Machine drive shaft

[colin hamiltonParticipant @colinhamilton16803]

I gave an old alpine tractor. Its getting very hard and expensive to source parts from Italy. I’ve just spent a small fortune buying this small drive shaft. I was wondering could I machine this from a billet? I have a chipmaster lathe and a bridgeport vertical mill. I’ve never tried milling splines so would appreciate any thoughts/guidance

[DC31kParticipant @dc31k]

That item has been (drop) forged. There is a reason for this – so the grain flow of the metal enhances the strength. It has also been heat-treated on the splines (you can see the oxide discolouration).

How long did the original part last? How hard was the machine working for that period of time?

Extrapolate from those two and work how old you will be when the next factory part is likely to fail.

Milling splines is as easy as milling six flats on a round bar to make a hexagon. Finding the specification of those splines will be less easy. Buying or making the cutter to produce splines to that specification will be even less easy. An inaccurate male spline will also damage (and thus shorten the life of) the female spline into which it seats.

[FulmenParticipant @fulmen]

Splines aren’t that hard, it’s just a lot of work. You simply cut each face with a slitting saw using a dividing head.

Not sure about the tolerances of the cross bore, so I’m assuming reamer territory. Buy one if you can. If not I’m sure you can pull it off with a boring head or perhaps line boring in the lathe.

Without any material specs I’d take a chance on something like pre-hardened 34crnimo6. it’s tough but machinable with HSS with some care and cooling.

[Robert Atkinson 2Participant @robertatkinson2]

What failed on the orginal part?

As DC31K says it appears to be forged with induction hardened splines. If the splines wore or yoke cracked you are unlikely to be able to produce a replacement of similar durability. I would only consider making a replacement for a part like this if there was an obvious design or manufacturing error that I could correct or it was totally unobtainable.

Robert.

[HOWARDTParticipant @howardt]

The splines are machined with a form cutter, side and face.  Threads are rolled not cut and as said splines are induction hardened.  Material such as EN16, equivalent needed for hardening and torsional strength. Most EN16 is round bar so some serious metal removal required.  As said forging is really the way to go.

[Howard LewisParticipant @howardlewis46836]

+1 for the commnents about grain flow. A forging retains this, machining from a solid billet does not, so weaker and more prone to fatigue cracking..

IF you go down this road, always ensure that all corners are not sharp, but have a substantial radius to minimise stress raisers.

If possible make the section larger, to compensate for the loss of grain flow, and fatigue strength.

(A cardan shaft experiences many variations in angular velocity, and therefore torque, during operation, so guarding against fatigue is important)

Howard

[old martParticipant @oldmart]

You may be able to get a ready made shaft which is close enough from a car breakers or autojumble. It is one end of a universal joint, and the cross member with the needle rollers is the most stressed part. If you are only going to use the tractor occasionally, there would be no need for all the heat treating and if I had to make one, it would be from some en19t or en24t chrome moly. This is partly heat treated and considerably stronger than mild steel, but still machinable.

For splines, I would use a rotary table on a mill and would singlepoint the threads on a lathe. The yoke would be milled last, but you need to have a think about the exact sequence of operations before cutting metal.

[FulmenParticipant @fulmen]

Is it really forged? If I were to guess I’d say cast, and either way it’s usually more to do with cost and efficiency.

[colin hamiltonParticipant @colinhamilton16803]

The part is the final part of the drive to the front wheel. One end is a universal joint for the steering. A hub, which carries the wheel fits over the splines. A single nut fitted on the thread holds the wheel hub onto this part. The threaded section sheared off at the end of the splined section. I have had two fail in exactly the same place (in 10 years!). I’m no expert but the part looks cast to me as it has that sort of grainy look to it.

How would I go about identifying the splines? Are they standards (like threads).

[JasonBModerator @jasonb]

With nothing to loose I’d consider drilling it out and tapping the hole then screw in a piece of suitably threaded steel. might need to make it a smaller thread than the exposed end and when you make it avoid that undercut at the end of the thread. Hopefully the induction hardening of the splines is not too deep.

Looks like a cast steel to me

[colin hamilton]

On colin hamilton Said:

Are they standards (like threads).

 

Yes. And the beauty of Standards is that there are so many to choose from.

Start here:

https://en.wikipedia.org/wiki/Spline_(mechanical)

Decide which type it is likely to be (not wishing to influence you, but involute splines might be a good choice). Decide which of the main Standards bodies (US or ISO) are likely to prevail in the country (Italy) of manufacture.

Measure the diameter of the shaft and number of splines – that will be part of the standard. There will always be a relationship between diameter and number.

Let your fingers do the walking on Google…

https://roymech.org/Useful_Tables/Keyways/Splines_Inv.html

Note that many Indian Standards are direct parallels of the corresponding ISO Standards and the Indian ones are often much more easy (i.e. freely available) to obtain in pdf form them paying for a BS or ISO version.

