gearbox output flange weldable?

[safe]

I'm installing a 78 915 box in my 69 tub.

The original driveshafts and CVs from my 69 are good, strong and usable. The 79 CVs less so.
I would like to use the older 108mm CVs, but finding a garbox output shaft with 108mm flange that fit my gearbox is nearly impossible or maybe not possible at all....

Could I have my machinist make a 108 mm flange and weld that to my 100 mm flange?
I don't think it would matter if I add a 10 mm thick flange, since the original 901 gearbox was wider than the 915 box

But, are these cast parts reliably weldable?


Left -79 100mm, right a -75 108 mm

[haasad]

Definitley, I converted a set of coarse spline and fine spline output flanges with careful machining and high quality welding. They have handled 1200 miles on my
200 wheel HP motor no issues.

Andy

[safe]

Originally Posted by haasad

Definitley, I converted a set of coarse spline and fine spline output flanges with careful machining and high quality welding. They have handled 1200 miles on my
200 wheel HP motor no issues.

Andy

I thought I seen a thread about it but forgot where (on DDK maybe?) was that your?

You don't have any pictures on how you did it?

[haasad]

Hi yes it was on my r car thread ddk. I will try to dig out any puictures but they are on line.

Andy

[safe]

Found it!
DDK - Die Deutschen Klassiker • View topic - 915 gear box drive flanges

I was thinking of doing it like that, but thought the weld might be to weak. Therefore my suggestion above adding a new flange to existing and thereby having a longer weld.

Maybe you are right and it will hold just fine!
I'm going to use a ~300 hp 3.6 engine.

[chris_seven]

It is always a concern to weld parts of this type without considering the detailed metallurgy and the position of the weld.

The failure of this type of component is concerned with probability.

The first question would be what is the probability of finding a defect in the weld and the second is what is the probability that this defect will propagate.

In this case the weld line corresponds to the position of maximum stress and some basic stress analysis wouldn't be a bad thing.

In the '20 Golden Rules' for the avoidance of fatigue I would try to avoid welds placed at the point of maximum stress.

Some idea of the material's composition would also help as then you could decide on the likely condition of the Heat Affected Zone and the resulting microstructure.

I guess the last question is the risk assessment in terms of failure.

If the flange breaks will the drive shaft damage the rear seat pan and any (unlikely) passenger.

It is unlikely to be 'weak' in the traditional sense but it may well be brittle.

I would worry that the probability of finding crack like defects in this type of weld could be quite high and if the material has a carbon equivalent of 0.4 or higher then its HAZ structure will probably allow any defect to propagate.

If the steel used is low in carbon and alloy content then it may not be an issue but I would like to know.

The last thing to worry about the impact of any failure. If the flange breaks will the drive shaft damage the floor pan and how badly.

[safe]

Originally Posted by chris_seven

It is always a concern to weld parts of this type without considering the detailed metallurgy and the position of the weld.

The failure of this type of component is concerned with probability.

The first question would be what is the probability of finding a defect in the weld and the second is what is the probability that this defect will propagate.

In this case the weld line corresponds to the position of maximum stress and some basic stress analysis wouldn't be a bad thing.

In the '20 Golden Rules' for the avoidance of fatigue I would try to avoid welds placed at the point of maximum stress.

Some idea of the material's composition would also help as then you could decide on the likely condition of the Heat Affected Zone and the resulting microstructure.

I guess the last question is the risk assessment in terms of failure.

If the flange breaks will the drive shaft damage the rear seat pan and any (unlikely) passenger.

It is unlikely to be 'weak' in the traditional sense but it may well be brittle.

I would worry that the probability of finding crack like defects in this type of weld could be quite high and if the material has a carbon equivalent of 0.4 or higher then its HAZ structure will probably allow any defect to propagate.

If the steel used is low in carbon and alloy content then it may not be an issue but I would like to know.

