Your Choice Of The Most Difficult Ordinary Model-engineering Operation?

[DiogenesParticipant @diogenes]

Skew gears are an interesting subject- please do share your experiences, er when you experience them..

[Nigel Graham 2Participant @nigelgraham2]

Thankyou Julie, for the encouragement.

I am part way through re-commissioning a small, ancient Denbigh H4 Horizontal Mill, and gear-cutting is one of my possible uses for it.

I do have a dividing-head and rotary-tables; and in the last resort of course it’s not difficult to make a divider based on borrowing change-wheels from one of my lathes.

Its table-feed is missing (or was never fitted) but the worm-wheel under the table is there, and I might be able to make a replacement worm, from mild-steel (I’ve never going to wear it out!). I think I managed to find how to calculate the pitch though have to assume the pressure-angle. If I have the change-wheels for the peculiar number, I think the Harrison L5 lathe with its 1/4″ lead should cope with it, with care.

A full worm and wheel is another matter though.

 

As for the hobby taking over the house, I think my model-engineering is the householder and I, the lodger! Even going to and from my computer in the corner of the front room involves avoiding falling over a 7.1/4″g loco awaiting repairs.

[Nigel Graham 2]

On Nigel Graham 2 Said:

 

The silver solder I called 1/32″ diameter is really 1mm dia, 455 grade.

The paste is the same 455, but of thick consistency, dispensed from a syringe. It cannot be brushed or wiped on. It’s probably better for assembling flat surfaces to each other. Pushing the fitting onto the pipe just scrapes all the paste off.

I tried sealing the gap but it is too large, probably around 0.01″ wide.

 

I cannot satisfactorily photograph anything so small. The mimimum range for my camera is about 1 metre.

The nipple is a flanged bush; flange 0.3″ dia X 0.06″ thick, stem 1/4″ o.d. X 5/16″ long. Pipe o.d. 3/16″.

Placed for soldering with the pipe vertical, flange face supported on the refractory slab. Held there by a small clamp-stand clipped to the pipe well away from the joint.

Proper refractory, not firebrick – one of CuP’s small silver-brazing hearths.

Hi Nigel,

The link to Johnson Matthey   says Silverflow 55 requires a gap of 0.05 to 0.15mm to flow properly, which is quite a rattly fit.

[old martParticipant @oldmart]

I found that silver soldering brass hinges onto brass tubes which had a wall thickness of 0.010″ with oxy acetylene using the smallest available nozzle size. The tubes were about 2″ diameter and needed great care not to burn a hole in the tube wall.

[SillyOldDufferModerator @sillyoldduffer]

Design.  Anything complicated that’s outside one’s experience, can’t be copied, and has to satisfy a need is difficult.   Especially if it has to be delivered to time and budget.

Dave

 

[Nigel Graham 2Participant @nigelgraham2]

Dave H –

Thankyou.

That’s no more than 0.006″, but I think I had a gap larger than that.

The difficulty part is drilling a hole slightly oversize hole but not too much.

 

Dave (SOD) –

Tell me about it! It’s hard enough from a few old and rather inconsistent photos and dimensions from contemporary trade advertisements!

[bernard towersParticipant @bernardtowers37738]

Old Mart if you use excess oxygen it reduces the temperature of the flame, that might help. By the way you know if you have excess oxygen as the flame gets noisy.

[Diogenes]

On Diogenes Said:

Skew gears are an interesting subject- please do share your experiences, er when you experience them..

My particular interest is straight tooth skew bevel gears, as used on the Climax logging locomotive.

I am familiar with the designs by Kozo, and they are very clever indeed. I understand they are based on hyperbolic geometry. The tooth forms are noticably asymmetric whereas on pictures of the gears on Climax locomotives the tooth forms seem to be symmetric. The interesting question is why the difference?

Julie

[Nigel Graham 2]

On Nigel Graham 2 Said:

…I might be able to make a replacement worm, from mild-steel (I’ve never going to wear it out!). I think I managed to find how to calculate the pitch though have to assume the pressure-angle. If I have the change-wheels for the peculiar number, I think the Harrison L5 lathe with its 1/4″ lead should cope with it…

Nigel:

In theory making a slow helix worm is simply an exercise in screwcutting. The difficulty comes about when looking at the worm/worm gear pair. One of them must have an irrational pitch. For example if the worm gear is based on an integer value of DP the circular pitch of the worm gear will be irrational due to the presence of pi in the calculations. Since the CP is related to the thread pitch on the worm that means the thread pitch on the worm will also be irrational.

