dazz,

Great work! Shame that you won't be able to see it bear fruit. It seems that the biggest problem for some FSAE teams is the lack of any continuity of design philosophy. New team members come with their own ideas, and the "baby gets thrown out with the bathwater".

Anyway, your package would fit perfectly in the sort of car I have been promoting here. The driver's feet on, or behind, the front axle. All major masses packed close to the CG and low down. Plus a smooth running flat-twin with equally spaced intake pulses. It has a lot going for it...

In fact, ...
given that there are about 400 teams worldwide ...
there might be a good business case to build, oh ... say...
about a dozen of them???

That's just to dip your toe in the water...
Then with about $1-2k profit each, based on cheap student slave labour doing the machining, etc. ...
that should cover the cost of the rest of the car...

Up to the school though, I guess???

Z

quote:

Originally posted by dazz:
I got as far through the development as concept design and a feasibility study/report (for final year project).

Whenever I see small boxer engines, I'm reminded of the Lotus microlite airplane engine. Monobloc construction--only split at the crank line, output drive combined with the cam drive--1/2 engine speed for slower turning propeller, many other unique & lightweight features. Some pics here, http://stargazer2006.online.fr...craft/microlight.htm Prototypes ran, but it was never fully developed for proper aircraft reliability. I believe the original concept was by Tony Rudd, but others contributed.

quote:

Originally posted by Jan_Dressler:
Hi dazz,
"The space is there".. Brings back horror... Believe me, the space will not be there - not enough. Been there, done that. "snip"

The SV isn't as bad as the Honda looks to be:

- Gearbox needs a supply to the centre of input and output shafts and to a spray rail, all from the same side of the case, and quite close to each other.

- Engine needs supply to the two main bearings (crank big ends are fed from these also), 2 piston cooling jets, and a single supply to each head.

- A multistage dry sump pump built into the bottom of the engine cases opposite the starter motor would scavenge the crank case, and the left and right heads, probably via external lines.

I'm not saying it's a 5 min job, I would have already done it if that was the case, and I reckon that there's probably a similar amount of time required to what I've already put into the project to work out all the small details.

For reference, the SV manual is pretty nicely detailed as you can see below.

I’m going back to Z’s essay 1 here.

I agree that the packaging of motorcycle engines is less than ideal, but wasn’t sold on the idea of building custom crankcases to reduce yaw inertia. I wasn’t convinced that the performance gains from increased yaw acceleration would be worth the extra effort over using an off the shelf engine. My reasoning was that high yaw accelerations occur for only a small percentage of an autocross lap; so how important can yaw acceleration be?

I started to think of simple ways to simulate the effect of improving yaw acceleration. The main difficulty is that increasing or decreasing yaw acceleration alters the path a car will take through a corner. A car with reduced yaw acceleration capability will need to begin turning in earlier on the straight, and travel on a tighter radius at the apex to compensate for the slower turn in. Since this car is traveling on a tighter radius at the apex; it must take the corner at a slower speed to remain within its lateral acceleration limits. Trying to calculate the line a car will take corner by corner through a lap sim is pretty complex; so I tried to think of a simpler alternative.

I came up with the idea of simulating a slalom with some simplifying assumptions: constant velocity through the slalom, and constant yaw acceleration followed by constant lateral acceleration once the lateral acceleration limit is reached. The reason for choosing the slalom is that this is arguably the section of track where yaw acceleration will be most important, and the line taken by the car is easier to calculate since there are some simplifying symmetries.

An object travelling at constant velocity and undergoing constant yaw acceleration will trace a spiral path; the radius of curvature will be ever decreasing and the path will soon curl back on itself. Surprisingly; a numerical ODE solver isn’t required to find the endpoints of the spiral segment, although some of the simulation calculation steps need a numerical root finding method. From here it is a case of solving the calculations and geometry to get a working simulation with lateral acceleration, yaw acceleration, and track as inputs. All up, I probably spent almost as much time writing this dribble as I did getting the sim running.

Here’s a plot of the lines taken by two cars with different yaw acceleration and all else (track, lateral acceleration) equal:

Red represents the cones, blue is high yaw acceleration, green is low yaw acceleration, x’s represent the change from constant yaw acceleration to constant lateral acceleration. I used a large difference in yaw accelerations to exaggerate the difference in the paths taken.

