Ideal Camshaft Lobe Centreline Angle
Discussion
Right, let's get into this in detail now. Margaret Thatcher once said that to get a message across you have to say it three times. First tell the people what you are about to tell them, then tell them, then tell them what you just told them.
So I'll repeat what I've already told you. Optimum LSA is essentially invariant (plus or minus a few degrees) across a wide range of engine sizes and configurations. It is NOT in any way determinant on valve diameter per cc of cylinder capacity. What it is determinant on we'll come back to later but I'll start by running through the maths of some different engine configurations that affect overlap flow by either changing valve diameter per cc or cam duration.
1) 2V versus 4V engines.
Taking a 2v engine of a given capacity and redesigning it with two smaller inlet valves of the same total area increases total valve diameter by a factor of 1.414 which is the square root of 2. If valve diameter per cc played any part in LSA selection this would be a massive effect. According to the DV graph it would require a 5 degree widening in LSA. So how does it really stack up?
Let's take a 42mm inlet valve Ford CVH head, flow test it and then create on the computer a flow curve for two 29.70mm valves with the same total area and discharge coefficient. The peak flow stays the same but the flow at any lift below this changes radically. Low lift flow and therefore overlap flow goes up by a factor of 1.414 obviously - the increase in total valve diameter.
So the simplistic (not trying to be catty here but you have to think much deeper than a single level down) theory that valve diameter per cc affects LSA says that LSA must widen by a lot because overlap flow has gone up. That ignores many further factors.
First we need to look at total flow area under the crankshaft duration curve.
With an average cam for that valve size of 400 thou lift we actually get some small increase in high lift flow with the 4v head simply because the small valves reach peak flow at a smaller lift value i.e a higher ratio of Lift/Valve Diameter than the single large valve.
The area under the curve, which is a good measure of potential power increase, has gone up by 16.5%. However that still ignores many other factors. 4v engines can withstand higher compression ratios than 2v ones on the same octane fuel. Usually you can increase CR by 1 or 1.5 points. That alone increases potential power up to about 21% on an average engine with the same valve area as a 2v one. However the 4v engine has further benefits to offer. With its better swirl, faster burning chamber and lower ignition advance requirement it extracts more power from the same airflow and we end up with about 24% potential power increase.
So we're still seeing a smaller potential power increase than the overlap flow increase and simplistic theory says we still need to widen the LSA but that ignores the overiding factor of time. The faster the engine revs the less time there is for overlap flow to evacuate the exhaust gas residuals and with the higher power from the 4v design it needs to rev higher. BUT it doesn't need to rev higher in proportion to its power increase because it produces more torque. On average a 4v engine produces 10% more torque per litre than a 2v one so instead of having to rev 24% higher to produce 24% more power it only needs to rev about 14% higher.
So now we have it all. The increase in overlap flow with the same cam profile is up by a factor of 1.414. The power increase times the decrease in the time available for overlap flow to actually work is 1.24 x 1.14 = 1.414.
Bazinga! The factors cancel out. The increase in valve diameter with the same valve area gave us just the right amount of extra low lift flow and overlap flow to cancel out the increase in power multiplied by the decrease in time. Everything stays copacetic. The LSA doesn't need to change.
2) Change the bottom end capacity for a given cylinder head.
Let's take a given cylinder head and put it on a bottom end with 20% more capacity. Simplistic theory says we now have 20% less valve diameter per cc and LSA needs to tighten up. What does deeper theory tell us?
If we increase the engine capacity we can expect two things. The peak power won't stay the same but it will also not go up in proportion to the capacity change. In practice we find that with average engines we get about half the percentage increase in capacity as extra power. However we get most of the percentage capacity change as extra torque because torque stays closely proportional to engine capacity. So power goes up by 10% but also peak power rpm goes down by 10%. We know that from the simple rule that torque x rpm = power. Torque up by 20%, power up by 10%, rpm down by 10%.
So Bazinga! Again everything stays in proportion with respect to LSA. Overlap flow stays the same, power (and airflow) goes up by 10% but time available also goes up by 10% because rpm drops by 10%.
The same principle applies if engine capacity decreases under a given cylinder head. Power goes down but so does time available as rpm increases and overlap flow requirement stays the same.
3) Increase cam duration.
