F1 engine RPM

[Brian Coat]

I recently noticed/calculated that the mean piston speeds touched in a TAFC engine (NHRA) are in excess of those on an F1 engine.

I surmise this is because long life is not a major consideration - these engines only run for a few seconds at maximum rpm levels. None the less, 35+ m/s is impressive.

[Tommy Cookers]

AGREED !! ........ We have too long been (unreasonably) frightened of high piston speed

the crucial factors in engine design are still and always were related to rpm (not to piston speed as such), that is ......

at what piston acceleration does the compression ring start to pant and break ?
at what acceleration do rods start to fail ?
at what rpm does the crank start to fail ?

For most of the last century motoring writers focussed on the easy target, piston speed, and for road cars it was important, concerning wear and friction. They ignored the benefit of stroke length to compression ratio (important with poor fuel quality).
In competition 'longstroke' engines often punched above their weight because rpm is limited by stress ie piston acceleration, not by piston speed. Jaguar, Norton, Offenhauser had rather high piston speeds, and it didn't hurt them.
For half a century race engines had mean piston speeds around 4000 fpm.
Progress in shortening strokes had the effect of stabilising piston speed as piston acceleration increased with progress in materials, design, and manufacture.

In recent years this progress has been dramatic.
F1 has gone to unprecedented bore:stroke ratios (now 2.46), and we have finally seen clearly that piston speed is unimportant.
While piston accelerations are very high in F1, due to the very short stroke piston speeds not very high.
In Nascar etc we have the same very high accelerations but also very high piston speeds as the stroke is not short, (B:S ratio of 1.4)

Many attempts at raising B:S ratios in eg 60s F1 failed because these engines couldn't combust well enough to make good use of their potentially higher rpm. Also the rods were (relatively) too long, worsening the effects of the slow combustion.
Some motorcycles eg the 3 cyl 500cc MV Agusta were conspicuously good, then Weslake and Cosworth drove progress.
Later we needed, and got, fast-combusting fuel, pneumatic valve springs, and internal coatings against heat pickup, but very high B:S ratios only appeared (as the remaining development path) with years of rule stability after the turbo era. This led to even more compact engines, helping aerodynamically. Valve area benefits,too.

One hundred years ago GPs had faded, and the racing focus was on French 'light car' classes that were technically related to the 'horsepower' road tax. This drove them to very long strokes (eg 80mm x 280mm Lion-Peugeot V twin), as the tax and the race regulations allowed much larger displacements with these extremely low B:S ratios.
Yes, they had very high piston speeds and it didn't hurt them.

This tax-related low B:S ratio became a fashion that rapidly revitalised the GP class, and stayed 50 years too long (it was useful with the very low octane fuel of those earlier days).

Designers and their employers like to follow fashion, whatever it is.

Anyway, my thanks to those developments in the 21st century that have finally shown us what matters and what doesn't !

Interesting times !!

[Agerasia]

Its not hitting the the RPM LIMIT, which as you said is typically under the max RPM value, that does the damage, its the effect of hitting the RPM LIMITER and the resulting sudden deceleration/acceleration of the various components (valve train, piston/rod assembly etc) from having either spark and/or fuel cut to the engine. It's not the same as a gear change or deceleration event. The engine drops rpm and reverses loadings very quickly and is then it is reversed and loaded up again as the RPM drops under the RPM cut in rapid succession multiple times causing severe loadings in opposing directions and resulting is huge stresses on various components.

I have seen first hand the effect bouncing an engine off the RPM limiter can do in a both a large capacity V8 and a modified Honda K20A with a 10,500RPM limit. The Honda sheared off the retainers and valve tips (quality titanium/inconel units) dropping a number of valves into the engine. Suffice to say there was little left to use again.

Then by your argument the pit lane limiter should cause engine damage?

[strad]

Then by your argument the pit lane limiter should cause engine damage?

I have always felt that the ones that randomly drop cylinders do in fact hurt the engine.
But the pit lane limiter is operating at so much below the strain limits that I don't think it does. The driver doesn't have his foot on the floor and the button pushed...I bet it would be tough to manually hold a modern F1 engine to a steady RPM in that low of an RPM range...What's idle anymore? 4000 or so?

