How to possibly prevent cylinder scoring
- Thread starter dpoynton
- Start date
Very interesting Baz and thanks for your ongoing research. So if coolant speed could be kept consistantly higher would that be a possoble cure? If so then how about fitting an auxillary inline electric pump in the coolant circuit to the front rads.
Although our low temperature thermostat is something that we think could make these engines less likely to score bores - I have also been trying hard to work out what driving conditions might be more likely to lead to scored bores in the first place "" to be able to offer some advice on how best to avoid the consequences. Feedback suggests it is not people driving around in track situations that are particularly any more likely to experience this problem than anyone else "" so it does not seem to be purely down to revs and hard driving entirely.
Furthermore "" for those only ever driving on public roads - for some to fail so soon while others never seem to "" or when some owners never have the problem while others do "" there must be some differences - and if we could identify them (and back them up with technical/scientific support) it could possibly help avoid the consequences altogether.
I think I may have worked out a typical scenario "" best avoided "" here is the evidence and connected technical issues.
Scoring was more commonly caused by full piston seizures in which the piston grew too big to fit in the bore with some clearance and the resulting additional force between the piston and cylinder wall squeezed out the oil film resulting in metal to metal contact "" the increased friction causing local melting of the piston surface and scoring both piston and bore on both sides. Dynamometer tests I have previously carried out to destruction (to measure temperatures etc) showed this cycle took about 4-5 seconds.
The odd thing about this current problem is that one side of the piston and bore looks exactly like a fully scored seized piston but the other side is perfectly OK and untouched. It is therefore clear that when the scoring occurs "" one side of the piston is sufficiently cool to lubricate the piston/bore face and support the piston whereas the other side has become excessively hot and the oil film cannot support the loads applied "" which result in metal to metal contact "" overheating of the surface and scoring the bores and piston on one side only.
It is always the thrust side that scores (proving the additional load between the piston and cylinder wall is the problem) but it cannot be the piston temperature overall "" or the other side would also seize. Although the thermostat controls overall coolant speed (from the radiator efficiency) the coolant flow in individual parts of these blocks is much slower than traditional engine designs (because it is split into 12 separate flows streams instead of one (each therefore at least 1/12 of the speed) "" the 6 through the block being very much slower than the 6 through the cylinder heads and the individual speed past any one cylinder many times slower) so the temperature of the coolant as it slowly passes over the outer cylinder wall increases more giving a greater temperature rise in any one area. It is almost always bank 2 (because on that bank the thrust face is on the hottest side of the block "" as the coolant is about to leave the cylinder having picked up all the heat on its travel through the block - whereas on bank 1 the coolest coolant enters on the thrust side).
For the cylinder wall to get so hot on one side that the engine piston scores must require (IMHO) some bubbling of the coolant on the surface of the outer cylinder wall and this usually only occurs when coolant speeds would be very slow indeed and bubbles would naturally rise to the top (the thrust side on bank 2 "" not on bank 1). But it would also need high piston to cylinder wall forces as well to squeeze out the oil film and make hard contact between the piston and Lokasil bore.
Centrifugal coolant pumps have relatively low pumping speeds until they reach a critical speed after which they suddenly increase flow dramatically.
Radiators cool more when the road speed is higher (when their efficiency increases).
The engines in question (3.4 Cayman S, 996 and 997 3.6 and 3.8) produce more torque at low revs than their predecessors.
There is a piston cooling affect from inlet charge when the throttle is open momentarily lost as the throttle is closed.
Spray jets in the block lubricate the cylinder wall but have ball valves that require some oil pressure to open and the oil pressure and delivery is low at low revs "" rising until the pressure relief valve opens when it is constant at much higher pressure and delivery providing much better cylinder wall lubrication.
Crankshaft speed also distributes more cylinder oil spray at higher speeds.
The relatively long distance from the engine to the radiators reduces cooling reaction time.
The thermostat is on the way back to the engine so if the engine is ticking over after a run the cooler coolant that has just come from the radiator will slow down as it reaches the thermostat and for a while "" will not encourage it to open more "" in fact it would probably start to shut a little.
