posted on June 23, 2014 13:31
Project Miatabusa #17
Miatabusa EVO II is Finally Here!
If you’re one of the 6 people in the world who isn’t familiar with this car, step back, go HERE, and read every word of the drama (or comedy, or tragedy, we won’t really know until the end) that is Project Miatabusa. For the rest of you who have been waiting to hear what we did about the sackfull of hammers trying to beat their way through the bottom of our car, wait no more:
Step 1: Redesign everything
First, the root of the death rattle problem was our initial decision to remove the sprung hub that normally lives in the Hayabusa’s clutch basket. This led to a horrific gear backlash rattle caused by a difference of opinion between the crankshaft and the flywheel. The crankshaft thought it would be really groovy to accelerate and decelerate with every piston stroke, while the flywheel figured it would be just swell to spin around smoothly and continuously. Usually in these cases, when they’re bolted directly to each other, the two parties find some compromise, but when you stick a gear in between, the gear gets the worst of the argument.
The gear with the red arrow is the root off all our challenges.
The obvious solution is to put the Hayabusa’s sprung clutch basket back in, (since it seems to have been designed to prevent exactly this rattle), and drive the flywheel through it. When the crank wants to accelerate, it can compress the springs in the hub instead of forcing the flywheel to speed up, and when it wants to decelerate a few milliseconds later, the springs can extend again, letting the flywheel keep going at its steady pace. If we’re lucky, this harmonious new living situation will keep the peace in our crankcase.
Solidly connecting the gear to the shaft didn't work, but using the factory needle bearing wouldn't fly either.
The first redesign this requires is letting the power takeoff gear float on the flywheel shaft instead of being brazed directly to it like it was on Miatabusa EVO I. This will let the gear shuffle back and forth with the crank, while the shaft rolls steady with the flywheel. Back when the flywheel shaft was just the stock Hayabusa’s transmission input shaft, the power takeoff gear rode on a big needle bearing for just this reason, so the obvious thing to do is just put the stock parts back in there.
Not so fast. In the stock application, the gear would also intermittently spin on the shaft whenever the squidmonkey resting his balls on the airbox grabbed a handful of clutch. In our application, that needle bearing will never spin, it will just rock back and forth as the springs negotiate the differences between the crank and flywheel. Needle bearings really don’t like this kind of job, since some spots on the bearing race will see lots of action while others will see none. This ultimately leads to failure modes that would dump tiny bits of hardened steel into our crankcase.
So instead, we replaced that fancy needle bearing with a bronze bushing which is much better suited to this kind of cyclical, short-travel business.
Yeah, so we took our bushing to the mall for a Glamour Shots appointment, you got a problem with that?
Now that the stock sprung hub and clutch basket could move on the shaft, we had to figure out how to connect the sprung side of that hub to the flywheel. The stock Hayabusa clutch is a standard multi-plate affair, with half the plates splined to the ID of the clutch basket, and half splined to the shaft. So our path was pretty obvious. Simply make a big, thick, clutch plug that splines to both sides at once, making it act like a perfectly solid clutch stack.
Two problems down, eleventy billion to go.
Tuesday, June 24, 2014 6:30 AM
Holy heavens, I thought *I* was doing something insane and being super fastidious on my project ('86 Mitsubishi L144 with 1UZFE swap). You Sir, are on a completely different level. Can't wait to see how it turns out. Good luck!
Tuesday, June 24, 2014 9:33 AM
Glamor shots of a bronze bushing...lulz
Tuesday, June 24, 2014 10:10 AM
The engineering here is amazing. Somehow it seems that jamming a busa engine into a Miata is more complex that stuffing in an LS.
Tuesday, June 24, 2014 10:49 AM
Wow, this is definitely much more complex than I imagined it would be!
Tuesday, June 24, 2014 10:58 AM
It's more complex than any of us imagined!
Tuesday, June 24, 2014 11:20 AM
Well, like I always say, I always need to make some sort of silly oversight on my first few design iterations so that the next one can be what I should have done in the first place anyway.
Tuesday, June 24, 2014 12:18 PM
So, does all of this fancy pants stuff mean it won't be feasible to make this as a kit?
Tuesday, June 24, 2014 3:59 PM
This one's actually slightly easier to make than EVO I, if you can use the word "easy" in relation to any of this...
