bazhart
Barcelona
- Joined
- 20 May 2009
- Messages
- 1,343
Because so many specialists (and some privateers) now have boroscopes and owners are keen to find out about their own engine's condition – some understandable confusion has cropped up over interpretation about bore scoring (where the word 'scoring" is a dictionary definition taken to mean 'a deep line cut into a surface").
To try and clear this up first we need to explain some different types of failures. This is our interpretation and explanation after numerous tests, measurements, observations and correlations over many decades and fits all the evidence we have of a great number of failures – but may not be the full story. (N.b. Photos attached to demonstrate differences).
(1) Seizing-up. This is the traditional full seizure cause by the piston expanding more than the cylinder diameter (usually caused by weak mixtures, lack of coolant, ignition problems, blocked exhausts etc).
It was very common in 2 strokes in the mid '70's and '80's. When the piston expands so big it tries to be bigger than the cylinder diameter it is running in - the extreme pressure breaks through the oil film leaving alloy material pressing hard against any bore material and rubbing up and down so fast the piston surface micro-melts and cools raising bumps that can micro weld for a few milliseconds to the piston and/or cylinder wall and the rough surface that results tears into both the piston and cylinder bore leaving vertical grooves on both sides of the cylinder (although you can only see those in the cylinder wall with a boroscope as it requires stripping to inspect the piston surface).
Similar to pot holes in road surfaces - the lumps elongate the grooves and dislodge more particles to join the damaged lumps making things worse. They usually go from OK to fully seized in less than 3 seconds and leave several grooves or scores in the surface on both sides of the cylinder bore.
(2) Scuffing. This occurs when a piston is travelling slowly but with a high force (in an engine delivering high torque - typically big diesel ships engines trying to disembark where revs must be kept low to avoid the propeller cavitating but torque must be high to get the mass moving – and also known in some large capacity 4 stroke thumpers like Ducati's).
When a piston converts the linear movement of the piston to rotating movement in the crankshaft the force caused by the angle of the rod connecting them together pushes the piston into the cylinder wall on the aptly named 'thrust side". The oil film present then gets squeezed out during the stroke until the piston changes direction and encounters more fresh oil.
When the piston is moving faster (at higher revs) it has more momentum so there is less inclination to press into the cylinder wall preferring instead to continue straight movement and because the piston is barrel shaped - the oil film acts a bit like a surf-board allowing that oil film to roll-up on the leading edge and stay between the piston and the cylinder bore. Oil viscosity measurement is time based and so the slower the piston is travelling (i.e. low revs) the longer time there is for the oil film to be squeezed out leaving metal to metal contact under load. This will often leave a thin patch of alloy stuck on an iron cylinder wall but the piston is often not hot enough to melt the surface and does not always scores the bores.
If it does score the bore it will only be on the thrust side since the piston is still smaller than the bore and it is under no load on the other side where the oil has no pressure to squeeze it out of contact – it only scores as a result of the resolution of forces back from the con-rod angle pushing the piston against the cylinder wall.
(3) Scoring. This is a relatively new phenomenon particularly in the engines with Lokasil bores so we need to discuss the differences to understand cause and effect. It is more similar to 'scuffing" but occurs with less-load on the pistons and more due to silicon particle release often at higher mileages.
Because the thermal expansion of pistons is a function of the diameter – bigger pistons have greater clearances cold to hot than smaller ones and since oil film thicknesses are the same for both – cold to hot clearances can get too extreme in large bore engines – and since iron has a lower expansion rate - large cold clearances were needed to ensure sufficient clearance when hot that were noisy and did not run smoothly until pistons had expanded up to size during fast road work or racing.
So around 45 years ago alloy cylinders (that expand and contract more similarly to alloy pistons than iron bores) were developed that were lighter, cheaper and had better heat transfer qualities with one weakness – the cylinder bores were too soft. So silicon was mixed with the alloy in such proportions that some of it could not be absorbed in the mix and grew into evenly spaced small very well bonded hard silicon particles throughout the crankcase casting – creating 'ALUSIL".