Now I needed to figure out a typical range of values for yaw acceleration. I dug out some old data from our 2008 car, unfortunately my team didn’t have a yaw rate sensor so I calculated yaw acceleration from rpm (for some reason wheel speeds weren’t logged at this test) and lateral acceleration. This is probably the least accurate way to find yaw acceleration, short of rolling a dice. To make matters worse; the logging rate for acceleration was only 10Hz at this particular test. All these issues combined make the resulting yaw acceleration trace pretty much useless.

Simulation Results:
The simulation results are inconclusive because I can’t narrow down a range for yaw acceleration. If I use a yaw acceleration value at the high end of what I think it might be; a sensitivity analysis shows that lateral acceleration is significantly more important than yaw acceleration. If I use a value from the low end; the sensitivity analysis suggests that some of the ideas from this thread may actually give worthwhile gains.

This sensitivity analysis looks purely at performance parameters (lateral acceleration, yaw acceleration, and track) without considering what design parameters would be needed to achieve this performance. As a result there won’t be peaks in the slalom time results since increasing performance will always reduce the time taken through the slalom. Track is a performance parameter in this simulation since it only changes how straight the line taken through the slalom is without altering the lateral or yaw acceleration. Changing track in this simulation is equivalent to offsetting the slalom cones laterally without altering the car.

A sensitivity analysis of these parameters doesn’t tell the full story. The other thing to consider is how easy it is to improve these performance parameters. Track is easily changed; but in reality will affect lateral and yaw acceleration. As Z points out; reducing yaw inertia has a first order effect on yaw acceleration, while reducing mass has only a second order effect on lateral acceleration for a non-aero car.

Summing up: oddball engine packages might give worthwhile gains; or maybe not. As you were.

quote:

Originally posted by nowhere fast

Summing up: oddball engine packages might give worthwhile gains; or maybe not. As you were.

I lol'd. What was the change in radius of curvature from one setup to the other? And then what was the change in velocity?

I reckon most simple laptime sims don't take into account yaw inertia, but then again they take a simplified approach to the track map that makes the car artificially slower around the course (eg. a constant radius path leading to a straight is slower than an increasing radius path). The decisions the driver makes arguably effects laptimes more than anything, and having a car that responds very quickly to change in steering input will probably go around a track much faster than a lethargic car would, even if simulations show little difference. I think the best way to do this would be to drive the car with ballast in the driver's lap, then again with ballast on the nose and tail. Then you can make a sure-fire decision as to whether inertia is an important factor to consider.