[Caito]

Pit limiter limts speed at 100km/h. It's by no means bouncing, like a rev limiter.

It's a soft cut, as I understand it, which just keeps throttle position constant to mantain 100km/h. You don't hear it bouncing like with a hard cut limiter.

[aussiegman]

Then by your argument the pit lane limiter should cause engine damage?

No because as said the engine is restrained by either throttle position and/or be dropping particular cylinders (as said above). The engine is also not under huge loads or seeing huge load oscillation such as it would see when running at max rpm when the limiter cuts ignition.

So the pit lane limiter does not force sudden, very rapid oscillations between full power and power cut. It simply holds the engine at a constant rpm to limit speed. In doing so via the throttle position or drag/power limitation from dropping cylinders, the engine and valve train does not see the huge loads derived from the sudden deceleration/acceleration oscillations of engine speed which the valvetrain is then require to also replicate.

[PhillipM]

So the pit lane limiter does not force sudden, very rapid oscillations between full power and power cut. It simply holds the engine at a constant rpm to limit speed. In doing so via the throttle position or drag/power limitation from dropping cylinders, the engine and valve train does not see the huge loads derived from the sudden deceleration/acceleration oscillations of engine speed which the valvetrain is then require to also replicate.

I haven't seen a rev limiter so crude it did that for about 60-70 years.

[PhillipM]

The engine drops rpm and reverses loadings very quickly and is then it is reversed and loaded up again as the RPM drops under the RPM cut in rapid succession multiple times causing severe loadings in opposing directions and resulting is huge stresses on various components.

What reverse loadings? As it's done by ignition cut the pistons are still compressing the same amount of mixture and taking loadings in the exact same way....

[aussiegman]

So the pit lane limiter does not force sudden, very rapid oscillations between full power and power cut. It simply holds the engine at a constant rpm to limit speed. In doing so via the throttle position or drag/power limitation from dropping cylinders, the engine and valve train does not see the huge loads derived from the sudden deceleration/acceleration oscillations of engine speed which the valvetrain is then require to also replicate.

I haven't seen a rev limiter so crude it did that for about 60-70 years.

Sorry?!?!?

How were they actively limiting engine RPM to specific tables based on throttle position separate to driver input back in 1942? How were they cutting spark to individual cylinders in sequential or non-sequential order to limit RPM back in 1942 using point and distributor ignition systems??? I'd really like to know.

Modern RPM limiters typically work by one of two methods or by sometimes combining them both.You can cut spark or limit electronic throttle openings with drive-by-wire throttles. Hardly "crude" when you can even integrate based on 3D mapping for target RPM limit allowing variations based on any monitored input such as gear, engine temperature, boost etc.

Most motorsport RPM pit limiters usually use spark cut as the transition to full throttle once released is fractionally quicker then if the throttle is limited as the engine already has maximum air flow and there is no delay in the throttle opening. There is also the allowance of extra fuel to be pushed through the engine to keep it cool after it has had reduced cooling flow to limit the effects of heat soak.

60-70 years ago there were no electronic throttles, no electronic ignition, no electronic fuel injection with variable pulse width modulation of piezo electric injectors. So how is the current method so crude by comparison??

So without using ECU's and drive-by-wire throttles to ignition cut and/or limit air flow, how do you proposed to now limit an engines maximum RPM?? AFAIK, there are three ways to limit an engines max RPM without just letting it continue till it explodes.
1: Cut/retard ignition;
2: Cut/limit fuel; or
3: Cut/limit air flow.

All RPM limiters I have seen are all just a variations of these three methods.

The engine drops rpm and reverses loadings very quickly and is then it is reversed and loaded up again as the RPM drops under the RPM cut in rapid succession multiple times causing severe loadings in opposing directions and resulting is huge stresses on various components.

What reverse loadings? As it's done by ignition cut the pistons are still compressing the same amount of mixture and taking loadings in the exact same way....

Yeah sure but the engine is forced to slow from 18,000rpm to 17,500rpm very quickly then is sped up to 18,000rpm again and then slowed again. So the "LOADINGS" that the engine components are subjected to reverse in direction very quickly. It is not the standard "load" that the engine sees under normal operation.