The last thing to worry about the impact of any failure. If the flange breaks will the drive shaft damage the floor pan and how badly.

Thanks for your insight.
I think I'll try weld it at a different location, but on the other hand it has worked well for Andy.

I can probably get the welds x-rayed, the neighbors to my garage is in that business.

[KTL]

What about having your machinist make an adapter from 100 to 108? But that would involve custom axles.....

I know you can also use a G50 axle flange. Rothsport modifies them in a way to fit the 915. Here's a pair that William Knight was selling

108mm Output flanges for 915 Fine spline


For others considering this in the reverse fashion (need a 108mm coarse spline flange to fit larger modern 108mm CVs) Albins makes a coarse spline axle flange. But I do not know what diameter it is. I would hope it is 108mm simply because I would guess they'd design it to fit the 930 or 934 CVs they also make?

You could also swap out your diff in the 915 for a coarse spline diff.

See page 5 for product "Porsche Components 915 Output Flange Coarse Spline LSD" http://www.albinsgear.com.au/productcatalogue.pdf

[Matt Monson]

I manufacture 108mm fine spline stubs. But if you have to ask how much they cost, they are probably more than you are willing to spend.

[KTL]

Originally Posted by Matt Monson

I manufacture 108mm fine spline stubs. But if you have to ask how much they cost, they are probably more than you are willing to spend.

Doh, didn't know that. Good to know. Hows abouts coarse splined 108s to accept the newer style "capped" CV joints?

I've currently got an old set of 108 coarse splines on the bench (just like the ones pictured above) I was going to fiddle with and see what kind of work is needed to make them work for me with the newer G50 style axles. But I don't like how the old flanges are not as robust/thick as the newer flanges.

[chris_seven]

I think that it may be difficult to weld at a different location and although it appears to have worked well so far we need to bear in mind that metal fatigue is an insidious process.

Damage accumulates over relatively long periods of time - and can be assisted by corrosion - before defects propagate in an unstable manner and a brittle fracture event occurs - normally with no prior warning.

Fatigue in one form or another is probably responsible for at least 90% of all engineering failures so needs to be considered.

1200 miles is not really much of a test and the SAE usually suggest this type of component should achieve a Weibull B10 life at a 50% confidence level and a significant amount of test data would be needed to 'prove' this method of manufacture.

Making one-off parts is risky and doesn't take any statistical or stochastic models into account and it is possible that the third shaft made using exactly the same methods could fail in a few minutes.


X rays are good and will find macroscopic defects but there may be some small defects that can propagate in brittle materials that could be difficult to find using this type of technique.

welding defects

Can your colleagues next door analyse the material as this may help to define the depth of the problem?

I do accept that I may be overcautious but the problem with using TIG welding for this type of components is the variability.

I would almost certainly buy new parts.

The $320 on the Albins Price List doesn't seem a bad price but I would make some locally as they are relatively straightforward.

[safe]

Originally Posted by chris_seven

I think that it may be difficult to weld at a different location and although it appears to have worked well so far we need to bear in mind that metal fatigue is an insidious process.

Damage accumulates over relatively long periods of time - and can be assisted by corrosion - before defects propagate in an unstable manner and a brittle fracture event occurs - normally with no prior warning.

Fatigue in one form or another is probably responsible for at least 90% of all engineering failures so needs to be considered.

1200 miles is not really much of a test and the SAE usually suggest this type of component should achieve a Weibull B10 life at a 50% confidence level and a significant amount of test data would be needed to 'prove' this method of manufacture.

Making one-off parts is risky and doesn't take any statistical or stochastic models into account and it is possible that the third shaft made using exactly the same methods could fail in a few minutes.


X rays are good and will find macroscopic defects but there may be some small defects that can propagate in brittle materials that could be difficult to find using this type of technique.

welding defects

Can your colleagues next door analyse the material as this may help to define the depth of the problem?

I do accept that I may be overcautious but the problem with using TIG welding for this type of components is the variability.

