Cam design/Duration/Lift/Valve Sizes
Discussion
Are you sure you want to get technical? 
Ok- so here we go:
I've designed and tested cams of different durations on the dyno and on simulation, playing with durations, phasing and lift.
I've never seen extra valve lift causing detriment to performance. The thing to remember however is that port flow and cam optimisation (both phasing and duration) are intimately related. Some people site that if a port flow maxes out at a certain valve lift there is little point in having a valve that lifts higher then that in an engine. I would counter that by saying that a port flow rig test is quite an artificial laboratory condition where a blowing rig is continually blowing or sucking- where as in a real engine has all manner of pumping and pulsating flows with crank angle. A flow bench will also usually be flowed at say, 10, 20 or perhaps 28 inches of water pressure drop. A real high performance road engine might see up to 70 inches of water pressure difference across an exhaust side for an instant. So flow rig data is great as a tool for comparisons of various ports but it is dangerous to draw direct conclusions to the engine. At any valve lift less then a quarter of the valve head diameter the flow will be limited by the annulus area formed by the valve seat and edge of the valve- above this the flow is more dictated by the valve size itself and then at still more lift the port size. Fitting as much valve lift in for a given duration is nearly always a good thing. Ultimately you're limited by valve train structural issues: including the diameter of the tappet (on a Direct acting bucket tappet design) which dictates your maximum "eccentricity" (this is to ensure that you don't run off of the chamfer of the tappet itself) - and of course your forces and Hertzian stresses- where you usually pay attention to the radius of curvature at the cam nose. It's important to keep entrainment and sliding velocities up for reasons of lubrication and wear.
Engine are very sensitive to both cam phasing and duration. If a cam profile has long duration ramps- a change in valve clearances will have a big effect on Volumetric Efficiency because you're effectively changing the duration for a little change in lash. I've seen this phenomenon on the dyno. Some engines are more sensitive then others and my current thinking is that better flowing head designs (for a given cylinder capacity) are effected more by this phenomenon (Pent roofs and hemis).
A longer duration cam fitted to an engine but timed in the same way ( with common Maximum opening point) will typically lose low speed VE due to charge-reversion- this is where the intake valve is kept open too long and at low engine speeds there is sufficient time around so that the fresh charge reverses back up the intake. On an engine with variable cam timing this is addressed by advancing the inlet cam at low speeds so that Intake valve closing (IVC) occurs earlier.
Alot of my work have been carried out on modern 4 valve pent roof chambered engines with variable cam phasing and negligible to poor exhaust tuning at high engine speeds. The rules change and become a lot more complicated with fixed cam phased engines. If we consider all the valve events
IVO, IVC EVO and EVC: I have found that IVC has the biggest bearing on performance. At low speeds you want to have IVC occur at around 30-35 degrees ABDC ( on a good flow four valve per cylinder engine)- in order to effective trap the charge). This will move up to about 60-65 degrees ABDC peak power engine speeds-assuming peak power is occurring at 6400- 7000 rpm. You can vary the duration but the above IVC optimums are generally true. If you go very short on duration and have bad flowing ports and/or high intake losses it is sometimes necessary to retard the intake cam further to perhaps an IVC of 70 degrees, When you start to optimise for a peak power ABOVE 7000 rpm the actual cam duration comes into play then and becomes the next constraint. This is why Hondas effectively vary their cam phase AND duration with engine speed. On a two valve engine where the flow into the cylinder for a given cylinder size is usually worse- a later IVC would be required.
Actual cam duration is much more important on the exhaust side due to blow down conditions. EVO is a very critical effect on top end/bottom end performance. At low speeds you need a late EVO (closer to BDC) event to get the most work out of the engine cycle. Too late however and the pumping losses increase. As the engine speed rises there isn't enough REAL time ( as distinct from crank angle time) for blow down to occur hence an earlier EVO is required. A late EVO is also beneficial to part load fuel economy.
If you have exhaust tuning, which is all too rare on production engines these days due to close coupled catalysts, you need overlap to encourage the scavenging process. Scavenging is when exhaust tuning successfully scoops out the remaining residual charge left in the engine cylinder while simultaneously encouraging fresh charge to fill the cylinder via overlap. On engines with variable cam phasing, because the IVC effect predominates all others for performance, advancing the cam at low speeds becomes a priority usually (not idle obviously-where you want to minimise overlap) – this goes hand in hand with high over lap at LOW speeds, hence you usually get more bang for your buck if you apply exhaust tuning at low speeds (higher overlap) – with modern VCP engines.