Watch an engine that is under load which then hits it RPM limiter. I have seen this on an engine dyno and chassis dyno and it can be quite violent. An engine will oscillate or "shake" on its mounts due to the loading and unloading of the engine as power is cut and then re-established very quickly.

As I said previously, I have seen the result of Honda K20 engines being bounced off their RPM limiters and they do not like it!! They will not cope with constant and/or prolonged bouncing of the limiter. Valve spring failures, valve retainer and collet failures, cam lobe failures, lifter failures, pistons ring failures, connecting rod bearings failures, gearbox failures (if held and consistent) can all happen due to the sudden very abrupt shock the engine and chassis are subjected to by the oscillation of the engine from full power to limited or zero power.

This is the type of oscillation in acceleration is manageable for some components and limited in others. But some are not able to cope with these vibrations and the harmonics that sometimes accompany them for extended periods.

Rod bearings fail when the sudden acceleration/deceleration forces punch through the oil film they ride on (same as happens under detonation) and the bearing surfaces make contract. A few times and you start to get high spots from micro welding, wear sets in, tolerances go out of spec and you "spin" a rod bearing. That's if you haven't sheared the top off a valve and dropped it into a cylinder first.

Another way to think of it is you are pushing a 100kg cart up to get up to speed (accelerating in one direction), then you try and stop it (deceleration in the opposite direction), then push it again, then stop it. The load on your arms and shoulders is reversing every time you transition from acceleration to deceleration. Your rotator cuff will hold for a while, but try and do it repetitively at a high frequency.

Hope you have medical insurance...

[PhillipM]

If your rev limiter control loop is so badly tuned that it oscillates the engine from 18000rpm to 17500rpm's then I'm not surprised it damaged things, especially if it's cutting everything so fast that then engine is varying from 100% torque output to 0% and reversing the loads through the chassis, mounts and block. But with the amount of variable retard and random cut algorithms available on even the most basic ECU's you'd have to be pretty damned hamfisted in your setup to let it do that, it shouldn't be anywhere even close to that with a load on it, hell even with no load on it you should be taking account of the rate of rpm rise anyway - and yes, you can even do that with a distributor!

You aren't trying to push something and then stop it, push it then stop it, you should be aiming to generate enough torque to hold the engine at the desired RPM's, and that's not even close to a zero value.

[aussiegman]

Let me start by re-posing the questions to you that you failed to address which were based on your extremely wide sweeping and IMHO optimistic statement that:

I haven't seen a rev limiter so crude it did that for about 60-70 years.

I simply asked:

How were they actively limiting engine RPM to specific tables based on throttle position separate to driver input back in 1942? How were they cutting spark to individual cylinders in sequential or non-sequential order to limit RPM back in 1942 using point and distributor ignition systems??? I'd really like to know.

So is there an answer in our future to these questions or are they to be ignored? No?? OK moving on, so to address your other concerns:

If your rev limiter control loop is so badly tuned that it oscillates the engine from 18000rpm to 17500rpm's, then I'm not surprised it damaged things,

Firstly, the arbitrary RPM values above were provided as an example only and are not an actual representation of RPM values for an F1 engine RPM limiter at work. Nor are they those which I have used in tuning said engines. These were an analogy and as such were only for exploratory discussion purposes. Perhaps, it seems, I should have used a disclaimer or perhaps it was that I (foolishly) expected a more adult discussion than has been forthcoming and perhaps even without the ad hominem personal attacks. But I digress.

Secondly, this whole discussion is centred on previous discussions surrounding "BOUNCING AN ENGINE OFF AN RPM LIMITER" and as you seem to feel the you can simply retard ignition to the point of limiting RPM, I'd like to see you try and "bounce" a racing engine off an ignition retardation based limiter. In my experience it doesn't work and only results in you overshooting your attempted RPM limit or a quickly overheating engine and engine components. So depending on the engine characteristics, one of two thing will happen depending on the engine configuration.

Where the engine had a natural RPM limit based on airflow (choke point) below its maximum mechanical RPM limit, it would simply stop at either the predetermined max air flow rate (under the mechanical RPM limit) regardless of said retardation limiter which had likely not been enforced OR stop accelerating where the torque of the engine was reduced to an equilibrium level whereby it equaled the force required to keep the engine moving at a constant velocity (zero acceleration). So there would be no "bouncing" as the engine would simply continue to hold a steady state RPM value based on either of these variables.

