47.5 rule for flat tappet camshafts

[HeinzE]

hoffman900,

Fascinating post on the Yamaha cams. Who generated that data for you? I have the .S96 files for my lobes as well as two unused cams. I would be very interested in having someone look at these and get some idea of whether or not there is any way to improve on them. Any suggestions?

HeinzE

[Stan Weiss]

hoffman900,

Fascinating post on the Yamaha cams. Who generated that data for you? I have the .S96 files for my lobes as well as two unused cams. I would be very interested in having someone look at these and get some idea of whether or not there is any way to improve on them. Any suggestions?

HeinzE

Since this is based on much simpler calculation than Jon / Bob's results it will be a little less accurate. But were done from data taken from a S96 file. If you want to email me you S96 files I would be happy to produce and sent to you the results.

Stan Weiss <srweiss@erols.com>

Stan

C:\Program_Files\Microsoft_Visual_Studio\VB98\CARFOR\cmm\COMP161.CMM
Smooth_=_NO

---_Cam
_______I_N_T_A_K_E
_________Cam_Lift________Velocity______Acceleration_______Jerk
__1___0.00000000000
__2___0.00051181000___0.00025590500___0.00012795250___0.00006397625
__3___0.00053325000___0.00001072000__-0.00012259250__-0.00012527250
__4___0.00059787000___0.00003231000___0.00001079500___0.00006669375
__5___0.00070866000___0.00005539500___0.00001154250___0.00000037375
__6___0.00086771000___0.00007952500___0.00001206500___0.00000026125
__7___0.00107008000___0.00010118500___0.00001083000__-0.00000061750
__8___0.00131437000___0.00012214500___0.00001048000__-0.00000017500
__9___0.00160577000___0.00014570000___0.00001177750___0.00000064875
_10___0.00196850000___0.00018136500___0.00001783250___0.00000302750
_11___0.00243930000___0.00023540000___0.00002701750___0.00000459250
_12___0.00304889000___0.00030479500___0.00003469750___0.00000384000
_13___0.00381890000___0.00038500500___0.00004010500___0.00000270375
_14___0.00475451000___0.00046780500___0.00004140000___0.00000064750
_15___0.00584871000___0.00054710000___0.00003964750__-0.00000087625
_16___0.00710212000___0.00062670500___0.00003980250___0.00000007750
_17___0.00854331000___0.00072059500___0.00004694500___0.00000357125
_18___0.01020559000___0.00083114000___0.00005527250___0.00000416375
_19___0.01211464000___0.00095452500___0.00006169250___0.00000321000
_20___0.01429134000___0.00108835000___0.00006691250___0.00000261000
_21___0.01673228000___0.00122047000___0.00006606000__-0.00000042625
_22___0.01942826000___0.00134799000___0.00006376000__-0.00000115000

---------------_Valve

C:\Program_Files\Microsoft_Visual_Studio\VB98\CARFOR\cmm\COMP161.CMM
RA_Ratio_=_1.65___Lash_=_0.0210__RPM_=_6000__Smooth_=_NO

_______I_N_T_A_K_E
_______Valve_Lift________Velocity______Acceleration_______Jerk________Velocity_fps
__1___0.00000000000
__2___0.00000000000___0.00000000000___0.00000000000___0.00000000000___0.00000000000
__3___0.00000000000___0.00000000000___0.00000000000___0.00000000000___0.00000000000
__4___0.00000000000___0.00000000000___0.00000000000___0.00000000000___0.00000000000
__5___0.00000000000___0.00000000000___0.00000000000___0.00000000000___0.00000000000
__6___0.00000000000___0.00000000000___0.00000000000___0.00000000000___0.00000000000
__7___0.00000000000___0.00000000000___0.00000000000___0.00000000000___0.00000000000
__8___0.00000000000___0.00000000000___0.00000000000___0.00000000000___0.00000000000
__9___0.00000000000___0.00000000000___0.00000000000___0.00000000000___0.00000000000
_10___0.00000000000___0.00000000000___0.00000000000___0.00000000000___0.00000000000
_11___0.00000000000___0.00000000000___0.00000000000___0.00000000000___0.00000000000
_12___0.00000000000___0.00000000000___0.00000000000___0.00000000000___0.00000000000
_13___0.00000000000___0.00000000000___0.00000000000___0.00000000000___0.00000000000
_14___0.00000000000___0.00000000000___0.00000000000___0.00000000000___0.00000000000
_15___0.00000000000___0.00000000000___0.00000000000___0.00000000000___0.00000000000
_16___0.00000000000___0.00000000000___0.00000000000___0.00000000000___0.00000000000
_17___0.00000000000___0.00000000000___0.00000000000___0.00000000000___0.00000000000
_18___0.00000000000___0.00000000000___0.00000000000___0.00000000000___0.00000000000
_19___0.00000000000___0.00000000000___0.00000000000___0.00000000000___0.00000000000
_20___0.00258071100___0.00129035550___0.00064517775___0.00032258887___3.87106650000
_21___0.00660826200___0.00201377550___0.00036171000__-0.00014173387___6.04132650000
_22___0.01105662900___0.00222418350___0.00010520400__-0.00012825300___6.67255050000
_23___0.01592604300___0.00243470700___0.00010526175___0.00000002888___7.30412100000
_24___0.02122609050___0.00265002375___0.00010765838___0.00000119831___7.95007125000
_25___0.02701386150___0.00289388550___0.00012193087___0.00000713625___8.68165650000
_26___0.03334966350___0.00316790100___0.00013700775___0.00000753844___9.50370300000
_27___0.04025153100___0.00345093375___0.00014151637___0.00000225431__10.35280125000
_28___0.04771332600___0.00373089750___0.00013998188__-0.00000076725__11.19269250000
_29___0.05571099150___0.00399883275___0.00013396763__-0.00000300712__11.99649825000
_30___0.06422385300___0.00425643075___0.00012879900__-0.00000258431__12.76929225000
_31___0.07327291500___0.00452453100___0.00013405012___0.00000262556__13.57359300000
_32___0.08288931300___0.00480819900___0.00014183400___0.00000389194__14.42459700000
_33___0.09307694100___0.00509381400___0.00014280750___0.00000048675__15.28144200000
_34___0.10382801100___0.00537553500___0.00014086050__-0.00000097350__16.12660500000
_35___0.11509902900___0.00563550900___0.00012998700__-0.00000543675__16.90652700000
_36___0.12685039500___0.00587568300___0.00012008700__-0.00000495000__17.62704900000
_37___0.13904907600___0.00609934050___0.00011182875__-0.00000412912__18.29802150000
_38___0.15168352200___0.00631722300___0.00010894125__-0.00000144375__18.95166900000
_39___0.16472251500___0.00651949650___0.00010113675__-0.00000390225__19.55848950000
_40___0.17810434500___0.00669091500___0.00008570925__-0.00000771375__20.07274500000
_41___0.19175816100___0.00682690800___0.00006799650__-0.00000885638__20.48072400000

