posted on August 02, 2015 14:01
Project Miatabusa: Part 18 - Research and Destruction
It has been so long since our last update that you could be forgiven for dismissing Project Miatabusa as dead and forgotten. It is not.
It has, in fact, moved forward in such complex bursts of productivity that there was little time to document it as it happened. And then, frankly, I forgot some of it. What follows is my best recollection of what the hell just happened.
Last time, we introduced you to Miatabusa EVO II, an almost completely new design intended to get rid of a ball-shattering death rattle from the primary drive gear (the one between the crank and flywheel). After an agonizing two years of work to make the new design come together, you can imagine the breathless trepidation when we first turned the key. At first, there was nothing but the sweet, smooth sound of Hayabusa, but after that first second, there was this deep, confidence shattering screeching, hammering noise that shook the whole car. And then it went away…
All that was left after that brief but horrible sound was the purr of Hayabusa. Oh, and a high pitched screeching noise. And a kind of Ducati-ish clattering noise. And an occasional Thor-hammering earthquake. But mostly it sounded like this:
In spite of the quantity of new sounds, most of the time Miatabusa EVO II sounded dramatically better than EVO I. Figuring out the source of the occasionally horrible sounds coming from deep inside a solid block of shaking metal is a tricky black box mental exercise where you must take the few little sights, sounds and smells we can detect from outside the engine, come up with some theory about what’s going on inside, and then test that theory. In a word: science. Here's what I can remember about our recent bout of science:
Playing the black box game, we mentally separated the symphony of destruction thusly:
Noise #1: Screech.
This sound correlated with the clutch pedal, appearing only when we pushed the pedal just hard enough to put the throwout bearing in contact with the clutch. Clearly this was just a rusty throwout bearing from the car sitting too long without an engine. Remember, the EVO II Miatabusa appeared after a long period of Miatabusa inactivity. Easy...
Noise #2: Ducati.
The characteristically Italian ringing clattery sound isn’t there on all Ducatis, only the ones with dry clutches. The rattle sound only happens when the clutch is disengaged and the floating stack of clutch plates can rattle around in the bell-like outer clutch housing. The solid aluminum plug we put in place of the Hayabusa’s original clutch discs had the same loose fit as the stock discs, and that little bit of backlash seemed to be ringing this particular bell.
Up in the blurry corner of this shot you can see the baclkash between our clutch plug and the Hayabusa's clutch basket. This is what we blamed for making the Ducati rattle.
Noise #3: Thor hammering you in the taint
This was the trickiest and most dangerous sounding noise. Every time the revs dropped below 2000 rpm, this shuddering earthquake would envelop the whole car. After some careful contemplation, we hatched a theory: The sprung hub we installed last time prevents gear rattle by letting six springs compress when the crank accelerates and the flywheel doesn’t want to. As the revs drop, the variations in speed between compression and power strokes get bigger, and the inertia of our stock Miata flywheel and clutch are dramatically higher than the tiny little transmission that was speeding and slowing with the crank in the Hayabusa. Perhaps all that inertia gives so much resistance that the springs compress until they bottom out. Lower revs give more time for the springs to compress at each stroke, so the correlation with RPM makes sense. There’s also a skull-hammering CRACK when you whack the throttle open from really low revs, which also kinda sounds like the crank accelerating so much faster than the flywheel that it bottoms out the springs in the hub. At least that was our working theory.
Figuring this problem out required removing the engine more than a dozen times.
Engine pull #1
So, only a week after our celebratory donuts, the engine came out again. We replaced the throwout bearing to fix noise # 1 and installed a new clutch plug machined to a tight press fit into the clutch basket to take care of #2. Based on our too-much-inertia-for-the-springs theory for noise #3, we also switched to the lightest conventional flywheel we could put our hands on, a roughly 8 pound aluminum Exedy flywheel found languishing under a layer of dust at MD Automotive.
Firing it back up, the screech was history, so noise #1 was solved. The Ducati sound was also dramatically reduced, so noise #2 was basically understood, if not completely solved (there was still a little backlash between the shaft and the clutch plug that we blamed for the remaining sound). But Thor continued playing with his hammer, though we did get to slightly lower revs before the hammer came down, lending credence to our theory that the flywheel inertia was causing the springs to compress too much. The lighter the flywheel, the lower the revs can go before the shaking occurs.
Completely shattering our theory that the springs were actually bottoming out, though, was the look of the springs themselves. Each spring has a bit of spray paint on it to help identify the spring spec in a bucket of nearly identical looking springs in that magical fairyland where Hayabusas are built. If the spring was actually bottoming out, as our theory suggested, adjacent coils should collide with each other, leaving a mark in that paint. There was no such mark.
If we were really bottoming these springs, there should be a stripe of missing paint where the adjacent coils touch each other. So much for that theory.
