What is the aim to design and how to assess the performance of restrictor ?

[RTAF]

What is the aim to design and how to assess the performance of restrictor ?

I use these boundary conditions.
Inlet : velocity 22 m/s
outlet : 101325 Pa


I got this result from flow simulator.

http://img232.imageshack.us/img232/176/11052757.png (http://imageshack.us/photo/my-images/232/11052757.png/)

[M. Nader]

are you asking what is the aim of designing a restrictor? or what do you understand from your results?

in both ways you need to answer, why did you make that simulation to begin with?

[RTAF]

i got this result from last year team. I newbie to CFD.

i think a good restrictor is low pressure drop and high velocity at throat. is that right?

i not sure about boundary conditions for simulate.

[M. Nader]

Just analyzing the restrictor will not be very useful to you, you need to know what for. What type of engine are you using, what would you want from your airbox/plenum to do for the engine.

how should the air be supplied to the engine cylinders, are these results accurate enough or not.

I cannot comment on the results themselves without knowing the side before and after the restrictor, the restrictor itself should behave as a nozzle (accelerating air) but this is because of the change of area before and after so you need to know that when strictly talking about the resrictor. calculations can be done easily using basic fluid mechanics knowledge with no need for a complex CFD.

Me personally, I would tell you not to spend so much time looking at the restrictor on its own but focus more on the plenum and the feed to the engine and if that is what you need or not. knowledge of fluid mechanics is crucial here so start reading on that first.

If you just want to do reverse engineering work on what the previous guys did then use a CAD model and get other CFD results and compare.

[mk e]

Originally posted by M. Nader:
Me personally, I would tell you not to spend so much time looking at the restrictor on its own but focus more on the plenum and the feed to the engine and if that is what you need or not. knowledge of fluid mechanics is crucial here so start reading on that first.



I would give the opposite advise. The reason he flow restrictor is there is to....limit the HP that can be achieved and limiting HP is bad if you want to go fast http://fsae.com/groupee_common/emoticons/icon_smile.gif

I think that you'll want the best flowing restrictor you can design, then design the rest of the intake track to work with it and that will give the best overall result.

[M. Nader]

Originally posted by mk e:
I would give the opposite advise. The reason he flow restrictor is there is to....limit the HP that can be achieved and limiting HP is bad if you want to go fast http://fsae.com/groupee_common/emoticons/icon_smile.gif

I think that you'll want the best flowing restrictor you can design, then design the rest of the intake track to work with it and that will give the best overall result.

One end gives the other so we are basically just suggesting different methodology. I would say your method would be better for a single cylinder or 2 cylinder engine.

for a 4 cylinder engine i think it would be better to start on the plenum ensuring that all 4 cylinders get equal supply, that would give a decent shape, then get to the restrictor and see if you can improve the plenum boundary conditions by improving the restrictor outlets.

I wouldn't mind getting lower Hp with a 4 cylinder than have cylinders un-matching, just get things working first then fine tune to improve. i guess it really depends on your situation.

[Owen Thomas]

Regarding your boundary conditions, your INPUT pressure (left side?) should be 10125 Pa, not the output. If you're analyzing the performance of an existing design, the output pressure and velocity will be what you are solving for.

As Nader mentioned above, a simple control volume analysis using some textbook methods might be easier, and will have the added benefit of assuring that you understand the solution.

You should also keep in mind that in that specific model (and for a wide range of your operating conditions) the restrictor flow will remain sub sonic (un-choked), which means it will not act as a nozzle. That being said, my quick interpretation is that I think you need more blue on the right hand side of your CFD model http://fsae.com/groupee_common/emoticons/icon_smile.gif

Edit: Tilman pointed out a mistake, and to clarify I should say that I meant the *diverging section* of the restrictor model will not act as a nozzle; the converging section will still accelerate the flow in sub sonic conditions. Thanks Tilman.

[Tilman]

Heyho,

with this restrictor you will have flow separation from the walls because the opening angle is really huge. If you are going to reduce pressure drop across the restrictor, your opening angle should be at a maximum of 7 degrees (rotational axis to wall).

However, ensuring that all cylinders get the same amount of air is something important to keep in mind.

The proper restrictor design aim is something you should think of as we do not know your overall aims and decisions. Safe, simple, reliable? Or high performance/low weight at any cost? Manufacturing options? Available time to do analyses?


You should also keep in mind that in that specific model (and for a wide range of your operating conditions) the restrictor flow will remain sub sonic (un-choked), which means it will not act as a nozzle.
No, it will not restrict the mass flow that much, but it will still act as a nozzle (which it always is).

[mk e]

Originally posted by M. Nader:

One end gives the other so we are basically just suggesting different methodology.

That's right. I find it's generally best to focus on making the part that is doing the most harm first then make the compromises in the parts that are less important. In this case, that is the flow restrictor since it's only job is to be a restriction. All the other parts in the system can be designed/sized as appropriate, not the flow restrictor.

My advice it get the restrictor as non-restrictive as you can, then live with the consequences. Part of optimizing the restrictors ability to flow is keeping at peak flow for as much of the time as possible and that means 1 or 2 cylinder options are right out unless a turbo or supercharger is used.