However, IMHO and experience, any engine used for racing (outside specific series restrictions) should have a natural RPM limit based on airflow (choke point) substantially above its theoretical maximum mechanical RPM limit. So where this is the case, retarding the ignition alone can only result in three basic outcomes:

1: The engine will stall and stop producing torque which defeats the purpose of using any system at all;

2: The use of maximum allowable (safe) ignition retard would be unable to prevent the engine reaching the mechanical maximum RPM limit at which it self destructs as it will continue to produce sufficient torque to exceed the theoretical mechanical maximum RPM limit regardless of the ignition retard available.

3: Where you keep pushing greater values of excessive ignition retard that has not caused sufficient reductions in torque production, severe overheating and subsequent damage to the engine will occur if implemented for extended periods. Excessive ignition retard can be as equally destructive as detonation.

especially if it's cutting everything so fast that then engine is varying from 100% torque output to 0% and reversing the loads through the chassis, mounts and block.

OK lets dance this dance again.

I never advocated 0% torque production, only substantial torque reduction that results in the ceasing of engine acceleration as well as the acceleration of the engines components AND the deceleration or reversing of acceleration of the engine. Is that more sufficient for you to understand?? As the engine is still spinning due to inertia, it can still produce torque due to this motion regardless of acceleration or deceleration, it just has a very limited amount of potential energy to dissipate

You also don't seem to either understand, want to discuss or do not know the differentiation between:

1: a soft RPM restrictor that slows an engines acceleration due to a reduction in torque production by for example retarding ignition (which you seem to understand)

AND

2: the separate and differentiated system of a HARD RPM LIMITER that decelerates an engine and usually takes the form of an ignition cut that drops spark to "individual cylinders in sequential or non-sequential order to limit RPM" (quoted from my previous posts) which determines an engines maximum RPM.

Such RPM limiters as that which F1 requires at 18,000rpm and is used in aftermarket ECU's are typically an "ignition cut" where ignition retardation either was not used for performance reasons or has failed to restrain the RPM's to whatever predetermined limit the tuner has set based on a theoretical mechanical maximum RPM for the engine and its components OR where the driver has simply continued to keep the loud pedal pushed firmly against the firewall for whatever reason.

Ignition retardation cannot in all/most instances stop an engine exceeding its theoretical maximum RPM and as such a secondary system must be employed if you wish to effectively govern the maximum RPM value the engine can reach, such as an ignition cut. There are also performance advantages to not using ignition retardation and instead implementing spark cut to individual cylinders in sequential or non-sequential order to limit RPM.

It should be noted that typically the effect of ignition retardation approaching the RPM limiter can be enough to warrant most drivers not proceeding much further into the RPM range due to torque (and therefore power) reduction. In simple terms they change gears and move the engine back into a lower RPM torque band. However, when a driver persists past the point where the ignition retard is effective, say during a burnout, a spark cut in some form OR throttle override control provided for by drive-by-wire systems must be used. The former is what has been the topic under discussion, "bouncing off the RPM limiter" or cut out and NOT an running through the RPM restrictor that ramps up to this provided for by ignition retardation.

Such an ignition cut may drop spark to individual cylinders in sequential, non-sequential or ramdomised order to limit RPM. Either way when this is hit it is a HARD limit that is reached and then continually "bounced off" as the engine acceleration is suddenly stopped as it is no longer accelerating and then the engine RPM's reduce by whatever amount is provided for.

Before you start bleating on about ignition cuts are not required, they are where an engine will keep producing enough torque to breach the theoretical maximum RPM determined by whomever as detrimental, undesirable or where damage maybe done, when ignition retardation is insufficient to stop this value being breached as discussed above. This could be (given that the natural RPM limit based on airflow is substantially above its theoretical maximum mechanical RPM limit) due to the speed at which the engine accelerates, momentum of the engine itself due to a higher mass flywheel/clutch assembly or a myriad of other reasons, you pick one. As such a HARD RPM LIMITER via ignition an cut (randomised, sequential, non-sequential or otherwise) can be used to stop the engine continuing to accelerate and exploding. Again, it is really that simple.