My Harrison M300 has a range of DP (8 to 56) and module (0.3 to 3.5) ‘threads’ that can be accessed using a special 56 tooth gear in the screwcutting chain. Fortunately I have the 56 tooth gear thanks to Ebay. Is your L5 the same?

Julie

[jaCK HobsonParticipant @jackhobson50760]

Keeping calm after making the mistake… then fixing it without scrapping the part.

[jaCK HobsonParticipant @jackhobson50760]

Set-up, or the subset: work-holding.

[jaCK HobsonParticipant @jackhobson50760]

Many jobs are easy when done right… but easy to do wrong.

[mike TParticipant @miket56243]

Doing up a left hand thread, nut and bolt. My spanners are pre-programmed to turn the wrong way.

Mike

[jaCK Hobson]

On jaCK Hobson Said:

Keeping calm after making the mistake… then fixing it without scrapping the part.

While I don’t have a huge amount of time for Adam Savage, I did read Every Tool’s a Hammer, which would have been worth it just for his advice on ones reaction to cock-ups. I am now just about able to accept them as a learning opportunity, and all in a day’s work. At worst, the scrap bin gains weight. Nobody died.

[old martParticipant @oldmart]

Bernerd, the excess acetylene reducing the flame temperature was exactly what helped with the thin wall brass tube, I am unlikely to do anything like that again, it was about 45 years ago.

[Speedy Builder5Participant @speedybuilder5]

Fat Fingers  and shakey hands!

Bob

[Diogenes]

On Diogenes Said:

Most difficult thing? – ‘proper’ engineering chisel-work; ‘square’, clean, crisp and accurate rebates, slots, pockets – even flatting..

The only time I ever saw that being done was remedial work on a keyway in a 2 inch shaft which had been assembled into a complex machine assembly and the keyway was then found to have been machined slightly off centre, so it was chiselled out with a cape chisel and filed out to make a fit – I know this because as a apprentice, I had to file it, and make the slightly oversize key.

[Nigel Graham 2Participant @nigelgraham2]

Julie –

Well, yes, “in theory”!

I measured the wheel as best I could without a proper gear-tooth micrometer, and determined its most likely form from the chapter on worm-gear in Ivan Law’s book. He gives a table of pitches, but still leaves you to try to calculate the blank diameter and depth of cut.

The circular pitch is 0.2618, giving 3.8197 TPI.

A lot of juggling of vulgar-fractions, the five-times-table and 127 gave me approximations I narrowed to [0.25 X (127 / 100) X (70 / 85) ] = 0.2615″ pitch!

3.8241 TPI.

Near enough, I suppose, over barely 4 or 5 turns. I don’t think there are any tenths of thous in my workshop anyway…. Still, this is to replace manually operating a feed-screw with a short handle, at very low speed, so not much torque and load.

 

I’ve only to ascertain I have those change-wheels: I am pretty sure I do.

 

The lathe should handle that although the manual does warn that trying to cut threads coarser than the leadscrew requires considerable care. It will amble round at about 60-70 rpm, and nice shallow cuts should be all right. Remembering I cannot use the half-nut release and thread dial indicator for an irrational-number pitch.

I have built a Hemingway Kits ‘Worden’ T&C Grinder so have no excuse for not forming the finishing-tool to the correct angle, or very closely so.

 

Ivan Law does give a method for calculating peculiar ratios to very high accuracy by a sort of extended long-division series, but I cannot make head nor tail of it! I simply use a calculator or spreadsheet to make successive approximations with the change-wheels I have, to find the best match.

[Nigel Graham 2]

On Nigel Graham 2 Said:

The circular pitch is 0.2618, giving 3.8197 TPI.

If I’ve understood correctly that equates to the worm wheel being based on 12DP. I assume the worm is single start?

Unlike spur gears worms and worm wheels are a law unto themselves. Even though there are standards they are often ignored by manufacturers for convenience.

To make the worm several pieces of information are required. The thread pitch, the thread flank angle, the thread depth and the pitch circle diameter/OD of the worm.

