When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
C6 Corvette ZR1 & Z06General info about GM’s Corvette Supercar, LS9 Corvette Technical Info, Performance Upgrades, Suspension Setup for Street or Track
This thread has some of the worse misinformation I have ever seen in my life among any car forum. There is not one single factor where a stock rocker will outperform a shaft mount setup. Not. One. Single. Thing.
They will control the valve more precisely, survive much higher springs pressures, allow adjustability both on the PR cup and on the valve stem, are more reliable, reduce valvetrain friction, and allow far more RPM potential. Anyone who has seen a spintron comparison or converted to shaft mounts before will know the benefits. Things like increased spring pressure, thicker pushrods, and expensive lifters are NOT mandatory with a rocker swap. Tougher springs are only needed if the ratio has increased since rockers not only increase max lift, but ramp rate as well which will make the same cam behave like a more aggressive lobe. They are not needed "just because" they are shaftmounts.
Vette owners are notorious for the "stock is best" mentality and often it is to their own detriment. Keep in mind GM built these cars to fulfill a LOT of different criteria like efficiency, cost, emissions, compatibility between different platforms, etc so not everything on them stock is built for ideal performance. It seems like many are completely happy to pay outrageous $2k+ prices on a stupid simple exhaust system that is $400 worth of parts, and on the same day claim that ~$1200 is too expensive a purchase for the MOST IMPORTANT part of your engine, the valvetrain. Makes no sense.
The basic argument myself and others are making is choose the rocker that is best for your application with the correct supporting mods to support that rocker. The light weight nature of the stock rocker, cost, reliability have advantages, and unless you want have a large budget to spend on all the extra modifications to run a roller rocker, work great. As far as Tony running them on his packages/setups, that is great, they are not low dollar near stock budget packages. He puts together a system with a bunch of other parts with those rockers in mind. No one is saying they can't work if you do that.
Just like if the OP wants to throw the TDs on his car, they can work, and be a great rocker I'm sure, but I wouldn't expect to bolt them on, and just go. He should be replacing the whole system to accommodate those rockers, and they are a class specific race car part, that probably have higher maintenance and potential issues than stock. When you up the spring pressure, etc etc you now lower the maintenance intervals, and have to inspect and check things more often.
Agreed.
I highly doubt I would throw on RR’s on my setup unless they will quiet my valvetrain (depending where money is at and if I want to pull these heads, which I do not unless something happens to them). Even then the only ones I would consider are shaft mounts. From what all I have gathered and been told, under .660 lift, stock with a trunnion upgrade will work great.
Tougher springs are only needed if the ratio has increased since rockers not only increase max lift, but ramp rate as well which will make the same cam behave like a more aggressive lobe.
There seems to be various kinds of anticipations what happens when rocker ratio is changed.
Spring force reserve (between the lobe and the follower wheel) 65 lb or 16 % of spring's rating.
Let's drop the rocker ratio to 1.7 and increase the lobe lift so that identical valve lift curve is achieved. We also assume no change in rocker inertia (so called "ideal rocker"):
Spring force reserve left 44 lb (11 %).
What just happened?
Short answer: We gave the spring less leverage so the result is expected.
Longer answer: Load from the valve train dropped from 473 lb to 468 lb (as seen by the cam lobe), but the change in spring's ability to deliver opposing force dropped more, from 584 lb to 551 lb.
Ok, let's give the mighty LS7 a mickey mouse 1.5 ratio and let the valve lift drop accordingly. Surely this should help things, right?
Result: Force reserve 18 lb (5 %).
WTF you might ask.
Yes, the load from the valve train dropped (compared to stock LS7) from 473 lb to 392 lb, but spring has even less leverage AND less compression due to which the opposing force dropped from 584 lb to 434 lb.
If we didn't have the the "ideal rocker" with a fixed inertia, the rocker inertia would be lower with smaller ratios and we would see a bigger drop in load generated by the valve train.
However, by keeping the rocker design similar in all these cases, I'm pretty confident that the net effect from the smaller rocker inertia couldn't reverse the results since the stock rocker contributes to the total effective valve train mass with only 9.6% share.
Edit: Let's add one more, 1.9 ratio and stock lobe lift (= increased valve lift) + assuming ideal spring (no coil bind) and same ideal rocker with no change in inertia:
Result: Force reserve 80 lb (19 %)
The load from the valve train increased from 473 lb to 503 lb, but also the opposing force generated by the spring increased from 584 lb to 638 lb due to more leverage AND higher compression.
