18 replies to this topic

[Pong]

I would like to know more about the effects of engine vibrations on the performance of a F1 car.
When the teams switched from 72 deg engines to 90 deg engines the vibrations increased but the lower centre of gravity more than compensated for this. Additionally the integration chassis-engine is supposed to be of critical importance for overall car performance. Why is this so?

[desmo]

I curious on what you base your assumption that the near universal change from 72 to 90 degree bank angles increased vibration.

[Pong]

Desmo,

my assumption is not based on anything. I just thought that vibrations increased going from 72 to 90 deg Vee. If the Vee does not influence vibrations - what does?

[Greg Locock]

V angle does affect the vibrations to some extent, because the firing intervals are no longer equal. This used to be common on V6s that were cut down V8s, and kept a 90 degree V. These were disgusting engines noise and vibration wise.

Perhaps the easiest way to think about it is to imagine each bank as an in-line 5 cylinder engine - if the V angle was zero you'd effectively have an I5, each 'thump' would be twice as big as if the V angle was 72 degrees. 90 degrees would be somewhere between the two.

Th damage that a vibration does is roughly (very roughly) proportional to the number of cycles times the stress^6, so doubling the size of the thump, and halving the number of thumps, would make it 32 times more fatigue prone.

However, I really don't think the vibration from a properly engineered V10 of whatever angle is a significant issue for an F1 car- it is probably about the same as that seen on your average in-line four at the redline (about 20g). The reason for that is that an in-line 4 is an even more disgusting engine than a 90 degree V6, and an I5 is not too bad, in that company.

[desmo]

Rest assured the bank angle will affect affect vibration, and in more than the obvious way of first and second order moments and forces of the reciprocating and rotating masses. The most critical engine vibrations in F1, unlike road cars I'm told at least from the engine designer's perspective, are crankshaft torsional vibrations, a springlike "twisting and untwisting" in the crankshaft in other words. In simple terms these vibrations occur in orders- fractions of the rotational frequency. As you might expect, these orders tend to occur at fractions of the crankshaft length from the node or PTO end corresponding to their order i.e. a second order TV occurs at twice the rotational frequency and the maximal torsional deformation will occur at or near the center of the length of the crank. If you play a stringed musical instrument, think of a guitar or vioilin string as an analogy with its harmonics analagous to the various order crank TVs. Some of these orders of TVs are more problematic than others within the working rpm range of an engine. All these twistings and untwistings at various orders are all going on simultaneously too. Picture that! Changing the bank angles will of course change the crankshaft geometry, the firing intervals, and the potential firing orders, all critical factors to addressing crank TVs. A change in bank angle might push a particular order TV that is problematic above or below the normal working range of the engine, and even if it looks worse on paper, might in fact be a real improvement.

Another thing to consider is that if one views a V-10 as a series of 5 V-twins (which the segmented crankcases make easier for me at least to visualise as such), a 90 degree bank angle will minimise the balance issues each cylinder pair, taken in isolation, creates. This is true of any V engine regardless the number of cylinders where paired cylinders ojn opposite banks share crankpins. Let me make clear this last point may not be a factor in the adoption of 90 degree bank angles as the de facto current F1 standard. But then again, it might. If the widening of bank angles were driven primarily by the intention of lowering CGs, it hardly seems likely that each team would independantly arrive at the arbitrary 90 degree figure as the best compromise.

I'd love to hear what others think. Why 90?

[Greg Locock]

I don't agree. 180 degrees (a boxer) is the best way to minimise the L/r forces away from each bay, because you'll have perfectly cancellation in the bore axis coordinate, and only have to worry about the piston side forces, which are a lot smaller.

Quite why they went to exactly 90 is a bit of a puzzler to me.

[desmo]

Greg, I didn't mean to imply that I knew that 90 or 72 or whatever degrees is the best solution. As you point out, a boxer design may well be theoretically, notwithstanding it's apperent utter unsuitablilty for F1 due to packaging/aero reasons. I just threw that out there because I can think of no other reason for the near universality (are Renault the only ones left not using a 90?) of 90 degree bank angles in F1. Wright wrote that "F1 engines have large out-of-balance forces and plus or minus 100g at 300Hz can be measured at typical electronics mounting locations, with higher levels at sensor locations on the engine itself."[!]

Wright has stated that the second order crankshaft TV is the most problematic in F1 engines and in fact is the primary factor limiting rpm to around ~19k at present. An engineer told me that 5th (seems fairly obvious for a V10), and 7th order TVs in particular also are a matter of concern.

[Wuzak]

90° is just a compromise, isn't it.

Between vibrations, compactness, stiffness in bending in 2 planes in its role as a chassis member, aero, etc, etc.

[Pong]

So the conclusion is that there are a complex system of vibration of several types and magnitudes...
My next question would be what the negative impact of these vibrations are?
I can imagine two:
1) the vibrations (specially the longitudinal ones) have a negative impact on the whole drivetrain. Wear increases with vibrations and to cope with these the weight increases.
2) all vibrations are sooner or later passed on through the tires to the ground. This will probably effect grip in a negative way.

Am I getting close here or is there more to it?