These are the reasons hard RPM limiters are required, which is primarily the simple fact that people will ignore the RPM restrictor and the reduced torque output and proceed straight on to the RPM cut and then proceed to BOUNCE THE ENGINE OFF THE RPM LIMITER, which is what this is all about. Not the most efficient and effective way to slow an engine down prior to reaching this maximum value.

So, again slowly, when a hard RPM limit via ignition cut is reached, the engine itself does not stop rotating from what ever arbitrary RPM value you decide (with the above caveats, this is an example only) to a zero RPM value (stationary), it just "stops accelerating", which is the rate at which the velocity of something changes over time, and then starts decelerating or accelerating in the opposite or reverse direction.

So simple physics to explain

When something is accelerating in one direction (increasing RPM's) then ceases this acceleration and starts decelerating, it has reversed this motion. Mathematically, deceleration it is acceleration in the opposite direction to that of the original motion. If something moves in an opposite direction of motion, it can be said to have reversed that direction. I hope you understand the difference.

But with the amount of variable retard and random cut algorithms available on even the most basic ECU's you'd have to be pretty damned hamfisted in your setup to let it do that, it shouldn't be anywhere even close to that with a load on it, hell even with no load on it you should be taking account of the rate of rpm rise anyway

There is a huge difference between the hard RPM LIMITERS you can bounce off and the soft RPM RESTRICTORS that ramp up to an RPM limiter as described above. An imperfect but simpl-ish analogy (perhaps only to my mind) is the way a turbo restrictor works. These can also be considered as an engine LIMITER and RESTRICTOR as they limit maximum power and usable engine RPM (dependent on engine capacity).

These simple devices have a maximum air flow LIMIT defined by the size of the orifice over which it will not flow more air. Prior to this maximum limit, it acts as an airflow RESTRICTOR and as it approaches closer to its choke point where air speed goes super sonic, it will restrict or slow the rate at which air can enter the engine. Once it reaches the maximum flow limit, it cuts off further air flow. Ignition cuts, while operating on very different principles, to my mind are somewhat similar in that they restrict air flow in steps similar to ignition retard by restricting but not stopping torque production, then it simply stops increasing air flow as an ignition cut simply cut spark in its many forms reducing torque production.

Aside from all these analogies, I am more than aware of the various algorithms used, as I am privileged enough know some of the people that wrote them for various ECU firms and even tested prototype systems some on my cars. Even these people agree that there are two (2) distinct stages to an RPM limit using ignition, the RAMPING stage and CUT stage. Both have VERY specific limitations as to effectiveness for certain engines. What they agree on is you CANNOT tune out the "stupidity" (sic) of some people who feel the need to hold an engine at the maximum RPM limiter and continue to bounce it off that limit for extended periods of time. They all agree this act WILL eventually result in, and I quote "damage to an engine regardless of what methodology you use to implement said arbitrary limit due to the forces involved in reducing the acceleration of the engine at the maximum RPM limit of operation that will see those forces reverse due to deceleration."

As for my "hamfisted" setups, you have no idea what I have or have not done or used or implemented, so I would be VERY careful throwing stones.

What I will say is when I have used either Motec, Autronic, Pectel, DTA or Vipec (I have used all these) they all allow various ignition retard/ignition cut, drive-by-wire throttle (DBW) or boost limitation RPM restriction methodologies. However in 99% of instances the only 2 ways to effectively stop an engine continuing to accelerate is close the DBW throttle or cut the ignition, regardless of the ramping rate methodologies used leading up to these two events. Due to the downside in engine performance from using other methodologies, ignition cut is typically the preferred method used. However it does require a modicum of self restraint and understanding for the end user (driver) not to abuse it.

and yes, you can even do that with a distributor!

Yes, I agree you can limit RPM with a distributor, but I am yet to see a distributor system from 1942 (your 60 to 70 year time frame) or that does not use a modern ECU intervention system somewhere (like an MSD box) that can replicate the action of cutting spark to individual cylinders in sequential, non-sequential or randomised order.

As an example of what a basic distributor system can do, look at the what distributor ignitions used in the form of solid state circuitry that interrupted the grounding of the ignition circuit, such as Ford used in the 70's with the Boss 302 V8 at 6250rpm via the Autolite RPM module. This effectively cut ignition after the mechanical/vacuum ignition retard systems failed to stop the engine accelerating past the maximum mechanical RPM limit. Sound familiar??? A two stage retard and cut system via ignition for RPM limitation similar to what I had said is used. Wow what do you know!!!