Because the small surface particles were so hard – and because after honing there were always some particles almost machined completely away and hardly held in the surface that easily broke loose initially - pistons only worked if they had a hard iron coating electroplated onto them (so any loose silicon particles could not impinge into the piston surface and were small enough to escape before causing any damage).
Alusil worked very well (924S, 944, 944 Turbo, 968) the only problem was the cost of machining the crankcase with hard silicon particles everywhere so Lokasil was an attempt to locate silicon (Local – silicon or Lokasil) just at the cylinder walls. Larger silicon particles (than in Alusil) were pre-formed into a porous tube (looking like a rough cast in iron liner) and similarly cast into the cylinder block - except the liner was very porous and the molten aluminium flowed under pressure into the voids to entrap the silicon just near the cylinder bore area. A great idea – if it worked. We think the variation in silicon distribution that resulted was insufficiently even to be foolproof in every example.
Early 96mm bore engines suffered from bore ovality cracking cylinders because the silicon pre-formed area of the bore was not as strong as Alusil and neither was the outer tube of pure alloy surrounding it so when the 93mm bore Boxster S casting was bored out 3mm larger (without increasing the O/D) the cylinders were weaker. Later engines probably had a different silicon mix (to make them stronger) and that probably included even larger silicon particles. Unfortunately at the same time the hard iron piston electroplating process was apparently banned on health and safety grounds in Europe so piston coatings became softer plastic instead.
We believe scoring occurs when small (but larger than previously) particles of silicon break free from the surface of the cylinder bore and become trapped between the piston and the cylinder bore, rubbing up and down between them until they either escape from the top or bottom of the piston to be taken away with the oil before it has done any damage or gets stuck on the piston surface to develop a score and at the same time knock a few more particles loose in a typical 'catastrophe theory" knock-on effect creating larger scores and more free particles.
We think that sometimes the particle can sit in the groove it has created causing no further damage for a while, or become imbedded deeper in the piston surface so it no longer 'sticks out" – again delaying further damage until another particle becomes loose.
Like scuffing – this only ever occurs on the thrust side of the piston (because it relies on the pressure under load to press the particles into the surfaces while the oil film is thin) and because the other side of the piston is unmarked they can be driven for many mils during early bore scoring – eventually suffering high oil consumption and/or knocking noises as the piston rocks too much at TDC.
Contributory factors are therefore high loads at lower speeds, thin oil (by choice or high temperatures), higher mileage examples (where the oil has gradually washed away some surface alloy exposing more silicon particles), increased bore clearances (due to increasing ovality resulting in pistons knocking and rocking into the loaded surface freeing up particles with more momentum and squeezing the oil out faster), uneven distribution of silicon particles during manufacture (resulting in some areas have more or less silicon in some areas), poor bonding of the silicon particles in the substrate alloy in some areas.
It occurs on bank 2 long before bank 1 where the main difference between banks is that the coolest coolant enters the bottom of both banks but the thrust surface is on the bottom of bank 1 (where it will be cooler) and the top of bank 2 (where our tests shows it is hotter – especially during heat soak while stationary after a run).
Hotter means the oil must be thinner on the bank 2 thrust face and therefore more likely to suffer the consequences of loose silicon particles than in bank 1 where the oil will be thicker and form a thicker layer in which the particles can float and/or escape before causing too much damage.
Our interpretation fits all our evidence as follows.
(a) Bank 2 suffers first and is hotter.
(b) Variation in silicon distribution shown by some cylinders in which the oil has washed away patches of the surface.
(c) Different distribution shown by some areas of the bore being different sizes than other where the hones have more easily removed material.
(d) Our hard black 'diamond like" coated test pistons clearly showing minute score lines of free silicon particles in the shiny black surface.
(e) The fact that hard iron coated pistons resisted scoring but softer plastic coated ones did not.
(f) That some leading specialist piston manufacturers that use plastic coatings now exclude their use in Lokasil and Alusil.