===============

_______I__N__T__A__K__E
Rocker_Arm_Ratio_=_1.650_________Valve_Lash_=_0.0210_

VALVE_____Lift______Opens___Closes__Duration
_________________Deg_BTDC__Deg_ABDC_____________Area
_________0.00000____47.69_|__84.37_|_312.05_|__61.01
_________0.00600____44.55_|__80.80_|_305.35_|__60.99
_________0.01000____42.73_|__78.67_|_301.39_|__60.98
_________0.02000____38.71_|__73.99_|_292.70_|__60.91
_________0.04000____32.32_|__66.84_|_279.16_|__60.69
_________0.05000____29.68_|__63.94_|_273.61_|__60.56
_________0.10000____18.96_|__52.59_|_251.56_|__59.70
_________0.15000____10.52_|__43.84_|_234.35_|__58.68
_________0.20000_____3.06_|__36.19_|_219.26_|__57.27
_________0.25000____-3.97_|__29.07_|_205.10_|__55.41
_________0.30000___-10.85_|__22.20_|_191.36_|__53.71
_________0.35000___-17.79_|__15.20_|_177.41_|__51.04
_________0.40000___-24.96_|___7.98_|_163.03_|__48.74
_________0.45000___-32.52_|___0.38_|_147.86_|__45.30
_________0.50000___-40.80_|__-7.88_|_131.31_|__41.46
_________0.55000___-50.19_|_-17.26_|_112.55_|__36.16
_________0.60000___-61.45_|_-28.50_|__90.05_|__29.76
_________0.65000___-76.72_|_-43.55_|__59.73_|__20.32
CAM
_________0.00600____56.01_|__92.95_|_328.96_|__39.04
_________0.01000____50.50_|__87.43_|_317.93_|__39.00
_________0.02000____41.86_|__77.65_|_299.51_|__38.83
_________0.04000____30.98_|__65.35_|_276.33_|__38.48
_________0.05000____26.89_|__60.95_|_267.84_|__38.29
_________0.10000____11.47_|__44.82_|_236.29_|__37.07
_________0.15000____-0.70_|__32.36_|_211.67_|__35.53
_________0.20000___-12.08_|__20.95_|_188.87_|__33.37
_________0.25000___-23.72_|___9.23_|_165.51_|__30.83
_________0.30000___-36.39_|__-3.48_|_140.13_|__27.21
_________0.35000___-51.53_|_-18.59_|_109.88_|__22.27
_________0.40000___-72.76_|_-39.65_|__67.59_|__14.31

[hoffman900]

Heinz (Karl),

Take up Stan's offer!

With a flat tappet, the first thing I would look for would be velocity and how much are you utilizing what you have?

Harold and Mike have both outlined here the velocity for a given flat tappet diameter. If you really trust the machining of the lifter bores and the cam's stiffness, you could run closer to the edge of the lifter, thus giving you more effective lifter diameter and thus increase the velocity. If you look through Comps lobe catalog, you'll see lobe families that require 'true machined lobes'. That's what's going on there.

Acceleration is likely to be variable and likely contingent on what your valvetrain can withstand (ie: weight, valve spring limitations, how long / flexible are the pushrods). Mike could give more information on that.

Mike Jones generated it for me with Audie's software. I have two more cams to compare in the near future. I'm really curious about the stock cam.

Mike did point out that the spike in acceleration at intake valve closing (as compared to just the lobe) is caused by the base circle being a bit off for this valvetrain - likely the lobe design was designed for another application. That's the type of information I am hoping to uncover with this exercise as well. For other's edification, this is a SOHC engine with rocker followers.

Here is an old Harold quote:

With a valve spring bucket 1.38" in diameter, you divide by 2(.690") and subtract your clearance from the edge(.690"-.020"=.670"), then divide by 57.3. Your answer should be .011693"/* max velocity. A higher design velocity gets closer to the edge of the spring bucket----How brave are you?
Finding out this number is most easily done using various cam analysis software, such as CamDoctor, etc. Trying to measure it with a degree wheel is nearly impossible. Most cam designers should be able to give you this number on any given design---For instance, all my current solid flat tappets for the SBC/BBC .842" diameter lifter are at .007041"/*. My old .960" diameter mushroom designs were around .00807"/*. Most of my roller designs are in the .0085"/*-.0095"/* range.
Velocity numbers are only important in flat tappet solid and hydraulic designs, acceleration numbers are important in roller designs.
The highest lobe lift I have designed a DOHC bucket follower for has been .575", and yes, it did not run off the tappet.

Mike:

Flat tappet cams are limited to a maximum velocity you can move the lifter. For a .842" diameter, the max velocity is just over .007" per degree. If you go over this, the lobe is no longer riding on the face of the lifter, and catches the edge of the lifter, and destroys everything.
Roller cams are not limited to a max velocity. This allows you to gain more area in the same duration, or the same area with less duration. In an apples to apples test, the roller will almost always beat the flat tappet.
NASCAR gets around the max velocity limits of the flat tappets by using very high rocker ratios. The old dodge NASCAR engines ran a 2.37:1 rocker, so the .0075" velocity at the lifter turn into a valve velocity of .017775", and that valve velocity would be about all you could get away with in a roller cam.