You aren't trying to push something and then stop it, push it then stop it, you should be aiming to generate enough torque to hold the engine at the desired RPM's, and that's not even close to a zero value.

Again this was merely an example of the forces, not an exact representation of the movements in the engine. Last I checked, engines did not contain rotator cuff elements. Try reading it again more carefully. In trying to "stop" but not actually stopping the trolley, you are stopping the acceleration of the trolley and decelerating it OR stopping the acceleration in one direction and reversing the acceleration in the opposite direction. Then repeat at great frequency. This is similar to that which happens in an engine that "bounces off an RPM limiter".

All that being said, the effect and forces on certain components in the engine are exactly these forces of being accelerated in one direction, stopping (in any event where a force reverses there is a momentary halt in motion however brief) and then being asked to cope with forces of acceleration in the opposite direction that have reversed due or a "deceleration" event.

Primarily, it is the valve train and rod bearings that suffer most from this which is why failures in these areas are common for an engine that has been "bounced off the RPM limiter".

[PhillipM]

It's very simple to make the distributor a random cut, you ran another set of counterweights run at a differentiating speed to the main setup (simple epicylic gear over the top of the standard setup), which provided the hard cut function, that gave you a semi-random spark cut setup, and you could still have the hard full cut setup on the standard parts which cut every cylinder in an overrev situation, there were thousands of those around, hell, I've got a few in the attic somewhere from some old rally Mini's. Obviously it's not truly random, but then neither is an ECU algorithm.

Regardless of the lengthy constant repetition of the exact same things you've just posted (yes, soft cut and hard cut, we know, it's not rocket science), a hard cut should never be implemented in way that would cause more torque differentiation than the driver could ever do by simply shutting the throttle, the only time then that the loadings would be more of an issue is if you have secondary problems with the rigidity or damping of the holdings/mountings. Which is probably what you saw on the dyno as most amateur engine dyno setups are not particularly stiff and are positively dire for damping movements.

Whilst this change in acceleration you are concerned about makes itself evident as jerk or lurch on the dyno, and can be felt in a vehicle as such, first principles ought to point out that it does not change the torque or forces upon the engine generated purely from the engine's output.
I.E - The engines you've witnessed falling apart on the limiter, we're either marginal on the limit to start with, or had secondary influences.

[aussiegman]

It's very simple to make the distributor a random cut, you ran another set of counterweights run at a differentiating speed to the main setup (simple epicylic gear over the top of the standard setup), which provided the hard cut function, that gave you a semi-random spark cut setup, and you could still have the hard full cut setup on the standard parts which cut every cylinder in an overrev situation, there were thousands of those around, hell, I've got a few in the attic somewhere from some old rally Mini's. Obviously it's not truly random, but then neither is an ECU algorithm.

Seen these used, observed their variable and inconsistent results and then watched them fail. They are in no uncertain terms total crap and massively unreliable. And as you said as well, they are definitively not a random, and certainly not a sequential cut system as well as excruciatingly unreliable. I expect that's why they are sitting in your attic and not on a car somewhere...

Regardless of the lengthy constant repetition of the exact same things you've just posted (yes, soft cut and hard cut, we know, it's not rocket science),

I was making sure i was as clear as possible so you could understand what I was saying.

a hard cut should never be implemented in way that would cause more torque differentiation than the driver could ever do by simply shutting the throttle,

So are you advocating a hard ignition cut should replicate as closely as possible the 100% closing of the throttle?
Firstly, aside from switching the engine off altogether by cutting the ignition system in its entirety, how would you achieve this??
Wouldn't this cause a greater interruption to torque production then the sequential or non-sequential ignition cut used and discussed and be self-defeating??

the only time then that the loadings would be more of an issue is if you have secondary problems with the rigidity or damping of the holdings/mountings.

Aside from the fact you are totally ignoring the fact that individual engine components have mass and inertia, are in motion within the engine casing and experience huge acceleration and deceleration forces, then yes if you have suboptimal mounting and/or damping then the rocking or oscillations of the engine will surface as a broken engine mounts or other failure in the mounting system or ancillary systems.