Nathan,

Good work! That is exactly the sort of analysis all the new teams should be doing, rather than "what is design your car, respond urgently..." (with added spelnig mistakes). As you noted, it takes similar time to do the analysis, as it does to write it up.
~~~~~o0o~~~~~

Qualitatively, the gain from low yaw inertia is visible in Nathan's graph, at least as applied to the slalom. The high inertia (green) car must take a tighter radius turn near the cones, so will be slower. The big question is how much?

I think one "number" all teams should have for their cars is the "Radius of Gyration in Yaw". (Does anybody want to share theirs?) Combined with wheelbase (in a crude bicycle model way), this gives a non-dimensional parameter Ca = 2*RG(yaw)/WB, being the "Coefficient of Agility". I just made that name and definition up, but it is a good indicator of how quickly the car can change direction (smaller is better).

"Ca" is the ratio of the yawing resistance (proportional to the length of the dumb-bell) to the yawing control force (proportional to WB). Note that the maximum practical range for "normal" vehicles is about 1/2 to 2/1. A 200kg total mass car with "point mass" wheel assemblies of 10kg each, on a 1.2m x 1.6m rectangle, and ALL the remaining 160kg mass at a point at the centre, is equivalent to a dumb-bell ~0.9m long, for Ca=~0.56.

At the other extreme is a short wheelbase, lightweight farm tractor with fully loaded front-end loader and rear carry-all. Thus with most of the mass at the extremities. This can approach Ca=~2 (ie. dumb-bell twice as long as WB) and makes for very leisurely driving. You have to turn the wheel a long way before the corner, and then patiently wait for the nose to gradually come around...

I think the best way to experience these differences in FSAE is as ZAMR suggested. A certain quantity of ballast (eg. barbell weights, say 50kg), mounted either under the CG, or split between the extreme nose and tail of the car. Then time yourselves through a slalom, and extrapolate backwards to see how a lower yaw inertia car would perform.

I also think that while low yaw inertia is mainly good for slaloms and chicanes, it also makes the car easier to drive in unpredictable situations, such as traffic, or for an inexperienced driver who might come out of a corner "pointing slightly the wrong way". The high Ca example above requires the driver to plan well ahead. If not, then the only alternative is to brake hard and hope you hit something soft...

Z

Sensitivity analysis, yes, OK, good for comparison but talking absolute numbers, does it make any difference at all? I believe that an easy-to-drive and predictable car can do much much better than a (theoretically) faster car which has "spooky" handling. That happens all the time in professional racing series, as there are numerous examples of drivers doing better laptimes with (again theoretically) lower-grip set-ups, because the car is easier to drive or better suited to their driving style. Now add the factor that FSAE drivers (ours at least) are nowhere near pros... My point is that from a pure design point of view, sensitivity analysis IS the way to go, but for real-life....hmmm, not so sure! For real-life comparison I would go with the ballast test Z mentioned. Anyway I'd personally prefer a car tested for 2 months than a car 0.5sec faster per lap and 2 weeks of testing...

quote:

Originally posted by Z:
If not, then the only alternative is to brake hard and hope you hit something soft...

Z

The famous "oh fuck" moment....

quote:

Originally posted by Z:
I think one "number" all teams should have for their cars is the "Radius of Gyration in Yaw".

Olley and others started looking at transient response in yaw in the 1930's and discussed it in his Notes. We include it in "Chassis Design" as Section 4.7 "Moment of Inertia and Wheelbase", page 247-253.
The lowercase "k" is radius of gyration (a distance). Rather than use the total wheelbase, the simple model is based on the front and rear portions of the wheelbase, "a" and "b". The nondimensional figure of merit is k^2/ab.
http://books.sae.org/book-r-206

Doug,

Yes, and "k-squared-a-b" is even more commonly used in pitch mode analysis (ie. ride comfort - Rowell and Guest papers, 1920s, which are also covered in "Chassis Design"). Over the years I have kept coming back to the "dumb-bell to wheelbase" ratio as being more visually intuitive. It is roughly the square root of k^2/ab, so not much difference. (Incidentally, I think point "P" in Figure 4.24 of C.Dsgn is in the wrong place.)
~~~~~o0o~~~~~

Now for some quantitative analysis.

Based on countless assumptions, gross simplifications, etc., etc., I reckon lowering yaw inertia gives a 1 minute advantage in Endurance! Well, roughly... All other things equal... This worked out in the margin of the TV times, late last night... You decide...
~~~

Consider a car with wheelbase Wb, mass M with 50% F:R, yaw radius of gyration Kz (so Ca=2.Kz/Wb), tyre coefficient of friction Cf, and gravitational acceleration G (racing is slower on the moon). The maximum horizontal force that front or rear tyre pairs can exert is;

Ff,r = (M/2).G.Cf ........... (1).
~~~

Now look at Nathan's curves and consider mainly the section from X=-3 to X=0 (with X-axis along centreline of cones).

At X=-3 the car is moving to the right and in steady-state "circular" cornering. So both front and rear wheels have a lateral road-to-wheel force to the car's right (ie. downwards in pic). These forces converge near the turn centre, somewhere on a vertical line below X=-3 (neglecting Ackermann, slip-angle drag, etc.). So picture a series of different radius arcs (ie. paths of the car's CG) with their centres on this line (below X=-3) and tangent to (-3,1). Maximum velocity the car can travel along these arcs of radius R is,

V = Sqrt(G.Cf.R) ............ (2).
~~~

Somewhere between X=-3 and 0 the circular right-hand-turn phase ends, and the car transitions towards straight-ahead (at X=0), and then on to the circular left-turn (near X=3). This requires a change from clockwise yaw velocity W (omega, in radians/second) at X=~-3, to same magnitude but anti-clockwise W at X=~+3. Therefore, this phase is a period of yaw acceleration dW/dt. Picture this "S-shaped" transition phase as simply a straight line tangent to the circular arcs and passing through (0,0).

Big question - what is the yaw couple applied to the car during this transition phase?

Let's assume that coming out of the "circular" phase the driver instantly turns the steering and front wheels to the left. The front centripetal force thus disappears, but the rear centripetal force remains at maximum. Conveniently, at X=0 the rear force disappears, and the front force suddenly rises to maximum, but now leftward for the turn around the next cone.

Of course, the forces would vary differently, and there is "tyre relaxation length" to consider, etc. But this simplification means we have a nice constant yaw acceleration given by;

dW/dt = G.Cf/(Wb.Ca^2) .......... (3).

(From dW/dt=T/I, and T=F.Wb/2, I=M.K^2, so dW/dt=(M/2).G.Cf.Wb/(2.M.K^2), etc.)

This yaw acceleration (Eqn. 3) is half the maximum possible, so I figure it is a reasonable guess. (Note, maximum is if both the front and rear wheels push laterally with max force, but in opposite directions.)
~~~