Wednesday, June 25, 2014 12:41 AM
im buying one even if i have to sell my organs. this is the best project on the entire interwebs
Wednesday, June 25, 2014 8:10 AM
Dave, why not run the engine on the bench and see if the death rattle is gone before putting it in the car?
Wednesday, June 25, 2014 4:58 PM
Because building a bench with wiring and cooling and mounting for the engine is harder than putting the engine in the car. Besides, we're visualizing success! Blind optimism all around!
Wednesday, June 25, 2014 5:04 PM
I don't think the gear ratios are going to work out too well... Unless I have my logic wrong, that primary reduction is just going to exaggerate the 'ratio gaps' of the already unsuitable Miata transmission. My rough calcs suggest if you rev the engine to 10500RPM in first gear, you will have a 4250 rev drop when changing into 2nd, then 3000 drop into 3rd etc. This will pull the engine completely out of it's usable 'powerband'. If you had a curb weight of 400Kg you might get away with it but it doesn't look like you will achieve numbers like that to me. Even with the closest ratio Quaife gear set available, things don't look much better.
Wednesday, June 25, 2014 8:16 PM
I highly doubt that you haven't (since you do work r&d), but are you guys tracking all the little changes that need to happen to make this kit production ready? Forget subframe, custom bell housing, and starters. Re reading the first 16 articles today (I know...) all sorts of little things like motor mounts, throttle cable mounting, and small bolts being changed out seemed to have happened in Dave's shipping container garage.
I would seriously consider a full kit for this over a ls powered miata but only if all the mess you dealt with showed up well engineered in the box.
Wednesday, June 25, 2014 10:52 PM
i cant speak for dr. coleman but the kit should only have the adapter itself and one "built not bought" sticker, nothing else. were trying to use that adapter on any sort of unsuspecting compact. if everyone had could have the same setup life would be boring
Wednesday, June 25, 2014 10:54 PM
everyone could have* sorry typo
Thursday, June 26, 2014 12:28 AM
In a kit like this, you're buying the parts you can't possibly build at home (like the bellhousing adapter, modded subframe, engine braces, custom shafts, etc.) and doing the rest yourself (getting engine mounts, pulling weight, wiring, etc.). If you think everything is going to come pre-packaged and ready to drop in, you should probably not go this route. I think it's a safe bet to assume that these kits will be made to order since they're pretty damn complicated.
That said, this is a PROTOTYPE. You have to figure out how to make things work before you can figure out how to make things work well. Prototypes are always messy and part of moving into production is cleaning up the loose ends. Obviously not everyone has a 9-pin mil-spec connectors lying around their garages. Once this particular car is finished I think we will be seeing some updates dealing with the kit itself.
@Adamw I think your logic is correct, but the powerband is still quite usable from 6000 to redline (11000). Also in a street car you don't relly want to be rowing through the gearbox between every stoplight. A turbo on this beast should really bring it to life as well.
Thursday, June 26, 2014 12:52 PM
@Adamw: Looking at the dyno of the 'busa engine in the first project post, first to second on the Miata 5-speed, shifted at 10,500rpm, drops it to 6300ish RPM. Which is right about the torque peak. Seems good enough to me. The 6-speed is the same 1-2, but better on all the other shifts.
Thursday, June 26, 2014 9:13 PM
@Dan DeRosia: There's a bit more to it than that, Although you might land near the peak torque at 6000RPM, at the same time the HP has dropped from about 160 to 110. Even though 110HP is still more than the original Miata engine would have you will experience a big drop in acceleration after a gear change which is not ideal for a performance car. I'm also not so sure that the crankshaft or primary gear will like being decelerated by something like 30-40% over a few tenths of a second every gear change for very long.
Thursday, June 26, 2014 10:30 PM
Of course HP drops, but torque at the wheels from redline at 1st drops only by something like 20% to where you end up in 2nd gear. But look at the Hayabusa gear ratios; they're not that much narrower. And the primary gear is no longer being forced to contend with clutchless upshifts, it's having shock loading buffered by use of the clutch and having a lot more time to deal with things than it originally had to live with.
Now, it's not ideal, sure, but the Miata boxes (especially the 6 speed) are about as good of a close ratio box you can find in a production car. And the 'busa engine doesn't have that narrow of a powerband.
