In a previous post, UDharold talked about the high accelerations associated with XE cams, and said that the Voodoo line is easier on the valvetrain. Please explain to me how this is true.

Lets look at the XE256, vs. the voodoo 256. Same duration @ .006", the voodoo has MORE duration at .050", and MORE total lift. So it would seem to me, that the valve is has a higher rate of acceleration from .006" to .050" (more deg @ 0.050" for the same deg @ .006"), and a higher rate of acceleration from .050" to max lift. After all, more lift for a given rotation = higher accelerations, no??

How is it possible that the Voodoo line is easier on the valvetrain? Are we talking about accelerations on the clearance ramp from the seat to .006"?

Harold, are you out there???

Novadude,

I don't want this to turn into a seminar on good cam design, after all, Comp Cams probably reads Team Chevelle, too.
The Xtreme Energy cam we measured had a peak acceleration rate of over .000510"/°2. ALL VooDoo cams have a peak rate of .000475"/°2. ALL VooDoo cams have a LOWER nose acceleration rate than my old UD 272H, which was considered well-rounded, and went over 100,000 miles in many an engine.
What we have is a higher AVERAGE rate of acceleration, which gives us the better high lift area, but it is ONLY the peak rate of acceleration that hurts valve trains.
ALL VooDoo hydraulics stay further away from the edge of the tappet than .019", a very safe number.
Did it take a lot of design? Sure, I worked on the VooDoo line for about 10 months, and designed HUNDREDS of cams, before I got the ones we made. I have also done hydraulic rollers and solid lifter cams in the VooDoo series, and some racing solids and rollers.
For your question about the seating rate, the closing side is 4° longer than the opening side at .004", where the valve touches the seat.....
This is enough information to start a number of wheels turning in Memphis.....
After all, this is what I do. And I've been doing it since December, 1972, so I've had a little practice....

UDHarold

Thank you , sir!

I am an engineer, so I tend to ask a lot of questions... I appreciate you sharing what you can. smile.gif

but it is ONLY the peak rate of acceleration that hurts valve trains I was under the impression that it was the rate- of-change of acceleration ("jerk" or "yank" as you may know it) that influences impact loading which in turn directly affects component life.

Without giving away too many trade secrets, does the Voodoo line have continous and/or finite higher-order derivatives (i.e., x-triple-dot, x-quad-dot...)? At what point (which order) do the higher-order derivatives have discontinuites or become non-finite?

70GS455,

This is about as far as I'm going to go in delving into my techniques of cam design, but I have design equations that control the 5th derivative...........
The numbers generated are probably below the accuracy of the master-generating equipment.
The rule of thumb is:
1st derivative/velocity====Significant number, 3rd decimal, .001
2nd derivative/acceleration==Significant number,
4th decimal, .0001
3rd derivative/jerk========Significant number,
5th decimal, .00001
4th derivative/snap========Significant number,
6th decimal, .000001
5th derivative/'crackle'===Significant number,
7th decimal, .0000001=one ten millionth of an inch!
I have not found the 6th derivative, 'pop', to be useful yet........

UDHarold

ALL VooDoo hydraulics stay further away from the edge of the tappet than .019" Hope you didn't give away too much there, with 19 thou of margin, that puts the max velocity at no more than 0.007016"/deg (on an 0.842").

70GS455,

It's more than .019", and less than .007016"/°...
Of course, that also means it's less than .020".

UDHarold

You can take a Cam Doctor and plot lift vs degree rotation and then use a program called SIMULQ using Fortran-4/What-5 and come up with an Nth-1 order polynomial from N number of data points that you can then manipulate to give the velocity and acceleration info IF you are Mathmatically inclined and have alotta time on your hands..

I would have to do some "head-scratching" to turn the aceleration numbers into strengths of the valvetrain componets using E, (Mod. of Elasticity of Steel), and F-MA stuff if not mistaken, but it's been years and years since I have considered this...

pdq67

I have absolutley NO idea what you guys are talking about...
buy hey, it makes for an interesting read! :D

Cam Doctor can get you close to what the actual design parameters (and ultimately, design equations) are. But the problem is that errors in the measurements taken show up as "noise" in the higher-order derivatives, and gets worse as the derivative orders get higher.