Putting all this together "" if we had a scenario in which the engine was under quite high load and road speed but then it suddenly went to tickover speed for a short while "" at this point the piston would be very hot (no inlet charge), the coolant speed would be very low (pump slow), the heat passing from the cylinder outer wall to the coolant would still be high (probably leading to bubbling which is very poor for heat dissipation), the oil spray cylinder wall jets output and crankshaft rotational speed would be very low (so reduced cylinder wall lubrication) and while the coolant in the block would be very hot the coolant near the thermostat would be quite cool.
If just then the car was driven off under high torque/full or nearly full throttle (not necessarily at high revs) the pistons would instantaneously heat up, the dynamic forces between the piston and the cylinder wall would be at their highest, the cylinder wall and oil film temperature would be very high, the thermostat would not be fully open, the car would not yet be at a high enough road speed for the radiator to cool the coolant effectively and the oil spray jet and crankshaft speed splash lubrication would not yet be good.
So "" put simply "" I think that if the owner of one of these models "" on a spirited drive - could not then resist the temptation to floor the car away from standstill (at say lights or junctions) - then I think we could have the worst possible scenario for piston/cylinder scoring at that precise moment.
This would also explain why "" once the models are up to temperature and having had the speed gradually increased around say a circuit or long unrestricted roads (say in Germany) they seem to be OK despite the extra strain put on the engines (because thermostats are more open, radiator cooling is good, engine speed is high so oil spray etc is good et)c.
I agree such a proposition is almost madness "" I can almost hear you all shouting "what the bl**dy hell do you think I bought it for"Âť - why would anyone buy such a car if they cannot at least feel that acceleration and enjoy that performance (not to mention the satisfaction of showing everyone just how quick the car is when in reality none of us can safely or legally demonstrate anything like the cars top speeds anywhere in the UK "" but we can out accelerate most others). Surely that exact scenario is partly why we bought it and high on the list of the pleasure we get from driving it and possibly the only occasion we can stuff it to other drivers! I am not against that "" in fact I also love it myself (truth be known).
So I am not proposing you all drive around like old men in a Ford Fiesta "" (I am of that age "" but still not a slow driver either)but I am just suggesting that "" it makes sense to me "" that weighing up all the evidence and connecting it to solid technical analysis - this is the most likely moment of failure and therefore "" it could be that if you could simply accelerate from a stationary rest - a little more gently "" under slightly less throttle opening -for a few seconds until the coolant speed, oil pressure and thermostat opening times and the balance of the oil and lubrication system have equalised again "" it may just save the expense and disappointment of an engine failure.
Baz
Furthermore "" for those only ever driving on public roads - for some to fail so soon while others never seem to "" or when some owners never have the problem while others do "" there must be some differences - and if we could identify them (and back them up with technical/scientific support) it could possibly help avoid the consequences altogether.
I think I may have worked out a typical scenario "" best avoided "" here is the evidence and connected technical issues.
Scoring was more commonly caused by full piston seizures in which the piston grew too big to fit in the bore with some clearance and the resulting additional force between the piston and cylinder wall squeezed out the oil film resulting in metal to metal contact "" the increased friction causing local melting of the piston surface and scoring both piston and bore on both sides. Dynamometer tests I have previously carried out to destruction (to measure temperatures etc) showed this cycle took about 4-5 seconds.
The odd thing about this current problem is that one side of the piston and bore looks exactly like a fully scored seized piston but the other side is perfectly OK and untouched. It is therefore clear that when the scoring occurs "" one side of the piston is sufficiently cool to lubricate the piston/bore face and support the piston whereas the other side has become excessively hot and the oil film cannot support the loads applied "" which result in metal to metal contact "" overheating of the surface and scoring the bores and piston on one side only.
It is always the thrust side that scores (proving the additional load between the piston and cylinder wall is the problem) but it cannot be the piston temperature overall "" or the other side would also seize. Although the thermostat controls overall coolant speed (from the radiator efficiency) the coolant flow in individual parts of these blocks is much slower than traditional engine designs (because it is split into 12 separate flows streams instead of one (each therefore at least 1/12 of the speed) "" the 6 through the block being very much slower than the 6 through the cylinder heads and the individual speed past any one cylinder many times slower) so the temperature of the coolant as it slowly passes over the outer cylinder wall increases more giving a greater temperature rise in any one area. It is almost always bank 2 (because on that bank the thrust face is on the hottest side of the block "" as the coolant is about to leave the cylinder having picked up all the heat on its travel through the block - whereas on bank 1 the coolest coolant enters on the thrust side).