In real world rocker inertia would be slightly bigger, but again it would not reverse the result with similar rocker design.
Spring force reserve (between the lobe and the follower wheel) 65 lb or 16 % of springs rating.
Let's drop the rocker ratio to 1.7 and increase the lobe lift so that identical valve lift curve is achieved. We also assume no change in rocker inertia (so called "ideal rocker"):
Spring force reserve left 44 lb (11 %).
What just happened?
Short answer: We gave the spring less leverage so the result is expected.
Longer answer: Load from the valve train dropped from 473 lb to 468 lb (as seen by the cam lobe), but the change in spring's ability to deliver opposing force dropped more, from 584 lb to 551 lb.
Ok, let's give the mighty LS7 a mickey mouse 1.5 ratio and let the valve lift drop accordingly. Surely this should help things, right?
Result: Force reserve 18 lb (5 %).
WTF you might ask.
Yes, the load from the valve train dropped (compared to stock LS7) from 473 lb to 392 lb, but spring has even less leverage AND less compression due to which the opposing force dropped from 584 lb to 434 lb.
If we didn't have the the "ideal rocker" with a fixed inertia, the rocker inertia would be lower with smaller ratios and we would see a bigger drop in load generated by the valve train.
However, by keeping the rocker design similar in all these cases, I'm pretty confident that the net effect from the smaller rocker inertia couldn't reverse the results since the stock rocker contributes to the total effective valve train mass with only 9.6% share.
Real eye opener, I would say.
Fancy. Now can you do the same for stock LS7 rockers versus shaft mounts? Or any/all roller rockers? I think that would be a neat comparison.
Fancy. Now can you do the same for stock LS7 rockers versus shaft mounts? Or any/all roller rockers? I think that would be a neat comparison.
Sure, BUT I would need the rocker inertia. Without going in too deep, I can generate quite adequate lobe lift curves (and hence their 1st, 2nd, 3rd and 4th derivates as well) from Comp and Crane lobe catalog data for spring dimensioning. It's amazingly good compared to Cam Doctor data which I have for a few cam lobes.
I already posted one comparison (see one of my earlier posts), in which I tried to model the Comp RR for LS7 without actually having it in my hand. That modeled rocker had 149% higher inertia and in the place of stock LS7 rocker it would result like this:
Force reserve 13 lb (3 %).
Load from the valve train increased from 473 lb to 535 lb and opposing force from the spring stayed the same at 584 lb (forces existing between lobe and follower wheel, just like in earlier post).
Sure, BUT I would need the rocker inertia. Without going in too deep, I can generate quite adequate lobe lift curves (and hence their 1st, 2nd, 3rd and 4th derivates as well) from Comp and Crane lobe catalog data for spring dimensioning. It's amazingly good compared to Cam Doctor data which I have for a few cam lobes.
I already posted one comparison (see one of my earlier posts), in which I tried to model the Comp RR for LS7 without actually having it in my hand. That modeled rocker had 149% higher inertia and in the place of stock LS7 rocker it would result like this:
Force reserve 13 lb (3 %).
Load from the valve train increased from 473 lb to 535 lb and opposing force from the spring stayed the same at 584 lb (forces existing between lobe and follower wheel, just like in earlier post).
It might actually help the folks if you explain what the graphs mean/represent. I know I don't quite understand what I'm looking at. And also, are these graphs generated by data input, or are they actual measurements on a mocked up valvetrain? Where do you get your data to input into the graphs? What tools are you using to produce these charts? Software? Please don't think this is a call-out. I'm just trying to alleviate confusion about variables and data sources. It's too easy for people to see a fancy graphic and think that person MUST know what they're talking about because that have fancy graphics and use big words.
In the above diagrams blue curve represents the force generated by the spring as function camshaft rotation (degrees). Since the force generated by the spring is in the opposite direction from deflection (compared to it's natural state = rest) => F = -kx, where the k is the spring rate and x is the amount of deflection = actual valve lift. When spring force is reduced to the lobe side of the rocker, rocker ratio is needed.
The green curve represents the sum of forces generated by each of the moving masses within the valve train at given engine rpm. Note that valve train operaters at half speed compared to crankshaft. Basically F= ma, where m is the mass of body and a is the acceleration of that body. Masses on both sides of rocker fulcrum have to be redused to the side on which the calculation is made, typically on the lobe side. Here the rocker ratio comes into play again. The rocker itself moves in circular motion, so mass moment of inertia must be known. The angular acceleration for the rocker can be calculated from the rest of the valve train dimensions and movements.