[Halfwitt]

Vibrations mean movement, movement means strain, strain means extra stress, which means things break if it gets out of control, specifically fatigue failures. Vibration also leads to dissipation of energy as heat which otherwise would be useful work.

[McGuire]

Originally posted by Greg Locock
I don't agree. 180 degrees (a boxer) is the best way to minimise the L/r forces away from each bay, because you'll have perfectly cancellation in the bore axis coordinate, and only have to worry about the piston side forces, which are a lot smaller.

Quite why they went to exactly 90 is a bit of a puzzler to me.

But won't a flat eight with four crank throws fire two cylinders at the same time, making the torque reversals on the crank roughly equal to a four?

[McGuire]

Originally posted by desmo
Another thing to consider is that if one views a V-10 as a series of 5 V-twins (which the segmented crankcases make easier for me at least to visualise as such), a 90 degree bank angle will minimise the balance issues each cylinder pair, taken in isolation, creates. This is true of any V engine regardless the number of cylinders where paired cylinders ojn opposite banks share crankpins.

Yep. To extend this model even further, we can then visualize this cylinder pair as one vertical cylinder, noting its vertical reciprocating force. Then when we splay it back out into two cylinders at 90 degrees to each other, the reciprocating force is the vector sum of the two forces working at right angles to each other. That is, at right angles to their combined angle. The former "vertical" forces cancel each other, and we are left with a horizontal force (parallel to the ground) at around 70% the magnitude of the vertical force of a vertical cylinder. But if we precisely halve the combined angle of this hypothetical cylinder pair from 90 degrees, deploying the cylinders 45 degrees to each other, then we have a Harley-Davidson, and we will be sudenly compelled to dress like leather queens. Apparently the vibratory periods resultant to this specific cylinder bank angle have some psycho-sociological effect on the human brain, though it has never been satisfactorily explained.

[hydra]

Originally posted by McGuire


But won't a flat eight with four crank throws fire two cylinders at the same time, making the torque reversals on the crank roughly equal to a four?

According to Taylor, a flat-plane crank (we're assuming 4 crank throws here) flat-8 would indeed have 2 cylinders firing simultaneously making the engine feel (and sound maybe?) like a BIG 4 cyl engine. A cross plane cranked flat-8 on the other hand has an even cylinder firing every 90 degrees. However it does suffer from a primary shaking moment of 3a(2ZcosÈ-ZsinÈ) where a = bore spacing and Z= piston inertial force (mrù²). I was wondering if there were any alternate combinations for a 4-throw flat-8 that didn't suffer from either problem? Any ideas gentlemen?

[Greg Locock]

You need to separate out the gas forces from the inertial ones. Typically in automotive engines the inertial forces dominate at high speed. The second order (and 4th, 6th etc) inertial force comes from the non sinusoidal motion of the piston, which comes about because of the short conrod length relative to the stroke (aka L/r ratio).

There are two main reasons why vibration is bad

1) the crankshaft will break.

2) the engine accessories will fall off

TVs don't really matter as much in the driveline as there is a lot of compliance in the clutch system (at least in a production car - those little coil springs etc)

Another issue is the load in the main bearings - that is how the counterweights are sized. In theory you could run an engine with no counterweights, and then balance at the flywheel and crank pulley. In practice this will break the crank or bearings as it has to react all the out of balance (first order) forces through to the balancing planes.


I'm a bit surprised they get to +/- 100g (70 g rms) - no wonder they are paranoid. Ah, acceleration scales as rotational speed ^2, they are running at 3 times the speed, so at 6000 rpm they'd be at a fairly civilised 8 g rms - our I6 is more like 5g rms.

[ciaoduc]

Throw me a bone please gentlemen...
Help me out with the various orders of vibrations. I'm sure it could easily get over my head, but how many are there and is there a relatively easy to understand them??? I've read some threads about harmonics associated with flat plane V8 cranks and the like but never understood them.
Thanks.

[Greg Locock]

Order means multiple of rotation speed.

half order - this is the firing frequency of a single cylinder. If you have even firing intervals then all the half orders convolve together to form 'firing' order (5th order for a V10). Also same thing with valves.

Errors associated with firing tend to excite all n+1/2 orders, to some extent

First order - static and two plane balance

Second, fourth, sixth (?) eighth.... order - from L/r ratio

If you want to see a spectogram showing some of these orders hit

http://members.optus...ck/cepstrum.htm

[turby]

Greg,

I'm quite impressed about your acoustic analysis and the engine speed evaluation. I see you use mathcad for this.

Now, my question is quite off-topic to F1: how do you get your data into mathcad? AFAIK mathcad can read formatted data and the newest version can also read WAV-Data. But you have an AVI-File. How do you extract the audio-signal from that AVI and how do you feed this in mathcad?

Thanks in advance

Uli

[Greg Locock]

Hack.

In this case I used goldwave, from www.goldwave.com , to strip the avi soundtrack into memory, and then wrote it out as .txt. I think I then had to delete a few of the header lines (probably in editpad, too big for notepad) and then picked that file up in mathcad, the (engineering) world's most wonderful program. Failing that I would have written a program in qbasic.

[turby]

Greg,

thanks for the link. I gave it a try and it worked. I converted .avi to .wav which I could read into mathcad - really cool stuff.