Wheels for the kitbuilder

[belstone]

I mentioned a while ago the idea of making a "generic" six-coupled chassis available as a platform for small manufacturers to develop new body kits using 3D printing and/or etched brass, and straight away the problem of driving wheels came up.  Nothing available (apart from 2mm finescale), horrendous tooling costs to make a range of different sizes, and the bit that always puts people off, getting the back to back and (especially) wheel quartering right. I was thinking about this on the way in to work this morning and came up with a idea.

The wheel itself is machined (from brass? or nickel silver?) with a hollow centre and a short, hollow stub axle to take a slotted brass bush which slides into a U-shaped cutout in the chassis side frame.  The axle is moulded in hard plastic, with either a gear or a plain larger diameter section in the middle, and square outer ends.  Wheel centres are either moulded or very possibly 3D printed with a square centre. 

Push the wheels onto the axles until the ends of the stub axles contact the gear or centre section - back to back set automatically.  Press the wheel centres onto the square outer axle ends - quartering set automatically.  The wheels can be knocked out on a CNC lathe for pennies, and possibly the only serious tooling cost is the moulded axles (assuming 3D printing of the centres is viable).  They will be solid backed of course, although there is nothing to stop someone drilling lots of holes in the wheel backs to give a fair impression of see-through wheels.  The centres can be made with a flat outer face so that the end of the axle cannot be seen, and the same wheels can be fitted with centres for spoked, Boxpok, LNWR pattern and any other wheel. Crankpin would probably just be a 14BA screw in a tapped hole in the centre.

Even better, if you put a groove in the centre of the plain axles to bear on a compensating beam, then "Flexichas" compensation becomes very simple indeed.  That means you don't even have to get the chassis frames (etched brass with plastic spacers) perfectly square.  Add in a simple fold-up etched motor mount, attached to one of the spacers to insulate it from the frames and you have a chassis that is really no harder to assemble than an Airfix kit.  No pickups to worry about, DCC-friendly and possibly not too expensive.

Feel free to point out the obvious flaws - I don't know enough about 3D printing to be able to say whether a spoked wheel centre is possible with the technology.  I can see something like this being the basis for a huge range of simple four and six coupled locomotives which are never likely to be available RTR.

Richard

[Dr Al]

I don't see how you'd deal with accurately citing, drilling and tapping for the crank pins for rods - this is as critical as quartering and you'd then need to screw rods into 3D printed plastics, which can be brittle. Certainly thin FUD prints can be.

Also, I'm not convinced by compensation in N for any standard type trackwork (e.g. Peco code 55) where the frog gaps are relatively large, as you'll just get huge frog drop, which is really bad for running (actually more at higher speed than low). I suspect that's why most chassis are a lot more rigid. Also, rigid chassis can perform fine if set up right - the Farish 04 shunter (effectively a steam loco in terms of mechanics) was acclaimed for its slow speed and pickup performance despite its diminutive wheelbase. Ok, a few QC issues of wiper pickups needing bent out, but fundamentally, design wise, it's spot on.

One would also question the demand - ABS/Beaver only did one run of various steam loco wheelsets precisely for the kind of use here, and apparently took years and years and years to sell them.

If this was a project that was to progress on, I'd approach Dapol and DJM to enquire if spares of existing or upcoming tooling (plenty of variations in Dapol range, and forthcoming from DJM) could be made available. e.g spare J94 wheels from DJM - I suspect DJM would be most receptive to this to get maximum out of his tooling.

Cheers,
Alan

[PGN]

Belstone -

I have been thinking along these lines for some while, too, but like you I lack the technical design and engineering skills to tackle it.

May I suggest a slight refinement, though: the stub axle needs to be an integral part of the wheel. The wheel doesn't push onto it: the wheel is permanently attached. The stub axle has a circular cross-section where it joins the wheel and passes through the frames, but this is transformed into a square cross-section at the "inside end"; and the muffs are made with a square-profile aperture for the axles to press into. That way, accurate quartering is guaranteed.