I have designed after market cams for a 2 valve fixed duration engine, which had a limit on overlap set, due to idle stability and the fuel injection air mass meter getting unsteady but it had an 80 Bhp/litre imperative. In this case I used a very late IVC of about 70 to 80 degrees ABDC, a mildly late IVO. This engine had poor Volumetric efficiency at low speeds due to poor scavenging from little over lap and poor trapping efficiency due to the late IVC, (about 60% VE) but "came on the cam" at around 2500 rpm and did make over 80 Bhp/litre thus meeting its criteria. It also gave good part load fuel economy due to the late IVC lowering the pumping losses at part load.
Ok- so here we go:
I've designed and tested cams of different durations on the dyno and on simulation, playing with durations, phasing and lift.
I've never seen extra valve lift causing detriment to performance. The thing to remember however is that port flow and cam optimisation (both phasing and duration) are intimately related. Some people site that if a port flow maxes out at a certain valve lift there is little point in having a valve that lifts higher then that in an engine. I would counter that by saying that a port flow rig test is quite an artificial laboratory condition where a blowing rig is continually blowing or sucking- where as in a real engine has all manner of pumping and pulsating flows with crank angle. A flow bench will also usually be flowed at say, 10, 20 or perhaps 28 inches of water pressure drop. A real high performance road engine might see up to 70 inches of water pressure difference across an exhaust side for an instant. So flow rig data is great as a tool for comparisons of various ports but it is dangerous to draw direct conclusions to the engine. At any valve lift less then a quarter of the valve head diameter the flow will be limited by the annulus area formed by the valve seat and edge of the valve- above this the flow is more dictated by the valve size itself and then at still more lift the port size. Fitting as much valve lift in for a given duration is nearly always a good thing. Ultimately you're limited by valve train structural issues: including the diameter of the tappet (on a Direct acting bucket tappet design) which dictates your maximum "eccentricity" (this is to ensure that you don't run off of the chamfer of the tappet itself) - and of course your forces and Hertzian stresses- where you usually pay attention to the radius of curvature at the cam nose. It's important to keep entrainment and sliding velocities up for reasons of lubrication and wear.
Engine are very sensitive to both cam phasing and duration. If a cam profile has long duration ramps- a change in valve clearances will have a big effect on Volumetric Efficiency because you're effectively changing the duration for a little change in lash. I've seen this phenomenon on the dyno. Some engines are more sensitive then others and my current thinking is that better flowing head designs (for a given cylinder capacity) are effected more by this phenomenon (Pent roofs and hemis).
A longer duration cam fitted to an engine but timed in the same way ( with common Maximum opening point) will typically lose low speed VE due to charge-reversion- this is where the intake valve is kept open too long and at low engine speeds there is sufficient time around so that the fresh charge reverses back up the intake. On an engine with variable cam timing this is addressed by advancing the inlet cam at low speeds so that Intake valve closing (IVC) occurs earlier.
Alot of my work have been carried out on modern 4 valve pent roof chambered engines with variable cam phasing and negligible to poor exhaust tuning at high engine speeds. The rules change and become a lot more complicated with fixed cam phased engines. If we consider all the valve events
IVO, IVC EVO and EVC: I have found that IVC has the biggest bearing on performance. At low speeds you want to have IVC occur at around 30-35 degrees ABDC ( on a good flow four valve per cylinder engine)- in order to effective trap the charge). This will move up to about 60-65 degrees ABDC peak power engine speeds-assuming peak power is occurring at 6400- 7000 rpm. You can vary the duration but the above IVC optimums are generally true. If you go very short on duration and have bad flowing ports and/or high intake losses it is sometimes necessary to retard the intake cam further to perhaps an IVC of 70 degrees, When you start to optimise for a peak power ABOVE 7000 rpm the actual cam duration comes into play then and becomes the next constraint. This is why Hondas effectively vary their cam phase AND duration with engine speed. On a two valve engine where the flow into the cylinder for a given cylinder size is usually worse- a later IVC would be required.
