Just to cover the basics real quick:

Just to cover the basics real quick:
Duration - The number of crankshaft degrees (360 being a complete cycle - a= ll the degrees) that the valve is open. When they say "Duration @ 0.05" of lift" that just means that the valve is open 0.05" or greater for xxx degrees of crankshaft rotation. Typical values are 0.05", 0.2" and 0.008 or 0.002 or 0.005 (all a= re sometimes called "advertised" duration, or seat to seat. Obviously the greater the lift number the smaller the number of degrees

<= o:p>

look at the graph ab= ove - this isn't a real cam profile, but approximates one. Duration @ lift would be like picking the lift you want duration numbers for (on the y axis), finding whe= re it intersects with the plot of valve motion, then looking on the x axis to = find the number of degrees between the two.

Now that's the b= asics. Now we get to the more complicated parts. First we need to remember that th= is is a 4 cycle engine - let's follow the valve events starting with the intake stroke.

The intake valve= is open/opening as the piston is moving away from the head. This motion create= s a vacuum which causes atmospheric pressure to push in the air. Now the earlier the intake valve opens the more air you can get in, but if you open it to e= arly then the intake charge will rush out the exhaust (since the exhaust valve is closing as the intake valve opens - look at the second graph on my excel document). If you open the intake valve too late then y= ou don't get the benefit of scavenging from the exhaust (the outflow through the exhaust valve helping to pull in intake charge) - so as you can= see it's a matter of balancing. Now we reference everything vs. crankshaft degr= ees, but in real life flow depends on the amount of time the stream has to flow a well. What amount of overlap might give enough time to flow (how long say 40 degrees of overlap was) at say 2000 rpm would obviously be not enough time = at 4000 rpm - since we would have half the time since 1 revolution happens twi= ce as fast.

The intake valve= will close and the piston will begin to come back up - this is the compression stroke - here obviously no valves are open or everything would be forced out them. The piston moves back up towards top dead center and some amount of degrees before it gets there (your timing advance) the spark plug fires. Pressure builds up and pushes the piston back down - this is the power stro= ke, and no valves are open for most of this since you want as much force concentrated on pushing the piston down and not expanding into the intake or header.

Next the exhaust= valve opens and the piston moves back up - this is the exhaust stroke. Here the piston pushes the spent mixture out the header. Some time before TDC (top d= ead center - piston all the way up) the intake valve will open - while the exha= ust valve is open - to help "Scavenge" exhaust gasses. As was pointed= out above you need a certain amount of flow - but this flow will depend not onl= y on degrees of overlap, but on how long each actual degree represents (rpm). Th= is is why different amounts of overlap are appropriate for different rpm range= s. Thus the cycle begins over again.

Okay, most of you probably already knew that - so what = was the point:

Just this - when= we talk about duration, lsa, etc. we are really ta= lking about byproducts of the valve events (opening, closing, etc.) - as well as a measure of how "steep" the lobe is (how quickly it opens and clos= es). Don't get me wrong, these are handy references, but they really only tell a small part of the story - you could have 2 cams with the same duration @ 0.05", lsa, and lift, but that had radical= ly different effects on powerband, drivability, et= c. Keeping this in mind when I talk about duration, lsa, etc. I am going to talk about them in ref= erence to their effect on valve events.

First a word abo= ut cylinder pressure - cylinder pressure is directly related to torque - which as we kn= ow gives us hp @ rpm. So the obvious thing to do is try and maximize cylinder pressure across our desired rpm range. (you can get a measure of peak cylinder pressure using = a compression tester) - it is a function of static compression= as well as cam timing (dynamic compression)

Cylinder pressur= e I primarily determined (as read by a compression tester) by when the intake v= alve closes - if it closes later than the compression stroke may actually push s= ome of the gasses back out the intake valve, and so you have less "stuff&q= uot; in the chamber, so the compression is less. If it closes earlier then you l= oose less, so you begin with more, and so the final compression is more. Now als= o realize if you close the intake valve to early at high rpm's you will not have enou= gh time to completely fill the intake - conversely if you close it to late the= n at low rpm' you will loose allot of your intake charge and be down on power. <= o:p>

This is where du= ration comes in - duration is basically how long the valve is open - duration in a= nd of itself is good - you want as much duration as you can to fill the intake= as well as expel the exhaust gases. Where you start running into problems is t= he example above - when the duration gets so big (exhaust closing) that you hu= rt your low rpm drivability. Now you can also increase duration by opening the intake valve sooner - this will keep your dynamic compression up, and give = you more time to fill the cylinder. But remember, the exhaust valve is closing = as the intake opens, so if you open it up to soon you will blow most of your intake charge out the exhaust - the lower the rpm the more pronounced the l= oss. More duration on the exhaust valve (closing it later) will also have the sa= me effect on overlap, and closing it to late will cause you to loose allot of = your intake charge. Closing it too early will cause you to loose the benefit of exhaust scavenging.
Opening the intake earlier gives you more duration, but doing this also saps power away from the power stroke - so you really can't open it much earlier= - esp. at lower rpm' - without hurting power production.

Now as you can s= ee just saying the duration is increased doesn't tell you how - it could be increased with an earlier opening or later closing - both having different effects. This is where lsa comes in - lsa is the number of degrees between peak lift on the intake and peak lift on the exhaust.

<= o:p>

Look at the graph above and find th= e 2 peaks. The number of degrees between them is the LSA.

LSA can confuse some people because they don't think about valve events. Narrowing the LSA will cause more overlap - just like increased duration ca= n, but instead of lowering dynamic compression it will increase it! When you narrow the lsa you are bringing the 2 lobes more inline, so your intake valve is actually closing sooner - while your exhaust valve opens later. So the intake closing - which determines dynamic compres= sion - increases it, while the exhaust closing, which effects overlap (and idle/charge quality) increases - giving more overlap and a rougher idle.

Peak lift really doesn't mean a whole lot - except as far as what kind of spring life you ge= t. It is an indication (only an indication though) of how "intense" = - or steep the lobes are. A steeper lobe can still give you the valve events you want while then opening faster to allow for airflow in non-critical areas (where more is always better) - this is why solid rollers make more power t= han hydraulic rollers.
Also, just because say your heads don't flow much more at a certain lift doesn't mean it's worthless to have a cam with that lift. This is called overlifting your heads, and it works because - event = though ay .6" of lift really doesn't help you much, the fact that your cam has that much lift means you spend a greater portion of time at higher lifts ac= ross the board - so overall you still have more flow (integral of flow vs. lift = as a function of crankshaft degrees). As long as your heads are truly stalling a= nd killing airflow you are good.
Lift is free - it will basically increase your torque across the board - the drawback is that it is harder on valvesprings. = Up to 0.6" of lift and you can get good life, after that things become iffy.