mustanghammer

iTrader: (3)

Okay I have assembled/built two engines for my race car. First was a stock port 12A and the second was a 13B Street Port. On both engines the tension bolts became loose after the engine had been raced for several hours. The hours of use were not consecutive, this is time spread over numbers heat cycles generated in racing conditions at different racing/testing events. These engines were always shifted at red line or slightly above and did see some temps that were over optimum from time to time. They are race engines....I beat on them.

On the 12A I noticed that the bolts were loose when doing a flywheel/clutch change. I re-torqued them back to spec and then ran the engine at one more event before selling it to another racer several years later. I don't know if the bolts were loose again after the re-torque

On the 13B I noticed that the bolts were loose during a recent disassembly to start a repair/rebuild process. The bolts were snug at best and had none of the resistance that I have encountered during the disassembly to stock engines. There was no leakage around the bolt heads.

In both cases, I believe I used anti-seize on the threads and no other chemicals/sealants except for a little assy lube under the bolt heads to protect the bolt seals. The bolts were torqued in order in two sequences using the FSM specs. The bolts were clean and so where the threaded holes in the front iron.

On the 13B I had allot of issue with bad driveshaft harmonics that I eventually fixed but it took several events to get this addressed completely. So it got the crap shaken out of it. The 12A suffered through the normal 100MPH harmonic when it was raced.

So....what am I doing wrong?

Howard Coleman

iTrader: (6)

HI Scott,

your raise an excellent important issue. here's coverage of the problem from my upcoming website:

"We are now close to buttoning it up. By buttoning it up we are about to clamp 5 plates together that will make between two and three times the power of the factory motor.

Before we reach for the stock thrubolts...

Let’s review the FSM torque spec… “24-28 ft pounds.”

Many of the incoming (20 years of age) engines have been previously rebuilt. Given the overall build quality it would be a real losing bet that the 24 to 28 torque spec was respected. My bet is that all were overtorqued… 30 to 35 perhaps.

Bolts gain their clamp force by being elastic. Most are aware that the proper way to tighten a rod bolt is by measuring… stretch! Guess what happens when you over-tighten a bolt? If you exceed proof strength all of the elastic disappears and you lose clamp. Forever.

The Mazda thrubolt doomsday number (if) the bolt is lubricated is 35 ft pounds. If it is not lubricated it would be less!

There is no way to know if a thrubolt was overtorqued.

I built a motor that was, a year or so later, in for a clutch job. Someone happened to put a wrench on one of the thrubolts and found a few were on the loose side. All had been properly torqued to 28 pounds by me. I am quite certain that the thrubolts had been previously overtorqued and had lost elasticity..

Whether overtorqued or not, the OE Grade 5 bolts, are not really up to the job when power levels are increased.

I see evidence of interplate movement on almost all incoming motors. Sure it is modest but shouldn’t exist.

I recommend, but not require, additional clamp. Various vendors offer differing fixes. There are 16 thrubolts around the engine perimeter and an additional two on the feet.

The feet have nothing to do with the internal forces so can remain OE.

I have two requirements for a stud kit:

It must use 16 studs. Some only stud the bottom/hot portion of the motor. I prefer equal clamp around the entire motor.

The stud diameters must fit in the motor plates with no machining. Some kits use larger diameter studs that require that all the holes be enlarged.

Maybe if you are building a 1000 hp monstor but at 600 or less Grade 8 studs at 10 mm will increase the clamp force 45% which works for me.

I recommend either the Elite Rotary Shop or Turblown kits and use both.

My install procedure is, Hondabond, OE thrubolt washer, studkit washer, studkit nut. 40 ft pounds and over 8100 foot pounds of clamp versus 5500 or who knows if the thrubolt was previously over-torqued."

bottom line is when building a motor use either new thru bolts, or thru bolts from a motor that has not been previously rebuilt (becoming harder to find), or switch to a 16 stud kit.

to the extent you are making more power studs become a more important build item

peejay

I have had one engine "loosen up", and it felt "weird" when I was assembling it. I torque all of my engines to 23ft-lb, period. I want them to live if and when coolant temps get in the 230-240F range, because backing off in competition is for losers.