During the opening (and closing) the inertia of mass (of all the components in the valve train) creates force towards the cam lobe (as does the spring, too), meaning all the components between the spring seat and cam lobe "see" their combined load (the distance between the green and blue curve). This is the phase which stresses the components most.
When the lobe rotates closer to the peak lift (located at 0 degrees in the diagrams), the speed at which the valve opens starts to slow down. This means that there is negative acceleration = deceleration (remember, F=ma, and now a is negative). If the (here negative) total force from the valve train exceeds the force the spring delivers (green curve goes below the blue curve), the follower wheel will no longer follow the cam lobe and in worst case coil bind can occur for this reason alone. Of course bad things continue to happen since lobe continues to rotate while the lifter wheel is airborne and when they meet again, there's nothing friendly about it.
The green curve also includes the worst case scenario of static and dynamic pressure difference over the valve.
The lobe profile is the "driver" behind all that takes place so lift curve must be known. The best way is to have a lobe measured by CamDoctor or equivalent. Note that the measurement is "noisy" (contains small errors from various error sources), so most use running average as a filter to be able to calculate smoother looking 2nd derivative (which is acceleration, see above). This is quite ok for spring dimensioning, but if one want's to dig deeper, Matlab or Octave are useful tools in order fit groups of polynoms in the measurement data. This allows me to filter out so called outliers and I'm able to get high quality 3rd derivate (known as jerk = change rate of acceleration).
Here is an example showing lobe lift (inch) and corresponding acceleration (inch/deg^2) with Excel. (Note the Y-axis units which are different than in the complete valve train force diagram):
Here is an example of LS7 intake lobe with Octave (all previous diagrams are made with Excel using simpler approach)
- CamDoctor data points (blue squares)
- Acceleration calculated from above data with running average filter (green line, just for visualization, 9 deg averaging window)
- Lift curve (blue continuous line) <= this is the one we need for spring force calculations
- Speed curve (small red dots, 1st derivative)
- Acceleration curve (dashed bold red line, 2nd derivative) <= this is the one we need for all other calculations
- Jerk curve (dashed red line, 3rd derivative)
Degrees shown in the header are durations at
- 0.006"
- 0.050"
- 0.100"
- 0.200"
- 0.300"
Executive summary, one needs
- to measure all masses in the valve train (this is the easy part)
- to know the rocker ratio (ok, this is even more easy)
- to be able to determine by 3D CAD tools or by measuring the rocker mass moment of inertia (this is the tricky part and no, there's no concept such as "MOI at the tip" nor "tip weight" has any relevance here)
- to measure spring rate and seat load (need to have right equipment)
- to measure actual lobe lift profile and use selected methods for data filtering
- to have an engine simulator printout showing pressure levels before the intake valve and in the combustion chamber + cfm/cylinder (in NA engines the effect is smaller, but in FI engines it's significant)
The rest is physics and how to apply it into real world case.
Hope that makes it a bit clearer.
Last edited by barum; 07-10-2018 at 01:56 PM.
Reason: more explanations and clarifications added, typos corrected
On my "to do" list somewhere there is FFT-analysis of the lobe profile (at various rpm points) which will show frequencies (and their harmoncis = multiples of the base frequency) which can excite nominal frequencies in valve train components. Spring is the obvious candidate. Modeling the multiple mass/multiple spring system including frictions and non-linear behavior is another story and would need Simulink.
It's good to remember that also the pushrod and the valve stem vibrate, but can have much higher nominal frequencies than spring.
On my "to do" list somewhere there is FFT-analysis of the lobe profile (at various rpm points) which will show frequencies (and their harmoncis = multiples of the base frequency) which can excite nominal frequencies in valve train components. Spring is the obvious candidate. Modeling the multiple mass/multiple spring system including frictions and non-linear behavior is another story and would need Simulink.
It's good to remember that also the pushrod and the valve stem vibrate, but can have much higher nominal frequencies than spring.
In the above diagrams blue curve represents the force generated by the spring as function camshaft rotation (degrees). Since the force generated by the spring is in the opposite direction from deflection (compared to it's natural state = rest) => F = -kx, where the k is the spring rate and x is the amount of deflection = actual valve lift. When spring force is reduced to the lobe side of the rocker, rocker ratio is needed.
The green curve represents the sum of forces generated by each of the moving masses within the valve train at given engine rpm. Note that valve train operaters at half speed compared to crankshaft. Basically F= ma, where m is the mass of body and a is the acceleration of that body. Masses on both sides of rocker fulcrum have to be redused to the side on which the calculation is made, typically on the lobe side. Here the rocker ratio comes into play again. The rocker itself moves in circular motion, so mass moment of inertia must be known. The angular acceleration for the rocker can be calculated from the rest of the valve train dimensions and movements.