Monday, August 03, 2015 3:45 AM
Let me be the first to say wow, that's a heck of a road to getting there. Congratulations! Look forwards to the track test bit.
Monday, August 03, 2015 6:15 AM
This is a good way to start a Monday morning. Work might just be tolerable today.
Monday, August 03, 2015 6:22 AM
It sounds like the Thor sound was the sprung clutch-flywheel system hitting the natural frequency. As I'm sure you know, when mechanical systems resonate without being designed to do so bad things happen (See the Tacoma Narrows bridge).
Interestingly, your solution raised the natural frequency above your operating conditions - the natural frequency of an undamped system is ~sqrt(k/m). The springrate of the hub (k) was increased, and the flywheel (m) was decreased. Neat stuff, looking forward to the track test!
Monday, August 03, 2015 7:18 AM
Wow, what a long journey just to get to this point. Happy to hear that everything is still moving forward. How is it that every time Project V8 Rx7 rears its head, Project Miatabusa makes a stunning comeback?
Also, did you guys modify the engine harness to make pulling the engine easier each time?
Monday, August 03, 2015 8:36 AM
Finally an update! Glad to hear you might be that much closer to making this a realistic swap.
Monday, August 03, 2015 8:51 AM
On dual mass flywheels they use a fluid grease in the spring assemblies Ive always thought that was part of the damping characteristics needed. Yes a true Path is always so straight LOL. Great work.
Monday, August 03, 2015 9:19 AM
This project is always a rewarding read!
When this kit gets on sale I will start to consider a mx5 😁 hope the track test brings mor WOT clips of this beast 😂
Monday, August 03, 2015 9:52 AM
Protodad, astute observation about the engine harness! Yes, we actually did relocate the ECU from the cowl to the engine bay to simplify engine removal. It now lives behind the Shorai battery.
Monday, August 03, 2015 10:21 AM
Do I get to be the engineer that points out that Tacoma Narrows wasn't simple mechanical resonance but instead aeroelastic flutter? Yes I do! It's a good day.
Dave, very cool stuff. It's impressive how far into uncharted territory you are.
Monday, August 03, 2015 10:42 AM
Dave is the steel ring a shrink fit? Or did you use loctite green and a few thou interference fit? Any kind of mechanical location on it? I think it would tend to walk toward the inside due to hoop deflection so I hope you machined a step in the drum so it cannot move inboard.
Monday, August 03, 2015 1:31 PM
MDR, thank you for being that guy...
Mike, the steel sleeve is a few thou interference fit, no loctite green, but yes, there is a shoulder to keep it from walking due to hoop deflection.
Monday, August 03, 2015 2:30 PM
Oh man...so excited to read this update!
Glad you've finally worked out the most major of the kinks in this swap. Nobody likes getting hammered in the taint.
I'm definitely interested as well, when/if this becomes a swap kit
Monday, August 03, 2015 2:37 PM
Not many people would have the resources or knowledge to come out of all that smelling like a rose. I can't wait to see a real shakedown now.
Monday, August 03, 2015 6:16 PM
Dave, at the expense of potentially sounding stupid since I am just going off pictures. Since you have access to nice machining, why didn't you just forgo the basket and the splined thingie to the input shaft and just make a once piece deal that the sprung hub can bolt to using doweling and proper bolts. The dealie can then go through the third bearing which should be fixed in the wedge thing, floating on the steel shaft instead, better for long term wear. After that it can become the output shaft. That would be stronger, have less gear rattling inertia, less failure points and less other places to make noise. Am I not seeing something?
Monday, August 03, 2015 6:57 PM
"I have not failed. I've just found 10,000 ways that won't work." - Thomas A. Edison
I am impressed that you keep pressing forward Dave. Most people would have given up by now.
Monday, August 03, 2015 8:36 PM
We're thinking along the same lines. Once we test a few more things, we'll likely settle on something like what you're suggesting. One reason we aren't there yet is because we had to fail parts fix-by-fix to figure out the root causes. The other reason is because we had so many bolt-related failures that we're convinced only a rivet will work, since we have to lock together four layers of moving parts, three of which are pre-existing Hayabusa parts.
Monday, August 03, 2015 9:36 PM
This definitely makes me question my desire for a bike powered car.
Monday, August 03, 2015 10:14 PM
I have an idea, use NAS shoulder bolts with the shanks countersunk into my proposed input shaft adaptor thing. This way the bolt tightening torque load will be on the shaft not the sprung hub. By sinking the shoulder bolts down, the shear loads will be on the hardened and thicker shank not the threaded areas as well and you can control how much clamp the sprung hub gets by how deep you make the countersink. Since the tightening torque is on the bolts, they will not be subjected to the fretting action that is probably loosening them now. You could also ream the holes in the sprung hub for a zero clearance fit on the bolt shanks. This will kick ass on rivets. You must let the shaft slide on the bearing, not have the bearing slide in the aluminum wedge. This will soon wear and make a racket. You could nitirde or WPC the shaft for super long term durability. Nitriding will probably last longer but affects tolerances. All easier than machining clutch drive dogs and flank drive splines!