Actual cam duration is much more important on the exhaust side due to blow down conditions. EVO is a very critical effect on top end/bottom end performance. At low speeds you need a late EVO (closer to BDC) event to get the most work out of the engine cycle. Too late however and the pumping losses increase. As the engine speed rises there isn't enough REAL time ( as distinct from crank angle time) for blow down to occur hence an earlier EVO is required. A late EVO is also beneficial to part load fuel economy.
If you have exhaust tuning, which is all too rare on production engines these days due to close coupled catalysts, you need overlap to encourage the scavenging process. Scavenging is when exhaust tuning successfully scoops out the remaining residual charge left in the engine cylinder while simultaneously encouraging fresh charge to fill the cylinder via overlap. On engines with variable cam phasing, because the IVC effect predominates all others for performance, advancing the cam at low speeds becomes a priority usually (not idle obviously-where you want to minimise overlap) – this goes hand in hand with high over lap at LOW speeds, hence you usually get more bang for your buck if you apply exhaust tuning at low speeds (higher overlap) – with modern VCP engines.
I have designed after market cams for a 2 valve fixed duration engine, which had a limit on overlap set, due to idle stability and the fuel injection air mass meter getting unsteady but it had an 80 Bhp/litre imperative. In this case I used a very late IVC of about 70 to 80 degrees ABDC, a mildly late IVO. This engine had poor Volumetric efficiency at low speeds due to poor scavenging from little over lap and poor trapping efficiency due to the late IVC, (about 60% VE) but "came on the cam" at around 2500 rpm and did make over 80 Bhp/litre thus meeting its criteria. It also gave good part load fuel economy due to the late IVC lowering the pumping losses at part load.
OK, with you there. Just to recap, I assume phasing is relating to the harmonic content of the lobe shape?
My key curiosity was how there is a current trend to high lift/less duration with performance engines whilst the long cams have dropped in favour.
The engine I was thinking about was the venerable old A series (Which has proved one of the most thoroughly developed/tested engines going). In the old days, due to classes, restrictions there seemed to be more interest in short stroke screamers, now longer stroke, bigger capacities are in favour BUT as a general rule the overlap has come down. As you say exhaust tuning has fallen from favour quite a bit. I was wondering about how bore/stroke play a part in the equation and if this influences rough cam choice.
I started (but didn't finish) a mathematical treatment of what was actually happening. It got rather complicated, but the area which suggested a relationship was "Area for flow" vs crank position for different bore/stroke configuration. Assuming the port velocities have not been topped, I figure that enlarging the valve has an identical bearing on the area for flow as increasing lift. (Ignoring the speed/direction/weight of the gas for the moment).
I am convinced that bore/stroke are critical factors but will chuck in a thought for good measure. Harmonic Profiles. I suppose in the recent years, a lot of work has been done regarding these profiles instead of using symmetric, simple patterns. So, perhaps a modern short duration cam actually works better!?!?! Still not convinced.
I completely understand the basics of torque curve behaviour vs duration and how you fine tune that for the size of your engine with lift/valve sizes. But not quantitatively.
I suppose then there is also the "high octane fuel" problem of recent years and general trend of declining compression ratios.
What made me think about this was the lack of information available for selecting a supercharger cam. Since my engine is going to be about 8-12PSI of boost and will be a unique large capacity. The proprietry cams are designed for N/A or turbo engines of a different capacity. I immediately saw the need for about another 15 degrees of exhaust duration and getting the valves as big as possible as well as ports. But since I am considering alternative bore/stroke ratios I wondered how these would tweak the final figures.
Since I wasn't 100% sure of the right answer, I thought I'd better understand the principles of how bore and stroke couple with "area available for flow" as I will call it.
I expect quite a bit is to do with the available fuel, the resulting compression ratio and how that is achieved and the revs/duration which favours it. Also I expect that high lift rockers have taken the place of vicious ramp rates in the interest of life. But what I have not considered are the properties of the moving gas. This has to be critical in the equation and relating that to piston speed has to give a massive insight.
What do you reckon?
(By the way, my background is Chemistry, not Physics and my Maths is about good A level standard..Don't get too complicated too quick
)
My key curiosity was how there is a current trend to high lift/less duration with performance engines whilst the long cams have dropped in favour.
