I’ve been slow to write anything on this blog, maybe because it isn’t as limited-audience as facebook, so I thought I’d try to generate some content out of my daily ruminations. I decided to post an email that I sent to a friend yesterday. It is informal and a bit carguy-technical, so I’ll offer some preface:
I was considering how someone might build a very light sports car that is easy-to-drive-smoothly. I am critical of the weight of vehicles for a variety of reasons, but mostly because it has holistic benefits. Whatever attitude you have toward cars, it would probably be either resolved or improved by simply decreasing weight in a non-superficial way. Almost any issue, performance, economy, safety, political, social, etc. Many cars have not gotten lighter because a lack of reconsideration of what is necessary to overcome practical design challenges of the overall vehicle once the decrease in weight has been applied in many parts of the vehicle but probably not enough in others, such as in putting in smaller, simpler drivetrains. For example, a number of mid-engined sports cars suffer from a handling trait called “lift-oversteer.” I will detail this more, later, but it is largely caused among commendably light sports cars, like a Lotus Elise, that are at least partly dependent upon drivetrains derived from other, larger, volume-sales vehicles. Like a toyota matrix. It could be a very long conversation if I were to lay out in detail so that the non-enthusiast would appreciate, but it’ll have to be later. I’ve got a bunch of other stuff to do right now. I passed the MPRE last night, and now I have to sort out what steps are next on the way to actually practicing as an attorney, like getting sworn in, signing up for CLE classes. So, here’s the email. It’s long, but it’d be longer to go through all this stuff so as to explain each little detail, like what a “variator” is:
From: Dan Becker
Date: September 10, 2012 12:45:35 PM EDT
Oh, I just wanted to ask you about something related to a thought I had about cars… nothing important.The question was how to find out what the safe-operation upper input rev limit was on an Audi-LuK 01J “multitronic” CVT transmission.
edit: well, I just couldn’t help myself, and I rambled out what I’ve been kicking around in detail below. Only bother reading if you need something to laugh at and have the time to do so. Otw, ttyl!
So… the whole enchilada:
I came to this issue when considering the possibilty of a renesis mated up to an audi 01J Cvt for a mid-engined installation that was optimized for:
-fairly high proportion of weight on front end compared to similarly sized existing sports cars (so, minimum large component overhang…no porsche transaxles)
-from used parts.
-smooth and flat and long-rev-span torque band
-rapid, zero-drama (manual) shifts, even if a manual
-ALso: minimum quantity of internally moving parts from the donor components.
[The point there would be to design around existing mechanical “fuses” that make breakdown modes predictable and limited in quantity/ complexity of diagnosis.For example, the Renesis chews its rotor tip seals, while the )01J wears out its clutch plates.]
Feel free to remind me of the early durability issues of that audi cvt. One durability advantage I saw in using the rotary is its significant torque deficit compared to the Audi 3.0L v6 that was used with the CVTs in production.So even with an early model, maybe the clutch pack would survive longer between any rebuilds that were originally caused by the underbuilt clutches.But it makes its peak power at ~8400 rpm, and I don’t know whether the audi cvt can spin its input side that quickly, even at fairly low torque compared to its original engines.
The other hard-to-lookup thing about this idea would be finding out the centerline differences between a renesis crank and a CVT’s input (automated multiplate metallic “shim-style” clutch pack.) One thought I had about CVTs was why they weren’t ever offered with a clutch or a dampened-resistance manual adjustment lever mechanism for moving the variator* between speeds without having to rely on computer shift algorithm or preset ratios. Essentially giving manual control over continuous ratio selection, not just the ability to force rapid ratio changes between discrete preset ratios. This would of course be only a secondary project. I’d want to make sure the mechanically-unaltered components could work together first!
