Feasibility of pushrod/cam-in-block engines in Formula One [OHV/overhead-valve]

All that has to do with the power train, gearbox, clutch, fuels and lubricants, etc. Generally the mechanical side of Formula One.
Tommy Cookers
617
Joined: 17 Feb 2012, 16:55

Re: Feasibility of pushrod/cam-in-block engines in Formula One [OHV/overhead-valve]

Post

the UK motorcycle dealership B.M.G. made such a thing in the 1960s for the (2 valve) Velocette motorcycle
which was competitive in the 500 cc class in production/endurance racing
they sent one to the Velocette factory - the package was eventually returned unopened

User avatar
coaster
16
Joined: 30 Jun 2012, 05:10

Re: Feasibility of pushrod/cam-in-block engines in Formula One [OHV/overhead-valve]

Post

Here in Oz we have improved production racing, motor from same manufacturer, same layout and number of valves, all else is free of design.
You hear of strange swaps in this category, frankenstein engines built to have an edge.
I can see merit in this, but for the sake any other reason,, uncertain.

Hoffman900
163
Joined: 13 Oct 2019, 03:02

Re: Feasibility of pushrod/cam-in-block engines in Formula One [OHV/overhead-valve]

Post

coaster wrote:
05 Jul 2020, 08:47
Jade Gurrs wrote a good book on the ilmor indy 2 valve V8, they built a valvetrain dyno rig and tested until destruction, once they reached indy 500 race distance they locked in the design.
These are old news now. Bob Fox, owner of Trend Performance built and sold the first commercially developed Spintron in the early 1990s. They now have been updated and have much better control / measurement software. Besides tracking valve motion, they also allow engineers to install strain gauges on valve springs, pushrods, valve spring seats, etc. They have been hugely instrumental in making pushrod valvetrains work very well, where they are mandated per rules. OHC valvetrains as well.
I agree, my point was that the inertia of the valve train has very little effect on the overall inertia of the engine in reply to Smokes' post.
As I have said in my previous posts however, the inertia of a pushrod system has a very large negative impact on engine performance due to the spring force requirement.
On the cam side of the rocker, it doesn't have that much effect. NASCAR, before the gear rule (so rpm restricted), was utilizing 110lb springs on the seat, ~500lb-ish open. Over 1" of valve lift, intake valve weight of ~65grams, and spinning to nearly 10,000rpm.

Again, when the top series switched to roller lifters from flat tappets, they didn't gain anything, as the valve spring, through rocker ratios in excess of 2:1 were already at their limit. Yes, the extra inertia of the roller required some tweaking, but that was sorted out.

A bucket OHC engine is just a giant flat tappet camshaft, and the diameter of the bucket is what dictates maximum cam velocity. There are tricks around this utilizing dwell at constant velocity, etc. but it's still a limit. OHC systems due to their stiffness are not as acceleration limited. Jerk is never really good. A problem people have had with making OHC buckets larger and larger in diameter is the bucket beats like a drum, and causes all sorts of issues.

Roller valvetrains are acceleration limited, not velocity.

The pushrod can be made to work very effectively. Is it better than a OHC finger follower set up? No, but a good pushrod valvetrain is better than a poor OHC valvetrain. The real issue is actuating 4 valves and the pushrods fouling the port / cooling jacket locations.
Last edited by Hoffman900 on 09 Jul 2020, 19:19, edited 1 time in total.

63l8qrrfy6
368
Joined: 17 Feb 2016, 21:36

Re: Feasibility of pushrod/cam-in-block engines in Formula One [OHV/overhead-valve]

Post

Hoffman900 wrote:
09 Jul 2020, 17:56
I agree, my point was that the inertia of the valve train has very little effect on the overall inertia of the engine in reply to Smokes' post.
As I have said in my previous posts however, the inertia of a pushrod system has a very large negative impact on engine performance due to the spring force requirement.
On the cam side of the rocker, it doesn't have that much effect.
Not much compared to the valve side at extreme rocker ratios, but compared to an OHV setup it is still a massive difference.

Hoffman900 wrote:
09 Jul 2020, 17:56
Roller valvetrains are acceleration limited, not velocity.
All valvetrains are both acceleration and velocity limited. The higher the acceleration the higher the forces.
Regardless of how the valve is driven there is a seating velocity beyond which the valve seat wears too quickly so all valetrains are also velocity limited in this regard.