There's no "right" answer I just like to grab the low hanging Fruit first I guess..... because I'm lazy http://fsae.com/groupee_common/emoticons/icon_smile.gif

[mk e]

Originally posted by Tilman:
Heyho,

with this restrictor you will have flow separation from the walls because the opening angle is really huge. If you are going to reduce pressure drop across the restrictor, your opening angle should be at a maximum of 7 degrees (rotational axis to wall).

However, ensuring that all cylinders get the same amount of air is something important to keep in mind.


I was about to suggest he try 6-8 degrees

Uniform flow is always the best, but I think uneven flow is a problem I'd much rather be dealing with than inadequate flow. With uneven flow the worst case is just give up and fix it by mapping each cylinder.

[Mbirt]

Originally posted by M. Nader:
One end gives the other so we are basically just suggesting different methodology. I would say your method would be better for a single cylinder or 2 cylinder engine.

for a 4 cylinder engine i think it would be better to start on the plenum ensuring that all 4 cylinders get equal supply, that would give a decent shape, then get to the restrictor and see if you can improve the plenum boundary conditions by improving the restrictor outlets. Ask teams the teams producing the most power for each engine configuration and you'll find that the restrictor design is most important for teams that can choke it on a time-averaged basis (hitting the theoretical mass flow limit at the speed of sound) and those teams likely have 4 cylinders. As the cylinder count drops (or firing orders become odd), the plenum design becomes increasingly crucial to ensure the restrictor does not choke during the induction event only and go underutilized for the other 500 crank degrees. Running a restrictor and diffuser at the end of an intake tract has been a consistent recipe for 25 hp on a 450 cc single in my testing experience.

[Z]

RTAF,

Are you in a "FSAE" team (ie. designing a small racecar)?

If so, why are you showing us CFD of a rocket nozzle? http://fsae.com/groupee_common/emoticons/icon_confused.gif

Z

[M. Nader]

Originally posted by Mbirt:
Ask teams the teams producing the most power for each engine configuration and you'll find that the restrictor design is most important for teams that can choke it on a time-averaged basis (hitting the theoretical mass flow limit at the speed of sound) and those teams likely have 4 cylinders. As the cylinder count drops (or firing orders become odd), the plenum design becomes increasingly crucial to ensure the restrictor does not choke during the induction event only and go underutilized for the other 500 crank degrees. Running a restrictor and diffuser at the end of an intake tract has been a consistent recipe for 25 hp on a 450 cc single in my testing experience.

I do agree with what you are saying, We actually have the same setup (diffuser and restrictor), the question is just where to start. I recommend looking at the plenum first, and that is not ignoring the restrictor as you and others have said is the limiting factor for the competition. Just build the plenum first then get a look at the restrictor as it is the part which will take the most time to optimize, and with a plenum already done giving equal feed to the cylinders improving the restrictor would only make things better without the need to drastically change anything. On the other hand i think that a properly designed restrictor with an inefficient plenum would be just wasted, if for example cylinders 1 and 4 get just 50% of the air needed the power will drop by 25% even with the best restrictor this is.


I am not saying that my method is better or anything, as i said it really depends on the goals and resources of each team as well as their manpower and time frame.

[Z]

Originally posted by Mbirt:
Running a restrictor and diffuser at the end of an intake tract has been a consistent recipe for 25 hp on a 450 cc single in my testing experience.
Mbirt,

Do you mean that a 450cc single with an intake pipe that consists only of throttle-restrictor-diffuser-head, and nothing more (no plenum), will only put out 25hp? I can see why this might be the case, but just checking if this is what you meant.

Furthermore, what size plenum would you recommend for a 450 single? (I guess I could use "Find", but while I'm here... http://fsae.com/groupee_common/emoticons/icon_smile.gif.

Z

[M. Nader]

I would say 1-1.5 liter should be enough for 450cc

[Warpspeed]

Seems to me, that to get best mass flow, there really needs to be an almost constant depression downstream of the restrictor.

Severe pulsing is going to be death to flow through a restrictor.

Four cylinders will beat one cylinder for pulsing, and a big plenum is always going to help.

But how about a positive displacement supercharger which should have an almost constant intake flow ? Most pulse on the discharge side, but that is quite another matter.

Not talking about generating huge silly boost pressures, just pulling more "average" steady airflow through the restrictor for a given engine, and recovering some of the restrictor pressure loss, plus maybe a bit extra.
If boost pressures are kept low, the Hp required to drive the blower will also be low.

Should pull a lot more air through the mid range torque region than the engine could do by itself. It may not produce much extra extreme top end power, but the mid range torque could really benefit.

[mk e]

Originally posted by Warpspeed:
Seems to me, that to get best mass flow, there really needs to be an almost constant depression downstream of the restrictor.

Severe pulsing is going to be death to flow through a restrictor.

Four cylinders will beat one cylinder for pulsing, and a big plenum is always going to help.

But how about a positive displacement supercharger which should have an almost constant intake flow ? Most pulse on the discharge side, but that is quite another matter.



yes
yes
yes

A 3 cyl with a decent cam and well done headers and intake comes pretty close to constant flow too.

