I hope that this type of connecting rod will be for sale one day for our toys, we do not stop the progress!

If someone would post the mass of a complete piston assembly (rings, wrist pin and circlips included) plus the mass of the connecting rod, I'll run the calculation for the max tensile and compressive force applied to the connecting rod at a given rpm and bmep. Then we will quantitatively know what alternating force we're dealing with.

I'll also need the mass of a complete connecting rod. Knowing this, I'll apply the 2/3 - 1/3 mass distribution convention...that is 2/3rd of the mass is at the big end, 1/3 mass is at the top end. The top end mass is lumped with the piston assembly to get the total reciprocating mass.

Another request - would be nice to know the X-section area of the production con rod at the C-to-C distance 2/3 up from the big end. This seems to be the approximate location where most connecting rods fail from fatigue.

Knowing the X-section area and alternating force applied, we now know the alternating stress applied to the con rod and can check this against the material properties to see if we are encroaching on the fatigue limit.

Let's use the 3.2 L engine as the specimen. I've wanted to do this and now have a good reason to do so, but didn't want to disassemble my engine just for the measurements. As an engineering consultant in the engine business, I have the software to make quick work of this problem and with the help of Chris on the metalurgical end, I think a real answer can be had. Should be fun!

Dave,

To model it well you may need to estimate the twist that occurs and factor in this value. Also peak cylinder pressures may help.

I would guess a peak accel in the order of 5000g as maybe somewhere to start.

By developing a loading spectrum it may be possible to do some rainflow analysis and come up with a life estimate.

We used to run nCode Software on our Fatigue Testing machines but this was during the time when it was being developed at British Rail in the late Eighties and early Nineties.

Compressive loads are not as bad as tensile loads (for example on high rpm deceleration with engine braking) where mass of rod/piston assembly counts the most.

Hence could do the calculation even without knowing the peak cylinder pressures.

I am not sure what you mean by this statement but it needs some examination.

If you are considering crack growth then the effect of crack tip closure on da/dN needs to be considered and by testing in a tension/tension test could result in significant crack growth rate errors.

Crack initiation is far more complex and is generally driven by resolved shear stresses, which when the become critical cause the intrusion and extrusion of slip bands that lead to the development of a crack. Twist and Euler Buckling contribute to this behaviour and are both caused by compressive loading.

If we consider a crystal that has a favourable orientation then tensile loading causes a positive resolved shear stress and when this reaches a critical value slip will occur.

The action of this slip will be to produce a residual stress and due to the continuity of the stress field within the crystal the residual stress relieves the positive shear stress not only locally but across the crystal. This means that the positive shear stress is very localised during tensile loading.

Effectively reducing relieving a positive shear stress has a similar effect to increasing the negative shear stress and during compressive loading the greatest negative shear stress will also cause slip.

This slip then relieves the negative shear stress which effectively strengthens the positive shear stress and makes slip in the tensile direction much more likely.

As this process repeats the magnitude of positive and negative slip increase monotonically with increasing cycles of loading and produce a slip band extrusion.

This is the beginning of crack initiation.

I do believe that the academic world has long accepted that fully reversed loading is more damaging that a tension-tension case particularly in terms of initiation.

I would agree that looking at models with and without compressive loading would be interesting but I would think compressive loading will affect fatigue life.

Developing a realistic loading spectrum would be a good first step

http://rs1234.wz.cz/img_m/ART_166.pdf

Hi,
I miss only the weight of segments.
The set was weighed with a precise balance of laboratory in 0.01g.

Piston : 443.95 g 98mm Mahle
Rod: 508.69 g R/R of LN engineering. Origin rod : 660 g
Axe: 127.94 g
Clips: 0.91 g/u (2)

Total: 1082.4 g

;)

I think Raceboy is saying, no one here makes enough power to bend a rod.

RPM? Sure, someone can mechanically overrev.

I "money shifted" mine hitting somewhere in the 8k range with stock rodbolts, and no issue with throwing the rod(s) out though....

Con Rods do bend - I think there can be no doubt about this - the forces generated may not cause plastic deformation but this debate really centres of elastic behaviour and my point is that tension, compression and bending all need to be modelled if we are to obtain good results and be able to come to good conclusions.

The entire debate about the 'life' of a Titanium Rod is to due to the accumulation of damage at a microscopic level.