However, to go back to the individual engine components and their mass and inertia, the issue centres on broken engine components not engine mounts. The oscillating loadings are a problem for internal engine components and has a variable effect on individual components. Any rigidity and/or damping of the mountings will have marginal bearing on the effects seen at the valve tips, pump drives or bearing faces.

Which is probably what you saw on the dyno as most amateur engine dyno setups are not particularly stiff and are positively dire for damping movements.

The engine dyno facilities I have used is certainly not an amateur facilities. They are used by professional race teams. The movements observed are not the engine jumping 30cm's in any direction. They were as indicated oscillations, or more simply rapid movements in opposite directions of a small order of magnitude, maybe 3 to 5 cms.

Whilst this change in acceleration you are concerned about makes itself evident as jerk or lurch on the dyno, and can be felt in a vehicle as such,

Absolutely, then string a series of them together in rapid succession and what do you have?? Rapid oscillation or movement of the engine due to the constant and rapid variations in torque production.

first principles ought to point out that it does not change the torque or forces upon the engine generated purely from the engine's output.

In the quiet words of someone holy, come again???

Firstly, are you saying that when torque from the only torque output is reduced, it will have no effect on the torque seen by the engine or its internal and external components?? How does that work and which first principle are you working too??

Secondly, are you saying that cutting spark (or by extension 100% or there abouts closing the throttle as per you example above) does not change the the torque or forces seen by the external and internal engine components that are generated from the engines acceleration and deceleration that is directly linked to its torque output and the reduction thereof?? Again, how does that work???

Simply put, I absolutely and 100% disagree. But that's OK we are allowed to disagree and that's what makes the world work.

Each component (as well as he engine as a whole) has mass and therefore inertia. When in motion or accelerating, when something with mass and inertia is accelerating in one direction, such as when an engine is increasing in RPM and then stops that acceleration, it and the individual components will stop the movement temporarily then start decelerating or accelerating in the opposite direction. This is a quantifiable reversion of its motion and mathematically definable. Deceleration is acceleration in the opposite direction to that of the original motion.

As the components have mass and inertia, when the torque output of the engine is reduced by any %, they undergo some form of deceleration which results in forces acting on them in an opposite direction to those that were seen previously to accelerate them.. Under conditions of a RPM limiter, these conditions are very different to those seen during a gear change, a normal deceleration event or other reason for the engine to slow.

The primary reason "bouncing off the RPM limiter" for an extended period can and does cause engine component damage is the frequency at which these events take place over any given time.

One off deceleration events, such as a gear change are exactly that, an event with a frequency of ONE. Same for a deceleration event, an event with a frequency of ONE. Bouncing off the RPM limiter can have a frequency of many thousands of times a second and it is this rapid oscillation and the harmonics it sets up that can and does cause damage.

As an arbitrary example, if the RPM limiter is set at 9,000rpm on a 4 cylinder engine and the RPM cut sequentially cuts every 4th ignition impulse to the 4 cylinders, then this occurs 562 times per second (rounded down). As such, if the engine is held, bouncing off the RPM limiter for 30 seconds, that is 16,860 ignition cuts. This results in 16,860 individual acceleration and 16,860 individual deceleration events in a 30 seconds time frame due to the engine torque value dropping and picking back up.

That is a lot of gear changes in a very short period of time..

I.E - The engines you've witnessed falling apart on the limiter, we're either marginal on the limit to start with, or had secondary influences.

Holy broken valve tip Batman, I better call Cosworth/Ferrera/Pankl/Arrow et al (pick a name) and tell them their engine parts are marginal at best!!!

I would again seriously disagree with your hypothesis. Given the weight of experience, anecdotal evidence and actual evidence from research done by again myself and others this is simply not true. The engine(s) I mentioned didn't "fall apart", they were not $100 eBay special rebuilds I assure you. They had specific component failures endemic of the stresses bought about by a known and quantifiable issue, that being they were "bouncing off the RPM limiter" for extended periods.

These engines had performed flawlessly under race conditions with general servicing and inspection after every race meeting. Components that were out of spec were changed for new. They had been typically run to their limiter frequently during each race (by myself and others), but they were not held there and did not fail under normal race conditions when run up to this limit under normal circumstances. So they are and were anything but marginal.