Any type of supercharger/turbocharger works pretty well, but I think you are right that a positive displacement would be optimal at getting air through the restrictor. I'm not so sure would be optimal at making HP because you need to do work to spin it...don't get me wrong, I LOVE positive displacement superchargers and I've done several custom installs over the years including 4 on Ferraris.....but they were all street cars were drivable hp was the main goal, not brute hp, for that I always go turbo.

I think the real options that make sense for fsae are 1/ 2 cyl with a trubo or 3/4 cyl without

[Warpspeed]

The restrictor is there for the sole purpose of limiting mass airflow and power.
And very efficient it will be at doing that job.

If all you want is max possible top end power, a centrifugal supercharger would probably be optimal.

But I believe there is far more to be gained by trying to build a mountain of a torque curve at mid rpm than trying to squeeze an extra top end Erg (or two) by sacrificing everything else.

[mk e]

Originally posted by Warpspeed:

But I believe there is far more to be gained by trying to build a mountain of a torque curve at mid rpm than trying to squeeze an extra top end Erg (or two) by sacrificing everything else.

You know....it really doesn't much matter. if you build an engine that makes say 50 ft-lbs torque at 5000 or an engine that make 25 ft-lbs at 10000 rpm becasue you still have the transmission and final drive sockets with the goal of turing the wheels at say 2500 rpm. The first engine gets a final drive ration of 2:1 and puts 100 ft-lbs to the wheels and the other gets a final drive of 4:1 and puts.....100 ft-lbs to the wheels. The 2 engines woulds have the same hp so you get the same final result.

In street cars it's often not so easy to change the final drive so it can make more difference what setup you choose, but on an fsae other things are more important. I would think that making it easy to drive would be right at the top of the list, so a wide smooth torque curve is what I'd be after to minimize the amount of shifting required and so the drive always knows what to expect. It's also nice when you can operate between the torque peak and the hp peak (or just beyond) because there is an inherent stability, the slowly dropping torque in this range means wheel spin will self terminate vs self start you can get on the rising torque portion.....you want it easy to drive at m
almost any cost I think.

[Warpspeed]

I don't think a wide smooth torque curve is really possible with a restrictor. As rpm rises the restrictor is going to strangle the Ve at an ever increasing rate.

Probably the best you can hope for is a fairly flat power curve, with max airflow being drawn through the restrictor over as wide an operating range as possible.

As you say, an engine like that would hit peak torque early, and have falling torque with rising rpm. It would not make much difference what gear you were in. Sounds weird, but it would be much easier to drive and probably require fewer gears and many fewer gear changes.
That could be a decisive factor with a rookie driver under race day stress.

But to draw in a lot of air, and make a lot of torque low down requires either a larger capacity (and heavier) engine, or a positive displacement supercharger.
I see the problem as designing the whole engine package around the restrictor in such a way as to maximise airflow through the restrictor over as wide an operating rpm range as possible, not just at peak rpm.

Most racing classes have an engine capacity limit, and the aim is always to achieve max horsepower within that limit. That usually results in a nasty peaky engine.

With the restrictor being the main real limitation for us, IMHO the whole engine selection and design problem takes on a very different aspect.

The whole engine thing really comes down to "horsepower under the curve".
As we always hit a brick wall with the restrictor at the top end, there are probably more useful gains to be had by boosting low end torque either with a larger capacity engine or a suitable blower.

[Mbirt]

Originally posted by Z:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Mbirt:
Running a restrictor and diffuser at the end of an intake tract has been a consistent recipe for 25 hp on a 450 cc single in my testing experience.
Mbirt,

Do you mean that a 450cc single with an intake pipe that consists only of throttle-restrictor-diffuser-head, and nothing more (no plenum), will only put out 25hp? I can see why this might be the case, but just checking if this is what you meant.

Furthermore, what size plenum would you recommend for a 450 single? (I guess I could use "Find", but while I'm here... http://fsae.com/groupee_common/emoticons/icon_smile.gif.

Z </div></BLOCKQUOTE>Z,

Yes, except the setup I tested was throttle-restrictor-diffuser-about 5" of intake runner-head. The setup you describe would make even less power because the intake runner, without a plenum volume into which it would normally terminate, acts as plenum volume itself.

As for plenum size, the general consensus in this thread is too small. In testing (on a 470 cc CRF450X) a 1 liter plenum, I found that intake runner length was still behaving more like plenum volume and less as a tuning tool. As the runner was shortened from a length that was good for the 2nd ramming wave at somewhere below 5000 rpm, power was lost everywhere and the shape of the curve did not change. An increase to 2 liters changed this and finally allowed intake tract tuning to behave as expected. Above 2 l, the volume of diminishing returns will depend on the engine system design and its VE. I would thus start in the 2-2.5 l range and go up from there.

Our WR450f stock bore/stock stroke likes 3.5 l, making 50 hp from 7800-11500 rpm with a nice peak at 60 hp at 9800.

[Z]

Mbirt,

Thanks. That makes sense.

The 25hp suggests the restrictor is sonic ~20% of the time (given that 100% is ~125hp). So sonic for a bit less than the full intake stroke. And 3.5 litres is ~8x swept volume, with more being better (space permitting). Although 50+hp sounds good. http://fsae.com/groupee_common/emoticons/icon_smile.gif

Z
(PS. Still wondering about that rocket nozzle???)