For the cylinder wall to get so hot on one side that the engine piston scores must require (IMHO) some bubbling of the coolant on the surface of the outer cylinder wall and this usually only occurs when coolant speeds would be very slow indeed and bubbles would naturally rise to the top (the thrust side on bank 2 "" not on bank 1). But it would also need high piston to cylinder wall forces as well to squeeze out the oil film and make hard contact between the piston and Lokasil bore.
Centrifugal coolant pumps have relatively low pumping speeds until they reach a critical speed after which they suddenly increase flow dramatically.
Radiators cool more when the road speed is higher (when their efficiency increases).
The engines in question (3.4 Cayman S, 996 and 997 3.6 and 3.8) produce more torque at low revs than their predecessors.
There is a piston cooling affect from inlet charge when the throttle is open momentarily lost as the throttle is closed.
Spray jets in the block lubricate the cylinder wall but have ball valves that require some oil pressure to open and the oil pressure and delivery is low at low revs "" rising until the pressure relief valve opens when it is constant at much higher pressure and delivery providing much better cylinder wall lubrication.
Crankshaft speed also distributes more cylinder oil spray at higher speeds.
The relatively long distance from the engine to the radiators reduces cooling reaction time.
The thermostat is on the way back to the engine so if the engine is ticking over after a run the cooler coolant that has just come from the radiator will slow down as it reaches the thermostat and for a while "" will not encourage it to open more "" in fact it would probably start to shut a little.
Putting all this together "" if we had a scenario in which the engine was under quite high load and road speed but then it suddenly went to tickover speed for a short while "" at this point the piston would be very hot (no inlet charge), the coolant speed would be very low (pump slow), the heat passing from the cylinder outer wall to the coolant would still be high (probably leading to bubbling which is very poor for heat dissipation), the oil spray cylinder wall jets output and crankshaft rotational speed would be very low (so reduced cylinder wall lubrication) and while the coolant in the block would be very hot the coolant near the thermostat would be quite cool.
If just then the car was driven off under high torque/full or nearly full throttle (not necessarily at high revs) the pistons would instantaneously heat up, the dynamic forces between the piston and the cylinder wall would be at their highest, the cylinder wall and oil film temperature would be very high, the thermostat would not be fully open, the car would not yet be at a high enough road speed for the radiator to cool the coolant effectively and the oil spray jet and crankshaft speed splash lubrication would not yet be good.
So "" put simply "" I think that if the owner of one of these models "" on a spirited drive - could not then resist the temptation to floor the car away from standstill (at say lights or junctions) - then I think we could have the worst possible scenario for piston/cylinder scoring at that precise moment.
This would also explain why "" once the models are up to temperature and having had the speed gradually increased around say a circuit or long unrestricted roads (say in Germany) they seem to be OK despite the extra strain put on the engines (because thermostats are more open, radiator cooling is good, engine speed is high so oil spray etc is good et)c.
I agree such a proposition is almost madness "" I can almost hear you all shouting "what the bl**dy hell do you think I bought it for"Âť - why would anyone buy such a car if they cannot at least feel that acceleration and enjoy that performance (not to mention the satisfaction of showing everyone just how quick the car is when in reality none of us can safely or legally demonstrate anything like the cars top speeds anywhere in the UK "" but we can out accelerate most others). Surely that exact scenario is partly why we bought it and high on the list of the pleasure we get from driving it and possibly the only occasion we can stuff it to other drivers! I am not against that "" in fact I also love it myself (truth be known).
So I am not proposing you all drive around like old men in a Ford Fiesta "" (I am of that age "" but still not a slow driver either)but I am just suggesting that "" it makes sense to me "" that weighing up all the evidence and connecting it to solid technical analysis - this is the most likely moment of failure and therefore "" it could be that if you could simply accelerate from a stationary rest - a little more gently "" under slightly less throttle opening -for a few seconds until the coolant speed, oil pressure and thermostat opening times and the balance of the oil and lubrication system have equalised again "" it may just save the expense and disappointment of an engine failure.
Baz
Nice article and a reasoned explanation. However all this theory and testing will come to nothing unless the hypothesis and results are fed back to the parent company in Stuttgart. But do they really care? The problem has been with us for more than a decade and still the sales of Porsche's exceed all expectations with fabulous profits to to line the management pockets. I would have thouight that if Porsche wanted to find a solution they could have done it years ago, but until the customer base start to falter then why should they 'waste' money on a solution.
