kg of brass/bronze was worth £30. But it's a round trip of 60 miles and takes hours to go round. Plus it's a bonus coming out with inflated tyres given the amount of scrap lying on the ground. It's easier, but still a pain, to go to the council tip. The idiots that claim to run things, and business, bang on about recycling but they sure don't make it easy.

Today I've debunked another model engineering myth. Rolling a cone by tilting the third roll on a set of pinch rolls plain doesn't work. Of course if one thinks about it, it would be obvious why it doesn't work. Slanting the third roll changes the diameter being rolled along the length. But the pinch rolls feed both ends of the pattern at the same speed whereas the larger diameter end needs to be fed faster than the small diameter end. Looking online it's clear that the professionals use special cone rolls to form proper cones.

I've managed to get an approximation to a cone with BF, BI and a copper mallet. But to get the final shape I think I'm going to have to make a former. A 500mm length of 100mm diameter steel ain't cheap, but isn't completely out of order!

Andrew

This morning I took a car load of stuff to the council tip; mostly metal for recycling and loads of non-working and obsolete light bulbs. Got rid of all the lights I've recently replaced with LED ones. Plus two big bags of aluminium swarf, two buckets of steel swarf and two bowls of brass/bronze swarf. There was a separate container for brass (didn't ask if I could put bronze in it as well!) but aluminium is now dumped along with steel. Seems a bit daft to me.

This afternoon I've been chomping up sheet metal on the guillotine. I used paper patterns to nibble out the chimney and final drive gear cover blanks from 1mm sheet. Blanks have been cut for the tender bottom from 2mm sheet as well as blanks for the tender sides and the spectacle and front plates from 3mm sheet. Here's a selection, with a 24" rule at the bottom for scale:

Should really be fitting the mirror in the bathroom this evening, but I also need to go to the supermarket. I suspect that some beer will be purchased.

Andrew

Edited By Andrew Johnston on 18/12/2019 17:16:07