Now, to get a feeling for the numbers, let's draw a table of these potential yaw accelerations for different grip tyres. The Cf=0.5 is for wet road, and Cf=3 is for an aero car, where downforce = more grip. Assume Wb=1.6m and G=10m/s.s.

Ca=1 ("long" dumb-bell with same length as wheelbase).
------------------------------------------------------
Cf...... =..... 0.5..... 1.5..... 3.0
dW/dt =..... 3.1..... 9.4... 18.8 (rad/s.s)

Ca=0.8 ("shorter" dumb-bell 80% as long as wheelbase).
---------------------------------------------------------
Cf...... =...... 0.5..... 1.5..... 3.0
dW/dt =...... 4.9... 14.6... 29.3 (rad/s.s)

So, the short dumb-bell car is quite a bit more agile (faster turning) because of "Ca-squared".
~~~

(Deep breath, we're almost there...)

Now we sketch a number of different radius steady-state corners, and draw the tangents to these through (0,0). This way we can estimate, by scaling, the transition distances S (quicker than using trigonometry). Note S/2 is half this distance, only up to (0,0). We now figure out how much constant dW/dt is needed for the transition. Some algebra (based on W in the corner, then this dropping to zero in the time taken to go S/2, etc...) gives;

dW/dt = G.Cf/(S/2) ............ (4).

And since equations (3) and (4) are equal (for a perfect driver), we get;

S = 2.Wb.Ca^2 ...........(5).

Again, just to check that this is making sense, we draw up some tables. Conveniently, a 3/4/5 triangle gives us a maximum radius R=5m, which requires infinite dW/dt for the transition S of zero length. Shorter radii give longer transitions.

Cf = 0.5, wet road.
-----------------------
R....... =..... 2........ 4........... 4.5......... 4.8......... 5.......... (m)
S/2.... =..... 2.4..... 1.45...... 1.1......... 0.75....... 0.......... (m)
V....... =.... 3.16.... 4.47..... 4.74....... 4.90....... 5.......... (m/s)
dW/dt =.... 2.06.... 3.45..... 4.55....... 6.67...... inf!....... (rad/s.s)

Cf=1.5, normal.
-----------------------
R....... =..... 2........ 4........... 4.5......... 4.8......... 5.......... (m)
S/2.... =..... 2.4..... 1.45...... 1.1......... 0.75....... 0.......... (m)
V....... =.... 5.48.... 7.75..... 8.22....... 8.49....... 8.66..... (m/s)
dW/dt =.... 6.17.... 10.3..... 13.6....... 20.0..... inf!....... (rad/s.s)

Cf=3.0, +aero downforce.
--------------------------------
R....... =..... 2........ 4........... 4.5......... 4.8......... 5.......... (m)
S/2.... =..... 2.4..... 1.45...... 1.1......... 0.75....... 0.......... (m)
V....... =.... 7.75.... 11.0..... 11.6....... 12.0....... 12.3.... (m/s)
dW/dt =.... 12.4.... 20.7..... 27.3....... 40.0..... inf!....... (rad/s.s)
~~~

Finally, we come to some conclusions;

1. Corner radii and transition lengths are only a function of Wb and Ca (for this particular slalom, see Eqn. 5). The long dumb-bell car (Ca=1) has a transition distance of S=3.2m, so has to turn a corner of R=~3.7m. The shorter dumb-bell car (Ca=0.8) can transition in only S=2.0m, so can turn a larger corner of R=~4.6m.

2. Corner speed (which is the "constant speed" through the slalom) is a function of R and Cf. So with the normal tyres of Cf=1.5, the long dumb-bell car travels at V=7.4m/s (27kph), and the shorter, more agile car at V=8.3m/s (30kph). The higher the grip (Cf), the greater the speed differential.