And the bunch I shared earlier in the thread. Looking at Harold's above numbers, he's using ~.017" to the edge of the lifter.

I remember Harold and Don Teweles collaborating in the 70's at General Kinetics Cams on quite a few really good flat tappet cam profiles ... regular and mushroom. Some great running OHC profiles for slider followers, too.

A championship with Grumpy Bill in NHRA Pro Stock, a championship with Shirl Greer in NHRA Funny Car, and Benny Parsons' 1975 Daytona 500 win during that time as well.

[novadude]

How does the "47.5" formula even make sense when it looks at VALVE lift and ignores rocker ratio? By that math, almost every BBC cam with published valve lift based on a 1.7 rocker ratio would "fail". Consider the 214/218, .461/.480 350hp 396 cam. 214/461 = 46.4%. Those cams went 100k miles or more in stock applications.

[hoffman900]

A quick lookup of the Aermacchi 350 shows a 10mm dia mushroom tappet, with 12mm ones being fitted as well.

That works out this way, using 0.018" to the edge of the lifter:

10mm = 0.393701"

((0.393701/2)-0.018)/57.3= 0.0031213"/*

12mm = 0.472441

((0.472441/2)-0.018)/57.3 = 0.003808"/*

or think of it this way, the 12mm tappet would allow 18% more velocity! That's like going from a .842" Chevy lifter to a 1.000" mushroom lifter with a lobe designed to match.

Additionally, here is a little exercise I did using the mean-square error between multiplying lobe lift * rocker ratio (minus lash) vs the measured valve lift. This is more a theoretical vs actual display. The red line being the theoretical. You can see how the measured valve lift deviates. This highlights the shortcomings of just simply mutliplying lobe lift by the rocker ratio (minus lash) to come up with the valve lift curve.
Intake:

Exhaust:


edit: I'll crop these and reload tomorrow. I also meant ‘true machined lifter bores’ in regards to Comps catalog.

[HeinzE]

hoffman900

The 10mm and 12mm tappet dimension you refer to are not the cam contact surface diameter but the shank of the lifter, which ride in bushes pressed into the case. The lifters are a nail head style with the cam acting on a flat pad with a diameter of .770". If you take a look at the Dyno Cams web site you'll see the style lifter I'm describing. The 10mm shank lifters would often break right at the joint of the shank and cam rubbing face when performance cams were intalled. The 12mm shank makes for less overhang between the shank and face. Both my engines use the 12mm shank lifters, but the contact pad is .770" on both.

HeinzE

[hoffman900]

Ah, even better then.

So with edge distance of 0.018”, you’re looking at a maximum velocity of 0.00640”/*.

[Stan Weiss]

Karl,
I know you do not want me to post the lift data. But I think showing these numbers will help the discussion.

Peak Intake velocity opening side = 0.00592446

Peak Intake acceleration opening side = 0.0002463

acceleration @ peak lifter raise = -0.0001495


Stan

[HeinzE]

Stan,

No worries. Feel free to post any information you feel will be usefull.

HeinzE

[HeinzE]

hoffman900,

A small correction to the lifter pad diameter I gave earlier. The original Aermacchi dimension for the tappet face diameter is 20mm or .7874". However, production variables being what they were, some lifters were completely flat out to the edge while others had a small bevel machined at the periphery. This bevel reduced the effective area down to the .770" diameter. The .770 diameter is what was used for the cam I have. There is probably enough room for a .800" tappet diameter or close to it. Would this make a enough of a difference to pursue?

HeinzE

[hoffman900]

Karl,

Going through the math.

Given:
Peak Intake velocity opening side: 0.00592446"/*
Tappet Diameter: 0.770"

Need:
Edge Distance: x

Equation:
x = -1((Velocity*57.3)-(Tappet Diameter/2))

Solution:
x = -1((.00592446 * 57.3)-(0.770/2))
x = .045528"

Effective Tappet Diameter:
.770 - (x*2) = 0.680"

I'd say you're giving up a lot.

Let's try this with 0.017" edge distance.

=((.770/2)-0.017)/57.3
= .006422"/* or... ~8%

Mike's comment on lobe chamfer:

Almost every Flat Tappet lifter comes from the manufacturer with a chamfer on the edge. It's to keep them from chipping.
On the NASCAR Cup engines, we would have the lifters made with no chamfer, so we could design the profile to use every bit of the .875" face.

That's the reason for having no chamfer, but if the cam profile isn't designed to go all the way out to the edge, it doesn't matter.
If I design a profile to go out to a max of .4175" from the center of the lifter, that's .020" from the edge of an .875" lifter(.875"diameter=.4375" radius), so if it has a .010" chamfer or not, the lobe will never see it.
If I design the profile to go to a max of .435" from the center of the lifter, it better not have a .010" radius.

So let's look at an unchamfered tappet:

=((.7874/2)-.017)/57.3
= 0.006574"/* or ~ 2% greater than .017" edge distance, but it's ~10% greater than what you have. You could also look at this as a chamfered .800 tappet.

Looking at a .800, with no chamfer:
=((.800/2)-.017)/57.3
= 0.006684"/* or ~11% greater than what you have.

Going from what you have to a a chamfered .800" tappet, it's just a bit less than going from a .842" Chevy tappet to a .960" vintage NASCAR mushroom tappet, but that's a pretty big difference. How much do you trust the location of the lifter bores in relation to the cam?

Keep in mind this is one facet of cam design though. The real art of the math is in the shape of the curves. Through improved equations, computers, and manufacturing ability, cam designers have been able to reduce peak acceleration and jerk, while maximizing the velocity of a given tappet and keeping or gaining lobe area.

It's really a question of how crazy do you want to go?

Peak velocity is easy, but what can your valvetrain withstand in terms of acceleration? As you know, what is happening at the lobe isn't always the same as what's happening at the valve. Take a look at the numbers I posted and you can see that for yourself. To quote Mike, "The trick is to get the area under the curve without a high rate of acceleration".