Hoffman900
163
Joined: 13 Oct 2019, 03:02

Re: Feasibility of pushrod/cam-in-block engines in Formula One [OHV/overhead-valve]

Post

Mudflap wrote:
09 Jul 2020, 18:29
Hoffman900 wrote:
09 Jul 2020, 17:56
I agree, my point was that the inertia of the valve train has very little effect on the overall inertia of the engine in reply to Smokes' post.
As I have said in my previous posts however, the inertia of a pushrod system has a very large negative impact on engine performance due to the spring force requirement.
On the cam side of the rocker, it doesn't have that much effect.
Not much compared to the valve side at extreme rocker ratios, but compared to an OHV setup it is still a massive difference.

Hoffman900 wrote:
09 Jul 2020, 17:56
Roller valvetrains are acceleration limited, not velocity.
All valvetrains are both acceleration and velocity limited. The higher the acceleration the higher the forces.
Regardless of how the valve is driven there is a seating velocity beyond which the valve seat wears too quickly so all valetrains are also velocity limited in this regard.
Well, yeah, you can't design anything with infinite jerk / acceleration / velocity, but as far as constraints go, on a flat tappet / bucket set up, velocity is your constraint or you run off the edge. On a roller, acceleration as there are radius of circle / pressure angles to contend with.

Maximum velocity is pretty easy to figure out for a bucket / flat tappet design:

Maximum velocity ("/*) = (((Tappet or Bucket Diameter)/2) - edge clearance) /57.3

Edge clearance is matter of the design, and how tight the tolerances are. Most production cams are in the .018"-.020" range. NASCAR guys ran as close as they dared. To quote Billy Godbold (Competition Cams designer), he would lose sleep worrying about engines blowing up on Sunday due to how close they were pushing it.

Your high end pushrod lobe profiles are assymetrical for dynamic reasons. Softer openings and closings faster (relative to the opening) yield the best compromise between performance and valvetrain dynamics. These are the types of designs being used in everything from the top Top Fuel teams, to Daytona 24hr winning prototypes, NASCAR, the Corvette factory team, etc.

Here is an older flat tappet NASCAR design on the left. Note the constant velocity sections. This is from Billy Godbold's AETC presentation in 2017.
Image

63l8qrrfy6
368
Joined: 17 Feb 2016, 21:36

Re: Feasibility of pushrod/cam-in-block engines in Formula One [OHV/overhead-valve]

Post

It's just semantics and I'll demonstrate:

I agree that for a fixed tappet diameter with a fixed curvature (I am not sure why we are still discussing flat tappets at this point, they were obsolete before I was born) you can't increase velocity past a certain point BUT:

What limits tappet diameter and tappet curvature - both of which increase the maximum allowable velocity?
Of course, the larger the diameter, the higher the mass so the higher the force for a given acceleration. So provided there are no physical constraints for diameter, the acceleration becomes the limiting factor.

Alternatively, to increase the allowable velocity for a given diameter, one can reduce the radius of curvature of the tappet. This will increase the hertz stresses which are also driven by inertial loads so again acceleration becomes the limiting factors.

So yes, in a spec series where the tappets have to be flat and they have do be a certain diameter the maximum velocity is a limiting factor but in the real world it is definitely not.

Hoffman900
163
Joined: 13 Oct 2019, 03:02

Re: Feasibility of pushrod/cam-in-block engines in Formula One [OHV/overhead-valve]

Post

Mudflap wrote:
09 Jul 2020, 20:14
It's just semantics and I'll demonstrate:

I agree that for a fixed tappet diameter with a fixed curvature (I am not sure why we are still discussing flat tappets at this point, they were obsolete before I was born) you can't increase velocity past a certain point BUT:

What limits tappet diameter and tappet curvature - both of which increase the maximum allowable velocity?
Of course, the larger the diameter, the higher the mass so the higher the force for a given acceleration. So provided there are no physical constraints for diameter, the acceleration becomes the limiting factor.

Alternatively, to increase the allowable velocity for a given diameter, one can reduce the radius of curvature of the tappet. This will increase the hertz stresses which are also driven by inertial loads so again acceleration becomes the limiting factors.