[Wesley]

Originally posted by mk e:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Warpspeed:

But I believe there is far more to be gained by trying to build a mountain of a torque curve at mid rpm than trying to squeeze an extra top end Erg (or two) by sacrificing everything else.

You know....it really doesn't much matter. if you build an engine that makes say 50 ft-lbs torque at 5000 or an engine that make 25 ft-lbs at 10000 rpm becasue you still have the transmission and final drive sockets with the goal of turing the wheels at say 2500 rpm. The first engine gets a final drive ration of 2:1 and puts 100 ft-lbs to the wheels and the other gets a final drive of 4:1 and puts.....100 ft-lbs to the wheels. The 2 engines woulds have the same hp so you get the same final result.
</div></BLOCKQUOTE>

Problem is you compress your usable torque curve significantly, which means your accel score lowers from increased number of shifts, your lap times suffer because newbie drivers can't keep the peaky engine in the power band, and you operate the engine at a less efficient flow range through the restrictor, increasing your pumping losses and decreasing efficiency.



Originally posted by M. Nader
On the other hand i think that a properly designed restrictor with an inefficient plenum would be just wasted, if for example cylinders 1 and 4 get just 50% of the air needed the power will drop by 25% even with the best restrictor this is.

If you can make more overall power with a biased plenum design, do it. You want more power always. You can tune out individual cylinder trims if it's a big problem, and you'll be losing some ultimate power, but at least youll be making more than if you used a rocket nozzle for a restrictor.

More power is more power, regardless. Yes, you should focus on a plenum with good cylinder distribution AS WELL AS an efficient restrictor, but there are reasons we have the ability to trim cylinders.

Plenum volume considerations are manifold. The larger the plenum the closer to steady state the restrictor acts, which means you aren't bringing it to choked flow every time a valve opens and wasting the rest of the time. It also increases drawing volume for the cylinder - you have a larger volume of air you can induct with less pressure drop within the plenum.

At the same time, that buffering effect that helps total top end flow through the restrictor by dampening pressure drops from individual valve opening events, it takes that much longer for pressure to start building in the plenum in high vacuum situations and you see a resultant drop in throttle response.

Most of our plenums were sized around 2-3x engine displacement, which didn't rob too much throttle response and allowed for steady WOT flow through the restrictor. We also used a larger divergent angle than a lot of teams thought was "correct." Because of packaging constraints the larger diverging angle allowed for a more gradual increase to the plenum interface instead of a dump into the huge volume of the plenum and the turbulence resulting from that, maintaining flow momentum within the plenum.

Like all things, it's a tradeoff. Either do some 1d simulation to determine what your pressures look like through the restrictor and investigate plenum pressure rise rate after sudden throttle position changes (throttle response) or build a few test plenums that you can test on a dyno.

[Warpspeed]

Never worry about the torque curve, it is the power curve that matters.
Gears are for multiplying torque, engines are for making power.

If you have the power (at ANY rpm) matching that to the road speed with suitable gearing will get you the tractive effort.

If you can keep your engine at peak power, and peak airflow, as limited by the restrictor, over as wide an rpm band as possible, you will be out in front.

[M. Nader]

Originally posted by Warpspeed:
Never worry about the torque curve, it is the power curve that matters.
Gears are for multiplying torque, engines are for making power.

If you have the power (at ANY rpm) matching that to the road speed with suitable gearing will get you the tractive effort.

If you can keep your engine at peak power, and peak airflow, as limited by the restrictor, over as wide an rpm band as possible, you will be out in front.

This kind of talk has always confused me, I understand that the torque values can be fiddled with various gear reductions as much as you need, but i have always thought of power as more of a tertiary property rather than the primary one, the engine produces torque and power is a function of torque not the other way around. for me i think a good torque curve is more important than a power curve is it is what actually does the driving

[Nick Renold]

While we are having this wonderful discussion of restrictors, I was wondering if any teams have investigated the effect of surface roughness on the restrictor performance. Is there a big difference between a powder rapid prototyped restrictor and a polished metal or smooth CFRP one? Are there any resources you can recommend on the effects of surface roughness on airflow? I know that RCVD has some brief sections on it, but those are more related to bodywork aerodynamics.

[Warpspeed]

Originally posted by M. Nader:

This kind of talk has always confused me, I have always thought of power as more of a tertiary property rather than the primary one,

If you could choose between fitting a V8 engine with 400 Ft/lb of torque, or a lawnmower engine that was heavily geared down with a gearbox to also generate 400 Ft/lb of torque out of the gearbox.
Which do you think would make a vehicle accelerate and travel faster.
Remember the torque is identical, but the POWER is hugely different.

Power is the primary factor.

[Dash]

I think Nick is more or less talking about looking at one motor. Do you look at the power curve, or the torque curve for tuning? Personally ( I'm not an engine guy), I'd probably look at the torque curve to see where to improve.

More power is more power though when looking at two choices, and gearing will help you out.