3. If we now assume that this 3kph speed differential applies for the whole Endurance (because the car is faster onto the straight, needs less braking for the next slalom, perhaps a bigger speed differential at hairpins, etc, etc, etc...), then if the long dumb-bell car takes 30minutes for its 22km, for an average of 44kph, the shorter dumb-bell car should average 47kph, and take only 28.1 minutes.

This, after conservative rounding, to cover all those wild assumptions, gives a 1 minute advantage! QED.
~~~~~o0o~~~~~

Corrections welcome (I doubt that I calculated/copied all those numbers right.....).

Z

PS. Anyone want to share how big their, err, dumb-bell is? That is, does your fancy CAE gizmology give a figure for your car's Yaw radius of gyration, Kz, complete with wheels and driver?

This message has been edited. Last edited by: Z,

quote:

Originally posted by Z:
PS. Anyone want to share how big their, err, dumb-bell is? That is, does your fancy CAE gizmology give a figure for your car's Yaw radius of gyration, Kz, complete with wheels and driver?

I'll throw out some numbers for our 2011 car that I have a medium level of confidence in. Both the driver and the aero are on the heavy side, but both are quite quick as well

Mass: 260 w/90 kg driver (235 kg w/o aero)
Wheelbase: 1.555 m
Rear Weight: 53%

MOI about Z:
Sprung mass 80 kg m^2 (60 kg m^2 w/o aero)
Unsprung mass 30 kg m^2
Total 110 kg m^2 (90 kg m^2 w/o aero)

Radius of Gyration: 0.65m (0.62m w/o aero)

Z's Ca Factor: 0.84 (0.80 w/o aero)

Milliken's k^2/ab factor: 0.70 (0.63 w/o aero)
or (k^2/ab)^(1/2) factor: 0.84 (0.80 w/o aero)

My last fsae car:
I_zz: 92 kgm^2 (measured on a rig at an OEM)
k: .587
Ca: 0.74

The current car I work on:
I_zz: 3500+ kgm^2
Ca: 0.91

Z, one thing I see missing here is the connection of these 'architectural' parameters back to the driver. One could make 2 cars with identical k^2/ab feel totally different with simple steering and tire changes. Including cornering stiffness normalized by steering ratio or some other metric will better project how the car will 'feel'. As Mr. Milliken writes, the linearized gains of sideslip angle, lat accel, and yaw rate vs steering angle can be quite valuable in developing a framework for fundamental vehicle architecture.

Both above cars are quite compact/agile, with Ca=0.84 & 0.74.

I think the last time production cars had Ca>1.0 was in the 1960's with some of the US "land yachts". Many pre-WWII cars, with front beam-axles necessarily in front of engine and little rear overhang, had Ca=~0.7 (k^2/ab=0.5).

So, can anyone beat the current FSAE record of Ca=0.74???
~~~~~o0o~~~~~

Ryan,

My feeling for these MoI effects developed way back when I was driving various very high Ca (=~2) "farm" vehicles, often on low grip surfaces like wet clay. This high-Ca, low-Cf meant that everything happened very slowly, so it was easy to watch and think about all the dynamics going on.

For example, turn the wheel for 45 degree slip-angle, then wait and watch as the front tyres gradually bring the nose around. Interestingly, while this is happening the rear tyres are pushing outwards! Obvious when you think about it (both pairs of wheels between the dumb-bell half masses). But probably overlooked if everything is happening too quickly, like on a high Cf racetrack.