When I was assembling that engine, the bolts were springing back about 1/8 turn after the torque wrnch clicked. I thought it was odd, but didn't feel like taking the stack apart to find out why, so I let it go. A month or two later I was pulling it back apart.

Ever since then, I chase all of the tension bolt threads in the front end housing with a 10-1.0 bolt with a couple flutes ground into it. Haven't had a problem, or experienced bolt springback, since. The engine that has been in my car since 2012 was built largely with that engine's parts.

mustanghammer

iTrader: (3)

Thanks for the responses. The 12A had been apart before but I am pretty sure the 13B had never been disassembled/rebuilt in the past.

Do you guys put any kind of lube on the tension bolt threads? Is blue loc-tite a no-no?

Andrew-s

iTrader: (1)

On my race engines I went along with the FSM instructions to lightly coat both the o-ring/sealing washer and tension rod threads in oil prior to installing. It's worked well for me so far.
I know that volumes of technical study have been dedicated to the relationships between fastener torque, clamping force, and thread lubrication. But it seems Mazda is pretty clear that their torque value includes (needs?) oiled fastener threads.

Turblown

iTrader: (12)

I've had customers have this exact issue, and switch to our studs, and problem goes away...

Turblown 13B Engine Stud Kit designed for high hp applications

peejay

Threaded Fasteners Torque-to-Yield and Torque-to-Angle - Engine Builder Magazine

Turblown, are you talking about drag engines that barely get up to operating temperature before load is taken away, or circuit racing applications where the temperatures get fully saturated and normalized?

That article explains succinctly why I think studs are a Bad Idea for engines that see a lot of thermal stress. Short (sub-30 second) bursts don't really stress an engine.

peejay

The relevant part is discussing clamping load when the engine is at operating temp vs. cooled down. Small diameter bolts have the advantage of allowing the engine to grow with heat and still not yield. Although in the rotary's case, it's a known that it is not the bolts that yield, it is the rotor housings. This leads to buckling of the chrome surface near the spark plug holes at best (byebye sealing, hello beaming stresses on the apex seal) and permanently shrunk (crushed) rotor housings at worst.

This is why I torque to the minimum spec. I want the engine to be able to stretch with heat, and then I want the bolts to retain their tension after everything cools off again.

This is the difference between high physical stress, like heavily turbocharged drag engines, and high thermal stress, like circuit racing.

Howard Coleman

iTrader: (6)

thanks for the article as well as the site. it looks like a very interesting resource.

i don't dispute anything in the article as it is just engineering. i do agree w most of your comments however i do disagree w your conclusion w studs. first off, to the extent that studs have a larger cross sectional area i agree that works against them. that is why i do not like any of the significantly larger stud options.

the article presents an example of clamp loss in relation to area using 7/16 and 9/16 studs. it does show a significant difference however a 9/16 stud has 66% more area than a 7/16.

according to my notes the OE thrubolts are 9.35 mm. the kits i use are 10 mm and are a scant 14% more area.

i will happily trade a miniscule give up due to slightly more area for 43% more clamp.

i find on the more than 130 BREW motors i have built that interplate movement is the norm. i also measure housing width on every motor and have found 7 motors that were not serviceable
due to having more than .0015 crush.

conclusion: more clamp to offset more CCP is needed and housing crush is not a big problem.

i do agree that the primary culprit is due to the fact we have approx 7 inches of aluminum that we are trying to corral. aluminum is near the absolute top of the llst as to expansion. i note the article states that an aluminum cylinder head expands .005 at 250F. since the typical head at the bolt boss is around 3 inches our two aluminum rotor housings should probably expand .01.

finally, i have lots of experience w studded motors. all my motors are dual purpose daily driver single turbos. i see no interplate movement and no change in housing width.

i have the Greatest respect for mazda engineering and there is no question in my mind that the OE thrubolts are entirely appropriate for OE power levels.... but we are dealing with a motor that can outflow per displacement almost any motor on the planet and with that comes huge CCP and CCH.