So yes, in a spec series where the tappets have to be flat and they have do be a certain diameter the maximum velocity is a limiting factor but in the real world it is definitely not.
We're discussing them because that's what OHC direct acting (bucket) type valvetrains are. They're just direct acting, flat tappet camshafts with a large diameter lifter (bucket).

The bore size is going to dictate the valve spacing and diameter, which is ultimately going to dictate the maximum diameter of the buckets. Valve cant plays a role too and I know some F1 engines had this.

You cannot increase velocity on these designs as I pointed out without running off the edge. This wipes out camshaft lobes. It's literally a physical constraint you cannot go past. However acceleration and the rate of acceleration (jerk) are still at play, and is dictated by design and material selections. You can tune fourier shape forces by spring shape and material selection, and in pushrod application; material, shape, thickness.

This is what top pushrod racing valvetrain engineers do and they actually tune the system with the pushrod, rocker, and spring shape / thickness / material. They tune the whole system to resonate at certain points, cancel out competing harmonics through the system. If the whole thing resonates at the same time, it breaks. This is on top of everything else they're dealing with. Good ol' boys in NASCAR need not apply and haven't for 35 years.

In the real world, you can only make lifters / buckets so big due to physical constraints, and thus your maximum velocity will also be dictated by this. Pushrod engines with rockers and finger follower type applications, there is a multiplication effect, so you have to account for what the spring can handle and work backwards through to design the lobe.
Last edited by Hoffman900 on 09 Jul 2020, 22:23, edited 3 times in total.

User avatar
PlatinumZealot
550
Joined: 12 Jun 2008, 03:45

Re: Feasibility of pushrod/cam-in-block engines in Formula One [OHV/overhead-valve]

Post

Rodak wrote:
25 Jun 2020, 03:11
PlatinumZealot wrote:
19 Jun 2020, 02:36
Well think of all the finer reasons why they did not continue to be used.
How does it go with wear?
Can four valves be actuated easily with that design? Two valve have interesting charecteristics but not as balanced as four valves.
Backlash In quick throttle changes?
Might be some reasons why....
Two weeks after the race USAC reduced the boost from 55 inches to 52 inches for purpose built pushrod motors. Four months later Tony George, the owner of the Indy track started the IRL and began the CART/IRL wars. He changed the rules to allow only 2.2 L turbo V-8 engines from 1996 and changed qualification so that the top 25 IRL teams would be guaranteed entry to the 500 regardless of qualification speed....... For the 1995 race boost was reduced to 48 inches, effectively killing the pushrod motor. Ilmor had orders from 17 teams for engines for 1995 and had started production when the new boost levels were announced; they had a lot of scrap castings....
I am trying to see the connection of how reducing the boost for all the engines killed the pushrod. How do you explain that one?
🖐️✌️☝️👀👌✍️🐎🏆🙏

User avatar
PlatinumZealot
550
Joined: 12 Jun 2008, 03:45

Re: Feasibility of pushrod/cam-in-block engines in Formula One [OHV/overhead-valve]

Post

Hoffman900 wrote:
04 Jul 2020, 20:35


Again, the engine doesn't care how the valve is lifted. It only sees the valve lift profile.
What about.. valve train losses and durability? The engine must care about this. Or are you referring to the combustion only?
With the push rod system, there is some disconnect between the lobe profile and the valve profile through the rpm range, due to system stiffness, this is all measurable and compensated for, and in NASCAR's case, it is used to their advantage through controlled loft that gains open valve area with increasing rpm. Regardless, top pushrod racing engines are valve spring limited, not anything else.
Can you explain how this is done? Do they rely on some sort of valve float?
We're talking pushrod actuated cam systems here. Pushrod does not have to mean 2 valve per cylinder.

A SOHC like Honda's CRF450, provides a narrow / compact top end like a pushrod actuated system does over DOHC, but with the system stiffness of a OHC platform as well as rocker multiplication through finger followers:
https://motocrossactionmag.com/wp-conte ... m-sohc.jpg
A lot of hondas and other japanese cars and even BMW used this sort of setup. SOHC with rockers. There is a tuner called Bisimoto (he tunes porsche now but used to tune honda) who capitalized on the light weight of the SOHC design, and once had the fastest "All motor" Hondas. He didn't care about indivually tuning the exhaust and intake cam because he could just build a custom cam, and he didn't care about variable valve timing of course... Sometimes I wonder why F1 doesn't use SOHC? Maybe because the want to avoid rockers? Though rockers are not necessarily that bad as you say.
🖐️✌️☝️👀👌✍️🐎🏆🙏

J.A.W.
109
Joined: 01 Sep 2014, 05:10
Location: Altair IV.