A good cam designer doesn't make enough for how talented they have to be.

Some Mike and Harold quotes:

Harold:

I started to give you some of my opening and closing rates at valve contact, then I realized I might be revealing too much, although I think I've mentioned seating rates for several years.
However, in the thread on 565s, I have a partial matrix of cams in terms of .020, .050. .200 durations and various lobe lifts. I did not put iin there all the cams in that series, but just the ones of interest for under-10 sec 565s.
ALL of those cams use the identical opening ramp, and the identical closing ramp. They only vary in duration, and lobe lift. Naturally, the higher the lobe lift, the higher the negative acceleration, and the bigger the duration, the lower the negative acceleration.
The PEAK acceleration rates on those cams is .0002975"/*2 on the opening side, and .000295"/*2 on the closing side. I have had cams with SLIGHTLY higher rates win class at the 24-Hours of Daytona, etc.
I use these ramps on my daily-driven street cars, like 540s, where I use a 257/265 at .050 with .725"/.725" gross valve lift. I have never had anyone report a broken valve spring with that cam, and I have been making it since the spring of 2006.
This is not my most aggressive cam, not by a long shot. However, it is a very successful series of bracket/race cams, with lobe lifts over .540".
My designs are all multi-segmented polynomials, which is similar to a B-Spline. I have done this way since April 1980.

I have been designing race-winning cams since December 1972, and doing it full-time since August 1974. I have been General Kinetics, Competion Cams, UltraDyne's, Lunati's, and Custom Camshaft Company's cam designer.
The two main things in determining maximum RPM are the negative acceleration on the nose(valve float or valve loft), and the valve seating velocity(valve bounce). None of this is related EXACTLY to power, but plays a part in it. Obviously, there has to be enough area-under-the-curve to make power at a given RPM.
This is why cams that are dynamically smooth enough to go 10,000 RPM may not do it; they just may not make enough horsepower.
In 2007 I designed a Pro-Stock cam that would turn 10,500 for over 90 seconds with no bounce or float. Why wouldn't it go any higher?
To get the required area-under-the-curve, the cam required high positive and negative acceleration rates(1.9:1 rocker ratio, .575" lobe lift), and therefore it also required very high spring pressures.
The Spintron's electric motor did not have enough HP to turn the valve train any higher. The cam was dynamically smooth and stable at 10,500.
And it runs OK, too. Just ask Alan Johnson........
Cams that have pointy noses have more negative acceleration than cams with more rounded noses. They require more spring pressure for the same RPM, and they are in Catch 22---A pointy nose is more likely to wipe out than a more rounded one, and the higher required spring pressure makes it easier to wipe out......

These are just some thoughts and observations of an old cam desiginer....

There are people in the cam design world that charge several thousand dollars for that information.
However, here it is a la UDHarold:
Almost all negative acceleration in hydraulics and solid flat tappets cause too small a radius of curvature on the nose if the negative accel rate exceeds -.000220, I do not go even that high. The open spring pressures necessary to control that rate of negative acceleration, and with that small a radius of curvature, cause almost sure wipe-out of the cam. Too much pressure concentrated in too small an area.
Roller cams are relatively OK on Negative acceleration, it just requires more open pressure, especially when the valve lift goes over 1.000". Eventually cam materials come into play, particularly when you have 1400 lbs at 1.000" valve lift, and a 1.9:1 rocker.
Flat tappet cams, both hydraulic and solid, run very high positive acceleration rates( .000400-.000500"/*2) compared to rollers. However. they are limited by max velocity before the contact point of lobe and lifter doesn't exist on the lifter anymore.
One of the signs of a good cam designer is how well you live with this limitation. I keep all my designs .017" away from the edge for over 28 years, and have had some go 100,000 miles on the street. An engine block with out-of-position lifter bores may wipe this clearance out, and therefore wipe out every cam put in it.
Roller cams, both solid and hydraulic, are controlled by positive acceleration rates, as well as base circle radius. Any acceleration rate over .000340"/*2 can cause negative raduii of curvature---The dreaded 'Inverse-Radius' cam, unless special math methods are used.
All of my 28* Major Intensity cams use such techniques.
Any positive acceleration rate under .000250"/*2 is almost a waste of time, although valve train life is tremendous.
An acceleration rate around .000300"/*2 is good for 24-hour racing, if the valve seating velocities are done right.
As is obvious, opening side rates and closing side rates may not agree in positive accelerations or opening/closing velocities, but the nose acceleration must always be the same, or there is a visible line, as well as unequal velocities, at the nose.
It took me 38 years of work to learn all this.
Good Luck!!!

I do not know if I should, or even want to, give a proper lesson on ramp design. After all, there may be other cam designers lurking about, wanting to learn something.....
However, your question is not about valve lash, but about ramp design. You can design ramps with wide valve lash--.026" to .036", that open valve SLOW. Those cams do not work very well. There are mathematical limits to how fast you can open a tight-lash (.012"-.016") cam. Trying to exceed this limits causes valve train problems.
Of course I have opened valves fast and shut them slow for 28 years now, since 1977. Anyone who doesn't know this hasn't been paying attention, I have been advertising it for this long.
I have never heard of any valve train problems with tight-lash cams, as far as heat-over-expansion of the valve train, etc. Valves that are too loose, such as .036" to .045", can cause battering of the valves and seats, short spring life, etc. I have had excellent valve train lift with ramps .0173" high--.026" for 1.5s, .030" for 1.7s. These ramps have been the ones that have gone 100,000 miles, so they aren't too bad. Tightening them down slightly, as tight as .018", have only helped valve train life.
Tightening them down increases seat duration, which may add a little top-end, but certainly takes away bottom-end. Tighter valve lash may improve TIRE life, though!
I have ran solids/solid rollers on hydraulics/hydraulic rollers, and as long as the valve lash is really tight, say .004" to .008", I have seen no trouble. All hydraulic cams start lifting off the seat by .004" ramp height, regardless of advertised duration heights, and running much more valve lash can put the point of rocker/valve contact deep in the cam acceleration curve, which isn't good. Kinda like a sledgehammer blow to the valve. You can also run hydraulics on solids, but they had better be tight-lash solids, or else you will have veeeery long hydraulic seat durations!!!
There are several reasons for what ramp height and velocities are chosen, but this may be getting into secret territory. Part of the Great Cam Design Mystery is learning how to increase velocity at lower and lower ramp heights. The reason for lower and lower ramp heights? It gives you longer and longer times to accelerate the valve....