[M. Nader]

Originally posted by Warpspeed:

If you could choose between fitting a V8 engine with 400 Ft/lb of torque, or a lawnmower engine that was heavily geared down with a gearbox to also generate 400 Ft/lb of torque out of the gearbox.
Which do you think would make a vehicle accelerate and travel faster.
Remember the torque is identical, but the POWER is hugely different.

Power is the primary factor.

I don't think this example is the primary choice here as these engines are on two ends of the spectrum in terms of RPM range and weight which are critical choices for choosing an engine. for an FSAE team the RPM "useful" range in terms of power and torque is necessary to know as shift times vary depending on the system used and a lot of teams desire to do certain events (autocross) in only 2 gears for example.

I still claim that having a neat torque curve is more important than having peak power, the power curve will also be quite good as it depends on the torque curve!

my support is that power depends on torque and not the other way around, the second thing i would look at after torque would be RPM range

[mk e]

Originally posted by Wesley:

Problem is you compress your usable torque curve significantly, which means your accel score lowers from increased number of shifts, your lap times suffer because newbie drivers can't keep the peaky engine in the power band, and you operate the engine at a less efficient flow range through the restrictor, increasing your pumping losses and decreasing efficiency.

High rpm hp does not = peaky any more than more than low rpm power = board power delivery.

Your assertion that the usable rpm range is compressed by gear reduction is correct and needs to be considered, but the assertion that using gear reduction = slow acceleration and poor lap times is not.

Have you ever seen an F1 car accelerate? http://fsae.com/groupee_common/emoticons/icon_wink.gif

I do agree with what I think were your under lying assertions though, balance is the key to success and in that balance focusing on easy of operation over absolute performance will probably produce a better competition result.

[Tilman]

Guys,

acceleration of the vehicle is a = F / m, accelerating longitudinal tire force is F = T / r with the torque T at the rear axle and the loaded tire radius r. Torque T at the rear axle is crankshaft torque T_c multiplied by all the gear ratios between the two, so one could assume that torque is the parameter that matters. But this only holds for instantaneous acceleration since crankshaft torque T_c is a function of crankshaft rpm which is a function of vehicle velocity which is the time integration of vehicle acceleration, so we have a first order differential equation with independent variable time t (neglecting shifts, partially engaged clutches, tire slip and maximum tire forces).

Since it is our goal to minimize the time needed to travel a given distance (which is the time integration of vehicle speed), we first need to solve our differential equation and this cannot be done without knowing the connection between crankshaft torque and rpm, hence the power curve.

Right now I do not see where engine power is introduced into the equations since I have not written it down properly but I think it is obvious that one cannot easily argue "flat torque curve is the best because ..." or "constant power over a long rpm range is best because ..."

There is a paper out there called "Shift-time Limited Acceleration: Final Drive Ratios in Formula SAE" (http://papers.sae.org/2004-01-3554/) which should cover some aspects that occurred here.

[Mbirt]

Does anyone have the link to the Powertrain Matching presentation made by John Bucknell at the SAE Collegiate Roadshow in Sept. 2006? It does a great job of erasing all doubt that power is indeed absolute and that the highest average power produced during an acceleration event will provide the lowest elapsed time.

[Tilman]

Originally posted by Mbirt:
Does anyone have the link to the Powertrain Matching presentation made by John Bucknell at the SAE Collegiate Roadshow in Sept. 2006? It does a great job of erasing all doubt that power is indeed absolute and that the highest average power produced during an acceleration event will provide the lowest elapsed time.
http://students.sae.org/compet...aseries/roadshow.htm (http://students.sae.org/competitions/formulaseries/roadshow.htm)

wasn't that hard to find ...

[Mbirt]

Originally posted by Tilman:
http://students.sae.org/compet...aseries/roadshow.htm (http://students.sae.org/competitions/formulaseries/roadshow.htm)

wasn't that hard to find ... D'oh! I searched these forums, google, the Detroit section SAE site, but not the obvious one. Thanks for the find.

[mk e]

Originally posted by Mbirt:
...It does a great job of erasing all doubt that power is indeed absolute and that the highest average power produced during an acceleration event will provide the lowest elapsed time.

This bit isn't exactly true, well true but maybe a touch misleading. It true as long as zero power is assigned during to shifts. This was the point being made about wide torque or more importantly wide power curves.

The way I explain it is:
power tells you what you can do with the torque you have available. If you want to know like how fast can you go or how fast can you accelerate, power is the number that tells you the answer.

[Warpspeed]

Torque is really a static measurement, but power introduces the time element.

The faster you wish to go, or the lower the lap time required, always demands greater available power to achieve it.

You will get more acceleration at the power peak with more torque multiplication in the transmission than at the engine torque peak pulling a taller overall gear at any given road speed.

There may be less ENGINE torque at peak power, but you get all that back and more, through the extra torque multiplication gained in the transmission.

It is the single reason why racing engine designers always try to maximize engine power output.

[Z]

It is indeed power that makes a racecar go fast.

BUT! if you have a high powered but peaky engine, then you also need a good transmission. Preferably IVT/CVT, or less effective is a multispeed gearbox with very fast shift times. The fact that most racing series have mandated crappy "change speed" boxes over the last 50+ years (with semi-autos only allowed recently in some series) has meant that most good engine builders have gone for a wide flat torque curve before maximum power.