If you ignore reversed loading during compression you will almost certainly overestimate fatigue life as you will lose around 50% of the stress amplitude, bending is part of this issue as is buckling and twisting which depends on the design and tends to by why H Beams are favoured in Turbo engines.


The stresses in rods generally has a basic form which will change depending on several variables but the basic shape of the distribution doesn't change.

http://i197.photobucket.com/albums/a...psnwcndefs.jpg

I have been interested in the design of con rods for some time and we have manufactured our own con rods for 2.0 litre race engines for the last 3 years and would like to make a wider range of products within the next 12 months.

If normal combustion is bending rods, isn't detonation a lot harder on them?

Love to hear 8K, any video clips?SmileWavy

I would think so but bending stresses should still be within the elastic limit - the closer to the elastic limit the more likely it is that fatigue damage will accumulate and the shorter the life.

The models being used by the leading research companies could add detonation to their load spectrum and estimate the impact of this problem.

It is still fair to say that fatigue failure of 911 rods is very rare so the design is conservative but the Ti rods that started this debate are more of a concern.

If the Ti is manufactured under controlled conditions and I would assume that material purchased to AMS specifications should be consistent then it is likely that the quoted 'life' is conservative.

If the rod is then not used to the limits that were used during the design stage it may be capable of much greater than the quoted life.

Without knowing all the details it is just difficult to judge.

It was a "money shift", mechanical overrev.

Redlined 3rd, then decided I'd like to put it in 2nd instead of 4th.

Damn G50's shift so well! ;)

Starting to wonder if motors that have thrown a rod (any motor, not just 911's) in the past possibly were from detonation?

Of course you have your standard oiling issues and overrevs.

Personally, I've never seen an actual rod failure from detonation in any engine (going back to 1962 when I started). :)

Usually, pistons/ring fail before stresses build on rods prior to outright failure.

JMHO, but the majority of rod failures stem from lubrication/bearing issues which trigger seizure and I've seen my fair share.

so you're saying to fix any stretching from over-rev, just turn up the boost?

http://forums.pelicanparts.com/suppo...eys/icon26.gif

That's what I've always believed too...

Pat - Thank you for the component mass information. I’ll start assembling the input data based on these numbers. I noticed that the piston mass is for a 98mm bore – the 3.2 L engine is a 95mm but what you provided will be fine for my purposes. I’ll use the original rod mass of 660g and 1/3 of this (220g) is considered the reciprocating mass contribution of the rod. Added to the piston assembly, this totals 793g for the recip mass. I have all the rest of the geometry of the 3.2 engine at hand, so just now a matter of preparing the input file for the analysis.

For cylinder gas pressures, I use the Vibe heat release model which has proven accurate to within a few percent of not only predicting the peak cylinder pressures encountered during normal combustion, but the pressure at every crankangle increment during the power stroke due to a realistic heat release calculation. Inputs are delivery ratio (volumetric efficiency) compression ratio, A/F ratio, engine speed, and target BMEP. Cylinder gas pressure force applied to the piston dome area is superimposed over inertia loading, and this comprises the basis for the dynamic model. Most of this input info I’ve collected on the 911 engine over the years or know where to go to get it.

Chris – When you say twist that occurs, do you mean a torsional deflection of the con rod? I realize this can and does occur in actual running engines but my code does not take this into account. Also, the radial outward force in the plane of the con rod that causes bending is not accounted for as this is usually small in magnitude as compared to the recip forces, and is ignored. However, I can factor this in as a separate calculation and it may become significant with engine speeds of 8000 rpm.

I’ll work on this as a “hobby project” over the next few days and will present results ASAP. Can’t drive or work on my 911 at the present time (winter in Wisconsin!) so I’m deskbound.

Still working on it - should have something to post in 24 hours.

Dave,

This may be interesting

http://www.eng.utoledo.edu/mime/faculty_staff/faculty/afatemi/papers/2006JMESShenoyFatemiVol220PartCpp615-624.pdf

Chris - thank you for the technical paper link - I will peruse this over the weekend - looks excellent. The analysis that follows does not consider the bending forces in the plane of motion of the connecting rod. Will have to provide this as a separate calculation.

This thread sparked an interest in learning more about the inner-workings of the Porsche air-cooled engine – namely, what kind of loads are the connecting rods subjected to in a modified engine. Looking at a worst-case scenario would allow one to put some numerical values on the maximum tensile and compressive loads the con rod feels. Undoubtedly the Porsche engineering department went through a detailed analysis during the design phase of the connecting rod and to my knowledge, the factory rods prove very durable and failure-free, even in high-revving, competition engines. Nevertheless, it’s just interesting to be able to assign numbers to what is generally accepted to be the most highly stressed mechanical component in the engine.