Re: Feasibility of pushrod/cam-in-block engines in Formula One [OHV/overhead-valve]

Post

Nearly 40 years ago, as a young, enthusiastic dirt biker, I bought as a hi-po project,
a (non-running) Honda CR 250 MX bike that'd had a Yoshimura-spec Honda XL 250
SOHC 4V 4T engine shoehorned into it..

I spent too much time/money in trying to do what Honda does today, by making a
lightweight high-revving 4T MX machine, it had a carefully ported-by-Yoshimura head
& their lightweight slipper piston (hi-comp) plus fat (high lift, long duration) cam,
along with very stiff multiple springs, & a massive (for the time) 36mm Dellorto carb
(with its 4T friendly accelerator pump, from a dead Ducati)..

Ok, when running, this device would rip, but at the same time, the high frequency
vibes would crack the chrome-moly chassis, & the stock Honda cam 'finger-followers'
would invariably snap under the combined strain of cam/spring/rpm stress, (even if
new/polished, media blasted, relieved/reinforced & heat-treated), yet, I pigheadedly
persevered (with a 4T-fan father's assistance) - 'til I dropped a valve through the piston.

Naturally, I went 2T - forthwith - despite dad's disdain...
"Well, we knocked the bastard off!"

Ed Hilary on being 1st to top Mt Everest,
(& 1st to do a surface traverse across Antarctica,
in good Kiwi style - riding a Massey Ferguson farm
tractor - with a few extemporised mod's to hack the task).

User avatar
coaster
16
Joined: 30 Jun 2012, 05:10

Re: Feasibility of pushrod/cam-in-block engines in Formula One [OHV/overhead-valve]

Post

I'll agree with your statement PZ, the Bugatti sohc and Rollsroyce Merlin sohc setup lends itself well to a narrow valve angle and seems like a design worth revisiting.
The only drawback is removing the cam to change a centrally located spark plug, the 2 examples had side plugs, obsolete from poor combustion.

63l8qrrfy6
368
Joined: 17 Feb 2016, 21:36

Re: Feasibility of pushrod/cam-in-block engines in Formula One [OHV/overhead-valve]

Post

Hoffman900 wrote:
09 Jul 2020, 22:17
The bore size is going to dictate the valve spacing and diameter, which is ultimately going to dictate the maximum diameter of the buckets. Valve cant plays a role too and I know some F1 engines had this.

You cannot increase velocity on these designs as I pointed out without running off the edge. This wipes out camshaft lobes. It's literally a physical constraint you cannot go past. However acceleration and the rate of acceleration (jerk) are still at play, and is dictated by design and material selections. You can tune fourier shape forces by spring shape and material selection, and in pushrod application; material, shape, thickness.

...

In the real world, you can only make lifters / buckets so big due to physical constraints, and thus your maximum velocity will also be dictated by this. Pushrod engines with rockers and finger follower type applications, there is a multiplication effect, so you have to account for what the spring can handle and work backwards through to design the lobe.
This is going nowhere.
This whole velocity limited thing started off for the simple reason that in NASCAR they imposed bucket diameter and curvature. I can't stress this enough but it not an issue in the real world!

To convince yourself go and pick a NASCAR roller lifter cam profile with the highest velocity you can find and back-calculate the required bucket diameter to see if it is a reasonable dimension. Until then we are just arguing over how long a piece of string is.

I'll work out an example for you:
In the lift profiles you showed above the peak velocity is 0.008 in/deg. With sensible clearance the required bucket diameter is 0.97in. The NASCAR max diameter was 0.875 in. In NASCAR that would be illegal, in the real world to make the diameter 0.095in larger is an engineering graduate's morning warm up job.
To put this into perspective, the 1 inch 1966 Cosworth DFV tappets would have been more than adequate for this profile.