It is a bit different in FSAE, as I ranted on about in the "Any way to objectively choose an engine" (http://fsae.com/eve/forums/a/tpc/f/125607348/m/824105905?r=56320469051#56320469051) thread. Briefly, much of FSAE is traction limited (ie. max power through the restrictor (~120hp) can spin the wheels almost everywhere), so the ideal engine has a wide flat torque curve, with only a single gear, which has a ratio that will only just spin the wheels at WOT.

Z

[M. Nader]

Originally posted by Z:
It is indeed power that makes a racecar go fast.

BUT! if you have a high powered but peaky engine, then you also need a good transmission. Preferably IVT/CVT, or less effective is a multispeed gearbox with very fast shift times. The fact that most racing series have mandated crappy "change speed" boxes over the last 50+ years (with semi-autos only allowed recently in some series) has meant that most good engine builders have gone for a wide flat torque curve before maximum power.

It is a bit different in FSAE, as I ranted on about in the "Any way to objectively choose an engine" (http://fsae.com/eve/forums/a/tpc/f/125607348/m/824105905?r=56320469051#56320469051) thread. Briefly, much of FSAE is traction limited (ie. max power through the restrictor (~120hp) can spin the wheels almost everywhere), so the ideal engine has a wide flat torque curve, with only a single gear, which has a ratio that will only just spin the wheels at WOT.

Z


+1

[ZAMR]

Originally posted by: Warpspeed
If you could choose between fitting a V8 engine with 400 Ft/lb of torque, or a lawnmower engine that was heavily geared down with a gearbox to also generate 400 Ft/lb of torque out of the gearbox.
Which do you think would make a vehicle accelerate and travel faster.
Remember the torque is identical, but the POWER is hugely different.

Power is the primary factor.


Originally posted by: Warpspeed
The faster you wish to go, or the lower the lap time required, always demands greater available power to achieve it.

You will get more acceleration at the power peak with more torque multiplication in the transmission than at the engine torque peak pulling a taller overall gear at any given road speed.


Originally posted by: Z
It is indeed power that makes a racecar go fast.

BUT! if you have a high powered but peaky engine, then you also need a good transmission. Preferably IVT/CVT, or less effective is a multispeed gearbox with very fast shift times. The fact that most racing series have mandated crappy "change speed" boxes over the last 50+ years (with semi-autos only allowed recently in some series) has meant that most good engine builders have gone for a wide flat torque curve before maximum power.

Based on these arguments, if I was simply trying to determine if power or torque was more important, I would remain unconvinced based on the amount of handwaving. I'm going to try a long example to see if it clears some things up for the uninitiated.

First I'd like to point out that a FLAT TORQUE CURVE IS THE SAME THING AS A PEAKY POWER CURVE. Any broad power curve means that your torques are decreasing with respect to increasing RPMs. This also means, inevitably, that you are sacrificing your peak hp in order to extend the range of your "peak."

What determines the acceleration of a car at any given moment in time (assume power limited) is the torque applied at the wheels. Take the lawnmower engine example: Our lawnmower powered car will accelerate at the exact same value as the 400hp V8 will provided we gear them to produce an equal number of torques. The catch is that the lawnmower car is geared so low it can only achieve this acceleration at say, 1/10 mph.

The following example requires the use of an IVT (engine RPM stays the same, so bhp and brake torque stay the same) and some simple maths:

Any attempt to move the lawnmower car faster will prove futile because the gear ratio is 10000:1 at 1/10 mph. To get to 10mph the gear ratio would be 100:1. Your torque has decreased 100-fold between the two speeds.

Now take the V8. Let's say it can accelerate at the same value as the lawnmower car, but at say 40mph. The overall ratio may be 10:1 at this speed. To change our speed to 50mph, the gear ratio changes to 8:1. Now we have only lost 20% of the torques.

At this point we have assumed constant engine RPM and constant power (torque at the wheels changes with speed but the power and torque produced by the engine is constant).

Now do the same thing but assume we have a gearbox. Assume constant engine power and that we are dealing with a single gear. Mathematically there is no difference! The lawnmower powered car can now reach 10mph by increasing the revs. However, since power is constant, the torque produced by this little guy decreases linearly with increasing RPM, and his ability to accelerate decreases 100-fold between 1/10 and 10mph.

The same happens for the V8, except its torque only decreases by 20% from 40 to 50mph so its ability to accelerate is not severely affected.

So yes, if we use a constant power model, brake power is what makes a fast car fast, or what makes it "fast" at usable speeds.

However, when you run your acceleration computer models, you are inevitably using torques in your f=ma maths. What matters, at least mathematically, is the torque your wheels receive throughout the various speeds your car will see during the race.

So I'll have to disagree with Warpspeed's second quoted comment. If you are traveling at any given speed (40mph), the only thing that affects acceleration at that instant in time (when you are at 40mph, crossing from 39 to 40mph) is the torque delivered by the wheels to the road.

Like Z said, and I'm sure he explained it in his other thread but I'm too lazy to go read the whole thread (that would take months!), the most desirable setup is a flat torque curve, which inevitably leads to a "peaky" power band. However you don't have to hold your engine at this peak power to get peak acceleration. You have multiple RPMS to choose from that all produce the same torques and thus the same acceleration for whatever speed your engine is geared for.

