Tales from the Batcave (RX-7, rallycross, misc. stuff content)

No pictures because working.

Rotor housings and replacement rear stat gear are washed.  Rotors are cleaned and rotor gears are smashed back down.  They have .007" clearance right now, have no berking clue what the spec is supposed to be.  Apex seals are clearanced for length and Superglued together.  Discovered that Atkins seals have nonstandardized apex corner pieces: the ones I got varied by .003" from longest to shortest.

If I hadn't forgotten my 10-1.0 thread chaser and my torque wrench at work, I'd still be in teh Batcave trying to assemble an engine.  Probably cursing the process of threading the intermediate housing over a slightly-lifted eccentric shaft.

So rotor clearance in the MFR engine building bible is .005-.007".  I'm on the happily loose end of spec, yay!

When stacking an engine, I prefer to have everything ready to go first, so I can work quickly.  I spent the evening loading the rotors with seals.  I forgot how much of a PITA it is to install the oil control rings when they have new O-rings in them.  But all seals save the apex seals/springs are loaded up in the rotors.

Then, I spent another good amount of time turning 2.4mm Viton O-ring material and 2.0mm neoprene O-ring material into inner and outer coolant seals.  $40 got me a 20 foot roll of the former and a 50 foot roll of the latter, enough to do two engines.  Probably three engines' worth of outers, actually.  The rough lengths needed are 30" per inner and 36" per outer.  So the inners basically cost about $3 each and the outers less than a dollar.

All tension bolt holes are chased and the tension bolts cleaned up and ready to go back in and torque up nice and smooth.

And then....  I decided that I want to try an adjustment to the porting, so I'm going to do that very quickly at work and I might actually assemble an engine tomorrow evening.

Obligatory "partially assembled engine" shot that everybody who builds an engine takes.

How do you seal the two ends of the O-ring material to each other to turn it into a ring?

codrus said:

How do you seal the two ends of the O-ring material to each other to turn it into a ring?

Superglue, works surprisingly well.

Or just let 'em float loose and use some sealant, like I have done for years.  Put the inner's gap against the coolant-dead area at the intake port, and put the outers' gap at the very top of the engine so you can see a leak if it develops.

When I used to use 18 gauge electrical wire, I'd have gaps of up to a half inch.  No problem.  Well, no seal gap related problem.

Went to the Batcave, sat down to munch on a burrito, and fell asleep.

I guess thrashing on the car after work and sleeping only four-five hours a night is not something I can keep up for very long.

The only thing i can think of when i see a rotary disassembled is how i’d be figuring out how to make a 6 rotor engine. 

I can get you in contact with someone who makes 4 rotor engines.

I don't know if he does the eccentric shaft himself or if he's using a (Guru?) shaft from Australia.  The trick is that assembling anything longer than two rotors requires a multipiece shaft and machining that is tricky.  Oh, and you're looking at the price of a modest house for the engine.  (IIRC the 4-rotor e-shaft alone runs close to $20k)  Given that and the fact that there are 2 rotor doorslammers in the 6s, I'm not terribly keen on the idea myself...

Anyway.  This brings up most/all of why I haven't put the engine together yet.  Stacking the engine is not something do to tired, or under time constraints.  In order to get the intermediate housing over the rear lobe on the eccentric shaft, you have to lift the shaft enough that you can gyrate the hole in the intermediate shaft past it.  At this point, what happens is the front rotor moves slightly, which makes one of the corner pieces of the apex seal ping out under spring pressure and disappear somewhere.  Then you spend a half hour looking for it, pop it back into place, try again, and it pings out AGAIN.

Any time I start to assemble an engine, it either goes very smoothly and I'm done in an hour, or it is a six-seven hour long ordeal.

I do have a plan for this one, though.  Not sure how good of a plan it is:  I am thinking of dropping the eccentric shaft and intermediate housing on as a single unit, with a disused rotor on top to stabilize the assembly.  Downside:  This will mean carefully maneuvering 70-odd pounds of iron into not one but two journal bearings.  I might'a been able to do that when I was in my 20s.  I also used to put transmissions in by hand.  Getting a little too wobbly to do that kind of precision gruntwork.

Hmm.  Consternation.  Decisions.  Time running out.

OMG I should have bought the engine stand adapter years ago.  Video is pending.

Okay, let's see how long it takes for this video to get blocked or the audio muted.

I always assembled engines over a bucket.  I should have bought this Mazdatrix engine stand adaptor a long time ago.  Bumping the e-shaft up with my knee makes swiveling the intermediate housing over the rear lobe a PIECE OF CAKE.

Assembling 2-piece apex seals backwards still sucks hardcore.  I think I had 45 minutes in trying to get the rear rotor's apex seals in.

Also, I'm going to find the guy who wired in this car and kick his ass.  I desperately want to replace all of the engine harness.  I settled for replacing some of the melted loom with new loom.