I would almost certainly buy new parts.

The $320 on the Albins Price List doesn't but I would make some locally as they are relatively straightforward.

I was thinking of making a "ring" drill and tap, then weld that to the existing flange. Then the weld would be in a less stressed location and I would get at least double the amount of weld.

Not sure if the guys can analyse the material.

I'm aware that it's sort of a gamble...

[ClickClickBoom]

Not to be a fly in the ointment, I doubt the those axles are cast. As a seat of the pants non-engineer, that application would indicate something forged and machined to spec. The real problem is with the fatigue and stress analysis, or lack there of. Welding and the resultant HAZ will introduce a whole 'nuther set of variables that most aren't able to answer. The question is why would Matt make a solution to the problem if simply welding and machining would suffice?
And the $96,000 question, if the axle breaks at a most inopportune moment, say in a 3rd gear, 6000 RPM sweeper what will happen, other than the the valves meeting the pistons in a most inappropriate fashion. Not to mention what will happen to the car.
The next big question is how much will any of the above cost to fix if the failure occurs, bet Matt's axles seem cheap now....

[lrodri64]

Just a suggestion....
Why not change also the differential spider gears?

[safe]

Originally Posted by lrodri64

Just a suggestion....
Why not change also the differential spider gears?

It's an option, maybe. I would probably need to change the whole diff.

Could anyone confirm that the right stub axle in the picture would fit an earlier 915 course spline diff?
As you see its not only the spline that is different....

[Shaun @ Tru6]

Originally Posted by ClickClickBoom

And the $96,000 question, if the axle breaks at a most inopportune moment, say in a 3rd gear, 6000 RPM sweeper what will happen, other than the the valves meeting the pistons in a most inappropriate fashion.

Can you explain why pistons will hit valves if an axle breaks.

[Fishcop]

Magnus, having done a similar conversion to a 915 in a 69 chassis; I have to ask what you'll be running that a set of SC axles/CVs can't cope with? For me it was infinitely easier to change the stub-axles in the trailing arms than changing gearbox internals or 'customising' drive shafts... If it's a cost thing, it may be false economy. Sell/swap the early driveshafts for good SC versions.

[safe]

Originally Posted by Fishcop

Magnus, having done a similar conversion to a 915 in a 69 chassis; I have to ask what you'll be running that a set of SC axles/CVs can't cope with? For me it was infinitely easier to change the stub-axles in the trailing arms than changing gearbox internals or 'customising' drive shafts... If it's a cost thing, it may be false economy. Sell/swap the early driveshafts for good SC versions.

Nothing special, a mostly stock 3.6.
But I have a nice set of drive shafts from the 69, that's among the strongest CVs Porsche did (according to Grady). The 100mm CVs are a bit on the weak side.

New SC drivshafts is ~$240 each, but the outer stub axles is $530. That's way more than I'd like to spend on this...

Haven't found any good used complete axles yet. I would consider that.

[ClickClickBoom]

Originally Posted by Shaun 84 Targa

Can you explain why pistons will hit valves if an axle breaks.

The flange, if it were to fail would most likely be under load, 6000RPM means a foot full of gas, any sudden decrease in loading while under that amount of power would cause the engine to accelerate instantly, into an over rev situation, enough over rev will cause valve float.

[Fishcop]

Originally Posted by safe

Nothing special, a mostly stock 3.6.
But I have a nice set of drive shafts from the 69, that's among the strongest CVs Porsche did (according to Grady). The 100mm CVs are a bit on the weak side.

New SC drivshafts is ~$240 each, but the outer stub axles is $530. That's way more than I'd like to spend on this...

Haven't found any good used complete axles yet. I would consider that.

Fair enough.

I located the stub-axles second hand for $100 each and cleaned up a second hand pair of shafts with new grease and boots (and a bit of paint) for $150 each. But I'm only running a 2.7 engine.

Good luck with the conversion.