Power, Torque and "X"

[Neal H]

Hi All,

I was thinking about engines and engine performance again the other day and I know there has been a lot written regarding torque and power and how these have an influence on vehicle performance. What I was thinking is that these are both static measures, i.e at "x" rpm an engine produces "y" torque and "z" power. Although important, these numbers would not determine how effective an engine is in a given application because to get from one velocity to another in a car you need the engine to rise in speed.

For example, if there were two engines of identical power and torque outputs at the same engine speeds (so on a dyno sheet both would appear identical), and one of the engines took half as long to get from one rpm point to another, it would be more effective engine and in essence, if everything else was the same, provide a faster car. Is there a known method of quantifying what I described above or am I talking total pap?

I know inertia would have an impact on this, but would other factors also influence how effective a given unit would be?

Your thoughts on this would be much appreciated,

Thanks,

Neal.

[Guy Croft]

You have a valid question. What you're referring to is a very real phenomenon - throttle response - the speed with which the cylinder condition responds to the new power demand. It is determined by transient (step) testing on dynamometers and engines do vary markedly. It is primarily a function of:

1. Valve - valve flow across the cylinder (purging and inlet tract inertial ramming)
2. Ex port pumping loss
3. Ex manifold back pressure and configuration (though this is really an installation criterion).
4. Inlet tract layout and size
5. Cam timing and other cam characteristics.

There are other things too, like fuel system etc etc. The list is long and the truth is you don't know how good it's going to be till you test and compare. Item 1 of course something that I deem critical in flowbench dev - optimising the valve-in response.

EG: In NHRA dev in the 90s I have done countless step tests on the 8V TC and compared recovery/response against much more powerful 16v Vauxhalls. The results in 3000-8500rpm range on full-throttle 250 and 500rpm steps showed that the TC had far better throttle response against the XEs being used at that time which is why they picked up out of corners and down the straight far better, but of course by the end of the straight the bhp of the Vauxhalls began to show.

GC

[Neal H]

Thanks for your reply, Guy. So when dyno testing you would measure the time taken for the engine to move from one specific RPM to another? Then to gather comparative data it would also have to be measured at the same gear ratio to compare different engines?

Thanks again,

N.

[NickRP]

You have a valid question. What you're referring to is a very real phenomenon - throttle response - the speed with which the cylinder condition responds to the new power demand. It is determined by transient (step) testing on dynamometers and engines do vary markedly. It is primarily a function of:
1. Valve - valve flow across the cylinder (purging and inlet tract inertial ramming)
2. Ex port pumping loss
3. Ex manifold back pressure and configuration (though this is really an installation criterion).
4. Inlet tract layout and size
5. Cam timing and other cam characteristics.

Very interesting points, indeed. Since factors 1-5 listed above also play a crucial role in torque curve shape, can it be assumed that strong and stable torque curve implies good throttle response (in the rev range where the curve is strong, logically)? I saw some transient tests evaluation study that made me believe that cylinders respond to the new power demand within a few engine cycles (given the ideal fueling and ignition maps), and that the step response is primarily a function of volume of air contained between throttle butterfly (or whichever throttling means employed) and intake valves (for the given engine design, of course). The tests were simplified in terms of always having atmospheric pressure on the throttle butterfly feed side (i.e. no air filter losses, no column of air that has to accelerate through ducting etc).

Best regards,
Nikola

[Guy Croft]

Neal, hi

re: Thanks for your reply, Guy. So when dyno testing you would measure the time taken for the engine to move from one specific RPM to another? Then to gather comparative data it would also have to be measured at the same gear ratio to compare different engines?

Yes. The load is held by computer control for as many seconds as it takes for the torque response and reading to stabilise and then reduced to allow the speed to increase to the next setting (250 or 500 rpmhigher). The time from say 4000-8000 can then be established. I don't know if it can be done easily on an ordinary rolling road.


Nik, hi

re: Very interesting points, indeed. Since factors 1-5 listed above also play a crucial role in torque curve shape, can it be assumed that strong and stable torque curve implies good throttle response (in the rev range where the curve is strong, logically)? I saw some transient tests evaluation study that made me believe that cylinders respond to the new power demand within a few engine cycles (given the ideal fueling and ignition maps), and that the step response is primarily a function of volume of air contained between throttle butterfly (or whichever throttling means employed) and intake valves (for the given engine design, of course). The tests were simplified in terms of always having atmospheric pressure on the throttle butterfly feed side (i.e. no air filter losses, no column of air that has to accelerate through ducting etc).

I don't have any exp of transient testing with varying throttle Nik. That's not by way of saying it's not a good way to test, because of course step tests with constant throttle - whilst a good test of how the engine responds to load - don't exactly simulate the real operating demands on an engine in the real world.

On the tests I described the throttle is held wide open and the load reduced to raise the rpm.
There are other ways of transient testing. You could for example drive the engine in 4th gear from near-stall to full throttle and see how long it takes to get to max wheel speed. Or you could drive the car up thru the gears.

On the former, the throttle is continuously open and on the latter it is periodically closing on gear changes. Either way a whole series of parameters are going to come into play. For example (a)the varying behaviour of pressure waves going out thru the inlet tract and back in past a plate that effectively presented a means of reflection (certainly at smaller openings - at full throttle it could be ignored). And (b) the complex behaviour (turbulence and mixing) of air as it traverses a throttle plate, in that it flows past the plate, which it does more-or-less as two independent jets either side of it. The mass of air betw throttle plate and valve at any given opening isn't, in my view hugely significant, since the negative wave at valve opening is going to force air (that is, at that time) slow moving/static upstream of the plate into the inlet tract to an extent governed by (a) above.

I'm not entirely sure one can argue that the mass of air 'trapped' between the valve and an instantaneously closed throttle is any significant criterion governing step response. I am not sure there is any proof of this.
Naturally, in the first instance, in order to evaluate it there has to be very fine tuning of the speed of re-application of the throttle (ie: from closed, to permit gearchange, to full open again). Assuming even that the pressure ratio across the valve (function of piston velocity) is sufficient to draw it in against closed throttle, you can be sure the engine will take the lot (or most of it) and the bigger the volume betw throttle and valve the longer the inlet tract recovery rate. Outboard of the throttle plate is atmospheric pressure (or above atmos if boosted) and thus the distance throttle - valve governs the time taken for the 1st positive reflected wave to get back to the cylinder. And very harmonic too of course. That distance is pretty critical.

And certainly, given that this is a race engine site, we are concerned with power, and there is strong evidence that locating the throttle plate close to the valve (as opposed to siting it a long way distant on the end of a plenum) gives beneficial fuel-air mixing.

A strong and stable torque curve doesn't necessarily imply good transient response. Torque - as you know - has to be measured at steady-state, a time period that may of course be only a few seconds long, whereas a transient evaluation has a time function. It might be a good indicator, might not be.

GC