[Weissach911]

Henry,

There are of course several interesting issues that come out of this history and it is worth looking at some of these issues on a point by point basis.

It would be standard practice for any designer to consider the influence of expansion on clamping forces and the effect that this would have on both studs, pull out loads and the potential of yielding to occur in compression.

If I were designing a joint of this type I wouldn't look for an increase in clamping due to expansion as I would want enough preload to eliminate the fatigue loadings produced by the peak cylinder pressures but I would want to be sure that I didn't cause any failures due to increases in the forces.

The sand cast cases were likell to have been a 'Eutectic' alloy which would be about 12% Silicon and this would have expansions in the order of 21ppm/degC.

The load bearing capability of the Elektron casing must also have been considered to be adequate for these loads.

I am not sure about the material used for the Nikasil barrels but if it were a typical wrought alloy its expansion could be as much as 24ppm/degC.

The 17% Hypereutectic Alloy developed by KS can be used without liners by etching away the surface aluminium and leaving a small surface of Silicon.

These alloys expand about 15% less than wrought alloys - about 20ppm/degC.

There is a potential issue with the cast aluminium/silicon engine cases and that there is the possibility of silicon migration within the solid phase at temperatures around 200 degC.

This means that silicon particles within the component can grow which depletes the silicon in neighbouring regions and causes a local weakening of the matrix.

(This can be one of the causes of failure of Eutectic and Hypereutectic Pistons and why 2618 is preffered for high end Turbo engines.)

I am sure there must be some detailed information about the behaviour of these alloys but it has not been extensively published.

The die cast aluminium cases would not be too different from the sand cast cases in terms of material expect that the Silicon Morphology is likely to be more even in terms of both size and distribution.

There is, of course some porosity issues with die castings compared to sand castings.

Gravity die castings and even low pressure die castings tend to have low levels of porosity but they can exhibit porosity evenly throughout a structure (many die cast wheels leak air and now the trend is to powder coat to seal them.)

The composition of the case is interesting and in general die castings would be made from a Eutectic Alloy.

This is a Silicon content of about 12.6%. This composition has two adavatages.

The eutectic composition gives the lowest melting point of any Al/Si alloy and hence heating cost is minimised. The second is that Eutectic alloys don't have a freezing range. They remain liquid to a specific temperature and the solidfy evenly and uniformly with a change of only a couple of degrees.

This means that components can be reaily knocked out a die as soon as they ahve cooled, you don't ahve to wait the hundred or so degrees (freezing range) of alloys such as 2618.

These two charcteristics significantly reduce casting costs.

I am sure the stud pulling issue was a feature of magnesium combined witj Nikasil and thermal reactors increasing engine temp.

There is no doubt that Dilavar suffers significant sensitivity to Stress Corrosion Cracking in the presence of Chlorides (salt on the road is a great way to introduce chlorides to the studs).

Stress Corrosion Cracking is a very specific form of corrosion. the chlorides attack the grain boundaries of the material and effectively form a very sharp crack.

When the crack has elongated sufficiently to reduce the remaining area of the stud it can no longer support the load and it will suffer from a brittle failure.

There is a whole science (Fracture Mechanics) devoted to this type of failure but the crack length, crack tip radius and the stress intensity all contribute.

I have to say that apart from manufacturing defects there is no failure mechanism that would result in the brittle fracture of a new stud which had been loaded a left in an engine stand.

It would be good to look at the fracture surface of such a failure - probably with a Scanning Electron Microscope as there is bound to evidence of a defect.

Porviding the clamping loads have been correctly established there is no real reason why Dilavar should give problems due and why heads should lift.

I am suspicious about casing stability and stress relaxation where studs would pull in magnesium cases. Increase in Silicon content is likely to make this worse.

I would agree that the 'all thread' stud looks terrible but the tip root radius of the thread doesn't really cause any significant stress riser in this application.

i different geometries tip root radii of about 0.2mm would be an issue but not in studs or bolts in the type of materials we are disussing.

In fact the probable reason for the all thread is to ensure that any deformation is evenly distributed along the length of the stud - something which the thread will help to achieve.

If you take a bar with a single notch and put into a tensile test machine it will clearly fail at the notch. You would also need to measure the elongation at failure.

If you take a length of the same bar (cut adjacent to the first test length) and machine a series of multiple notches the failure load will be identical - within experimental scatter - but the elongation will be significanltly higher.

This does, of course, also need a material to have some ductility so ceramic bars wouldn't behave this way.

The all thread also have a slightly larger minor diameter than the plian stud so for a given prload will be slightly less stressed.

I am not sure why you would change the way you the nut on a stud from simple torque to a torque + angle and this needs some thinking about.

It may be to try to have a more even preload .

Torque to axial force using torque wrenches tend to give axial force variation of +/- 25% and this could be part of the issue with increased capacity and hence combustion loads.