Thursday, June 26, 2014 11:20 PM
the drag race boys run turbos on top of nitrous and launch on wrinkle wall slicks with that motor. stock gears too because they are already factory trick enough to be straight cuts (most every bike has helical gears at the crank and clutch
Friday, June 27, 2014 8:15 AM
@8695Beaters - I guess I didn't mean kit like it would show up as an ikea package, but more so for the Miatabusa specifically is everything being documented for 3rd, 4th, etc. prototypes. This way when you use a different chassis you don't have 4 years of construction, just a motivated month or two. The MotoIQ articles are great but hardly supply enough detail to follow and get a swap done.
Friday, June 27, 2014 11:03 AM
Adamw, you are wise to look at all the data with a skeptical eye. That's how I always approach something ridiculous like this as well. That said, the drop in acceleration between between gears is the natural and inevitable nature of any transmission. If we were running a 650cc engine, our ratio gaps might be too large, but with the incredibly wide, torquey power and of this engine, we drop right back at the torque peak, which is right about where you want to be. If the gears were any closer, you'd be past peak torque all the time.
Keep in mind that the difference between horsepower and torque is largely one of semantics. If you're looking at post-gearbox output, torque numbers need to be multiplied by the gear ratios to get at-wheel torque. If you're talking about horsepower, it's just horsepower, regardless of where you're looking at it. So, when you're looking at a drivetrain, pick a unit and do all your analysis in the same unit. I promise you that you'll get the same result whether you look at horsepower or torque (it is a mathematical inevitability), but you'll confuse yourself and all around you if you try switching units halfway through the discussion.
And no, the crank and gear won't mind the rev drop from shifting. That's a much smaller load than the violence of accelerating a piston and rod from dead stop at 11,000 rpm, and it does that all day long.
Also, the primary reduction gear does not impact the rev drop at each shift. Do the math. What it does is impact the actual RPM the transmission has to deal with, and the speed range each gear can cover. The true risk with the primary gear is that the overall vehicle will be geared too short. When that becomes obvious, there are alternative final drive ratios available to fine tune that.
Friday, June 27, 2014 11:10 AM
Protodad & theneil, don't worry! We know every millimeter of what's going on in this car and how to reproduce it, and are already working on a new engine mounting scheme that will be a little more IKEA-like than what we have now. Final result will surely be more built than bought, so it can stay flexible enough to go into other rear-drive cars, but there will be a well-worn path to Miata success for you to follow.
Friday, June 27, 2014 10:27 PM
is that "a well-worn path to Miata success for you to follow." bit mean the V2.0 is a success??!!
Saturday, June 28, 2014 2:03 AM
I think it's kind of funny you got hung up with the KISS principle regarding the crankshaft bearings, and finally "bowed to the wisdom" of the OEM. I don't know if I'm right, but after reading that my initial instinct was that the internal tolerances/play of the bearings would negate any crank alignment issues. If you can make them fit, there seems to be no such thing as too many load bearing surfaces.
It's always great reading your stuff Dave. FYI, I'm the guy that suggested the mid engine rear drive V8 Metro back in the days of SCC project car lore, also the guy that wrote in and argued the merits of 'merican "pony" cars and V8 torque back when drifting was just getting mainstream. I definitely remember Kojimas published rebuttal about how superior Japanese chassis dynamics were.
Saturday, July 05, 2014 11:46 PM
Tuesday, July 15, 2014 1:42 AM
Oh I am sooooo happy you guys are working this stuff out!! This really opens up a RWD option for us Classic Mini enthusiasts who don't want to cram something bigger than 1300 cc under the hood!
Monday, September 01, 2014 11:48 AM
so it was unsuccessful?
Monday, September 01, 2014 9:42 PM
Miatabusa EVO II has been just successful enough to keep us really busy fine tuning the sprung hub tuning to banish all the rattle noises. We've pulled the engine at least 8 times since this story, each time learning one more lesson about what is causing each strange noise. I think we've finally found the right combination of springs and shims and flywheel weights. The engine sounds and feels amazing at 10,000 rpm!
The weekly engine pulls have kept me too busy to write #18, but I should be able to get to it soon. Meanwhile, if you want to see and hear the car in person, it will be at Miatas at Mazda Raceway this weekend (Sept 6-7) at Laguna Seca.
Sunday, September 07, 2014 10:31 PM
I just saw your most recent work at Laguna Seca...impressive. :) I can't find your contact information on this site for the life of me--please contact me through the Catfish site, bauerltd.com . tks, Cord