You say "One could make 2 cars with identical k^2/ab feel totally different...". True, but you can make any two identical cars feel different by, say, letting the air out of one car's tyres. What I don't think is possible is to make a high Ca car feel agile. It is a bit like trying to teach an elephant to dance.
~~~~~o0o~~~~~

(Edit.) Using the same calcs as my above post, a Ca=0.6 car (probably the lower limit) would be 0.9kph faster than the Ca=0.8 car. So only about a half-minute faster over Endurance (law of diminishing returns).

However, adding a lot of front and rear overhung aero might be costly. Going up to Ca=1.2 makes the car 6kph slower than Ca=1.0, or ~2-4 minutes slower. The aero downforce might overcome that, or not???

Z

This message has been edited. Last edited by: Z,

It should if it's well designed. A downforce package on an FSAE car should be light and produce a lot of downforce at low speeds. One of the parameters we used to define our wing performance was the "break even speed" or the speed at which you create enough DF (hence greater available yaw forces) to overcome the added inertia of the wings. For our car we estimated it to be ~25mph, 30 if being conservative, which is the low end of the track speeds. However we were only able to accomplish this because of a very lightweight, high DF package. Heavier, less aggressive wings will not "break even" until higher speeds.

ZAMR,
At first I was hesitant to post numbers since I would rather teams make the effort to do this type of analysis for themselves. Although thinking about it now, quoting a number from a forum to a design judge without any work to back it up wouldn’t go down well. I tried to drop enough hints in my post that teams could start on creating a sim like mine, but without giving it all away. The biggest hint is that an fsae slalom is not the only application of this particular spiral curve, the most difficult parts of the math can be easily found online if you know what you’re looking for. For this sim; figure out what type of spiral you’re dealing with and you’re halfway there. Z’s method of simplifying the transient period to a straight line is even easier and not unreasonable since the length of the transient section is inversely proportional to yaw acceleration.

Harry,
Good point about the benefits of driveability and testing time. Ironically, sensitivity analysis can help prove your point. If a simple analysis like this shows that your car should score within say 10 points of the winner, but at the competition you end up hundreds of points behind the winner; then you know that the performance of your car is not what’s holding you back. This can help put into perspective the importance of issues like driveability, testing time, manufacturing time, project management etc.

Z,
I ran some of your numbers through my sim, the results are pretty similar. I used 9.8 m/s.s for g, I just realised you used 10 m/s.s which accounts for some of the difference. t is the time taken to travel from X=0 to X=3.

Cf=0.5, v=4.78 m/s, t=0.671 s, R=4.67 m, dW/dt=3.45 rad/s.s

Cf=1.5, v=8.28 m/s, t=0.388 s, R=4.67 m, dW/dt =10.3 rad/s.s

Cf=3, v=11.71 m/s, t=0.274 s, R=4.67 m, dW/dt =20.7 rad/s.s

This message has been edited. Last edited by: nowhere fast,

quote:

Originally posted by nowhere fast:
I tried to drop enough hints in my post...
The biggest hint is that an fsae slalom is not the only application of this particular spiral curve...

Nathan,

So I'm guessing that Leonhard must have been a horny bastard, because a horn is spiral shaped, and in Latin is "cornu". Except that Fred was a Frenchman??? On the other hand, in Greek mythology the Fate that was the "spinner" of the thread of life was Clotho.

Z

Yep.

http://fahrzeugtechnik.fh-joan...-Joanneum-Racing.pdf

Take a look at that...It seems like a neat little engine, developed by AMG Mercedes, Joanneum Graz and KA-RaceIng. Basics: 2cyl, 596ccm, 75kW, direct fuel injection, endurance consumption < 3,5l, integrated differential, first competition Formula Student UK 2012!

Wonder why this is not a "lean-rearwards cylinder" type of engine though....it seems that it could be built to package really neatly with integrated engine-chassis mounts etc. with a little more effort. Nevertheless it looks fantastic, cannot wait to see it in action! Any news from the other anticipated custom engines on FSAE? (Aucklands' and Brookes')?

quote:

Originally posted by mech5496:
http://fahrzeugtechnik.fh-joan...-Joanneum-Racing.pdf

Take a look at that...It seems like a neat little engine, developed by AMG Mercedes, Joanneum Graz and KA-RaceIng. Basics: 2cyl, 596ccm, 75kW, direct fuel injection, endurance consumption < 3,5l, integrated differential, first competition Formula Student UK 2012!

Wonder why this is not a "lean-rearwards cylinder" type of engine though....it seems that it could be built to package really neatly with integrated engine-chassis mounts etc. with a little more effort. Nevertheless it looks fantastic, cannot wait to see it in action! Any news from the other anticipated custom engines on FSAE? (Aucklands' and Brookes')?

Finally someone integrates a gear drive to diff in a transaxle setup...

But then doesn't integrate suspension pickups!

Regards, Ian

Ian, that might be because the engine is developed by 2 teams with AMG. Having suspension pickups (and chassis mounts) integrated, basically means that both teams will share rear halves of the car....

If I heard the guys from Karlsruhe right, they have more torque than basically anything ever seen in a combustion FSAE car. In every driving situation...

If it works as promised, it probably will be a great thing to watch!