I selected the 3.2 liter engine for this study. Here are the following specifications used for this analysis:

Bore – 95.0 mm
Stroke – 74.4 mm
Rod Length – 127.0 mm
Comp. Ratio – 9.5:1
Ign. Advance – 26 deg btdc
Mass Piston Assy – 572 g
Mass Con Rod Small End – 220 g
Mass Con Rod Big End – 440 g

The numerical model consists of a dynamic simulation of the slider-crank mechanism (crankshaft-conrod-piston) utilized in virtually all piston engines. Knowing the mass of the various components and the crankshaft rotational speed, it is straightforward to predict the inertia-induced forces in any of the components of interest. Superimposed on the inertia force is the cylinder gas pressure loading that is applied to the piston dome. Having accurate gas pressure data during the compression and power strokes, is paramount in obtaining a realistic dynamic loading of the components during the 720 degree rotational interval of the 4-stroke engine cycle. Such a simulation is standard practice for the design of a new engine, yielding important information on the dynamic loads of both the wristpin and crankpin bearings.

Not having access to any actual cylinder pressure data taken on the 3.2 engine under laboratory testing, the techniques given in Ref. 1 (Vibe combustion model) are employed to generate a theoretical gas pressure file. Inputs, in addition to engine geometry, are 1) compression ratio, 2) delivery ratio, 3) ignition timing, 4) air/fuel ratio, 5) ignition delay, 6) burn rate in crankangle degrees, 7) engine rpm and 8) BMEP. 1, 3 and 4 are known from Refs. 2 and 3. 2, 5 and 6 are based on experience and collected data on other air-cooled, 2-valve, naturally aspirated engines. Engine rpm and BMEP are based on the target performance we arbitrarily select for this study. Based on an earlier study posted several years ago (in which I probably confused more than informed, sorry), I looked at realistic BMEP outputs of the 911 engine in various displacements and states of tune.

http://forums.pelicanparts.com/911-engine-rebuilding-forum/645505-accurate-approximation-porsche-911-engine-power-output.html

For peak power I chose 188 psi BMEP @ 6100 rpm where the 3.2 L would be producing 280 bhp. For the peak engine speed desired of 8000 rpm, we’re way past peak power, so an educated guess has to be made regarding output here – I think 200 bhp is realistic with a corresponding BMEP of 102 psi.

Running the combustion simulation for the peak power point results in the following PV diagram with an air standard cycle shown as reference to a “theoretical maximum”:

http://forums.pelicanparts.com/uploa...1489783084.jpg

The sharp discontinuity at max volume is due to the “instantaneous” blow-down of the cylinder pressure, as the exact opening timing of the exhaust valve and associated pressure history is not considered nor is necessary for this analysis. Highlights here are maximum cylinder pressure developed – 923 psi.

With cylinder pressure now approximated, this can be applied to the kinematic model to obtain the force applied to the connecting rod beam, this being a function of the inertia force of the recip assembly with gas pressure force superimposed, for a full 720 degrees of rotation (1 full engine cycle).

http://forums.pelicanparts.com/uploa...1489783172.jpg

0 degrees of crank angle corresponds to tdc at the beginning of the inlet stroke – here the maximum tensile load is encountered as there’s no appreciable gas pressure on the piston dome to offset the inertia load of the recip mass. 180 degrees is bdc of the inlet stroke – moderate compressive loading due to inertia. 360 degrees is tdc on the power stroke – notice that prior to tdc the tensile force rapidly changes to compressive loading due to combustion initiating and providing the dominating force around 20 degrees atdc. 540 degrees is the exhaust open point where the cylinder “instantaneously” blows down to atmospheric pressure with very little change in loading. 720 degrees is the end of the exhaust stroke, with tensile load again at a maximum.

Here is the important take-away on the con rod forces at the max power point:

Force Max Compressive – 6790 lbf
Force Max Tensile – 3492 lbf
Load Cyclic Frequency – 50.8 Hz (@ 6100 rpm)

I don’t know the rod X-section area so can’t calculate the stress. If someone can provide an accurate approximation of the beam area at the 2/3 center-to-center distance (up from the big end) then we know the cyclic stress.

More to follow, including references.