During the opening (and closing) the inertia of mass (of all the components in the valve train) creates force towards the cam lobe (as does the spring, too), meaning all the components between the spring seat and cam lobe "see" their combined load (the distance between the green and blue curve). This is the phase which stresses the components most.
When the lobe rotates closer to the peak lift (located at 0 degrees in the diagrams), the speed at which the valve opens starts to slow down. This means that there is negative acceleration = deceleration (remember, F=ma, and now a is negative). If the (here negative) total force from the valve train exceeds the force the spring delivers (green curve goes below the blue curve), the follower wheel will no longer follow the cam lobe and in worst case coil bind can occur for this reason alone. Of course bad things continue to happen since lobe continues to rotate while the lifter wheel is airborne and when they meet again, there's nothing friendly about it.
The green curve also includes the worst case scenario of static and dynamic pressure difference over the valve.
The lobe profile is the "driver" behind all that takes place so lift curve must be known. The best way is to have a lobe measured by CamDoctor or equivalent. Note that the measurement is "noisy" (contains small errors from various error sources), so most use running average as a filter to be able to calculate smoother looking 2nd derivative (which is acceleration, see above). This is quite ok for spring dimensioning, but if one want's to dig deeper, Matlab or Octave are useful tools in order fit groups of polynoms in the measurement data. This allows me to filter out so called outliers and I'm able to get high quality 3rd derivate (known as jerk = change rate of acceleration).
Here is an example showing lobe lift (inch) and corresponding acceleration (inch/deg^2) with Excel. (Note the Y-axis units which are different than in the complete valve train force diagram):
The rest is physics and how to apply it into real world case.
Hope that makes it a bit clearer.
It does. And to me it also illustrates a high coefficient of friction increasing the effort/energy to initiate valve movement in the first peak as the lobe is coming up. This in my mind shows that friction is a major problem in the OEM design because if it were an ideal valve event, it should mirror the lobe profile (minus the force exerted by the spring and lifter of course) represented on the down travel of the lobe. Notice how it just sets the valve down (sharp downward peak) at about 55 deg? The remaining force is likely just the pressure in hydraulic lifter returning to base circle. If the rocker was more efficient, the opening peak should more resemble the closing peak. That's why I'm curious if you are able to model shaft or rollers.
The SPR rockers are vaporware from Comp. Never released due to some problems found in testing. Someone put the PR cart before the horse. Really REALLY disappointed as I was looking forward to running them. Reducing rocker MOI is something that interests me greatly and I had zero interest in running heavier than stock aluminum rockers which on top of that have less favorable geometry increasing the MOI.
What's funny about this thread (it has been a doozy, I love it) is that no one has posted weights of any of the rockers listed. The TD rockers for the COPO engines are near identical in weight of the factory rocker arm. Is there more weight over the nose due to the rocker tip? Yes. Is it lighter and stronger than the majority of the aluminum rocker arms out there? I'd put money on it. Would I love to have the stainless pro series Jesel shaft rockers, duh...but if people are taking stock rockers to 8000 rpm without hammering the bolts out of the head then I see no reason why the TD pedestal mount rocker won't be sufficient.
In my mind they're the most robust, lightest and have the least inertia weight of any of the bolt on adjustable rockers on the market at a very reasonable price point. Roller tips are required for a solid roller approaching 800 lift so the argument of stock vs roller tip is really not an issue.
Dismissing the TD pedestal mount as crap is just dumb.
Weight of two rockers with for all intents and purposes identical adjusters, roller tips and fulcrum shafts with their only difference being material and shape of the rocker body is irrelevant information?
Stiffer and tougher material, more centralized mass around the fulcrum, and more favorable geometry with regards to MOI calculations. I’ll take the steel over aluminum any day without needing a cad model to decide what is better.
The SPR rockers are vaporware from Comp. Never released due to some problems found in testing. Someone put the PR cart before the horse. Really REALLY disappointed as I was looking forward to running them. Reducing rocker MOI is something that interests me greatly and I had zero interest in running heavier than stock aluminum rockers which on top of that have less favorable geometry increasing the MOI.
What's funny about this thread (it has been a doozy, I love it) is that no one has posted weights of any of the rockers listed. The TD rockers for the COPO engines are near identical in weight of the factory rocker arm. Is there more weight over the nose due to the rocker tip? Yes. Is it lighter and stronger than the majority of the aluminum rocker arms out there? I'd put money on it................