If we assume unequal, but flat torque curves for two different engines, and that we geared the car such that the torque produced at the wheels is equal regardless of which engine is in it, the engine with the higher torque can sustain the application of torque to the wheels for a greater speed range, all other things being equal. So between identical cars, one with 70hp and the other with 80hp, they can both accelerate at the same rate up until the car with 70hp has to change gears.

Does this make sense?

-Zach

[Warpspeed]

Zach,

Lets compare two engines of equal top end power. One has constant power, the other constant torque.

A flat torque curve will give a constant rate of acceleration (neglecting aero !) up to the max rated power.

Now consider a flat power curve. Torque becomes massive at the lower rpm, and gradually falls to minimum at the top end power peak, where it will be the same for both engines in this hypothetical example.

The constant power engine will have much more torque everywhere except right up at the peak power, where both engines would be the same.

I would beat you, at any speed, simply because I had more torque everywhere.

[Tilman]

Heyho,

@ ZAMR:

That is exactely what I was trying to say: There is a huge difference between an instantaneous consideration in which torque at a given rpm is the only thing that matters. But since a race is not instantaneous, there is no other way than (numerically) solving the differential equation to get some precise numbers.

By the way: There is a huge difference between accelerating a car from 0 kph until it went 75 m and accelerating a car from 50 kph until it went 75 m because you cannot neglect aerodynamic drag in the second example. You need some power/torque to overcome drag and this cannot be used to accelerate the vehicle.

@ Warpspeed:
You are comparing completely different engines! If I had an engine (one!) and two pairs of fuel and spark maps, one pair that gives me constant power over all rpms and one pair that gives me constant torque over all rpms with top power being equal to the power of the constant power pair than I would use the constant power pair for the reason you mentioned. But the point is: The constant torque pair is artificially throttling the engine! That is the reason why your example works, but it has nothing to do with the engines we are using.

However, I do not have an idea which condition was acceptable to declare a constant power engine and a constant torque engine as "equal" or comparable.

I do not see another way than defining some test conditions like the 75m acceleration from 0kph or some 50m acceleration on an autocross straight, taking measured power/torque curves and doing some numerical simulations.

[Warpspeed]

That is the reason why your example works, but it has nothing to do with the engines we are using.

My example was an EXTREME hypothetical case.

But something like a gas turbine car with only one fixed overall transmission ratio may come close to a constant power source.

But to drag us back to reality, consider two very different internal combustion engines pulling air through an identical FSAE restrictor.

One is a very highly tuned small capacity engine that roughly approximates the peaky power curve and constant torque ideal.

The other is an absolute monster capacity engine... with a positive displacement supercharger.... This engine produces no more top end power, simply because the restrictor strangles it.
But it develops a huge amount of low end and mid range torque, simply because the sheer volumetric displacement keeps the restrictor well into the choke region over the entire rpm operating range.

If such an engine could be built that was not excessively large or heavy, might it not be superior ?

Again we are talking hypothetical extremes here just to illustrate a concept.

Building an engine where the rules require a fixed diameter restrictor, is very different to building an engine of maximum allowed volumetric displacement, as is often the case in some racing classes.

And to take us right back to the thread topic, there may be more to be gained by re thinking the whole engine, rather than just trying to build some super efficient restrictor that is .001% more flow efficient.

As in many forms of competition, there are sometimes subtle ways of gaining a large advantage unrealized by other competitors, while still working strictly within the rules.

[Wesley]

Originally posted by mk e:
High rpm hp does not = peaky any more than more than low rpm power = board power delivery.

Your assertion that the usable rpm range is compressed by gear reduction is correct and needs to be considered, but the assertion that using gear reduction = slow acceleration and poor lap times is not.

Have you ever seen an F1 car accelerate? http://fsae.com/groupee_common/emoticons/icon_wink.gif

I do agree with what I think were your under lying assertions though, balance is the key to success and in that balance focusing on easy of operation over absolute performance will probably produce a better competition result.

The impact that the intake and exhaust design on the power peaks is small compared to the impact the factory decisions regarding valve and port size and cam specifications have on the power curve. Given our small displacement, high revving engines, trying to create a power peak at low RPM using factory heads and cam specs is generally an exercise in futility. The cams, heads, and exhaust are designed from the factory to make power at high RPM and you are not going to "undo" this with a new intake and header and create a power peak with a very substantive change.

As a result, most teams optimize power levels from 9-12000RPM (on 600cc 4-cyls) because above that and you get closer to choking the restrictor, and below that, the cam and port geometry drops the motor on its face no matter how many millenia you spend simulating and dyno testing your intake and header.

Tuning for a power peak at any specific RPM results in a "peaky" engine whether low or high, but within the bounds of this competition, our power peaks are created at high RPM because that is the nature of the engines we start with; most teams do not do extensive port or cam work.

My assertion that lap times and acceleration times suffer is based on anecdotal assumptions:

- Shift times are in the 50-100ms range. Whether this is because the car has a manual shifter or because Johnny Fatfingers can't let go of the upshift button. Every shift you require adds that time proportionally to your acceleration score.