Let's take our 1950s traditional fast road cam with 30/60 - 60/30 timing and 270 degrees duration on a 105 degree LSA and increase duration by 10% which is a huge amount. We now have a 297 degree full race cam with 43.5/73.5 - 73.5/43.5 timing on the same 105 degree LSA. Overlap duration has gone up from 30/30 = 60 degrees to 43.3/43.5 = 87 degrees - a 45% increase.
What happens to power and overlap flow requirement?
Obviously the 10% extra cam duration should give us 10% extra area under the camflow curve for the same cylinder head. BUT this ignores that cam lift should also go up as duration increases. In fact it should go up in proportion to duration with a well designed cam and when we run the calculations the 10% extra lift gives us 4% to 8% extra flow area depending on how well the head flows at high lift. A good average is 16% extra total flow area from 10% extra duration. However the longer duration cam can also withstand more compression ratio and in fact needs it to work properly. We can increase CR by at least 2 points which increases power potential to about 21%.
Finally we must include the time factor because rpm must increase, and time go down, to gain the extra power. At the same torque value rpm must also go up by 21% to gain the extra 21% power. So we now have 21% extra power, airflow and overlap flow requirement but in 21% less time. 1.21 x 1.21 = 1.46.
Overlap duration goes up by a factor of 1.45 and flow/time area by a factor of 1.46.
Bazinga! Everything stays almost exactly in proportion once again.
So we can change almost any engine parameter - cylinder size, valve size, 2v or 4v, cam duration and everything relevant to LSA cancels out which is why LSA stays so constant over such a wide range of engine designs. Some things do indeed affect LSA including rod / stroke ratio as I've already said but I'll get into those in another post. This is probably enough for you to absorb for starters.
So I'll repeat what I've already told you. Optimum LSA is essentially invariant (plus or minus a few degrees) across a wide range of engine sizes and configurations. It is NOT in any way determinant on valve diameter per cc of cylinder capacity. What it is determinant on we'll come back to later but I'll start by running through the maths of some different engine configurations that affect overlap flow by either changing valve diameter per cc or cam duration.
1) 2V versus 4V engines.
Taking a 2v engine of a given capacity and redesigning it with two smaller inlet valves of the same total area increases total valve diameter by a factor of 1.414 which is the square root of 2. If valve diameter per cc played any part in LSA selection this would be a massive effect. According to the DV graph it would require a 5 degree widening in LSA. So how does it really stack up?
Let's take a 42mm inlet valve Ford CVH head, flow test it and then create on the computer a flow curve for two 29.70mm valves with the same total area and discharge coefficient. The peak flow stays the same but the flow at any lift below this changes radically. Low lift flow and therefore overlap flow goes up by a factor of 1.414 obviously - the increase in total valve diameter.
So the simplistic (not trying to be catty here but you have to think much deeper than a single level down) theory that valve diameter per cc affects LSA says that LSA must widen by a lot because overlap flow has gone up. That ignores many further factors.
First we need to look at total flow area under the crankshaft duration curve.
With an average cam for that valve size of 400 thou lift we actually get some small increase in high lift flow with the 4v head simply because the small valves reach peak flow at a smaller lift value i.e a higher ratio of Lift/Valve Diameter than the single large valve.
The area under the curve, which is a good measure of potential power increase, has gone up by 16.5%. However that still ignores many other factors. 4v engines can withstand higher compression ratios than 2v ones on the same octane fuel. Usually you can increase CR by 1 or 1.5 points. That alone increases potential power up to about 21% on an average engine with the same valve area as a 2v one. However the 4v engine has further benefits to offer. With its better swirl, faster burning chamber and lower ignition advance requirement it extracts more power from the same airflow and we end up with about 24% potential power increase.
So we're still seeing a smaller potential power increase than the overlap flow increase and simplistic theory says we still need to widen the LSA but that ignores the overiding factor of time. The faster the engine revs the less time there is for overlap flow to evacuate the exhaust gas residuals and with the higher power from the 4v design it needs to rev higher. BUT it doesn't need to rev higher in proportion to its power increase because it produces more torque. On average a 4v engine produces 10% more torque per litre than a 2v one so instead of having to rev 24% higher to produce 24% more power it only needs to rev about 14% higher.
So now we have it all. The increase in overlap flow with the same cam profile is up by a factor of 1.414. The power increase times the decrease in the time available for overlap flow to actually work is 1.24 x 1.14 = 1.414.