I went through the excersise of finding an alternative roller tip 4 years ago for my personal car. I have not had good luck with aluminum rockers in in the past, so I stick with steel. Ended up going with the Crower system. Wish I could have done some spin tron testing, but that just wasn't something I could justify, financially. I can say that they didn't like 130 seat load. I beat the hell out of the seals in under 4k miles of use. I bumped seat load up to 160, and that amount appears to be adequate. I started a thread 4 years ago detailing my findings, cautioning others about using oe rockers and high lift cams. https://www.corvetteforum.com/forums/c6-corvette-zr1-and-z06/3488299-info-oem-ls7-rockers-and-max-lift.html
Weight of two rockers with for all intents and purposes identical adjusters, roller tips and fulcrum shafts with their only difference being material and shape of the rocker body is irrelevant information?
Stiffer and tougher material, more centralized mass around the fulcrum, and more favorable geometry with regards to MOI calculations. I’ll take the steel over aluminum any day without needing a cad model to decide what is better.
Steel is the stiffest and most durable rocker material for the cycle life of a street engine. I don't think I would ever consider an aluminum body rocker on a street engine again. One expert referred to YT's as "diving boards" on the spintron.
Great info these past few posts. MTPZ06 summed it up perfectly with one picture.
It does. And to me it also illustrates a high coefficient of friction increasing the effort/energy to initiate valve movement in the first peak as the lobe is coming up. This in my mind shows that friction is a major problem in the OEM design because if it were an ideal valve event, it should mirror the lobe profile (minus the force exerted by the spring and lifter of course) represented on the down travel of the lobe. Notice how it just sets the valve down (sharp downward peak) at about 55 deg? The remaining force is likely just the pressure in hydraulic lifter returning to base circle. If the rocker was more efficient, the opening peak should more resemble the closing peak. That's why I'm curious if you are able to model shaft or rollers.
The hydraulic lifter is not modeled and the lobe measurement is done with same diameter tip as the wheel diameter is. Cam profiles are not totally symmetric as can be seen from the following diagrams where I have inverted the Cam 2 data series'.
Comp LXL is the most symmetric, Cam Motion clearly more asymmetric and OEM has biggest differences in opening vs closing ramp hardness.
My measurement showed 143-144 grams, but that was for the bone stock rockers. You have trunnion kits on them and so just a little longer fulcrum shaft (and possibly minor differences in the design) + two circlips increase the total mass 22..23 grams.
EDIT: My measurements, first stock:
I have a second set of BTR prepared rockers with their trunnion kit and I will measure them today to get one more data point and see how much those have gained weight due to that kit.
EDIT: Stock with BTR trunnions:
EDIT: Stock, body only:
Anyway, my point is that even a small difference in fulcrum shaft diameter or length effects these total mass measurements, so without actually dismantling them and measuring each part there's just too much room for error when visually estimating MOI. Especially when the OEM rocker body only weighs 84..85 grams vs. total mass including fulcrum shaft and related components.
If you still have these in your possession, 2D drawings with dimensions would be interesting. From these it would be possible to try to model them in 3D CAD. Of course the best option would be to have them 3D scanned and then process further in 3D CAD.
I've been thinking a test rig in which a low power motor would be used the accelerate rockers (attached from the fulcum with shaft & bearing dismantled) to a given speed and then decelerate back to zero. Knowing the motor torque, used gearing and needed times for acceleration & decelaration one would be able to calculate the MOI. Of course a base level needs to be made with empty rig, but the time differences would directly indicade the change in MOI.
Just the time differences would give the percentual change over the stock rocker, which would be adequate as well (if you the MOI of the stock rocker, that is).
Finally, I have to say I am positively surprised about the TD rocker weight and if one were to make longer fulcrum shafts joining the adjacent intake & exhaust rockers, these could be quite interesting choise. Generally I believe the simplest structure to get the job done is quite often also the most robust.
Last edited by barum; 07-11-2018 at 07:15 AM.
Reason: Measurement pictures added
The hydraulic lifter is not modeled and the lobe measurement is done with same diameter tip as the wheel diameter is. Cam profiles are not totally symmetric as can be seen from the following diagrams where I have inverted the Cam 2 data series'.
Comp LXL is the most symmetric, Cam Motion clearly more asymmetric and OEM has biggest differences in opening vs closing ramp hardness.
Okay, lifter not a consideration in the depicted graph. The peak I mentioned is still friction increase, is it not?
07-09-2018, 10:57 PM