We know the thread pitch aka the circular pitch of the worm wheel. The error you give will be inconsequential.The method given in the book by Ivan Law is that of continued fractions, been around since Euclid and extended since then. Not something I’ve used but doesn’t look too difficult. Apparently it is useful in number theory but I never came across it when I was teaching myself number theory way back when. The same method is detailed in the manual for my universal dividing head for calculating gears trains for a particular lead when cutting helical gears.

The flank angle is most likely to be 14.5 degrees, ie, 29 degrees in total. In other words Acme form. For Acme form the total depth of the thread is 0.6866xPc. In this case 0.6866×0.2618=0.1798″.

The OD can be whatever the manufacturer desires. An estimate can be got by measuring the curve on the top of the teeth on the worm wheel and adding two times the addendum.

If this were me I’d model the worm in CAD and 3D print one to see how it fits before committing to cutting metal.

My go to book for gear design is:

Gear Design Simplified

Franklin D. Jones & Henry H. Ryffel

ISBN 0-8311-1159-3

It is an American book so concentrates on DP style gears.

When I first started cutting gears I bought a gear tooth vernier and spent ages mucking about with it tweaking the depth of cut. I found it very difficult to get consistent measurements. In the end I put it back on the shelf and simply cut to the depth given on the particular gear cutter.

Julie

[Nigel Graham 2Participant @nigelgraham2]

Well, I’ve sorted both problems…

Hopefully.

I have made a new pipe assembly, but used a slightly unconventional approach. The two union nipples are of the flat-face type.

The pipe is for a boiler pressure-gauge moved to a lower position for visibility, on a bracket on the bunker to one side of the boiler. This is part of modifications to aid driving the loco, a Ken Swan 1/3 version of the Kerr-Stuart ‘Wren’, on a ground-level track. (I made my feelings very well known when other club members suggested hacking big ‘oles in the cab roof, having been one of the original builders!)

I ascertained first that the two 1/8″ BSP union nuts could be accommodated loosely, behind the 3-jaw chuck’s jaws.

Rather than turning the two connectors complete, I pilot-drilled them about 1/16″ or so through, then drilled the stop-ended hole for the pipe, turned the body and flange outside diameters, parted off.

I had estimated the pipe’s length from the formed first attempt, with a generous allowance for the siphon, but left it straight so far. Putting it in the lathe I used a file to reduce its o.d. at each end very slightly, to slide-fit the nipples.

Then cleaned all the parts thoroughly and this time worked the 455-grade silver-solder paste into each fitting before inserting the tube…. ensuring the two nuts were on and facing the right way before fitting the second!

‘Tippex’ all over the nipples. I read the label to ensure it is the solvent not aqueous type, before actually taking it to the Co-op counter.

The cartridge-gas plumber’s blowlamp I use gives plenty of heat.

Cleaned and pickled the assembly – this time it looks sound but I ought still give it a hydraulic test before fitting it to the loco. There is some other work on that first anyway.

Put the assembly in the lathe, faced the parted ends, drilled for the connection spigot on the gauge and banjo connection. I made the banjo from solid.

Only now formed the pipe to a close match to the first attempt: I won’t be able to finish that until I can get back to the loco itself. I use a small pliers-action, 3-diameter pipe bender for such work.

….. So fingers crossed!

 

………

 

The other tale of two troubles was how to turn a worm – encouraged on this thread(!) by Julie.

I calculated the gearing for my Harrison L5 lathe supposedly giving a pitch within a few thou of right. Then a stock-take this afternoon revealed I don’t have 3 of those 4 change-wheels as I had thought, so back to the PC, wind up ‘Excel’ and start again.

This time, I found a sensible combination even closer, using the wheels I do have!

worm p = 0.2618″. leadscrew 0.25″

0.25 X 55/60 X 80/70 = 0.2619″.

Can’t really argue with that.

This snippet gives part of the operation. The spreadsheet is not as elegant as it might be, but it works. The working is actually on another part of the page, and I copy the values to this little table which derives the error as [0.2618-(formula result)].

(I learnt to drive ‘Excel’, ‘Word’ and MS-Win-assorted at work, where I became quite familiar with making arithmetical spreadsheets complete with graphs, to a modest level of maths. There is nothing magic with the method here though, just plugging estimated tooth-counts from the list into the ‘Excel-ese’ version of the standard change-wheel formula, until hitting the best selection. )

 

[Author]

Posts