Mike:

Normally, Max acceleration occures before Max velocity.

Here are some examples of distance from lash point to max accel, and max Vel on different cams.

Comp Cam, Pro Cup roller: Max Accel=14 deg, Max Vel=46 deg
Crane cam, 410 sprint roller: Max Accel=30 deg, Max Vel=70 deg
Schrick Cam, BMW overhead bucket: Max Accel=10 deg, Max Vel=48 deg
LSM Cam, Pro Mod roller: Max Accel=32 deg, Max Vel=78 deg
GM 604 crate cam, Hyd roller, Max Accel=27 deg, Max Vel=53 deg
Jones Cam, "Banjo"'s roller: Max Accel=17 deg, Max Vel=59 deg

I just designed a roller cam that gets to max velocity in about 10 degrees, but it's only designed to turn 2,000rpm

BTW, you can have a higher velocity, without increasing the max acceleration. You accelerate to max acceleration, and then decrease the acceleration rate toward max velocity, then accerate. If you change the shape of the acceleration curve from max acceleration to max velocity, you will get a higher max velocity with the same max acceleration.

Here is a more modern .904" design.
It's designed for a class that has to run stock rockers, and stock size springs(130# on seat), and they turn about 6,800rpm.
It's a .325" lobe lift, and the lobe runs to within .008" of the edge on the .904" lifters
260 @ .020"
230 @ .050"
199 @ .100"
146 @ .200"
68 @ .300"

[hoffman900]

Additionally, a great article from one of my favorite tech writers.

TIOC May 2005
Eaten Alive by Parasitic Oscillations
by Kevin Cameron

As a new hi-fi amplifier is designed by an electronic engineer,
particular care must be taken to be sure that unforeseen combinations
of resistance, inductance, and capacitance do not permit so-called parasitic
oscillations; to build up, destroying the music.

Each property of electrical circuits has a mechanical analog, and one
mechanical system that is often rife with its own parasitic
oscillations is the valve train. Rocker arms bend, pushrods compress
and expand, camshafts deflect between their bearings, and valve stems
flop from side to side while valve heads deflect like trampolines or
floppy disks. All of this is invisible to us. Its symptoms are valve
seat recession and valve spring and valve breakage. In many cases,
after tests with a new cam that gives really good power, we have to
reluctantly back up to a previous, less powerful set-up because we
can't afford the DNFs and breakages that the hot set-up
produces. Valve train failures are hit-and-miss, trial-and-error, a
mystery. Lighter valve train parts and stronger springs sometimes just
seem to make everything worse. Is there any truth?

Back in the 1920s Percy Goodman decided it was time to put aside
clattering pushrods and adopt trendy OHC valve drive on the TT
Velocettes he manufactured. Soon he was driven crazy by erratic valve
motion and sensibly resorted to the use of a strobe light to reveal
what was happening. But naming the illness is not the same as a cure.

When I was recently at the NHRA drag nationals in Houston, I had the
opportunity of conversation with Byron Hines, whose purpose-designed
giant 160-cubic-inch V-twin engine has recently begun to win Pro Stock
races. I asked what had made the difference after some uncompetitive
seasons.
The Spintron, was his answer. It showed us that our valve
motion was nothing like what was in the cam profiles.

When I was recently at the NHRA drag nationals in Houston, I had the
opportunity of conversation with Byron Hines, whose purpose-designed
giant 160-cubic-inch V-twin engine has recently begun to win Pro Stock
races. I asked what had made the difference after some uncompetitive
seasons.
The Spintron, was his answer. It showed us that our valve
motion was nothing like what was in the cam profiles.

Spintron use a big electric motor to spin your engine while a variety
of instrumentation is used to measure the actual trajectories of the
parts you are interested in. This is different from using a
strobelight in that the information you get is detailed enough to
allow the flexing of each part to be isolated and understood.

Byron went on to say that valve train dynamics are particularly
difficult to control in big twins because of the large variations in
crank speed as each cylinder fires. The cam profiles were originally
developed to work at a particular maximum rpm, provided that the
camshafts turn smoothly. They do not turn smoothly because the crank
does not turn smoothly; it advances in a series of fairly violent
jerks. 80 percent of the recoverable energy in the hot combustion gas
does its work between 10-deg ATDC and 80 ATDC. Out of the 720 crank
degrees in the engine cycle, power is given during only 10%. This
means that the instantaneous speed of the cam can often be much higher
than its average speed. This also means that as a cylinder fires and
the crank accelerates suddenly, an open valve in the other cylinder
may be tossed right off its cam profile, or dropped prematurely onto
its seat.

This reveals why tuners of singles and twins found that their bikes
top-ended better and faster the heavier their cranks were made. A
younger generation of tuners has rejected this idea as turn-of-the-century
dirt-track nonsense, reasoning that physics requires lighter flywheels to
result in faster acceleration.
Vintage racer Todd Henning learned the heavy crank truth
in back-to-back testing of his highly tuned Honda twins, as did Rob
Muzzy in 1981-83 with 1025-cc Kawasaki in-line fours. The heavier the
crank, the smoother its rotation becomes, and the less power stroke
disturbance, or is transmitted to the cams. Where
does the lost power go when a light crank is used? Erratic valve
motion is one answer, and big valve bounce after closing is another.