- Our cars are traction limited already, even with small displacement engines. Because of gearing and tire limitations, you can only front end load your acceleration by so much before you start reducing accel times, even with infintesimal shift durations. Tractive force diagrams will show you impressive numbers until you consider tire data in the critical launch region.

-Working within the gearing limitations of motorcycle gearboxes, the split in ratios is very unfavorable for our 3-4x reduced speed range from original design intent. Because of this ratio split, a broader torque curve and larger axle ratio bandaids this design deficiency. This is of the largest impact on the auto-x and endurance, and a lot of corner exit power is lost to this lack of variability (combined with drivers that don't always hit that downshift just right.)

No number of gear shifts can put you in gear 3.5 unless you've had a really bad day. This is clear on a tractive force diagram of any typical FSAE car, with long stretches in 2nd and 3rd gear and almost complete disuse of 5-6th gears.

All of my posts are based under the assumptions of an OEM motorcycle-engine powered FSAE car, and do not apply to purpose built race cars not working within the design boundaries of OEM consumer products.

[Wesley]

Originally posted by Warpspeed:
I would beat you, at any speed, simply because I had more torque everywhere.


Bingo. Has nothing to do with power. It is important to note that power is a calculated number from torque itself. Torque is measured. Cars are unique in that they operate across a wide speed range, but with sufficient gear adjustment and manipulation (with transmissions, ratios, etc) that can be overcome.

[Buckingham]

When can we finally stop arguing 'power vs torque' or 'which really accelerates the vehicle'???

Torque [J] = Work [J] = Energy [J}

Power by definition is the time-derivative of Energy.

POWER [J/s] = dENERGY [J] / dTIME [s] !!!!!!

so...


Bingo. Has nothing to do with [the change in energy over time]. It is important to note that [the change in energy over time] is a calculated number from [the change in energy] itself. [the change in energy] is measured.

sounds kinda funny when you say it like that http://fsae.com/groupee_common/emoticons/icon_smile.gif


POWER [J/s] = dENERGY [J] / dTIME [s]

If we want a known change in speed (dENERGY) to occur in less time (dTIME) we need more POWER.
It's really as simple as that.

"But what if I increase X Torque at this RPM for a sacrifice in Y Power at that RPM, is that better?"

Plot both on Power vs RPM curves and calculate which has more average power over the operating RPM range. Whichever has more, is better.

[Wesley]

Originally posted by Buckingham:
Plot both on Power vs RPM curves and calculate which has more average power over the operating RPM range. Whichever has more, is better.

Absolutely. You can't have power without work. We're arguing about rate dependent lap times (power) but there are a lot more considerations than that (torque and tractive force) that actually make you go fast.

[Mbirt]

Originally posted by Buckingham:
When can we finally stop arguing 'power vs torque' or 'which really accelerates the vehicle'???

Torque [J] = Work [J] = Energy [J}

Power by definition is the time-derivative of Energy.

POWER [J/s] = dENERGY [J] / dTIME [s] !!!!!!

so...

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content"> Bingo. Has nothing to do with [the change in energy over time]. It is important to note that [the change in energy over time] is a calculated number from [the change in energy] itself. [the change in energy] is measured.

sounds kinda funny when you say it like that http://fsae.com/groupee_common/emoticons/icon_smile.gif


POWER [J/s] = dENERGY [J] / dTIME [s]

If we want a known change in speed (dENERGY) to occur in less time (dTIME) we need more POWER.
It's really as simple as that.

"But what if I increase X Torque at this RPM for a sacrifice in Y Power at that RPM, is that better?"

Plot both on Power vs RPM curves and calculate which has more average power over the operating RPM range. Whichever has more, is better. </div></BLOCKQUOTE>Forreal!

This isn't that hard for us to grasp, but the description of torque and power as seperate entities in automotive literature and among gearheads in general definitely breeds confusion among the impressionable. Correcting this should be the next crusade after squashing AGW theory and draconian CAFE standards.

[Z]

Originally posted by Buckingham:
Torque [J] = Work [J] = Energy [J}
To All Above,

Firstly, good to see this sort of discussion of fundamental principles.

Secondly, I must be a bit of a PITA and point out that Torque is NOT EQUAL to Work or Energy.

Work, and so also mechanical energy, is loosely defined as a "force by a distance moved". But, more accurately, it is the SCALAR PRODUCT of the force and distance vectors, hence giving a scalar quantity (work or energy have no obvious direction).

On the other hand, Torque (I prefer "couple") is the VECTOR PRODUCT of a force vector and distance (or displacement) vector. Torques thus have very definite directions in 3-D space (although they are "free" vectors).

So to find the amount of mechanical work, or energy, coming out of a crankshaft, we have to scalar (or "dot") product its torque vector with its angular displacement vector (magnitude measured in radians), thus getting a scalar quantity. But note that "angular displacements" are "pseudo"-vectors, so have to be treated cautiously (non commutative).

Or we can get the power by scalar producting the torque vector with the angular velocity vector (radians/second). Angular velocity and acceleration vectors are well behaved vectors in that they commute.

(Gotta get back to work ...)

Z