(g) Manufacturers own publicity differentiates the silicon particle size in earlier and later production types.
(h) Our internal tests on Alusil, alloy and Lokasil showed up the Lokasil as susceptible to crumbling under pressure, releasing surface particles. For this reason we do not plate Nikasil onto failed scored Lokasil bores in case the scoring damaged the underlying substrate or the pressure of thrust loads gradually breaks down the substrate that could eventually result in the Nikasil becoming detached. Also our cylinders stabilise the liner by converting it to a closed deck to stop the top going oval - which direct plating does not accomplish (Gen 2 not needed due to already being closed deck).
(4) Nikasil is completely different to any other cylinder bore. It is an electroplated surface bonded very well to an aluminium cylinder wall forming a thin tube of homogeneous material that will not crumble or release particles – and is oleophilic. It is also very hard and difficult for anything to score into it's surface. It is however very reflective and shows up polishing lines that have no depth (and are therefore by definition not scoring) and are caused by piston rings bedding in.
All surfaces (even finely ground ones) under enough magnification look like rough saw tooth like edges and how 'smooth" something is becomes largely a matter of how much magnification is involved.
If piston rings had chamfered edges they would not work effectively but minute surface differences do initially rub on the cylinder bore slightly differently in some areas compared to others. Because the rings are not perfectly round (under microscopic measurement) performance for many years benefits from 'running in" (just like many other parts of an engine and for the same reasons) but on the highly reflective bore surface this can show up as polished lines of no depth and with no deterioration of the surface – just an ability to reflect light differently making them look like scoring lines.
To help 'run-in" rings the Nikasil surface (like other bore materials) needs some roughness created by honing cross hatch lines that retain oil and help that initial running in period minutely wear the ring into a compatible shape where they fit each other perfectly.
The evidence of 'scoring" would therefore be if the surface had deeper grooves that cut deeper under the reflective surface or the honing marks (which would be scoring) or just simply looked different but had no depth (which would not).
(5) Gen 2 9A1 seizing of Alusil blocks.
Because the failures we have seen with this model were caused by the bore shrinking and reducing the piston clearance they create full seizures marking both sides of the piston and cylinder bore.
The difference with the GT3, Turbo and Hartech Nikasil plated cylinders is that they are plated onto strong dense aerospace alloy thick tubes (that have no surface porosity therefore support the plated Nikasil perfectly) forming a reliable hard naturally lubricated surface that holds shape, hardly ever wears and expands and contracts more in keeping with the piston enabling the piston clearances to remain tight during different running conditions and temperatures and in engines that can run well at both modest and high power outputs.
Nikasil was also plated onto most of the later air cooled 911 cylinders with similar lomg lasting reliable performance. The only downside is when the polishing marks in the cylinder bores result in the inexperienced using boroscopes interpreting them as bore scoring.
Thin alloy Nikasil plated dry liner tubes fitted into a hole machined in the original bore surface can suffer from differences in the fit and differential expansion that sometimes cause them to move or become lose.
Hartech Nikasil plated cylinders are thicker and are wet closed deck liners so do not incur any fitting tolerance problems and have full contact with coolant while are thick enough to retain roundness and size (and also incorporate ribbed exteriors to increase surface area to aid cooling rates).
CONCLUSIONS.
Generally Early Lokasil engines (from 2.5 Boxsters up to and including Boxster S 3.2's) had hard iron coated pistons and thick cylinders so rarely cracked or scored and were actually reliable.
Early Lokasil 996 3.4's also had iron coated pistons but thinner cylinder walls and eventually often cracked.
Later Lokasil 996 and 997 engines had thinner cylinders as well but stiffer (due to larger silicon particles and a different mix) and went oval more slowly and rarely cracked but had plastic coated pistons that didn't resist the impact of any released (probably larger) silicon particles and scored.
Nikasil plated cylinders (on solid tubular thick liners) are most expensive – highly reliable but may show polish marks that some may misinterpret as scoring but have no depth and are running in polishing marks of no consequence.