Bazinga! The factors cancel out. The increase in valve diameter with the same valve area gave us just the right amount of extra low lift flow and overlap flow to cancel out the increase in power multiplied by the decrease in time. Everything stays copacetic. The LSA doesn't need to change.
2) Change the bottom end capacity for a given cylinder head.
Let's take a given cylinder head and put it on a bottom end with 20% more capacity. Simplistic theory says we now have 20% less valve diameter per cc and LSA needs to tighten up. What does deeper theory tell us?
If we increase the engine capacity we can expect two things. The peak power won't stay the same but it will also not go up in proportion to the capacity change. In practice we find that with average engines we get about half the percentage increase in capacity as extra power. However we get most of the percentage capacity change as extra torque because torque stays closely proportional to engine capacity. So power goes up by 10% but also peak power rpm goes down by 10%. We know that from the simple rule that torque x rpm = power. Torque up by 20%, power up by 10%, rpm down by 10%.
So Bazinga! Again everything stays in proportion with respect to LSA. Overlap flow stays the same, power (and airflow) goes up by 10% but time available also goes up by 10% because rpm drops by 10%.
The same principle applies if engine capacity decreases under a given cylinder head. Power goes down but so does time available as rpm increases and overlap flow requirement stays the same.
3) Increase cam duration.
Let's take our 1950s traditional fast road cam with 30/60 - 60/30 timing and 270 degrees duration on a 105 degree LSA and increase duration by 10% which is a huge amount. We now have a 297 degree full race cam with 43.5/73.5 - 73.5/43.5 timing on the same 105 degree LSA. Overlap duration has gone up from 30/30 = 60 degrees to 43.3/43.5 = 87 degrees - a 45% increase.
What happens to power and overlap flow requirement?
Obviously the 10% extra cam duration should give us 10% extra area under the camflow curve for the same cylinder head. BUT this ignores that cam lift should also go up as duration increases. In fact it should go up in proportion to duration with a well designed cam and when we run the calculations the 10% extra lift gives us 4% to 8% extra flow area depending on how well the head flows at high lift. A good average is 16% extra total flow area from 10% extra duration. However the longer duration cam can also withstand more compression ratio and in fact needs it to work properly. We can increase CR by at least 2 points which increases power potential to about 21%.
Finally we must include the time factor because rpm must increase, and time go down, to gain the extra power. At the same torque value rpm must also go up by 21% to gain the extra 21% power. So we now have 21% extra power, airflow and overlap flow requirement but in 21% less time. 1.21 x 1.21 = 1.46.
Overlap duration goes up by a factor of 1.45 and flow/time area by a factor of 1.46.
Bazinga! Everything stays almost exactly in proportion once again.
So we can change almost any engine parameter - cylinder size, valve size, 2v or 4v, cam duration and everything relevant to LSA cancels out which is why LSA stays so constant over such a wide range of engine designs. Some things do indeed affect LSA including rod / stroke ratio as I've already said but I'll get into those in another post. This is probably enough for you to absorb for starters.
Edited by Pumaracing on Tuesday 7th February 14:52
David Vizard said:
First place to start – Dave B’s Peugeot numbers.
So, Dave B, let me have the flow figs, CR, rod length, valve sizes, rocker ratios, valve lift exhaust system spec etc for your Peugeot engine/s and let’s see what Cam Master comes up with. This might reveal a flaw in the program or a problem I have yet to uncover downstream of that when making the graphical simplifications. I am sure between all the switched on PH posters we have this can be figured out???? Throughout this though don’t forget your theory has to fit what happens in practice. You cannot change practice results to fit your theory.
That's not really where we need to start. We need to start with the A series which you undoubtedly have more flow data on than any of us and see what LSA your theory predicts for various engine configurations. We already know that the graph method based on CID/inch of valve diameter predicts about 116 degrees LSA which we know is at least 10 degrees too high.So, Dave B, let me have the flow figs, CR, rod length, valve sizes, rocker ratios, valve lift exhaust system spec etc for your Peugeot engine/s and let’s see what Cam Master comes up with. This might reveal a flaw in the program or a problem I have yet to uncover downstream of that when making the graphical simplifications. I am sure between all the switched on PH posters we have this can be figured out???? Throughout this though don’t forget your theory has to fit what happens in practice. You cannot change practice results to fit your theory.