—

If the female spline in the hub is worn, leading to slight relative rotational movement between it and the male, that will put considerable cyclic loading on the securing nut. Which will lead to the fatigue failure you appear to have seen (grainy, crystallised surface – put ‘fatigue failure surface’ into Google and look at images). If the securing nut is not properly torqued and staked over into the groove in the shaft (or if the nut is removed and then you attempt to re-use it) that will contribute to the problem.

The transition between end of spline and end of thread is not good – the relief groove they have made is square edged, which is a stress-riser. That transition should be as smooth and rounded as possible for best performance.

If the threaded part has sheared off, instead of remaking the whole thing, why not drill and tap the remainder for a shouldered stud? Loctite it in and if it fails, drill it out and replace it.

[Ian PParticipant @ianp]

If as you say the two failures are a break at the base of the thread then I think you may be able to make a modification and re-use the failed parts.

From the first picture it looks like the thread has been rolled but the relief at the base of the thread looks to be deeper than needed (and maybe not a big enough radius). My suggestion would be to through-drill the yoke and splined section and then tap (say M16x1.5 pitch) so you can use a HT bolt instead of a nut.

The splined section has been hardened but will not be too hard to drill and tap if you take your time, in any I would aim to make most of the hole clearance and have the threads at the yoke end. You can be a bit generous with the tapping drill size if you get a long thread engagement (say 2xdia).

Without know the dimensions of your part I’m only guessing that M16 would be suitable as there is a good source of long bolts. They are used on the rear wheel hub assemblies on several models of VW group cars.

Ian P

 

[Ian PParticipant @ianp]

DC31k beat me to it.

Through drilling and tapping though is easier than tapping a blind hole

Ian P

[Fulmen]

On Fulmen Said:

Is it really forged? If I were to guess I’d say cast, and either way it’s usually more to do with cost and efficiency.

I’m not so sure.  The part is highly stressed and likely should be well-made, either by drop forging or by casting.

Well-made means expensive – drop forging dies are costly, and although casting is cheaper, it only gets full strength if the part is heat-treated properly and tested for voids etc. Then the splines would be ground, rather than machined.

Unfortunately this is the sort of expensive part that gets faked.  Much cheaper to replace drop forged with a casting, especially if the casting isn’t heat treated to maximise strength.

As Colin has broken two of these in 10 years I’m wondering if they met the original manufacturing specification.  Apart from outright fakes, it’s possible the tractor manufacturer no longer makes spares, and they’re produced by a third-party to a lower specification.  Difficult to tell without lab equipment.

Could be machined at home, but will be weaker than a properly made factory part: might not matter if breakdowns are inconvenient rather than critical.

CNC is a possibility, but makers would need an accurately measured CAD model.

Dave

[FulmenParticipant @fulmen]

I think Ian’s plan is worth a try, it shouldn’t cost more than a quality tap and a few hours of work.

[Howard LewisParticipant @howardlewis46836]

Got to be worth trying drilling and tapping the shaft to screw in a replacement stepped stud for the retaining nut

FWIW a thread size midway / two thirds the size of of the failed thread.

Retain with an anaerobic sealer

Makesure that there is a good radius between the two threads, to avoid stress raisers

(Did you examine the failed surfaces?  A fatigue faiolure will show mas a series of striations across the surface.)

What about the way in which it failed?

If the thread has failed on two occasions, (Presumably at the undercut at the end of the thread?) it sounds as if is being overstressed, either in tension or in bending. Needless to say that the thread shouild not be subjected to bending.  If it is, find out why and remove the reason.

If it takes several years before failure, it sounds as if the loads are slightly above the fatigue limit, so try to remove any factor which can increase the cyclic applied loads.

The thread is there to provide an end load which rertains the hub, and once tightened should only see the tensile load applied by tightening.

Being part of a U J, the loading will be cyclic, most likely to result in failure in fatigue.

Possibly the hub is subjecting the thread to bending loads. Maybe a slight mod (shortening or lengthening with a shim to prevent any end float?) to the hub might cure the problem.

Are the splines loose fitting which will increase loads? If so, assemble hub to shaft with a gap filler (again an anaerobic?) to remove clearances.

HTH

Howard

[Howard Lewis]

On Howard Lewis Said:

…try to remove any factor which can increase the cyclic applied loads.

Taken at face value, that made me smile. What could be done? Only drive forwards and never reverse? Never drive down a hill?

—

I am unconvinced with the ‘bolt on’ idea above. Not so many cars have wheel bolts these days. If the threads in the hub fail, you have to replace the hub. With wheel studs, you knock out the stud and install a new one.

A deep clearance hole with threads at the bottom risks removing too much material from the centre of the yoke such that it fails in a different manner (e.g. torsion).

[Fulmen]

On Fulmen Said:

I think Ian’s plan is worth a try, it shouldn’t cost more than a quality tap and a few hours of work.

To be fair, Jason suggested drilling and tapping before I made my reply.

Ian P

[Howard LewisParticipant @howardlewis46836]

I am not referring to driving forwards or backwards, so don’t grin too broadly.

All Hooke’s Joints deliver a fluctuating angular velocity once the joint is no longer straight, despite a steady input.