Power pulsing is not the only disturbance to the valve train. Consider
a parallel twin, a flat twin, or an in-line four. In all of these, all
the pistons stop simultaneously. This means that as pistons decelerate
to a stop, the crank must accelerate rapidly because there is nowhere
else for the pistons' energy to go ; its conservation of
energy. Peak piston speed is about 1.5 times mean piston speed. This
means that all the kinetic energy in the pistons, moving at near 100
feet per second, is suddenly dumped into the crank. This is especially
bad for in-line fours, which have small-diameter cranks and little in
the way of flywheel mass. This sudden crank deceleration/acceleration
cycle is performed twice per revolution. No wonder new engine
development usually involves coping with cam drive breakage. No wonder
there are mystery failures of valve train parts.

Allan Lockheed is the man behind the engine design software Engine
Expert, and he talks to engine people all over the world. He has
tales of engines whose valves were stable with a chain or belt cam
drive, but which mysteriously began to break valve springs as soon as
a much more rigid gear drive was put in its place. The slight
flexibility of chain or belt took the sharp edge off the sudden crank
speed variations, preventing high frequency motions from reaching the
valve train. This may be why Yamaha retain a chain cam drive in their
M1 in-line four MotoGP race engine. The oil film between each of the
cam chain's pins, bushings, and rollers can be thought of as a kind
of viscous damper. Such fluid film damping is one of the motivations
inclining the designers of rocket engine turbopumps to give up rolling
element shaft bearings in favor of plain journal bearings. A gear cam
drive has fewer than 10 oil films between the crank and cam, but a
chain has a great many more.

Phil Irving, designer of Vincent motorcycles, suggested construction
of parallel twins with crankpins not at the traditional 360 degrees,
but separated by 76 degrees. With usual rod ratios, the piston reaches
maximum velocity at about 76-deg ATDC. At this point, the crank arm is
at right angles to the con-rod; the condition for maximum piston
velocity. This crankpin angle would therefore cause one piston to be
stopped when the other was at maximum speed. The result would be that
the two pistons would exchange their kinetic energy only with each
other, and not with the crankshaft, eliminating one important source
of cam drive disturbance.

Wide-angle Vee engines are better in this respect than parallel twins
but even they have their problems. The classic Cosworth DFV V8 GP car
engine of 1967 was estimated during design to have no more than a 35
pounds-foot torque peak in its cam drive, but actual testing revealed
peaks ten times greater; leading to drive failure. As the engine
was already near production when this was discovered, designer Keith
Duckworth had to scurry around and design a compact spring drive
(analogous to what is found in clutch cush hubs) that could be
incorporated into one of the gears in the drive.

Another approach is that seen on certain WW II German V-12 aircraft
engines, and on late race versions of Honda's RC30. In these and
other cases, small flywheels have been attached to the cams
themselves.

It is interesting to note that both Velocette and Ducati have found
that changing the stiffness of cam drive towershafts can be used as a
tuning measure to adapt an engine to a given race track a stiff
towershaft on short courses, and a more limber one for longer
tracks. The parts seem stiff and strong only in our imaginations and
in our not-very-stiff protein hands. In fact, at speed, everything in
engines is flexible because the amounts of energy moving from part to
part can be so large.

Byron Hines noted that as useful as Spintron's instrumentation is,
even more so is the experience and advice of Spintron personnel. One
of the first things they suggested was that he replace his aluminum
rocker arms with steel, for accurate motion depends more on stiffness
than on weight. He also said that full benefit from Spintron requires
making several laps through their process. A first lap involves re-configuring
the cam profiles and valve train to settle the motion and eliminate float
and excessive bounce. A second lap becomes necessary when it is realized
that after lap one, the engine's power curve has sagged in some places.
This is because before, cam timing and lift had been unwittingly chosen to
at least partly compensate for the uncontrolled valve motion. Once the valve
motion is settled, timing and lift are wrong. Now lap two consists of finding
new optimum cam timings and profiles to again maximize power. That, in turn,
brings to view new dynamic problems to be solved in lap three, and so on.
The people with the greatest sophistication in all this are, naturally, NASCAR engine
builders.

As an extreme example of what can happen, airflow
pioneer Jerry Branch was once called upon to dyno a special V-twin
that was conceptually a slice off a small-block Chevy, crank and all.
Being intended as a motorcycle engine, it had no flywheel other than
its little piece of V8 crankshaft. Branch said that although the engine
had plenty of displacement and hot tuning parts that suggested an
easy 100-hp, it never made over 35 horsepower. Each time it fired,
with almost no flywheel mass to smooth the pulse, it launched its
valves into orbit.

Another aspect of unplanned valve motion is the vibratory modes of
valves themselves. Particularly with rocker-arms (which exert some
side-thrust), a valve can be excited laterally as it lifts, the head
of the valve whipping from side-to-side on the stem; the stem
possibly made more flexible by undercutting to squeeze out that last
CFM of airflow. When this flopping valve approaches the seat, one edge
can hit first, causing a motion not unlike the final stages of a spun
coin's motion. Or, approaching its seat squarely, the rim of the
valve stops but the stem and center of the valve head keep right on
going; the trampoline mode of valve flex. When the motion
at the center finally stops, the valve head is quite deformed and it
now snaps back, tossing itself back up off the seat in a cycle that
can make several hops; with considerable lift being reached in the
process. This is a cautionary tale for those who wish to carve away
valve head mass in search of ultimate lightness. Or for those who wish
to replace existing valve shapes with something quite different. Think
about it if things don't go right. You may have made transformed
those elegant chunks of metal from valves into springs.

Spintron is not cheap but all of us can afford imagination. If you
have valve or spring problems, think about what has worked in your
experience with your particular engine, and what has not. Careful
mental sorting here can reveal a lot. Besides, what else is a body to
think about while waiting for a dental appointment? Sit with the
engine and rotate it through its cycle, thinking about what is
happening at each point. Over time you can develop a mental picture of
what may be happening and what might correct the problems. There is
more to valve trains than light parts and heavy springs."