Designing cams to .000001 sounds good on paper, BUT, how accurate can these things be machined? It would seem to me that Comp and Voodoo are targeting mass-production cam applications with these parts, and I wonder how close you can practically get to "the edge" while keeping in mind practical (cost effective) manufacturing techniques. Let's face it... the lift curve will never exactly duplicate the ideal once a cam is in a block, and you account for tolerances in cam manufacturing, lifter bores, etc.

This is interesting stuff... gotta dig out my engineering texts, as I have never looked into cam design. smile.gif

68SS454 pretty much nailed it for me. Comp Cams has treated me very well. I had issues with rocker geometry on my Twisted Wedge heads, they exchanged half of my Pro Magnum rockers for "special" ones to correct the problem. And they replaced a defective lifter (and cam) that would have been LONG out of warranty from sitting on my shelf. I have some loyalty to them as a result...but dang...I'm really impressed with you, Harold.

My Pontiac-building buddy had nothing but good things to say about Ultradyne. And he won the points total in his dirt track class every year for as long as he was involved.

It's a damned fine situation to be in: More than one reputable company deserving my dollars. Usually, I have to buy stuff based on "this company isn't as terrible as that other company".

Thanks, Harold, for your input on this board!

How do these new ones compare to the similarly sized legacy Lunati profiles, i.e., the H235, H240, etc.?

Snap, crackle, pop? Was that a joke that flew over my head or is the rate of change of jerk really called "snap"? What is that useful for? And am asking these questions in the most generic way (not just cam design but anything.)

I used to tune servo motors for huge, contouring milling machines, trying to match acc/dec curves between axis to cut perfect circles (right, perfect) and we never got any deeper than jerk... granted the CNC probably internally took care of it.

Just a curious question from a curious guy.....

I always understood the rate-of-change (derivative) of jerk to be yank.

Oops, that's not correct according to:

http://math.ucr.edu/home/baez/physics/General/jerk.html

Originally posted by UDHarold:
70GS455,

This is about as far as I'm going to go in delving into my techniques of cam design, but I have design equations that control the 5th derivative...........
The numbers generated are probably below the accuracy of the master-generating equipment.
The rule of thumb is:
1st derivative/velocity====Significant number, 3rd decimal, .001
2nd derivative/acceleration==Significant number,
4th decimal, .0001
3rd derivative/jerk========Significant number,
5th decimal, .00001
4th derivative/snap========Significant number,
6th decimal, .000001
5th derivative/'crackle'===Significant number,
7th decimal, .0000001=one ten millionth of an inch!
I have not found the 6th derivative, 'pop', to be useful yet........

UDHarold So....... anyone else understand what the hell is going on here?

Harold....do you have a link to the solid and race solid line of cams? Im in the market and have one picked out already but id love to see what you have to offer. If you dont want to post it for some reason my email is NickLingua@hotmail.com

Originally posted by 70GS455:
Cam Doctor can get you close to what the actual design parameters (and ultimately, design equations) are. But the problem is that errors in the measurements taken show up as "noise" in the higher-order derivatives, and gets worse as the derivative orders get higher. Is there any way of estimating measurement error and removing it from the equation?

Back in the 80's or there abouts Crane came out with a family of lobes call the " Hi Low No Pop " series of drag raceing lobes.... I " think " not really sure they were Harvey's designs.... anyhoo back then someone must have thought "POP" to be important but honestly when you get into the 10 millionths and such type numbers just how the bleep can a given valve train even detect such movements ?.... Or is it perhaps some unfelt to humans " Harmonic " amplitute thing that we cant see but the valve train
likes or dislikes ?.... I can certinally see where some of those derivitaves can easily get one in trouble , but those real small numbers ?... Or is there some other reason that non math experts like me can fathom ?

Here's some plain english explanations of this stuff.

http://www.powermechanics.com/define.html

I designed a writing instrument, "cap and barrel", incoming raw parts checker that used a proximity(Sp?) gage using a master and the checked part once that was good to the .00001" if I remember right when I was up at WA Shaeffer Pen in Ft. Madison, IA years ago for a short while so know that you can check a lot closer then my 1/10th's feeling thumbnail!!!