Hoffman900
163
Joined: 13 Oct 2019, 03:02

Re: Feasibility of pushrod/cam-in-block engines in Formula One [OHV/overhead-valve]

Post

Mudflap wrote:
10 Jul 2020, 14:38
Hoffman900 wrote:
09 Jul 2020, 22:17
The bore size is going to dictate the valve spacing and diameter, which is ultimately going to dictate the maximum diameter of the buckets. Valve cant plays a role too and I know some F1 engines had this.

You cannot increase velocity on these designs as I pointed out without running off the edge. This wipes out camshaft lobes. It's literally a physical constraint you cannot go past. However acceleration and the rate of acceleration (jerk) are still at play, and is dictated by design and material selections. You can tune fourier shape forces by spring shape and material selection, and in pushrod application; material, shape, thickness.

...

In the real world, you can only make lifters / buckets so big due to physical constraints, and thus your maximum velocity will also be dictated by this. Pushrod engines with rockers and finger follower type applications, there is a multiplication effect, so you have to account for what the spring can handle and work backwards through to design the lobe.
This is going nowhere.
This whole velocity limited thing started off for the simple reason that in NASCAR they imposed bucket diameter and curvature. I can't stress this enough but it not an issue in the real world!

To convince yourself go and pick a NASCAR roller lifter cam profile with the highest velocity you can find and back-calculate the required bucket diameter to see if it is a reasonable dimension. Until then we are just arguing over how long a piece of string is.

I'll work out an example for you:
In the lift profiles you showed above the peak velocity is 0.008 in/deg. With sensible clearance the required bucket diameter is 0.97in. The NASCAR max diameter was 0.875 in. In NASCAR that would be illegal, in the real world to make the diameter 0.095in larger is an engineering graduate's morning warm up job.
To put this into perspective, the 1 inch 1966 Cosworth DFV tappets would have been more than adequate for this profile.
So dismissive...

Peak velocity is around .00745 in/deg. It's the purple line. Working backwards, this gives an edge clearance of .010". As I said before, they push the boundaries with these things, and they can because the tolerances, very strong CGI blocks, and no chamfer on the lifter edge.Since you love the real world so much, these are real world figures.

Also, this is at the lobe, not the velocity seen at the spring / valve which is a whole other matter. With a OHC bucket type valvetrain, this is 1:1, what you see at the lobe is what you get at the valve. NASCAR, with the FT camshafts, were using in excess of 2.37:1 rocker ratios. .00745" velocity at the lobe would be almost .01778in/deg at the valve.

Of course a 1" lobe would be adequate, and considering the vintage, it would probably need that margin due to tolerances. NASCAR builders did use 1" mushroom tappets up until they were outlawed in the 1980s. This was the maximum they could go as they already had to narrow the lobes to keep the lifters from fouling on each other.

A 1" bucket / tappet, using the same tolerances as that NASCAR engine would be .00855 in/deg.

Again, as mentioned above, with mushroom lifters of 1", you run out of real estate to the adjacent lobe. Real world.
On a OHC application, the bucket can only be so big before you foul on the adjacent bucket. Real world.

If you worked backwards checking velocity, a roller cam may have an effective tappet diameter of 50in+. A single bucket / tappet alone would be the size of the entire engine.

However if you look at acceleration, they are similar and since bucket / flat tappets are not acceleration constrained, they can reach peak velocity sooner. However, since the roller isn't velocity constrained, it can keep increasing to the point that the system needs to also slow it down to get over the nose. With faster acceleration, you can move the valve faster and spend more time slowing it down, thus increasing open area at the top of the lift curve.

If you look at the lobe profile from Honda's Third Generation Formula 1 engine, you saw maximum intake valve velocity of 15mm/rad (.0103"in/deg) and acceleration of around 55mm/rad^2 (.03779in/deg^2). This doesn't account for multiplication through the finger follower. An older NASCAR application when multiplied through a 2.37:1 ratio rocker arm, might only be .01778in/deg (velocity) and .00209in/deg^2 (acceleration). Pneumatic "springs" are nice. This gets back to my initial post in this thread.
PlatinumZealot wrote:
10 Jul 2020, 01:57
Hoffman900 wrote:
04 Jul 2020, 20:35