[hoffman900]

Also, read this about the Honda's Development of Valvetrain for Formula One:
www.f1-forecast.com/pdf/F1-Files/Honda/F1-SP2_09e.pdf

Take note of what Kevin wrote about WWII German V12 aircraft engines and Honda's RC30 (early '90s), then see how they fixed fluctuations in angular velocity of the camshaft .

For reference, the maximum intake valve acceleration of 55mm/rad^2 is 0.0377925625"/*^2 , and that also meant they had the ability to spin to 20,300rpm.

Compare that to Mike's NASCAR restrictor plate cam circa 2006:

"Aggressive" depends on the application.

Here's the numbers at the valve on a 3 year old NASCAR restrictor plate cam.
300 @ .0001" Valve Lift(Seat)
262 @ .050" Valve Lift
.000882" Peak opening valve acceleration
.000512" Peak nose valve acceleration
.000797" Peak closing valve acceleration

That cam was mild compared to some others it ran against

Not being limited by valve springs is a nice thing.

Nothing is new...

[swampbuggy]

Camking wrote---we have positive flow before TDC because the exiting exhaust flow causes a low pressure area below the intake valve during overlap. This siphon, suction or whatever one wishes to call it was discovered in the mid 1950's by none other than the legendary Ed Iskenderian. This was given the name 5th cycle, which was waaay before most cam companies came into existence. So for those of you folks who don't know the history ? The term 5th cycle is a lot more than a catchy name, it was a discovery !! Now you know the rest of the story Mark H.

[Stan Weiss]

Karl,

Going through the math.

Given:
Peak Intake velocity opening side: 0.00592446"/*
Tappet Diameter: 0.770"

Need:
Edge Distance: x

Equation:
x = -1((Velocity*57.3)-(Tappet Diameter/2))

Solution:
x = -1((.00592446 * 57.3)-(0.770/2))
x = .045528"

Effective Tappet Diameter:
.770 - (x*2) = 0.680"

I'd say you're giving up a lot.

Let's try this with 0.017" edge distance.

=((.770/2)-0.017)/57.3
= .006422"/* or... ~8%

Mike's comment on lobe chamfer:

Almost every Flat Tappet lifter comes from the manufacturer with a chamfer on the edge. It's to keep them from chipping.
On the NASCAR Cup engines, we would have the lifters made with no chamfer, so we could design the profile to use every bit of the .875" face.

That's the reason for having no chamfer, but if the cam profile isn't designed to go all the way out to the edge, it doesn't matter.
If I design a profile to go out to a max of .4175" from the center of the lifter, that's .020" from the edge of an .875" lifter(.875"diameter=.4375" radius), so if it has a .010" chamfer or not, the lobe will never see it.
If I design the profile to go to a max of .435" from the center of the lifter, it better not have a .010" radius.

So let's look at an unchamfered tappet:

=((.7874/2)-.017)/57.3
= 0.006574"/* or ~ 2% greater than .017" edge distance, but it's ~10% greater than what you have. You could also look at this as a chamfered .800 tappet.

Looking at a .800, with no chamfer:
=((.800/2)-.017)/57.3
= 0.006684"/* or ~11% greater than what you have.

Going from what you have to a a chamfered .800" tappet, it's just a bit less than going from a .842" Chevy tappet to a .960" vintage NASCAR mushroom tappet, but that's a pretty big difference. How much do you trust the location of the lifter bores in relation to the cam?

Keep in mind this is one facet of cam design though. The real art of the math is in the shape of the curves. Through improved equations, computers, and manufacturing ability, cam designers have been able to reduce peak acceleration and jerk, while maximizing the velocity of a given tappet and keeping or gaining lobe area.

It's really a question of how crazy do you want to go?

Peak velocity is easy, but what can your valvetrain withstand in terms of acceleration? As you know, what is happening at the lobe isn't always the same as what's happening at the valve. Take a look at the numbers I posted and you can see that for yourself. To quote Mike, "The trick is to get the area under the curve without a high rate of acceleration".

A good cam designer doesn't make enough for how talented they have to be.

Some Mike and Harold quotes:

Harold:

I started to give you some of my opening and closing rates at valve contact, then I realized I might be revealing too much, although I think I've mentioned seating rates for several years.
However, in the thread on 565s, I have a partial matrix of cams in terms of .020, .050. .200 durations and various lobe lifts. I did not put iin there all the cams in that series, but just the ones of interest for under-10 sec 565s.
ALL of those cams use the identical opening ramp, and the identical closing ramp. They only vary in duration, and lobe lift. Naturally, the higher the lobe lift, the higher the negative acceleration, and the bigger the duration, the lower the negative acceleration.
The PEAK acceleration rates on those cams is .0002975"/*2 on the opening side, and .000295"/*2 on the closing side. I have had cams with SLIGHTLY higher rates win class at the 24-Hours of Daytona, etc.
I use these ramps on my daily-driven street cars, like 540s, where I use a 257/265 at .050 with .725"/.725" gross valve lift. I have never had anyone report a broken valve spring with that cam, and I have been making it since the spring of 2006.
This is not my most aggressive cam, not by a long shot. However, it is a very successful series of bracket/race cams, with lobe lifts over .540".
My designs are all multi-segmented polynomials, which is similar to a B-Spline. I have done this way since April 1980.

I have been designing race-winning cams since December 1972, and doing it full-time since August 1974. I have been General Kinetics, Competion Cams, UltraDyne's, Lunati's, and Custom Camshaft Company's cam designer.
The two main things in determining maximum RPM are the negative acceleration on the nose(valve float or valve loft), and the valve seating velocity(valve bounce). None of this is related EXACTLY to power, but plays a part in it. Obviously, there has to be enough area-under-the-curve to make power at a given RPM.
This is why cams that are dynamically smooth enough to go 10,000 RPM may not do it; they just may not make enough horsepower.
In 2007 I designed a Pro-Stock cam that would turn 10,500 for over 90 seconds with no bounce or float. Why wouldn't it go any higher?
To get the required area-under-the-curve, the cam required high positive and negative acceleration rates(1.9:1 rocker ratio, .575" lobe lift), and therefore it also required very high spring pressures.
The Spintron's electric motor did not have enough HP to turn the valve train any higher. The cam was dynamically smooth and stable at 10,500.
And it runs OK, too. Just ask Alan Johnson........
Cams that have pointy noses have more negative acceleration than cams with more rounded noses. They require more spring pressure for the same RPM, and they are in Catch 22---A pointy nose is more likely to wipe out than a more rounded one, and the higher required spring pressure makes it easier to wipe out......