[Dorsetmike]

ABS  used to produce some excellent wheels, but found it was not commercially viable, it took him years to sell one production run, one problem with them was quartering, no square holes, I believe an earlier version by ABS Beaver did have square holes, both types occasionally appear on Ebay.

[PGN]

ABS  used to produce some excellent wheels, but found it was not commercially viable, it took him years to sell one production run

Quite so - but that is precisely why 3D printing is such an attractive alternative solution. With 3D printing there are no "production runs" and no capital tied up in stock lying idly on the shelves - the goods are printed to order.

[Rabs]

I think that there are big differences between the market then and the market now.  At the time that ABS and beaver made their models a producer of wheels was catering only to scratch builders (a vanishingly small number in N) or hoping that they could persuade some of the small number of kit manufacturers (Langley or P&D marsh for example) to design using their wheels.  Unfortunately they never did that, preferring to make use of the nearest equivalent farish chassis for their kits.  It seems to me that there have been two major changes in recent years:
1) We expect a higher level of fidelity in our models, so compromising on the nearest off the shelf chassis isn't something kit designers look to do any more
2) There seem to be a much larger number of minor kit builders making use of 3D printing for body shells

Together I feel that these two factors means that things are sufficiently different today for the 'old don't bother, it's been tried' argument about N gauge wheel sets to need a careful rethink.

I like Belstone's concept but I don't quite follow the detail of the idea.  Could you stick a drawing up?

[Roy L S]

Belstone -

I have been thinking along these lines for some while, too, but like you I lack the technical design and engineering skills to tackle it.

May I suggest a slight refinement, though: the stub axle needs to be an integral part of the wheel. The wheel doesn't push onto it: the wheel is permanently attached. The stub axle has a circular cross-section where it joins the wheel and passes through the frames, but this is transformed into a square cross-section at the "inside end"; and the muffs are made with a square-profile aperture for the axles to press into. That way, accurate quartering is guaranteed.

I understand that 3D printing is moving on such that metal items or even composites of metal/plastic (for example) are possible at least in theory. If a wheel/stub axle could be printed in brass and the tread skimmed that would seem a possibility but cost would be the factor, I don't think such prints are that cheap yet but like  other contributors here I have insufficient knowledge of the current state of the art in the 3D print world.

Roy

[belstone]

Alan, I think the "rigid vs compensated" argument was beaten to death with a big stick a long time ago in 4mm, and no-one on here is going to thank us for reopening that particular debate   A lot of the 4mm Flexichas designs were rather too flexible IMHO - I remember seeing Iain Rice demonstrate one of his locos running over a matchstick placed across one rail, which probably did more harm than good in showing what the system was supposed to do. 0.5mm vertical movement per wheel should be plenty for N gauge and might not need jointed rods given the amount of slop in the crankpins on a typical N gauge model. Interestingly the J39 has at least that much movement due to the bushes being a sloppy fit in the chassis.  All it needs is a compensating beam between the first and second axles, and the rear axle bushes secured to stop them moving.  Hmmm...

I agree with you about the difficulties of getting precisely accurate crankpin positioning: 3D printing is still a fast-developing technology and accuracy is getting better all the time. Something that is not possible right now might well be in six months time. Demand is impossible to predict. Back when the Beaver / ABS wheels came out, etched brass was still fairly unusual, and minidrills were only for posh people.  If you wanted to build a chassis you'd be looking at cutting, shaping and drilling a pair of brass frames by hand to a very high standard of accuracy, and even if you managed to get that bit right, usually the only way to acquire a motor was to buy a complete N gauge loco.  And then you had gears, pickups etc to sort. No wonder the wheels sold so slowly - they were about thirty years ahead of their time.

Nowadays if I want a pair of accurate frames and conrods to match I can draw them with a CAD package, email the drawing to PPD and have them back within a week.  And if they work, I can order another twenty at a very low cost per set.  Motors - we're spoilt for choice.  Gears - 2mm Assoc and several commercial suppliers.  All we need is wheels.  I don't think there are a huge number of people out there who want to build their own mechanisms, but there is only one way to find out.  Hence my thinking about developing a wheel that doesn't need expensive tooling and uses split-frame pickup.

