It is very confusing and yet the general thrust of what has changed is relativley straight forward - perhaps the following will help.
Basically all models run hotter than many older model designs to assist with the reduced emission standards. The 2.5, 2.7 and 3.2 Boxster S had relativley thick cylinder walls but the 996 3.4 was made by boring out the 3.2 Boxster S block bigger in bore diameter (without increasing the outside diameter) - so that model had thinner cylinder walls.
The 3.4 Cayman S, 3.6 996 and 3.6 and 3.8 997's retained about the same cylinder wall thickness as the 996 3.4 (thinner than the first engines again but similar to each other).
The 3.4 Cayman S, 3.6 and 3.8 engines also had a different head gasket that allowed more coolant to flow into the cylinder heads (and therefore less into the cylinder blocks) - so they run slightly hotter cylinder wall temperatures.
The IMS bearing was made bigger for later models (and is much more reliable) but there is a confusing mixture of new bigger engines fitted with old smaller IMS bearings - so there is a mixture that is a little unpredictable.
The general flow of changes is that as the engines got bigger and more powerful - while they gradually fitted the larger IMS bearing (a benefit) the cylinder wall temperatures at maximum output would be higher and some scoring started.
If we forget the IMS issue for now - it is clear that as the engines had bigger pistons or stroke (more capacity) and worked harder (if they were driven to their limit) the cylinder wall temperatures were at the same time moving in the direction of going slightly higher.
When a piston scores it is usually because the friction has increased and with it the temperature of the piston surface - suddenly reaching the limit - and within a few seconds melting at the surface and setting again - the lumps created scoring the bore.
The oil between the piston and the cylinder wall is thinner if the temperature is higher and so as the friction increases - the ability of the oil to protect the running surfaces diminishes and as it does so the oil gets even hotter and therefore even thinner until damage occurs.
The side of the piston that gets hottest is the side running with the most friction and this is the "thrust side" where the piston pushed against the cylinder wall to drive the car along.
On bank 1 this is the bottom of the engine and on bank 2 it is the top.
The coolant enters at the bottom on both sides and the oil spray jets (that lubricate the cylinders and pistons) are directed at the lower part of the pistons - so in both cases - bank 2 has both less oil on the thrust face and runs at a higher temperature - and this is why that side has all the problems.
It is obviously very marginal because only a small % fail and anyway driving modestly will run the enginess further away from the critical temperatures.
Because the cylinder wall thicknesses of all the later engines are similar to the 996 3.4 (that cracks and "D" chunks) I would expect them to eventually do the same (and we have repaired some accordingly) but scoring seems to occur before that mileage has been reached (in those that have scored).
Piston loads and lubrication are worst at low revs and high loads (or throttle openings) and also at full throttle maximum revs runs. Because the engine sits with horizontal pistons - oil does not drain away from the cylinder bores when the car is idle. Because the pistons have to be designed to cope with expanding to their maximum at full throttle (and huge horsepower) they must be made with more taper than if they were less powerful and because of these 2 facts - sometimes - on start-up - the combination of that additional taper and sitting horizontally allows the oil that has collected on the cylinder bore to allow a brief puff of oil to burn (but has very little effect on the overall oil consumption).
Cylinders on all models are "open deck" and creep oval with time and this increases the piston clearance allowing increasing blow by to heat up the piston more than when the engine was newer and therefore gradually move the engine more towards running at a cylinder wall temperature that approaches critical limits.
Because the radiators are at the other end of the car to the engine and because the thermostat is fitted at entry to the engine (rather than exit from it) - sudden changes from modest driving to flat out can result in a relatively slow cooling response and temporarlily raise cylinder block temperatures higher than they may settle down to (even resulting in some localised boiling or bubbling).
Therefore (apart from the cloudy IMS issue of which bearing was fitted on 2002 to say 2005 models) you have engines that gradually increase bore ovality, runn slightly hotter cylinder wall temperatures than earlier engines, have the potential to run with higher piston to cylinder wall loads (either at low revs as a result of increased torque or high revs as a result of increased bhp) - so if the owner utilises all that potential torque and power - will run with higher cylinder wall loads and temperatures and may eventually score, crack or "D" chunk a cylinder - in a relatively small number of engines (increasing numbers with age).
Several businesses can now rebuild the engines (fixing the fault often with cheap short lived solutions) or new engines are available (with the same spec as had the failure).
Only one business adresses ALL the problems listed and optionally offers to re-round the oval cylinders (while stabilising them so they will not "D" chunk or crack", to replace a scored, cracked or "D" chunk liner (with a special alloy as strong as steel) - coated in Nikasil - the preferred finish for turbos and GT3's etc), to reduce the cylinder wall temperatures *by increasing cylinder block coolant flow), reduce running temperatures (by offering a lower temperature thermostat) and replaces the smaller IMS bearing with a better lubricated one with stronger spindle. That same business also carries out all that work "IN HOUSE" - in their own precision machine shop - but their contribution to this problem does not stop there.
Presently they are testing oversized pistons in Lokasil bores (to reduce rebuild costs). This has involved high investment in machinery and test cars - but ultimately will be able to provide the most cost effective solution, and manufacturing their own replacment IMS shaft (to suit earlier roller chain systems or later Hivo cahin systems) fitted with the larger improved IMS bearing. They have also successfully tested long term replacement parts for crankshafts and bearings (etc) in case critical parts ever dry up (or become too expensive) - to they can continue to rebuild customer engines to their own exact requirements and incorporate improvements in every area suspected of contributing to the problems some suffer from.
That business is Hartech and I am proud of the extent of the work we do and the options available (to fit in with customers future intentions for their car and their budgets) and the fact that we also provide a scheme to limit the cost of an engine rebuild (if it even happened) while gradually paying monthly for the future servicing and labour for repairs.
There is no single cause (hence the large number that are still perfectly OK) and instead a small number of issues conspire together to contribute to a failure in a variable range of circumstances that is difficult to make exact predictions about of a set of rules to try and avoid.
We continue to try and increase the range of solutions and reduce the potential for problems in every engine we rebuild - still at prices between a third and a half of that generally charged for replacement with a new engine with all the same features as the one that failed.
Further information available on our web site www.hartech.org
buyers guide section 4 and 5.