Tuesday, August 04, 2015 3:48 AM
I would get purpose made engineered bolts.. from your experiences so far even quality off the shelf stuff probably wont cut it. I think were in the range of forces where measuring stretch will be required to get things torqued well. And these kind of bolts would be good for one use most likely so rivets might still be better but getting rivets right in this kind of application is no small matter you may need to make up some nice specialty tooling. I always wish off the shelf would work better.
Tuesday, August 04, 2015 6:05 AM
A flutter is a resonance caused by the deflection of a member in a fluid flow. It's still an unintended resonance, which was the point I was trying to make.
Technicalities aside, it's neat to see how you solved the problem.
Tuesday, August 04, 2015 8:14 AM
"Seriously (Babe), like 20 minutes from frowny face to a hot shaft in your hands."
Great pick-up line Dave!
Tuesday, August 04, 2015 9:01 AM
The bolts are coming loose and falling out, not breaking. No need for custom bolts, the right fastener for the job is made already.
Tuesday, August 04, 2015 9:13 AM
YESSS!!!! Thank you dave i cant wait to read more and to buy an adapter!!
Tuesday, August 04, 2015 10:02 AM
Super great about fasteners you are happy with, Im sure they will work for you. Ive been down this kind of road before and yes falling out is more often the problem. Standard stuff is great at some things but engineered fasteners bolts etc work through way more than just how strong they are and can save a project. Just something to think about. I actually am a rivet fan. Been in a Hayabusa modified race car great fun. So look forward to see your results.
Tuesday, August 04, 2015 10:04 AM
And I thought I was stubborn. Holy shit, you guys take it to epic levels. :)
Tuesday, August 04, 2015 1:53 PM
I miss one day of checking this thread and of course that is the day you finally post up. Wow! What a journey. Hope its not another year before we get some more fix for our addiction....
Tuesday, August 04, 2015 6:08 PM
Looks like it's still a fun and challenging project, guys! Love the R&D with torsional vibrations, natural frequencies, mode shapes and gear rattle. Every change has some sort of unintended consequence. Keep at it!
Tuesday, August 04, 2015 6:57 PM
An NAS bolt is an engineered fastener.
Tuesday, August 04, 2015 7:17 PM
Shoulder bolts with pinned castle nuts.
Wednesday, August 05, 2015 5:25 AM
This is such a great project and I spent some time looking over the Hayabusa motor so I am just trying to be helpful. Even a mil spec nas bolt is still engineered to a standard. Just going through all the things that can be done with threads and all the thread shapes is neat stuff, then clamping force, coefficient of linear thermal expansion (Inconel), stretch, true torque, finish, weight, materials/heat treating, and more. Standard is great but it also sometimes makes it too easy, there is so much fun to be had with fasteners you can even get into making your own that will be better than standard stuff. You could even come up with stuff that hasn’t been done or has been forgotten. Oh then of course how is the joint designed and working as a system Physics is physics but designing this stuff is art.
Thursday, August 06, 2015 1:46 AM
Hi, is there any reason you can't use good 'ole motorsport standard safety wire on the bolts?
Thursday, January 07, 2016 9:49 AM
Seems time for an update.....
Wednesday, January 27, 2016 6:02 AM
The whole stock sprung hub needs to be trash canned and a purpose built sprung coupler should be used in it's stead. Preferably a coupler made of steel.
In automation they use zero backlash spurs
that sometime have a narrow sprung piggyback gear that rides behind the main drive spur ring and provides constant back force on the primary (pinion) gear while the main driven gear is free to get down to business.
The basic problem I have with your intent on using that stock Suzuki hub is that it is not designed for the loads you will encounter under compression braking. It was designed to handle forces of about 800lbs of bike and rider, not 3 or 4 times that weight and inertia. There is a lot of force you will be placing on those 3 little rivets and aluminum connection that you simply haven't experienced yet. Would you trust that weak link when you go canyon bombing and put it to 100mph+ downhill compression braking loads?
Wednesday, January 27, 2016 6:16 AM
Follow up: Please check out the damped PTO that Yamaha uses in their RZ1 (RX1) and Apex based snowmobile engine designs for some inspiration. Those are designed to swing heavier clutches and snow machines experience heavy loading from jumps and uneven terrain.
Thursday, June 16, 2016 10:39 AM
At least let us know this hasnt been forgotten......