Also also, running a bridge ported rotary with a really "off" tune in a closed garage:  Do not recommend.

Feeling a bit woozy still. 

The pinion angle was BERKED.  There was about 4 degrees of front U-joint relative angle and 7 degrees of rear.  More than one degree different front and rear and 3 degree at either end is Bad.   I moved the pinion up a bunch and now it is three and three.

Discovered that ever since the incident last year or so when I had a coil failure and had to convert to two MSD coils, I have been running leading-only.  I sort of forgot to connect the trailing coil.  No wonder it ran like crap.

Did some math, came up with some req_fuel numbers that would get me in the ballpark if not actually correct (was running 6.4ms with 680cc injectors, am now running 1200cc-rated injectors at Mazda standard 3.5 bar, so I guesstimated 1100cc actual, and figured 4.0ms req_fuel is a good start point), said "well, who the hell knows?" for injector deadtime since apparently the company I got my injectors from is made out of failure and lies, so I just kept using my existing deadtime of 1.2ms arrived at by shrugging and saying "well, why the hell not?".  This annoys me and I really want to throw these injectors out and get a pair of Injector Dynamics injectors, but for now I just want the dang thing running.  Then drove around a bit.

Hey, that horrendous highway speed shudder is gone!  Amazing what fixing the drivetrain angles will do for that when the driveshaft is spinning at 6000-ish RPM.

Tune needs a lot more fuel in the cruise range and less fuel in other areas.  I know the tune was crap to begin with, so I don't know how much is due to poor initial tune and how much is due to porting changes and how much is due to the injector constants being FUBAR.

If you get the car running semi-decently, you can measure the injector deadtime by playing with pulse width and measuring AFR at idle and then graphing the results:

https://www.miataturbo.net/ecus-tuning-54/how-find-your-injector-dead-time-56061/

That is how I normally do it, but you can't measure AFR at idle on an engine that doesn't run smooth.  At idle the engine is jumping from 1400-1600rpm and the O2 is getting so much raw oxygen that it is useless for tuning.

Works a treat on engines with wimpy cams though.

Another problem is that it doesn't really work with large injectors.  If I doubled my squirts/cycle then my injection time would drop to under 1ms.  The very low open-time range is one of those places where injectors are very nonlinear, which is why more sophisticated computers will have separate maps to model the injectors at super low openings.

One of the reasons I want ID injectors.  Paul Yaw (yes, THAT Paul Yaw, if you're an old rotorhead) models his injectors for you and he will give you all the relevant numbers.  I don't see it so much as paying twice as much for good injectors, I'm paying twice as much for good injectors and information.

Oh yeah, no question the ID injectors are an excellent choice.  I have a set of the id1000s in my Miata.

So:  Wednesday.  NAPA Autotech class after work, so no Batcave action.  Got my learn on about R1234yf and also learned some interesting diagnostic tricks.  (We're probably going to get a R1234yf capable machine pretty soon, and no other shops in the area have such a beast, so the timing of the class was great)

Thursday.  Threw the laptop in the car and went for a drive.  I datalogged and the car was damn near undrivable.  Dead lean in certain cruise ranges and it was none too happy at high throttle either.  So, I pulled off to the nearest gas station and loaded up MegaLogViewer to see the log...  And the program wasn't coming up.

Note:  If you have a dual screen setup for your laptop, and you use MLV on the second screen, it isn't smart enough to restart on the main screen if the second monitor isn't plugged in.

So, on with Autotune, which doesn't really work all that great, but any port in a storm.  Got it running "better" but now it is pig rich under cruise conditions because Autotune isn't all that great.  But it's drivable.  After I got home, I connected the second monitor, started MLV, and moved its window to the primary display, so maybe it will work the next time I try to use it on the road.

Also discovered that the drivetrain shudder/hum is still there and is REALLY bad at about 100mph indicated, which is about 80mph actual, give or take.  It's bad enough that I will NOT visit a dragstrip until I get this fixed since I had been trapping in the 102-103mph range with the old setup and worn engine.

Friday. 

Err, right.   The other thing I did on Thursday was throw the car on the alignment rack after work to see how square I had the rearend in there.  Toe was .05 degrees on the left and -.04 on the right.  As close to square as I could reasonably want given how much toe steer there is going on back there, and anyways I know for certain that the housing is straight too.  I noticed while I was under there that the driveshaft looked bottomed out at ride height.  Erk.

I'm wondering if some of the shudder/hum is because the driveshaft is bottomed out and transferring natural U-joint length wobble to the... everything.