The failures occurred when they were specifically and knowingly "bounced off the RPM limiter" for an extended period, not just run up to them as per normal. If an engine can run to its limiter for half a season or approx 250km in racing with only general servicing without a failure, only to then fail when it is purposely held at the limiter for an extended period due to a driver keeping his foot on the throttle as we were forced to use a wrong, shorter final drive which resulted in us running out of RPM's near 60% of the main straight distance and resulted in us to bouncing it off the RPM limiter to try and maintain position. We knew it was dangerous for the engine, but we accepted the risk and ran it anyway. It failed approximately 25% into the race distance.

So if a failure occurred only after this treatment and not under normal racing conditions as detailed previously, then I definitively wouldn't classify those engines as marginal on their limit and I would absolutely place the failure down to a result of it being knowingly and purposefully bounced off the RPM limiter for an extended period.

This was later backup by a professional engine builder (who stripped and reviewed the engine with me next to him) and further analysis of the telemetry and review of the individual components that failed.

Perhaps you have found the holy grail of engine management or can produce engine components from unobtainium that do not suffer from fatigue stress. If you have then you should be either rich or be working in aeronautics, space or racing engineering and I will be first in line to purchase them. However I suspect this is not the case.

Either way, I do not advocate that "bouncing an engine of an RPM limiter" as I have evidence that it is extremely detrimental to an engines longevity.

[PhillipM]

Whilst this change in acceleration you are concerned about makes itself evident as jerk or lurch on the dyno, and can be felt in a vehicle as such,

Absolutely, then string a series of them together in rapid succession and what do you have?? Rapid oscillation or movement of the engine due to the constant and rapid variations in torque production.

first principles ought to point out that it does not change the torque or forces upon the engine generated purely from the engine's output.

In the quiet words of someone holy, come again???

Firstly, are you saying that when torque from the only torque output is reduced, it will have no effect on the torque seen by the engine or its internal and external components?? How does that work and which first principle are you working too??

You see, this is why no-one can be bothered to write a long reply to you, you deliberately take the statements out of context, misinterpret them, then apply them to a different condition anyway, brilliant.

Jerk is a rate of change of acceleration, it can be seen, it can be felt, it can be measured, it gives no change in the internal or external forces THAT ARE ALREADY ENCOUNTERED BY THE ENGINE UNDER NORMAL CONDITIONS, that's linked to the acceleration of components NOT, as you seem to believe, the acceleration of the acceleration....
These rapid oscillations you are experiencing should be nowhere even close to a variation from 0% to 100% torque output because your vehicle no way in hell has so much power that it has to cut spark to every cylinder and then reapply it to them all instantaneously to keep the vehicle at maximum velocity - you have a load applied and therefore inherant damping/preload on the system, so that would need infinate amounts of horsepower. I'm pretty sure you're not close to that.

The only concern I would ever have for running on the limiter would never be from the limiter itself, but simply normal fatigue from engines not designed to be held at such rpms for long periods of time, such as your example, that's not an inherent issue with the limiter, as I say, you have problems elsewhere, you are not fatiguing the engine any more from varying the engine speed than you are with the compenents buzzing away 1rpm short of the limiter. The torque reversals you keep banging the drum about happen anyway within the engine as it functions, that's why we get timing drift and crank flex and end up running torsional monitors and the rest of that crap to dial it out (or slap a big damper on it if it's not a race motor).
Torque output variation is happening all the time, at much higher frequency, and the loadings you're worried about are less than the peaks the engine will exhibit under normal operation, the loadings on the bearings, rods, cams, lifters, crank, etc, would all be much higher from the driver simply shutting the throttle from high rpms.

...And the reason the distributors are sitting in the attic is they're from a couple of race winning cars that have sentimental value, who the hell would fit a distributor to a race car these days? Certainly nothing to do with the rpm limiting system, as far as I know they never had any issues with them...

I've certainly seen plenty of F1 motors recently geared too short and running on the limiters for over half the straights when they cocked the weather or DRS up. I haven't seen any spontaneously explode there yet either.

[hardingfv32]

Is there any reason the F1 ECU can not just allow the engine to race up to the rev limit and then maintain it there with proper fuel and timing adjustments? Why should the exhaust note sound different?

Brian