So we are left with an uncomfortable conclusion. Either the CamMaster program comes up with completely different answers to the CID/inch method but you still think both are right or it comes up with the same answers which we already know are wrong.
Then try out any theoretical even smaller engine like a four cylinder motorbike one into both methods and see if either comes up with anything close to 105 degrees. We already know the CID/inch one doesn't so we must conclude the CamMaster doesn't either.
The deal here is that your main recent research has been on engines with cylinders of 700cc capacity or more in most cases and when we look at engines with much smaller cylinders the valve diameter / low lift flow theory goes straight out of the window because the actual optimum LSA stays very much the same as the big engines.
Until you resolve the small engine / big engine variance and agree that your method doesn't work for both we can't go into it deeper and find out what LSA is really responding to in the universal situation.
Extrapolate to extremes and then come back to the middle to check that everything still works. Only then do you have a viable theory.
David Vizard said:
Yes you are right in the an A Series of some 1300 cc does, in real life, need a LCA of 105 -106 or there-abouts. Not as predicted by my chart at 114 or so.
But I did say that the A series appeared to be a somewhat special case. However if we reduce the capacity of the long block to about a 1000 cc and re-do the LCA dyno test guess what happens???? The optimum LCA spreads to about 110.
Hmmmm. That's not even remotely what the big yellow book says. Taking various information but mainly from pages 271 and 292.But I did say that the A series appeared to be a somewhat special case. However if we reduce the capacity of the long block to about a 1000 cc and re-do the LCA dyno test guess what happens???? The optimum LCA spreads to about 110.
With single pattern cams a 1275 wants 102 degrees LSA (p292). Bigger engines especially over 1400 want 100 degrees (p271). A 1 litre based on a short stroke big block wants 4 (to maybe as much as 6 degrees) more than a 1480 on the same block which would indicate 104-106 degrees (p292). No mention of anywhere near 110 degrees in any size of engine.
So once again we're back to the same range as any other engine. Somewhere near 105 degrees plus or minus a bit.
David Vizard said:
At this point in time I am very sure of my two valve stuff as it relates to typical US V8's such as those by Chevy Ford and Chrysler (it even seems to work well with the latest Chrysler Hemi.
Oh boy are you about to get burned then on 2V engines from your own books. I just found your big block Chevy tuning book on Google books and the mandated LSA per CID/inch bears no resemblance to that for the small block Chevy with the same CID/inch numbers.All your BB Chevy cams run between 104 and 109 LSA even though the CID/inch numbers are far higher than for the SB Chevy which supposedly run the same 104 to 109 LSA on smaller CID/inch numbers.
So CID/inch doesn't predict the same LSA on different engines which is what I've been saying since the start.
Every engine seems to default back to a common start point which is 105 degrees plus or minus a bit.
Now we just need to find out why and I'll come back to that later.
ivanhoew said:
1) 2V versus 4V engines.
Taking a 2v engine of a given capacity and redesigning it with two smaller inlet valves of the same total area increases total valve diameter by a factor of 1.414
Let's take a 42mm inlet valve Ford CVH head, flow test it and then create on the computer a flow curve for two 29.70mm valves with the same total area and discharge coefficient.
is that feasable in the real world bearing in mind the shrouding to flow from the other inlet valve ?
The peak flow stays the same but the flow at any lift below this changes radically. Low lift flow and therefore overlap flow goes up by a factor of 1.414 obviously - the increase in total valve diameter.
is this truly born out by a real world graph?
Yes. If I look at the actual flow curves of both standard and modified 4v heads the discharge coefficient (Cd) throughout the lift range is very similar to 2v ones except somewhat down at mid lift flow with valves that are very close together. This mainly affects big valve heads obviously. Stock ones don't suffer too much. However there appears to be relatively little interference to the flow through one valve from the one adjacent to it at high lift or low lift.Taking a 2v engine of a given capacity and redesigning it with two smaller inlet valves of the same total area increases total valve diameter by a factor of 1.414
Let's take a 42mm inlet valve Ford CVH head, flow test it and then create on the computer a flow curve for two 29.70mm valves with the same total area and discharge coefficient.
is that feasable in the real world bearing in mind the shrouding to flow from the other inlet valve ?