[*I’d probably have to repurpose a steering-assist pump from another car to generate the requisite engagement forces in sync with lever movement. Well… wait, if the point of most of these sorts of kit cars would be to use as few donors as possible and try to maximize weight reduction, I might try an RX8 donor, with the intent to get down to ~1800 lbs, and with a rearward weight distribution at that weight, maybe I could disconnect the assist mechanism from the steering and use it for this separate mechanism… although I’m not sure how well a steering rack designed for assist can be used without assist, by closing the loop on the hyrdaulics or running without hydraulics, if that’s even possible without damaging such a rack. hmmm.. using an Rx8 donor would get me all-round wishbone suspension, as opposed to struts, so that’s better for a lower car intended to have much curvier fenders and a more inset cabin, but then the wheelbase would have to be more than 100 inches to really be able to take advantage of the stock suspension geometry. 106 inches stock is much longer than I was hoping to use.This rotary-to-01J thought experiment was independent of the typical “one donor only” limitation I usually force on these ideas, since it was starting off with engine and transmission from entirely different sources.]
I had a thought that maybe some sort of reduction gearset, which also allows correction for a crank/transmission offset, like on a Lexus LFA, but the other way around, to allow the transmission to sit lower than the engine, could be a solution. It could mate up the renesis with the 01J at their natural heights, and fit between two in the gap between them within the span where some sort of adapter plate would be necessary in the first place. So that way both the engine input speed could be lowered for the transmission (if that’s necessary) and the transmission could sit lower to account for the potentially-too-high crank of the rotary. But who wants to hear a crank-speed straight-cut-gearset whining all the time? I bet a fairly quiet synchromesh setup would either have to be a custom setup costing as much as a used renesis or 01J or I’d have to take parts from some random transmission to repurpose…
Of course, the easier solution may just to try to get at much power out of a renesis at lower revs with different intake and exhaust, and then set a rev limit fuel cutout at a speed likely near the audi’s input rev limit. The problem with that could be simply the fact that if it can’t get anywhere near the rotary’s peak power, it may defeat a lot of the rationale** for using a rotary in the first place. Holy cow, this thought experiment is wayyyy more complex in text than it feels mentally.
[**the reasons why I was considering using a rotary in the first place is because of the surprising and surprisingly typical handling trait of lift-oversteer in minimum-wheelbase mid-engined cars (I had a “moment” in an elise without an LSD a few years ago). Most of those are far less stable than something like a boxster because of the dimensions desired only being practical with a transverse motor. The transverse packaging then usually limits engine choice to an inline-four or a v6. Those don’t fire evenly, so in a lightweight car with the engine right up against the cabin, manufacturers seem to have allowed relatively soft engine mounts to keep NVH from being unbearable. So that allows for both a significant angular displacement with braking or even with lifting throttle, as well as having poor control over the ensuing seconds’ (pitch axis) angular oscillation of the entire drivetrain. In a type of car with a weight distribution typically around 40/60 (elise s an example)! Using a rotary would get a lot of the same lengthwise/wheelbase advantages of a transverse engine, as well as lowering center of gravity,fires evenly and mounts transversely in the same space, so tighter engine mounts, longer moment arm facilitated by locating the mounts further apart along the length of the car, and far lower vibration would significantly cut handling detriment with throttle-lift and braking. The mass is also pretty close or lower than most four-cylinders. Around 250 lbs. So the weight distribution might able to be closer to 45/55 than 40/60 in a tiny sports car at roughly the same weight.]
…and to think that I still haven’t managed to explain to shannon just how much time and energy savings could be had by simply using a boxster for a donor for a mid-engined kit car design with identical footprint. Yeesh…
I think there have been some rotary-to-transaxle setups in the past, but they weren’t very graceful. There’s a VW-based rear-engined kit car called the Sterling that’s been in production for like 30-40 years and uses VW transmissions, and I think I’ve seen a rotary fitted to one, but that certainly could have only been possible to potentially more convenient dimensions (compared to a renesis) of an earlier rotary. I’ve also seen a lotus europa with one, but the transaxle in that car has a relatively long section of its transaxle ahead of the axles, probably facilitating some upward tilt or something. Again, maybe more convenient dimensions of an earlier rotary. I also think I saw a delorean with a 13B from an RX7. Or maybe it was that rare JDM 3-rotor that was never sold here.
I’d love to talk about any of this with you, as soon as any or all important stuff lets up for a minute, whenever.