So, Friday.  I had a good look at the rear links and decided to move the lower link up to the middle hole.  I'll be giving up some anti-squat but I'll also be getting rid of some roll oversteer (meaning the rear axle steering the car) which is better for predictable handling.  Also, since I had to move the upper link higher on the axle, and my lower links were parallel with the ground before, I still have more anti-squat than before.  Moving the links to the next hole up also moved the axle about 1/4" back because of how the holes were drilled.  Then, for good measure, I rotated the link adjusters three turns.  16 threads per inch means 1/16th of an inch per turn of thread, so three turns means 3/16" of length at each end of the link, or 3/8" longer in total.  So I basically moved the rearend back about 5/8", giving myself more driveshaft plunge.

Now I had to fix the pinion angle again.  It only took two and a half turns of the upper link (which is also 3/4-16 threads) to make it right, as far as I can tell with my crappy bubble level.   So the angle was either wrong before, or it's wrong now, because in theory the links should have needed to be extended the same amount...

Today:  Ferried five rally tires down to Columbus to stick on the Perfect Trailer.  Which needs rewiring again.   Aaand... it's raining now, so I can't do more tuning on the car just yet.  It's "drivable" but I am a perfectionist and I want to eke as much economy as I can get.  Need to leave Cleveland by 7pm or so and it's 4:30...

Currently luxuriating at EvanB's place, after 140 miles and two tumblers of iced Absolut Peppar (in that order).

The rearend is LOUD.

My math was absolutely correct according to the GPS.  140km/h on the speedo is exactly 70mph according to the Garmin.

Most of the drivetrain hum/rumble/shudder is actually the transmission!  At 75mph actual, I can shift to 4th gear and most of it goes away.  However, this does result in a cruise RPM of 6000 or so (or 100 revs/sec if you like) which is not conducive to anything good.

Estimated MPG is about 15 at 70.  Ug.

Now watching Binky 17 on the bigscreen!  GTG

We're still mobile, March Hare.

Well, I'm mobile enough that I made it back to Evan's place.  Should be okay to drive back home.

Autopsy pictures Monday or Tuesday.

On the bright side, my 13B is apparently strong enough to busticate Ford 9" rears...

Okay so:  Home.

Audited my GoPro videos.

I did, in fact, have three runs before my fourth run.  VINDICATED.

I want to see carnage pictures...

There aren't any.

What I found was interesting.  The backlash when I set the gears up was .009".  It felt like about .002-.003" when I pulled it out.  What the hell???

This is what I drained out:

I wasn't able to get good pics of it, but there were large FLAKES of metal everywhere, including inside the diff carrier.  I'm not sure what was coming apart, but it was coming the everloving berkeley apart.

Here's the ring gear, where some of the old layout paint was still stuck.

WHAT the HELL.  I did not set it up that badly,  The mesh was nowhere NEAR that deep!!!

Anyway, I took the Locker apart.

Here's the internals.

All this looks okay...

This looks okay too...  a little wear but not "OMG stick a fork in it"

Now this is a little concerning.  I don't remember there being these deep marks in the carrier/case/housing/whatever.

The things that I did find that aren't photogenic:  Both axle bearings are "indexing" something fierce.  They were new, from wherever Dutchman got them, so who knows what quality.  The plan is to buy new Timken bearings from NAPA, unless the price makes me wince, in which case it will be whatever is cheapest from Summit.  Or maybe I will ignore for now.  Probably that one.

The carrier bearings are worn, and have very minor pitting.

Didn't disassemble the pinion yet.

Unphotographed is that the paint looked burned off of the springs in the diff.  Wondering if I wrecked the temper by overheating the fluid.

The Plan, such as it is, is to put the original springs back in the Locker. clean EVERYTHING, replace the ring and pinion, maybe with something different ratio this time, and see what happens.

Well, hell, one of the reasons I wanted a 9" was because of the world of options available as far as gearsets and diffs is concerned.

Couldn't log in to eBay at work and I kind of forgot once I got home.  Turns out the gearset I was looking at was marked down 15% this morning.  New ultralite REM polished gearset for under $150 shipped.

Nice.

Look what the FedEx man put in my hands today!

Mmm... such lightness and polished-ness

i

This will surely last more than a week behind massive rotary power.

HA!  We shall see!

This is what is left of the old pinion bearings.

I forgot to take a picture of the thrust side outer bearing race, but it is thoroughly berkeleyed.

The diff carrier bearings are wracked as well.

BUT, these are nonstandard bearings, they are 3.25" OD instead of 3.0something  and IF I have to go to a spool because the metal is coming from the diff and not the ring and pinion, and IF I start breaking 31 spline axles with a spool, THEN the large diameter carrier bearings WILL allow me to run 35 spline axles, which is absolutely ridiculous for a mid 13 second car, but then I am not drag racing I am rallycrossing, which is way more elite and hard on parts.

At least, it's hard on parts the way I do it, which is to add more throttle until you stop sucking, and if you are out of throttle, then make more power.