The engine I was thinking about was the venerable old A series (Which has proved one of the most thoroughly developed/tested engines going). In the old days, due to classes, restrictions there seemed to be more interest in short stroke screamers, now longer stroke, bigger capacities are in favour BUT as a general rule the overlap has come down. As you say exhaust tuning has fallen from favour quite a bit. I was wondering about how bore/stroke play a part in the equation and if this influences rough cam choice.
I started (but didn't finish) a mathematical treatment of what was actually happening. It got rather complicated, but the area which suggested a relationship was "Area for flow" vs crank position for different bore/stroke configuration. Assuming the port velocities have not been topped, I figure that enlarging the valve has an identical bearing on the area for flow as increasing lift. (Ignoring the speed/direction/weight of the gas for the moment).
I am convinced that bore/stroke are critical factors but will chuck in a thought for good measure. Harmonic Profiles. I suppose in the recent years, a lot of work has been done regarding these profiles instead of using symmetric, simple patterns. So, perhaps a modern short duration cam actually works better!?!?! Still not convinced.
I completely understand the basics of torque curve behaviour vs duration and how you fine tune that for the size of your engine with lift/valve sizes. But not quantitatively.
I suppose then there is also the "high octane fuel" problem of recent years and general trend of declining compression ratios.
What made me think about this was the lack of information available for selecting a supercharger cam. Since my engine is going to be about 8-12PSI of boost and will be a unique large capacity. The proprietry cams are designed for N/A or turbo engines of a different capacity. I immediately saw the need for about another 15 degrees of exhaust duration and getting the valves as big as possible as well as ports. But since I am considering alternative bore/stroke ratios I wondered how these would tweak the final figures.
Since I wasn't 100% sure of the right answer, I thought I'd better understand the principles of how bore and stroke couple with "area available for flow" as I will call it.
I expect quite a bit is to do with the available fuel, the resulting compression ratio and how that is achieved and the revs/duration which favours it. Also I expect that high lift rockers have taken the place of vicious ramp rates in the interest of life. But what I have not considered are the properties of the moving gas. This has to be critical in the equation and relating that to piston speed has to give a massive insight.
What do you reckon?
(By the way, my background is Chemistry, not Physics and my Maths is about good A level standard..Don't get too complicated too quick
)love machine said:
I figure that enlarging the valve has an identical bearing on the area for flow as increasing lift. (Ignoring the speed/direction/weight of the gas for the moment).
The curtain area will be proportional to the square of the valves radius, and proportional to the lift. Not sure hoe this relates to flow though.
Practicaly (and especialy with the A series), as the valve size goes up, shrouding by the combustion chamber or cylinder becomes a bigger problem.
Phasing is the same as cam timing- just another term for it. Technically speaking the reason I shouldn't use the term cam timing- is that it CAN encapsulate variable duration. Put another way Variable Valve timing could mean variable valve duration And/Or changing the phasing. Trouble is, when you use the technically correct "phasing" term no one knows what you're talking about!
This trend toward shorter duration has come about for two reasons-
1) it goes hand in hand with better flowing pent roof 4 valve cylinder heads. For the usual 1000-6500 rpm rev range with reasonable torque throughout you don't need the old 2 valve norms of 260 period, 240 works just fine.
2) The more widespread use of VCT means that the duration can be shortened further. As outlined before with fixed cam timing engines the period and timing is a compromise. IVC event being the most important factor- this can NOW be varied and less periods are now used.
Overlap is restricted because customers are much m ore critical of things like idle speed stability then say they were in the 1970s. Valve lift has increased because valve train design is often a lot stiffer. Compare a direct acting valve train of a K series Rover to you’re a series for instance. Extra Valve lift is usually beneficial.
Larger bore engines- such as say the Porsches favoured 100mm bore-78.9mm stroke- are falling out of favour for emissions reasons mainly. Over square engines tend to kick out more hydrocarbons due to more crevice volume/traps in the ring lands. Over square engines also tend to not have the combustion efficiency (which is something quite different to breathing and Volumetric efficiency) of a square design due to adverse surface to volume ratios and heat loss. Their burn rates can be effected in an adverse way too- long flame paths-bad for knock.