Again, the engine doesn't care how the valve is lifted. It only sees the valve lift profile.
What about.. valve train losses and durability? The engine must care about this. Or are you referring to the combustion only?
What I mean is you design the valve lift curve and work backwards through the valvetrain to determine lobe shape. If the geometry was such that the lobe had to be square to get the valve lift shape, then so be it (obviously this is a hypothetical scenario). You can really get a sense of this with OHC rocker valvetrains / finger follower. To get a symmetrical valve lift profile (which isn't exactly ideal, but for conversation sake), the lobe will be asymmetrical due to the varying rocker ratio. If you look simulated mass flow into the engine, you'll see the mass flow doesn't exactly follow the valve lift curve exactly, so that's a whole other matter.
PlatinumZealot wrote:
10 Jul 2020, 01:57
Hoffman900 wrote:
04 Jul 2020, 20:35
With the push rod system, there is some disconnect between the lobe profile and the valve profile through the rpm range, due to system stiffness, this is all measurable and compensated for, and in NASCAR's case, it is used to their advantage through controlled loft that gains open valve area with increasing rpm. Regardless, top pushrod racing engines are valve spring limited, not anything else.
Can you explain how this is done? Do they rely on some sort of valve float?
They use the pushrod to pole vault the valve as rpms increase. This can all be controlled so where it's done, how it's done, and how / where the valve lands on the lobe. In a typical pushrod valvetrain, valve open area (as expressed as duration) is typically lost as rpms increase due to bending.With controlled loft, they are doing their best to maintain the duration figure but using the pole vaulting to increase peak valve lift, thus producing a net gain in open area. Obviously this takes A LOT of engineering to do, and from the valve side, it's a bit scary as they run very tight P-V clearances.
PlatinumZealot wrote:
10 Jul 2020, 01:57
A lot of hondas and other japanese cars and even BMW used this sort of setup. SOHC with rockers. There is a tuner called Bisimoto (he tunes porsche now but used to tune honda) who capitalized on the light weight of the SOHC design, and once had the fastest "All motor" Hondas. He didn't care about indivually tuning the exhaust and intake cam because he could just build a custom cam, and he didn't care about variable valve timing of course... Sometimes I wonder why F1 doesn't use SOHC? Maybe because the want to avoid rockers? Though rockers are not necessarily that bad as you say.
Bisi had the fastest SOHC in drag racing, but most were using DOHC for obvious reasons and were much faster. They were just in other classes. He did good work on that, but DOHC makes tuning infinitely easier and cheaper as you don't need to have a new cam cut every time you want to change centerlines. Most serious blank sheet performance engines (F1, MotoGP) and hot production engines (Sportbikes, etc.) are all using finger followers for reasons I have mentioned in other posts. It's the best of both worlds - very stiff / low friction valvetrain and rocker multiplication (which gets around the limitations of the bucket diameter).
Last edited by Hoffman900 on 10 Jul 2020, 17:06, edited 5 times in total.

User avatar
PlatinumZealot
550
Joined: 12 Jun 2008, 03:45

Re: Feasibility of pushrod/cam-in-block engines in Formula One [OHV/overhead-valve]

Post

Ok thanks, but Hoffman.

You lost me on the pole vault jargon, so I had to google...

https://books.google.com.jm/books?id=nl ... in&f=false

Sou are saying at high rpm, the deceleration of the valve lifters, and the inertial load, and the valve spring, compresses the push-rod causing it to buckle like the pole in pole vault, this stored energy is released again "launching" the valve, increasing valve-lift, when it would be otherwise slightly reduced at high rpm?

Image
🖐️✌️☝️👀👌✍️🐎🏆🙏

Hoffman900
163
Joined: 13 Oct 2019, 03:02

Re: Feasibility of pushrod/cam-in-block engines in Formula One [OHV/overhead-valve]

Post

What you see on a Spintron is that they take a baseline reading at the valve at something like 3000rpm.

Then as the rpms climb, you see seat to seat duration decrease due to flexing, but you'll see the valve come "unglued" to the baseline profile and carry over the top of the lift peak and land back on the lobe. The problem is you don't want seat area to decrease with increasing rpm, so you try to make up for it with increased peak area.

Here is a bad example. They're doing it without the bounce and controlling where it lands.

Image

Obviously, this type of development is out of the reach for most people / teams / series.

Start reading on Page 38: http://www.engineprofessional.com/EPQ1- ... 8f93e3.pdf

This is a 436ci engine Roush Yates developed for dirt track racing in the US.
Last edited by Hoffman900 on 10 Jul 2020, 18:49, edited 2 times in total.