Note the thick bottom parts of the casting in the photo where the shrinkage occurs.
Baz
To try and clear this up first we need to explain some different types of failures. This is our interpretation and explanation after numerous tests, measurements, observations and correlations over many decades and fits all the evidence we have of a great number of failures – but may not be the full story. (N.b. Photos attached to demonstrate differences).
(1) Seizing-up. This is the traditional full seizure cause by the piston expanding more than the cylinder diameter (usually caused by weak mixtures, lack of coolant, ignition problems, blocked exhausts etc).
It was very common in 2 strokes in the mid '70's and '80's. When the piston expands so big it tries to be bigger than the cylinder diameter it is running in - the extreme pressure breaks through the oil film leaving alloy material pressing hard against any bore material and rubbing up and down so fast the piston surface micro-melts and cools raising bumps that can micro weld for a few milliseconds to the piston and/or cylinder wall and the rough surface that results tears into both the piston and cylinder bore leaving vertical grooves on both sides of the cylinder (although you can only see those in the cylinder wall with a boroscope as it requires stripping to inspect the piston surface).
Similar to pot holes in road surfaces - the lumps elongate the grooves and dislodge more particles to join the damaged lumps making things worse. They usually go from OK to fully seized in less than 3 seconds and leave several grooves or scores in the surface on both sides of the cylinder bore.
(2) Scuffing. This occurs when a piston is travelling slowly but with a high force (in an engine delivering high torque - typically big diesel ships engines trying to disembark where revs must be kept low to avoid the propeller cavitating but torque must be high to get the mass moving – and also known in some large capacity 4 stroke thumpers like Ducati's).
When a piston converts the linear movement of the piston to rotating movement in the crankshaft the force caused by the angle of the rod connecting them together pushes the piston into the cylinder wall on the aptly named 'thrust side". The oil film present then gets squeezed out during the stroke until the piston changes direction and encounters more fresh oil.
When the piston is moving faster (at higher revs) it has more momentum so there is less inclination to press into the cylinder wall preferring instead to continue straight movement and because the piston is barrel shaped - the oil film acts a bit like a surf-board allowing that oil film to roll-up on the leading edge and stay between the piston and the cylinder bore. Oil viscosity measurement is time based and so the slower the piston is travelling (i.e. low revs) the longer time there is for the oil film to be squeezed out leaving metal to metal contact under load. This will often leave a thin patch of alloy stuck on an iron cylinder wall but the piston is often not hot enough to melt the surface and does not always scores the bores.
If it does score the bore it will only be on the thrust side since the piston is still smaller than the bore and it is under no load on the other side where the oil has no pressure to squeeze it out of contact – it only scores as a result of the resolution of forces back from the con-rod angle pushing the piston against the cylinder wall.
(3) Scoring. This is a relatively new phenomenon particularly in the engines with Lokasil bores so we need to discuss the differences to understand cause and effect. It is more similar to 'scuffing" but occurs with less-load on the pistons and more due to silicon particle release often at higher mileages.
Because the thermal expansion of pistons is a function of the diameter – bigger pistons have greater clearances cold to hot than smaller ones and since oil film thicknesses are the same for both – cold to hot clearances can get too extreme in large bore engines – and since iron has a lower expansion rate - large cold clearances were needed to ensure sufficient clearance when hot that were noisy and did not run smoothly until pistons had expanded up to size during fast road work or racing.
So around 45 years ago alloy cylinders (that expand and contract more similarly to alloy pistons than iron bores) were developed that were lighter, cheaper and had better heat transfer qualities with one weakness – the cylinder bores were too soft. So silicon was mixed with the alloy in such proportions that some of it could not be absorbed in the mix and grew into evenly spaced small very well bonded hard silicon particles throughout the crankcase casting – creating 'ALUSIL".