These are just some thoughts and observations of an old cam desiginer....

There are people in the cam design world that charge several thousand dollars for that information.
However, here it is a la UDHarold:
Almost all negative acceleration in hydraulics and solid flat tappets cause too small a radius of curvature on the nose if the negative accel rate exceeds -.000220, I do not go even that high. The open spring pressures necessary to control that rate of negative acceleration, and with that small a radius of curvature, cause almost sure wipe-out of the cam. Too much pressure concentrated in too small an area.
Roller cams are relatively OK on Negative acceleration, it just requires more open pressure, especially when the valve lift goes over 1.000". Eventually cam materials come into play, particularly when you have 1400 lbs at 1.000" valve lift, and a 1.9:1 rocker.
Flat tappet cams, both hydraulic and solid, run very high positive acceleration rates( .000400-.000500"/*2) compared to rollers. However. they are limited by max velocity before the contact point of lobe and lifter doesn't exist on the lifter anymore.
One of the signs of a good cam designer is how well you live with this limitation. I keep all my designs .017" away from the edge for over 28 years, and have had some go 100,000 miles on the street. An engine block with out-of-position lifter bores may wipe this clearance out, and therefore wipe out every cam put in it.
Roller cams, both solid and hydraulic, are controlled by positive acceleration rates, as well as base circle radius. Any acceleration rate over .000340"/*2 can cause negative raduii of curvature---The dreaded 'Inverse-Radius' cam, unless special math methods are used.
All of my 28* Major Intensity cams use such techniques.
Any positive acceleration rate under .000250"/*2 is almost a waste of time, although valve train life is tremendous.
An acceleration rate around .000300"/*2 is good for 24-hour racing, if the valve seating velocities are done right.
As is obvious, opening side rates and closing side rates may not agree in positive accelerations or opening/closing velocities, but the nose acceleration must always be the same, or there is a visible line, as well as unequal velocities, at the nose.
It took me 38 years of work to learn all this.
Good Luck!!!

I do not know if I should, or even want to, give a proper lesson on ramp design. After all, there may be other cam designers lurking about, wanting to learn something.....
However, your question is not about valve lash, but about ramp design. You can design ramps with wide valve lash--.026" to .036", that open valve SLOW. Those cams do not work very well. There are mathematical limits to how fast you can open a tight-lash (.012"-.016") cam. Trying to exceed this limits causes valve train problems.
Of course I have opened valves fast and shut them slow for 28 years now, since 1977. Anyone who doesn't know this hasn't been paying attention, I have been advertising it for this long.
I have never heard of any valve train problems with tight-lash cams, as far as heat-over-expansion of the valve train, etc. Valves that are too loose, such as .036" to .045", can cause battering of the valves and seats, short spring life, etc. I have had excellent valve train lift with ramps .0173" high--.026" for 1.5s, .030" for 1.7s. These ramps have been the ones that have gone 100,000 miles, so they aren't too bad. Tightening them down slightly, as tight as .018", have only helped valve train life.
Tightening them down increases seat duration, which may add a little top-end, but certainly takes away bottom-end. Tighter valve lash may improve TIRE life, though!
I have ran solids/solid rollers on hydraulics/hydraulic rollers, and as long as the valve lash is really tight, say .004" to .008", I have seen no trouble. All hydraulic cams start lifting off the seat by .004" ramp height, regardless of advertised duration heights, and running much more valve lash can put the point of rocker/valve contact deep in the cam acceleration curve, which isn't good. Kinda like a sledgehammer blow to the valve. You can also run hydraulics on solids, but they had better be tight-lash solids, or else you will have veeeery long hydraulic seat durations!!!
There are several reasons for what ramp height and velocities are chosen, but this may be getting into secret territory. Part of the Great Cam Design Mystery is learning how to increase velocity at lower and lower ramp heights. The reason for lower and lower ramp heights? It gives you longer and longer times to accelerate the valve....

Mike:

Normally, Max acceleration occures before Max velocity.

Here are some examples of distance from lash point to max accel, and max Vel on different cams.

Comp Cam, Pro Cup roller: Max Accel=14 deg, Max Vel=46 deg
Crane cam, 410 sprint roller: Max Accel=30 deg, Max Vel=70 deg
Schrick Cam, BMW overhead bucket: Max Accel=10 deg, Max Vel=48 deg
LSM Cam, Pro Mod roller: Max Accel=32 deg, Max Vel=78 deg
GM 604 crate cam, Hyd roller, Max Accel=27 deg, Max Vel=53 deg
Jones Cam, "Banjo"'s roller: Max Accel=17 deg, Max Vel=59 deg

I just designed a roller cam that gets to max velocity in about 10 degrees, but it's only designed to turn 2,000rpm

BTW, you can have a higher velocity, without increasing the max acceleration. You accelerate to max acceleration, and then decrease the acceleration rate toward max velocity, then accerate. If you change the shape of the acceleration curve from max acceleration to max velocity, you will get a higher max velocity with the same max acceleration.

Here is a more modern .904" design.
It's designed for a class that has to run stock rockers, and stock size springs(130# on seat), and they turn about 6,800rpm.
It's a .325" lobe lift, and the lobe runs to within .008" of the edge on the .904" lifters
260 @ .020"
230 @ .050"
199 @ .100"
146 @ .200"
68 @ .300"

Bob,
UDHarold is talking about a larger diameter lifter than Karl is running and Karl has said that valve train life is tremendously important to him.

I do not know what springs and rocker arms Karl has to run. Karl's cam has a .300 Intake lobe and has greater duration number than those other than .300".

Stan