The peak flow stays the same but the flow at any lift below this changes radically. Low lift flow and therefore overlap flow goes up by a factor of 1.414 obviously - the increase in total valve diameter.
is this truly born out by a real world graph?
In any event the overlap flow takes place at low lift and you wouldn't expect there to be much interference taking place at low flow numbers as indeed there isn't.
If there is interference it seems to occur mainly at mid lift figures around the 0.12 to 0.20 L/DV value. This can be explained by the fact that the air at those lifts is still trying to use the whole valve circumference, or at least most of it, and so at the sides of the valve seat the two valves do restrict each other somewhat. On a test I did of std 32mm and big 34mm valves in a heavily modified VW 16v head the low lift and high lift numbers went up exactly as predicted from the increase in valve size but the mid lift numbers stayed about the same so clearly restriction was taking place there with the big valves. I suspect that port shapes that biased the flow more to opposite sides of the two ports could alleviate some of this problem but there was no budget to investigate this.
At higher lifts the air starts to exit more through the long side of the valve seat directly towards the exhaust valves and the Cd picks up again.
At 50 thou lift the 32mm valves flowed 35.3 cfm and the 34mm ones 37.9 cfm - pretty much in line with the extra valve circumference. In fact the big valves were slightly better size for size.
At 500 thou lift the 32mm valves flowed 203 cfm and the 34mm ones 231 cfm - near as dammit exactly in line with the extra valve area. You can check both calculations for yourself obviously. So clearly the Cd was staying the same at both low and high lifts and just suffering somewhat in the middle where from 150 to 250 thou lift it stayed much the same.
I normally target a high lift cD of 0.6 on both 4v and 2v heads of similar downdraft angle and find both types of head achieve that in similar fashion.
Edited by Pumaracing on Sunday 12th February 09:26
ivanhoew said:
i applaud DV on his amazing levels of patience !
If I didn't give him a hard time when he gets something wrong he'd start to think I was going soft on him and worry that I'd had an aneurism or summat. You might not understand it but he respects the fact that I tell it like it is warts and all so we can both improve our knowledge much more than having yet another yes man just accepting everything as gospel and no one being any the wiser.In fact on that note you'd better not include that dodgy graph of effective low lift flow area with 30 and 45 degree valve seats in your new cylinder book DV or it'll be time for another reaming out. I sent you the correct figures and they'd better be in there!
ivanhoew said:
not to mention the good cop/bad cop/troll performance brings in more attention for his lectures ? hope your getting your cut db ? ;-)
Oh dear, and there I was hoping you might contribute sensibly to all this. I don't give a flying f**k about his lectures and think it would actually be best if I wasn't at any of them in case I kept interrupting and saying something was wrong to the consternation of the assembled masses.Please don't try and guess what relationship DV and myself have when you don't know either of us in the slightest detail.
I've reported their posts and if all they want to do is try and derail this thread maybe the mods will delete their posts so people who actually want to learn something can plough through the thread more easily. I can't see how two pages of stupid pictures and a "whoosh" are helping this.
ivanhoew said:
and i assumed from db's comments to others on other threads, particularly that poor american chap
You mean Stan Weiss? He's a friend of DV and I've known him from other online forums for several years. In fact it's a blessing he's still with us after the front porch collapsed when he was a toddler playing on it but luckily only the three hound dogs sleeping under it were killed. His parents were so proud when his first words as a baby were "Attention K-Mart shoppers". Recently when friends of his moved into a new house he was the first to volunteer to help them remove the wheels but it's a scurrilous rumour that he thinks a Volvo is part of the female anatomy.Stan Weiss said:
Pumaracing said:
ivanhoew said:
and i assumed from db's comments to others on other threads, particularly that poor american chap
You mean Stan Weiss? He's a friend of DV and I've known him from other online forums for several years. In fact it's a blessing he's still with us after the front porch collapsed when he was a toddler playing on it but luckily only the three hound dogs sleeping under it were killed. His parents were so proud when his first words as a baby were "Attention K-Mart shoppers". Recently when friends of his moved into a new house he was the first to volunteer to help them remove the wheels but it's a scurrilous rumour that he thinks a Volvo is part of the female anatomy.I now see how people believe that you have missed you true calling. You are a master at the art of creating a great fictional tale.
Stan
PS - you taken the "wide load" sign off the back of your house yet?
Gassing Station | Engines & Drivetrain | Top of Page | What's New | My Stuff