Because the small surface particles were so hard – and because after honing there were always some particles almost machined completely away and hardly held in the surface that easily broke loose initially - pistons only worked if they had a hard iron coating electroplated onto them (so any loose silicon particles could not impinge into the piston surface and were small enough to escape before causing any damage).
Alusil worked very well (924S, 944, 944 Turbo, 968) the only problem was the cost of machining the crankcase with hard silicon particles everywhere so Lokasil was an attempt to locate silicon (Local – silicon or Lokasil) just at the cylinder walls. Larger silicon particles (than in Alusil) were pre-formed into a porous tube (looking like a rough cast in iron liner) and similarly cast into the cylinder block - except the liner was very porous and the molten aluminium flowed under pressure into the voids to entrap the silicon just near the cylinder bore area. A great idea – if it worked. We think the variation in silicon distribution that resulted was insufficiently even to be foolproof in every example.
Early 96mm bore engines suffered from bore ovality cracking cylinders because the silicon pre-formed area of the bore was not as strong as Alusil and neither was the outer tube of pure alloy surrounding it so when the 93mm bore Boxster S casting was bored out 3mm larger (without increasing the O/D) the cylinders were weaker. Later engines probably had a different silicon mix (to make them stronger) and that probably included even larger silicon particles. Unfortunately at the same time the hard iron piston electroplating process was apparently banned on health and safety grounds in Europe so piston coatings became softer plastic instead.
We believe scoring occurs when small (but larger than previously) particles of silicon break free from the surface of the cylinder bore and become trapped between the piston and the cylinder bore, rubbing up and down between them until they either escape from the top or bottom of the piston to be taken away with the oil before it has done any damage or gets stuck on the piston surface to develop a score and at the same time knock a few more particles loose in a typical 'catastrophe theory" knock-on effect creating larger scores and more free particles.
We think that sometimes the particle can sit in the groove it has created causing no further damage for a while, or become imbedded deeper in the piston surface so it no longer 'sticks out" – again delaying further damage until another particle becomes loose.
Like scuffing – this only ever occurs on the thrust side of the piston (because it relies on the pressure under load to press the particles into the surfaces while the oil film is thin) and because the other side of the piston is unmarked they can be driven for many mils during early bore scoring – eventually suffering high oil consumption and/or knocking noises as the piston rocks too much at TDC.
Contributory factors are therefore high loads at lower speeds, thin oil (by choice or high temperatures), higher mileage examples (where the oil has gradually washed away some surface alloy exposing more silicon particles), increased bore clearances (due to increasing ovality resulting in pistons knocking and rocking into the loaded surface freeing up particles with more momentum and squeezing the oil out faster), uneven distribution of silicon particles during manufacture (resulting in some areas have more or less silicon in some areas), poor bonding of the silicon particles in the substrate alloy in some areas.
It occurs on bank 2 long before bank 1 where the main difference between banks is that the coolest coolant enters the bottom of both banks but the thrust surface is on the bottom of bank 1 (where it will be cooler) and the top of bank 2 (where our tests shows it is hotter – especially during heat soak while stationary after a run).
Hotter means the oil must be thinner on the bank 2 thrust face and therefore more likely to suffer the consequences of loose silicon particles than in bank 1 where the oil will be thicker and form a thicker layer in which the particles can float and/or escape before causing too much damage.
Our interpretation fits all our evidence as follows.
(a) Bank 2 suffers first and is hotter.
(b) Variation in silicon distribution shown by some cylinders in which the oil has washed away patches of the surface.
(c) Different distribution shown by some areas of the bore being different sizes than other where the hones have more easily removed material.
(d) Our hard black 'diamond like" coated test pistons clearly showing minute score lines of free silicon particles in the shiny black surface.
(e) The fact that hard iron coated pistons resisted scoring but softer plastic coated ones did not.
(f) That some leading specialist piston manufacturers that use plastic coatings now exclude their use in Lokasil and Alusil.
(g) Manufacturers own publicity differentiates the silicon particle size in earlier and later production types.