The answer to that wouldn't score many points in scrabble
]
vitesse
Active member
A friend of mine has a relative who works for a research consultancy in the Midlands where considerable work has been put into the causes of the "kettle " engine failures. There is no doubt that failure can be put down to piston interference within bores but their conclusions generally were not putting it down to the thermal performance of the water cooled engine ,rather the thermal performance of the engine blocks, The castings employed "chills" near the top of the bores to freeze the bore metal to an appropriate metallurgical condition during production-research indicated that in failed engines ,these "chills" had moved during production,allowing distortion of that part of the bore under certain engine running conditions leading to piston pick-up. As is known,the IMS bearing problem was down to using a sealed bearing ,possibly slightly too small ,where engine lubrication wasn't able to help out once the inbuilt lubricant dried out.
The castings are far too accurate to be subjeted to core movements and perform identically to each other. Lets see more of that report and the models involved so I too can judge the full quality and implications rather than just suggest that because they are a research company they know more than someone intimately involved in the marque who has analysed the results of hundreds of failures over many years. Actually I thisnk I am better off if I stop trying to help people reduce scoring and just fix the many more than will fail. Baz
Please keep up the good work Baz. You didn't respond to my pump idea, what do you think?
vitesse
Active member
Baz, Don't be so touchy-I was purely contributing to the scene-I didn't suggest that they were more qualified to make their comments than you-just reporting information I was told.Nor did I mention cores moving,only chills-it is a fact that crystal structure can be affected by casting cooling rates ,that castings can creep,that varying metal sections and/or density can alter the metallurgical properties & as far as I have read,the interference problems are at the top of the bore where the casting sections vary considerably & temperatures are highest. My contribution to the "mystery" does nothing to invalidate your solutions & reputation-I am Nick Appleton's father & know full well the great work you do
Sorry Vitesse - the difficulty is that the vast majority of owners are blinded to serious shortcomings in some of these engines by Porsche's past reputation and so I am constantly fighting against the stream when I am offering good solid honest and accurate advice - which does p*ss me off at times. It makes it worse when others with excellent general technical knowledge in a smaller area of specilaisation (like your good self) offer a contradictory comment (but not specific to this actual engine) that further encourages people to ignore advice specific to the problem from those immersed in it day to day. You see this engine only ever scores bores at the lower cylinder, the top never does (largely because the piston is always reduced in diameter above the rings) - and the ovality that creeps in is greater towards the top - increasing clearances (something you should have realised) and the top of the bores is an open deck design and captivates a cast in composite cylinder liner that has no need to have any metalurgical adjustment to fulfill its role in the engine. It is only close to where the cylinder casting is joined metalurgically to the rest of the block (about half way down) that the bore is smaller and the thrust loads from the psiton under acceleration are greater (due to the gas pressure reaching its maximum after TDC and the angle of the connecting rod increasing as the crankshaft rotates). At the top of the stroke the resolution of the forces places no thrust load to the cylinder bore at all and most between about 30 and 110 degrees after TDC by which time the piston is well down the bore. Then reduce the contact points further by the height of the reduced top diemater of the piston and you can see why marking only occurs down the bores. The above is not only pretty basic but anyone who has ever looked at either the evidence or the photos (of which there are many) would realise this. This is why Vitesse I can react against those that I know mean well, and do know their stuff in a smaller or less relevant field and frequently try to undermine my posts by writing about their own theory - when it is clear to me that they are not close enough to the problem I am writing about to have thought it all out. I am not actually bothered on a personal basis (I have achieved enough in my life to not need reassurance about my capabilities and am retiring soon) but because it undermines the confidence of those I am simply trying to help and therefore has a negative affext when actually it is largely irrelevant to the precise problem I have thought about, experienced and am reporting