I've done quite a bit of extensive studying on bore stroke relationship effect on performance. I found that in terms of airflow/volumetric efficiency if everything was kept equal-we're talking valve sizes, and rod length to stroke ratios, the bore-stroke relationship effect on a VE curve was minimal- but simulation revealed that a bigger effect came from surface to volume ratios- which effects combustion AND air flow- as the heat transfer to the incoming charge-which effects your VE. If all other things are kept equal (cylinder capacity) the tuned exhaust lengths shouldn't be effected by bore-stroke ratio. Although your mean piston speed does change- the rate of change of volume in the cylinder from the swept movement of the piston will stay the same. It is this downward movement of the piston which defines the tuning pulse.
Did your mathematical treatment of this treat the flow as compressible or non compressible? To encapsulate tuning effects it needs to be thought of as compressible fluid flow. Enlargening the valves will increase your curtain area which will effect your low valve lift flow- when youre restricted by the annulus area I outlined before. With extra valve lift there will inevitably come a point where the valve heads is no longer a restriction- as it is out of the way and the next bottle neck is somewhere in the port- so here the extra valve lift will show little gain. As for where an increase in flow (low lift or high lift) benefits performance- I'm not sure, it would need investigation but wouldn't be difficult to do.
"I suppose then there is also the "high octane fuel" problem of recent years and general trend of declining compression ratios. "
There was a period where CRs dropped down due to octane ratings.
Of late I would actually say that CRs have increased. 1980s BMW E30s 325i for instance started at 9.7:1 and tumbled down to 8.8:1 now the latest Bimmer engines are always 10 plus and the M3 CSL is up at 11.6:1. This is allowed due to central spark plugs and better combustion chambers with faster burn combined with sophisticated knock control system which are able to adapt.
The thing to watch with Supercharged engine is too much overlap and you'll get "short circuiting" of fresh charge straight out of the exhaust due to so much pressure on the intake side. You're right in that exhaust period is very very important on a supercharged engine and has a big impact to performance.
>> Edited by Marquis_Rex on Monday 10th January 16:43
This trend toward shorter duration has come about for two reasons-
1) it goes hand in hand with better flowing pent roof 4 valve cylinder heads. For the usual 1000-6500 rpm rev range with reasonable torque throughout you don't need the old 2 valve norms of 260 period, 240 works just fine.
2) The more widespread use of VCT means that the duration can be shortened further. As outlined before with fixed cam timing engines the period and timing is a compromise. IVC event being the most important factor- this can NOW be varied and less periods are now used.
Overlap is restricted because customers are much m ore critical of things like idle speed stability then say they were in the 1970s. Valve lift has increased because valve train design is often a lot stiffer. Compare a direct acting valve train of a K series Rover to you’re a series for instance. Extra Valve lift is usually beneficial.
Larger bore engines- such as say the Porsches favoured 100mm bore-78.9mm stroke- are falling out of favour for emissions reasons mainly. Over square engines tend to kick out more hydrocarbons due to more crevice volume/traps in the ring lands. Over square engines also tend to not have the combustion efficiency (which is something quite different to breathing and Volumetric efficiency) of a square design due to adverse surface to volume ratios and heat loss. Their burn rates can be effected in an adverse way too- long flame paths-bad for knock.
I've done quite a bit of extensive studying on bore stroke relationship effect on performance. I found that in terms of airflow/volumetric efficiency if everything was kept equal-we're talking valve sizes, and rod length to stroke ratios, the bore-stroke relationship effect on a VE curve was minimal- but simulation revealed that a bigger effect came from surface to volume ratios- which effects combustion AND air flow- as the heat transfer to the incoming charge-which effects your VE. If all other things are kept equal (cylinder capacity) the tuned exhaust lengths shouldn't be effected by bore-stroke ratio. Although your mean piston speed does change- the rate of change of volume in the cylinder from the swept movement of the piston will stay the same. It is this downward movement of the piston which defines the tuning pulse.
Did your mathematical treatment of this treat the flow as compressible or non compressible? To encapsulate tuning effects it needs to be thought of as compressible fluid flow. Enlargening the valves will increase your curtain area which will effect your low valve lift flow- when youre restricted by the annulus area I outlined before. With extra valve lift there will inevitably come a point where the valve heads is no longer a restriction- as it is out of the way and the next bottle neck is somewhere in the port- so here the extra valve lift will show little gain. As for where an increase in flow (low lift or high lift) benefits performance- I'm not sure, it would need investigation but wouldn't be difficult to do.