(At Rolls Royce, we used to use this feature to calibrate torsiometers)

This is why the joints at the two ends of a cardan shaft are carefully aligned, so that the angular velocity variation produced by one joint is cancelled by the second joint.

Wheel bolts are used as a cost saving.   One tapped hole in the hub, per fixing, and a fairly straightforward bolt as opposed to a stud with two threads, and a wheelnut, as well as the tapping in the hub.

Leylands went a lot of trouble, and cost, to secure the studs in the hubs of their vehicles. On a fleet of over 60 vehicles, I only had to replace one stud in three years. Bristols used an oversized version of what some cars used. All four hubs and all the studs on one vehicle had to be replaced when the nuts had been overtightened!

In both cases the thread was 7/8 BSF, with the nuts tightened to 450 lb ft.

High volume manufacturing cuts costs wherever possible. (The corners of the flanges on the bonnet of the original Ford Fiesta were sheared off, for the scrap value of the bits of steel!)

My concern with this problem is to find why the thread run out is being loaded to just beyond the fatigue limit, to produce eventual failure. Once the root cause is known, finding a solution becomes easier.

If the shaft is drilled and tapped, the undercut betwen the two threads of the stud needs to be radiused on both sides, to reduce the possibility of stress raisers. Preferably made from a high tensile bolt, to maximise strength, and tightened to just below yield.

The thread engagement of the repair stud needs to be no more than 2D to provide adequate strength.

Howard

[DC31k]

On DC31k Said:

On Howard Lewis Said:

…try to remove any factor which can increase the cyclic applied loads.

Taken at face value, that made me smile. What could be done? Only drive forwards and never reverse? Never drive down a hill?

—

I am unconvinced with the ‘bolt on’ idea above. Not so many cars have wheel bolts these days. If the threads in the hub fail, you have to replace the hub. With wheel studs, you knock out the stud and install a new one.

A deep clearance hole with threads at the bottom risks removing too much material from the centre of the yoke such that it fails in a different manner (e.g. torsion).

Only drive in a straight line, could be be added to the list of cyclic load reducing suggestions. A better idea would have been for the manufacturer to use a CV joint rather than a Hooke joint which is very unsuitable for a steerable wheel.

Without knowing the details of the hub design and the diameters involved its hard to give definite advice but it would appear that the thread is only needed to keep the splines engaged and maybe clamp the bearing inner races. There will be lots of repetitive angular stress on the splines, yoke and the rest of the transmission (because of misaligned Hooke joints) but I dont see why that would cause a thread to shear.

Ian P

 

 

[Howard LewisParticipant @howardlewis46836]

The failure mode needs to be found  Maybe the terminilogy used is not precise?

It may not be shear, the thread is there to retain the UJ on the hub, so will be in tension.

To shear off the thread off implies a force at right angles to the axis of the shaft.

Since the shaft is rotating, it could be fatigue faure in bending.  But what is imposing a lateral force on the thread?

To fail at the undercut of the thread, suggests that the thread is somehow subjected to a bending load,

(Because of friction between the nut, or backing washer and a hub which is a loose fit on the splines?)

The thread would be unlikely to be subject to a cyclic torsional loading, since it will only experience torsion caused by friction between the threads when the nut is tightened, to apply a tensile load.

Maybe the hub subjecting the thread via the washer to a cyclic torsional load, again because of a loose fit between nthe hub and the splines.

None of us yet have enough detail to be able to accurately diagnose the reason for the failures, so can only speculate.

The fact that a second has ocurred suggests that a very detailed examination of the parts needs to be made to determine if it is caused by a design, or a manufacturing fault, or even excess loading in service.

King pin failures were reported on certain Massey Ferguson tractors; until it was found that some dealerships lacked ramps, and tractors were being driven off the side of the trailers!

Howard

 

[DiogenesParticipant @diogenes]

Howard is right;

..under what circumstances have the failures happened..?

[Martin CargillParticipant @martincargill50290]

One thought. Series land Rovers have a similar(ish) arrangement on their front half shafts. The tightening torque for the half shaft nut is only 10-15 lb/ft. Your nuts are not being over tightened?

 

Martin

[colin hamiltonParticipant @colinhamilton16803]

The nut holds the hub onto the shaft. The tractor is used on very steep land (over 30 degrees). The dealer thinks that the loads on the lower wheel when turning at very steep angles with something heavy like a topper fitted could be causing very high loads. Especially if the wheel lock is set allowing the wheels to turn too far. Unfortunately I had to send the broken part off so they could identify it properly so will not be able to get pictures of the failure. The tractor has no suspension and my land is very rough so I do spend most the time bouncing around so everything takes a bit if a pounding.

[Michael GilliganParticipant @michaelgilligan61133]

It looks like an impressive little machine, which is probably being used at the limits of its ‘design envelope’ … I fear you may just have to accept that the shaft is a consumable !

MichaelG.

.

Edit: __ https://ctmltd.co.uk/

 

[Author]

Posts