(h) Our internal tests on Alusil, alloy and Lokasil showed up the Lokasil as susceptible to crumbling under pressure, releasing surface particles. For this reason we do not plate Nikasil onto failed scored Lokasil bores in case the scoring damaged the underlying substrate or the pressure of thrust loads gradually breaks down the substrate that could eventually result in the Nikasil becoming detached. Also our cylinders stabilise the liner by converting it to a closed deck to stop the top going oval - which direct plating does not accomplish (Gen 2 not needed due to already being closed deck).
(4) Nikasil is completely different to any other cylinder bore. It is an electroplated surface bonded very well to an aluminium cylinder wall forming a thin tube of homogeneous material that will not crumble or release particles – and is oleophilic. It is also very hard and difficult for anything to score into it's surface. It is however very reflective and shows up polishing lines that have no depth (and are therefore by definition not scoring) and are caused by piston rings bedding in.
All surfaces (even finely ground ones) under enough magnification look like rough saw tooth like edges and how 'smooth" something is becomes largely a matter of how much magnification is involved.
If piston rings had chamfered edges they would not work effectively but minute surface differences do initially rub on the cylinder bore slightly differently in some areas compared to others. Because the rings are not perfectly round (under microscopic measurement) performance for many years benefits from 'running in" (just like many other parts of an engine and for the same reasons) but on the highly reflective bore surface this can show up as polished lines of no depth and with no deterioration of the surface – just an ability to reflect light differently making them look like scoring lines.
To help 'run-in" rings the Nikasil surface (like other bore materials) needs some roughness created by honing cross hatch lines that retain oil and help that initial running in period minutely wear the ring into a compatible shape where they fit each other perfectly.
The evidence of 'scoring" would therefore be if the surface had deeper grooves that cut deeper under the reflective surface or the honing marks (which would be scoring) or just simply looked different but had no depth (which would not).
(5) Gen 2 9A1 seizing of Alusil blocks.
Because the failures we have seen with this model were caused by the bore shrinking and reducing the piston clearance they create full seizures marking both sides of the piston and cylinder bore.
The difference with the GT3, Turbo and Hartech Nikasil plated cylinders is that they are plated onto strong dense aerospace alloy thick tubes (that have no surface porosity therefore support the plated Nikasil perfectly) forming a reliable hard naturally lubricated surface that holds shape, hardly ever wears and expands and contracts more in keeping with the piston enabling the piston clearances to remain tight during different running conditions and temperatures and in engines that can run well at both modest and high power outputs.
Nikasil was also plated onto most of the later air cooled 911 cylinders with similar lomg lasting reliable performance. The only downside is when the polishing marks in the cylinder bores result in the inexperienced using boroscopes interpreting them as bore scoring.
Thin alloy Nikasil plated dry liner tubes fitted into a hole machined in the original bore surface can suffer from differences in the fit and differential expansion that sometimes cause them to move or become lose.
Hartech Nikasil plated cylinders are thicker and are wet closed deck liners so do not incur any fitting tolerance problems and have full contact with coolant while are thick enough to retain roundness and size (and also incorporate ribbed exteriors to increase surface area to aid cooling rates).
CONCLUSIONS.
Generally Early Lokasil engines (from 2.5 Boxsters up to and including Boxster S 3.2's) had hard iron coated pistons and thick cylinders so rarely cracked or scored and were actually reliable.
Early Lokasil 996 3.4's also had iron coated pistons but thinner cylinder walls and eventually often cracked.
Later Lokasil 996 and 997 engines had thinner cylinders as well but stiffer (due to larger silicon particles and a different mix) and went oval more slowly and rarely cracked but had plastic coated pistons that didn't resist the impact of any released (probably larger) silicon particles and scored.
Nikasil plated cylinders (on solid tubular thick liners) are most expensive – highly reliable but may show polish marks that some may misinterpret as scoring but have no depth and are running in polishing marks of no consequence.
Note the thick bottom parts of the casting in the photo where the shrinkage occurs.
Baz