"I suppose then there is also the "high octane fuel" problem of recent years and general trend of declining compression ratios. "
There was a period where CRs dropped down due to octane ratings.
Of late I would actually say that CRs have increased. 1980s BMW E30s 325i for instance started at 9.7:1 and tumbled down to 8.8:1 now the latest Bimmer engines are always 10 plus and the M3 CSL is up at 11.6:1. This is allowed due to central spark plugs and better combustion chambers with faster burn combined with sophisticated knock control system which are able to adapt.
The thing to watch with Supercharged engine is too much overlap and you'll get "short circuiting" of fresh charge straight out of the exhaust due to so much pressure on the intake side. You're right in that exhaust period is very very important on a supercharged engine and has a big impact to performance.
>> Edited by Marquis_Rex on Monday 10th January 16:43
I must say I found the previous text, and I'm sure whatever follows, fascinating. I'm afraid I'm unable to offer too much on the ins and out of cam timing, beyond the obvious stuff which has been stated here in more detail than I ever could.
What I'm also wondering though, is if there are any other mechanisims that could allow a faster port opening than a conventional poppet.
I'm aware of the compressed air, presumably computer controlled, stuff they use on F1 engines. I also wondered if there were other ways.
I thought about a rotating disc similar to that of two stroke, but obviously it's not that practical for the top of a four stroke engine.
There's the side valve, but I guess we won't go there!
I also wondered about a system like a ballfix valve (the rotating stainless ball with a hole through it).
>> Edited by dilbert on Thursday 10th February 05:30
What I'm also wondering though, is if there are any other mechanisims that could allow a faster port opening than a conventional poppet.
I'm aware of the compressed air, presumably computer controlled, stuff they use on F1 engines. I also wondered if there were other ways.
I thought about a rotating disc similar to that of two stroke, but obviously it's not that practical for the top of a four stroke engine.
There's the side valve, but I guess we won't go there!
I also wondered about a system like a ballfix valve (the rotating stainless ball with a hole through it).
>> Edited by dilbert on Thursday 10th February 05:30
Many thanks for that.
It all seems to lead into the whole hydrogen issue too. I'm just so amazed to find such detail, I guess this is one of those things where I have to say, why didn't I think of that. I'd like to have a go at building an I.C. engine one day (probably a small one) but having seen that I think I might try to go the rotary valve route.
I guess that not using oil in these valves is a good idea, you wouldn't want too much oxidization building up on the rotor (assuming it got the chance). I suppose that's a good reason for using a ball valve with hydrogen (cleaner burn).
Slightly off topic.....
I read a pretty big tirade against shipping hydrogen around.
I would have thought that it would be more practical to electrolyse the hydrogen at the filling station, or perhaps distribution points, than actually shipping it about.
Someone also mentioned that it would take 1 nuclear power station to drive 1M cars. I don't think that's a bad ratio. Hopefully information technology will reduce the need for transport.
I realise this is off topic, and I apologise. I'll not bother to post a new thread, but I suggest if you do want to reply to the hydrogen stuff, you start a new thread yourself. I'll keep my eyes open.
It all seems to lead into the whole hydrogen issue too. I'm just so amazed to find such detail, I guess this is one of those things where I have to say, why didn't I think of that. I'd like to have a go at building an I.C. engine one day (probably a small one) but having seen that I think I might try to go the rotary valve route.
I guess that not using oil in these valves is a good idea, you wouldn't want too much oxidization building up on the rotor (assuming it got the chance). I suppose that's a good reason for using a ball valve with hydrogen (cleaner burn).
Slightly off topic.....
I read a pretty big tirade against shipping hydrogen around.
I would have thought that it would be more practical to electrolyse the hydrogen at the filling station, or perhaps distribution points, than actually shipping it about.
Someone also mentioned that it would take 1 nuclear power station to drive 1M cars. I don't think that's a bad ratio. Hopefully information technology will reduce the need for transport.
I realise this is off topic, and I apologise. I'll not bother to post a new thread, but I suggest if you do want to reply to the hydrogen stuff, you start a new thread yourself. I'll keep my eyes open.
Gassing Station | Engines & Drivetrain | Top of Page | What's New | My Stuff