PGN - the stub axle is integral with the wheel.  The two wheels are joined by a hard plastic axle with a squared outer end, and the hubs and spokes (with a square centre) push onto the square axle ends.  Machining the wheels with an accurate square profile on the end of the stub axles is much trickier than just having them as a simple turned item.  What matters is that the crankpins are at 90 degrees to each other, and I think that making the wheel centre separate from the rim and stub axle should do this if the tolerances are tight enough. But that's a big "if". I see the attraction with your way of doing things, but it requires the centres to be fitted to the rims securely and 100% accurately at the manufacturing stage. You could do it easily enough with a jig, and use the same jig to drill the crankpin holes at the same time.

Richard

[Nigel Cliffe]

Richard,

there's something in your idea and, I think, a few deep holes which could undo things.   

Your ideas are not a million miles from a design I've been kicking around with a 2FS colleague for a few years. 
The difference I have is that the crank pin hole and boss are metal, machined at the CNC stage, and the finished wheel is see-through.  I don't think ultra-detail 3D print material is up to holding crankpins, and if you're joining a 3D print to a crankpin and to a wheel, that's two joins which have to be highly accurate at assembly - joins are where errors creep into most small assembly jobs.
In my design, the 3D print/moulded insert is purely cosmetic, it has no mechanical use, and the wheel would function without it.  I, and my 2FS colleague, have got as far as identifying a CNC production company who could produce the mechanical parts to the required quality, plus a rough prototype. 

There may be scope to collaborate - the cosmetic centres could be the same for both, just differences in the CNC'd wheel tread (N vs. 2FS) and possibly how the centre for joining axles, gears (etc) is arranged.   Our planned CNC process already has different centres produced in a single production run, so different tyre profiles should not be a problem.

You may find that the idea of "pennies" for CNC production isn't credible for the numbers in question.  A few pounds per wheel is possible, with tied up capital of a thousand or two in CNC'd parts which have to be sold. Plus the cosmetic spokes, plus parts to join the two wheels, plus whatever is required to cover investment costs, stockholding costs, and any profit.   

If serious, get in touch and we could meet up.  A number of the 2FS wheel designers/makers, including me, live in Suffolk and Essex.


-  Nigel

[PGN]

I understand that 3D printing is moving on such that metal items or even composites of metal/plastic (for example) are possible at least in theory.

TBH, for the time being I wouldn't mind if the wheels were non-conducting, and the locomotives were freewheeling with either a Terrier-chassis tender drive, or a powered vehicle in the train providing the motive force.

[belstone]

Nigel, I'm not in a position at the moment to do much more than dream up new ideas.  I certainly don't have any spare money to put into this right now, just wondering if the basic concept is feasible.  One possibility would be to have the hub moulded in the same material as the axles (I'm thinking a hard plastic like Acetal) - more tooling investment, but the same hub would serve for the entire range of wheel sizes (provided people can live with all the wheels having the same crank throw) and it might even be possible to mould in the crankpins.  Then a 3D printed (or etched?) insert for the spokes and rim, purely cosmetic as you suggest. I suspect that with my design the same axles and hubs would serve for N and 2mm - wider gauge but narrower wheels in 2mm, so the width over the axle ends should be near enough the same.  The hubs would need to be narrower for 2mm - possibly get round this by making all the hubs to 2mm dimensions and fitting a spacing washer behind them for N gauge use.

Richard

[steve836]

I still can't visualise from your descriptions, but would certainly be interested if you get as far as production, especially if you could do some extended axles for outside frame locos.

[NTrain]

Each time you add an injection moulded item, the cost goes up in the thousands.

The trick is to find someone with a veryyyyy good CNC lathe/mill, who is willing to do small runs and add cosmetic 3D printed infills.

Getting wheels done in China usually requires an order for 5000 axles................