And I figure if I want a big solid cam, I will use mushroom lifters and have it ground on a MOPAR curve that would approach a solid roller except fot all-out lift....

pdq67

I wish Crower still sold their drop in mushroom jr lifters...

graemlins/waving.gif Paul

Gentlemen,

To answer a lot of questions.....

Petersen-Bercos are only accurate to .0001", but by use of these higher-order derivative curves, you can design cams that make over .001" difference(or more) in the lift profile. The gain is a culmative one, and is mostly in high-lift areas, where it does the most good.
As far as using larger-diameter lifters, yes, you can get a big increase in high-lift areas, but to demonstrate the advantages of superior math, lood at this:
UltraDyne 292° flat tappet, master F4, designed and made in April 1980---292 at .020, 259 at .050, 170 at .200, .3677" lobe lift.
Designed for the .842" tappet, identical in wear as the 276/284--They share the same ramps.
Crane F-259/3733 Chrysler tappet, designed and made around 1990---294 at .020, 258 at .050, 166 at .200, .3733" lobe lift.
Their .904" tappet design is not even superior to my 1980 .842" tappet design, the 292 F4 profile being 4° fatter at .200.
It's all in the math......

UDHarold

BTW--I have a VooDoo 290 solid cam that is 290 at .015, about 285 at .020, 257 at .050, 170 at .200, and .372" lobe lift. Still on a .842" tappet, and over .018" away from the edge.....

Petersen-Bercos are only accurate to .0001", but by use of these higher-order derivative curves, you can design cams that make over .001" difference(or more) in the lift profile. The gain is a culmative one, and is mostly in high-lift areas, where it does the most good.So you are saying the use of higher order curves allows you to design a cam that is well within the tolerances of automatic griding equipment? That begs the question: Why doesn't everone use 5th and 6th order curves to design cams? Do most companies only design to the 4th order? If so, why? You would think that modern software would eliminate a lot of the mystery & painstaking calculation behind these higher order derivitives. Maybe I am wrong?

Hey Harold,

Do you have any new mechanical Street Rollers in that new Voodoo line? If so what kind of specs.

NovaDude,

Why do you think that most companies have gotten as far as the 4th order? Almost ALL of the cam companies using commercially-available cam design programs do not have designers that are math freaks. The program responds to whatever numbers are entered, whether or not they are the best.
On my new oval/bracket roller series, which I call the K-series, I went through A to J before I got a cam I thought was the best I could do at this time. The A design was quite a bit better than my old 288R roller, the most popular roller I have ever designed.

Darren,

Yes, I do, but the numbers are not finalized yet. We start at a 271° at .015, 235° at .050, and .370" lobe lift, although that may change. Highest lobe lift is .390", and there may be a seperate series for BBCs.
Let me know what you're interested in.....

UDHarold

Originally posted by UDHarold:

Darren,

Yes, I do, but the numbers are not finalized yet. We start at a 271° at .015, 235° at .050, and .370" lobe lift, although that may change. Highest lobe lift is .390", and there may be a seperate series for BBCs.
Let me know what you're interested in.....

UDHarold Harold,

I was just curious to what new you may be designing.

I have been kicking around the idea of trying a 502A3 ground on a 108LSA, but I'm not sure if I have enough P-V clearance and I not sure if my valve springs will work (Isky PN# 9005, 185# @ 1.875" seat, 465# open, .650" max lift).

I am running an Isky street roller now; (290/290, 248/252, .602"/.646", 110 LSA, 4* advanced). The 502A3 is abit more aggressive which I think would reduce p-v clearance and a 108lsa would reduce it also.

Harold, do you have any idea how much the p-v would change, if any?

What about the springs, do you think they would work? :confused:

Oh, this is interesting. Do these newer cam designs reduce harmonics in rats?

Harold,

You know I am anal about not using a solid roller on the street b/c of longevity issues SO please design a milder version of the great old 288 such that we can live with it on the street using SOTA lifters and milder spring pressures..

Is it possible to hold everything the same EXCEPT lift AND possibly the duration at 200 number???