Touched the wrong key before I have finished. I think it is fair to say that I am the only contributor to these forums who actually has the qualifications, experience and hands on knowledge of working with more of these engines than anyone else Worldwide - that bothers to contribute anything - and therefore am probably the only source that is offering quality technical advice - so I do at times expect a little more respect from others that so quickly contradict my findings when they are usually wrong and their posts just provide coverage for their skills but nothing usefull to the issue being addressed - sorry. It takes me hours writing this stuff basically to help people caught up in a problem they should never have had to experience and then I have to spend more time countering contradictory posts that offer nothing useful to the issue - by defending my original contributions (when they usually turn out to be right) and the whole experience simply confuses most readers and results in them taking less notice of my original advice that would - absolutely would - help them reduce the chances of suffering the missery and expense of an engine failure. I know this sounds like I am claiming to be always right and therefore arrogant - but look - I actually fix these things and have a history of fixing manufacturers problems going back 45 years - and am just amazed how quickly others cannot wait to dissagree when they have nothing like the same pracrtical experience, proven capability nor close knowledge of this particular engine and its problems. I actually hate having to respond like this and would prefer to find ways to modestly get my point accross. I try to write initially in a more modest matter of fact style - but then it just seems to encourage others to jump in all the time - so sometimes I get frustrated at the rather unjust situation and - you are right - I get touchy and tired of it all - for which I am truly sorry!. Baz
Hi Baz, as a recent arrival i was quite surprised that a few garages advise using a cheap semi synthetic 10W-40 in these engines. Outside of the book advice and a minimal saving for an engine that uses minimal amounts of oil. Could this be a contributing factor as well? i immediately changed mine to a high quality synthetic 10W-50 with a decent ZDDP level. Haven't got round to doing the coolant change on mine yet but still thinking about the low temp thermostat. Tim '92 928gts '03 996 Cab Tip
Good morning Baz Thank you for your explanations, I think you are right about the coolant flow at tickover when hot. In old turbo cars the practice was to keep the revs up for a minute or 2 after a fast run to cool the turbo. We use to have similar problems at work about 30 years ago with Maybach diesels. P
Baz, years ago I bought a 924S off you which got me totally hooked on Porsches and I think it will be a loss to the Porsche community when you do retire although I'm sure you won't be able to let go completely. Thanks for a good introduction to Porsche. A really good explanation of the water cooled flat 6 and interesting about the cylinder oil jets but at what % engine speed do the valves generally open? I suppose governed by oil pressure. Not knowing the oil system is there one valve controlling this or one per bank? The engines give such good low down torque that it is not necessary to drive on the cam as in the old carburetor engines we learned to drive on when we had to stay above 3,000 rpm or whatever to have any pull at all and as you say if the cars are running at higher revs the problem isn't a concern until traffic or legal compliance cause the engines to run low speed and then a quick getaway. Would it help as lights change for instance to raise engine revs to 2,000 before pulling away or blipping the throttle? To give the bores an extra squirt.
DivineE
New member
Great to read your latest findings. Also interested to hear your thoughts on the electric water pump suggested above and did I understand correctly that Porsche have added an electric oil pump? If so that is interesting, perhaps they agree with you and are trying to work to solution (without ovbiously changing the design at the end of the engines production life that might have opened the flood gates to a crippling number of claims!). I am very keen to see what the new freshly designed series of engines look like/perform like after a few years. I wonder how the design will compare. Regards, Ben
AndrewEllinas
New member
The majority of my journey work each day is in crawling traffic. I've often wondered whether this stop-start journey was responsible for the IMS failure in my 996 that went to Hartech to be fixed - and they are highly recommended.
To be honest I have not yet properly tested the pressure at which the valves open (there is one to each cylinder) but that is not really the issue - which is that at low revs you all already know the oil pressure is less than it is at higher revs when the relief valve opens - and whatever pressure it opens at - the flow and jet volume will always be less while the oil pressure is under maximum. Intelligent comments recently - it is (IMHO) exactly because these engines produce so much torque at very low revs that they are being used while both oil flow and pressure and coolant flow are relatively low and IMHO this should have been thought about and increased in the design to compensate. Good idea about keeping the revs up at standstill (or blipping the throttle like at a racing start) that really would help. Good contributions all adding to the preventative issue and worthwhile trying to adopt. thanks Baz
The possibility of the cooling system struggling is interesting. I've been working on a 400Hp diesel this morning and despite the engine to rad hoses only being about 70cm long they are still 75mm diameter. Flow capacity in the rad must be 100 times that in the 996. Tim
Posts made and opinions expressed are those of the individual forum members
Use of the Forum is subject to the Terms and Conditions