So that it will be more street friendly.....

pdq67

PDQ67,

Since I last posted tonight, I tried a nice little street roller design.
293.4° at .015
288.4° at .018
285.4° at .020
254.4° at .050
174.0° at .200
.394" lobe lift
Opening ramp of old 288R is 27.2% faster....
Closing ramp of old 288R is 17.7% faster....
New cam has Major Intensity(.050 to .020) of 30.97, old 288R was 33....
Over 3° offset at .020...
Not bad for a first attempt, now to see if I am allowed to make it......
This would be a family, with all cams in the family sharing the same ramps and relationships. The only changes would be in lobe lifts and in whatever duration we wanted, say in 4, 6, or 8 degree steps.
BTW, it's coming off the seat on the opening side at a lower velocity than the 276/284F cam, and shutting 15% slower than the 276/284F cam...


I'm outa here for tonight, the computer is now turned over to my daughter....

UDHarold

UDHarold! Aw shucks, just when things were starting to get interesting! I dream of the day either you design or recommend a cam for my BBC! Can hardly wait till your next brain storm! :D

Would you recommend that cam for a daily driver???

B/c I want a daily driver, 288 solid roller cam!!
That's why I am asking about lift vs longevity.....

pdq67

ttt...

Harold,

How about that grind on say, a .600" gross lift cam???

Will it last longer in a daily driver setup using GOOD solid roller lifters and not all that high pressure valve springs AND still run fine??

I guess I want ta run the sucker like a good old regular solid lifter cam on the street by just checking her every third oil change is all for like miles and miles... If GM has finally figured out how to make their high lift, hi-po hy-roller cams live in a daily driver mode, why can't somebody do the same thing with a solid roller??

It just to me wouldn't be a "balls-out" cam is all but still would be a good bit better then a hy-roller, imho....

pdq67

Can I get in line for that cam too?

Paul, I take it you don't like the new Schubeck lifter or it's price.
My main concern with it is dirty oil. How will it handle that? I can eventually get over the price.

I would rather use the Schbeck, raduis foot lifter vs their axle-less roller lifter until guy's come back and endorse them for daily use on the street is all.

But I figure a solid cam ground using a MOPAR dia. lifter or a cam and kit using a GM mushroom lifter with both of them having the little EDM holes in the lifter feet will be fine for me when I decide to go for a more "balls-out" motor combination..

That's why I keep bugging Harold about creating a solid roller cam that will live at least 50,000 daily driven miles before I have to do surgury on my motor to check it out.

B/c GM has figured out how to make the high lift hy-rollers live in a daily driver deal in the new SB motors!

So what if it isn't "BALL-OUT", SOTA, MAX. EFFORT solid roller cam if you can actually drive it a long time without the worry of when she's gonna let go?????..

I guess I'm too old-school b/c the only thing i want ta worry about is feeding my motor!! Not whe ta check it to see how long I have before the roller lifters crap out.. A street driven solid roller had better run like a good solid cam, imho to be practical to me....

pdq67

pdq67,

Just as an exercise, tonight I ran off a:

294.42° at .015",
286.44° at .020".
255.47° at .050",
169.77° at .200",
and .35300" lobe lift.

This one gives .600" gross valve lift, and with 1.7 rockers and .018 hot valve lash, about .580" net valve lift.
As I stated earlier, it would be easier in opening- and closing-rates than my old 276/284F10.
This series of street rollers hits the seat about 15% gentler than the 276/284F did, and the rule of thumb among cam and spring designers is 10% gentler doubles the spring life.
Lunati is supposed to have their Landis installed around March. When that happens, I will be able to make masters as I wish--I have demanded 8 hours a week just for making masters.
I have over 100 cam designs awaiting masters, all solids or roller cams, plus some LS1 and 4.6L stuff.

UDHarold

Some quy has actually dynoed some engine with these cams.

http://www.chevytalk.org/threads/showflat.php?Cat=0&Number=1053051&an=0&page=0&gonew=1#UNREAD

btw. you have a great forum =)

Looks like the Voodoo 268 will be my next cam. smile.gif

I spent my 11 weeks at Holley working with Doug F. on the dyno, etc. All the VooDoo cams have the same valve train dynamics, and lack of valve train noise.
We dynoed the 256, 262, 268, 276, and 284 in the same engine, along with the corresponding Brand 'XE' pieces.
I also drove around in the 1988 Camaro that had the VooDoo 268 in it. I really liked the way it performed in daily street driving, although I did not get to drive it home---327 miles, EACH way.
I will let Doug post the dyno sheets, and hopefully Holley will soon have them up on their website.

UDHarold