add a single turbo and the equation greatly changes as does the need for more clamp:

Rear Wheel Power

2018 Corvette ZR1 1.7 hp/ cubic inch

2018 Corvette ZO6 1.47

2018 Porsche Turbo S 2.23

2018 McLaren 570S 2.06

2018 Merc Benz AMG GT C 1.91

2018 Accura NSX 2.0

Nissan GTR NISMO 2.2

Average Rear Wheel Power per Cubic Inch 1.94

Now that we have a frame of reference let’s take a look at our turbo rotary:

Rear Wheel Power

(Mazda lists displacement at 80 cubic inches. There are a couple of legitimate ways to derive displacement. In order to be conservative I am going to use 160 cubic inches so as to better compare apples to apples.)

1993 Mazda RX7 OE 217 hp 160 cubic inches 1.36

1993 Mazda RX7 350 hp 160 cubic inches 2.19

1993 Mazda RX7 400 hp 160 cubic inches 2.5

1993 Mazda RX7 450 hp 160 cubic inches 2.81

1993 Mazda RX7 500 hp 160 cubic inches 3.13

1993 Mazda RX7 550 hp 160 cubic inches 3.44

1993 Mazda RX7 600 hp 160 cubic inches 3.75



This is profound. A “lowly” 350 rwhp RX7 is within spitting distance of the highest output boosted motors offered for 2018!.......

2.19 V 2.23

While all this is very interesting, the primary takeaway of this important thread/question is that rotary engine assembly w re to thrubolts requires a respect and understanding re the elasticity of fasteners and the fact it can be lost by over-torquing.

j9fd3s

iTrader: (3)

i use the factory bolts, chase the threads and follow the factory assembly method (oil on the thread). i used to torque to the higher end of the factory spec, but i use the low end too. no problems.

i have done some rearing about fasteners, and studs, in theory, would give you a more even clamp load, which is good. other than that though, i see no real benefit. given some of the things we know about the rotary, i am inclined to think that more clamping, is just going to bend the housings. remember Mazda's solution for using the engine as a stressed member of the chassis was a thick dry sump plate using more and larger fasteners, and a top plate that bolts to all the housings.

peejay

If a 13B is 1.3 liters then an LSA is 3.1 liters.

I'd like to see a studded engine get its water temps run up to the 220-230F range and then parked at WOT on a steady state dyno for a few minutes.

I'm not denying that it maty work for street engines or strip engines.

mustanghammer

iTrader: (3)

I reached out to an engine builder that has had a fair amount of Runoffs success and described my situation. Turns out I am doing it wrong. His process is to torque OE tension bolts to 28ft lbs in 10 steps not the 2-3 steps that I have been using. This is on the kind of engine I am trying to build for the same intended purpose - street ported 13b for club racing.

Regarding studs, is anyone using them in a N/A club racing application for E Production, GTL or GT3? Like Peejay described, even in the best of circumstances, engine cooling can be an issue. Temperature spikes on the track and on the way back to my paddock spot have happened.

To control plate movement, what are everyone's opinion of the use of a one-piece dowel on the bottom of the engine?

Thanks again for all of the responses!

peejay

For an N/A application I figure, let 'em move around. Better they move around controllably than warp.

BLUE TII

iTrader: (9)

I am 99% sure this is the problem.

I just clean the tension bolt threads and wipe the tension bolts with very very light oil (WD40) and make sure the tension bolt holes do not have coolant, oil or anything else in them. I have never had tension bolts loosen, but they do take a bit of force to "break-away" when loosening them for a rebuild.

Kenku

I might have to look at stuff like this since it's not like there's good bolt solutions for the short crank 3 rotor. I'm sure as hell trying not to see temperature excursions like that though. Seems like rallycross screws with things in different ways than most anything else.

neit_jnf

iTrader: (17)

what's your take on studs that come out the front plate and have sealing washers and nuts at both ends?

Turblown

iTrader: (12)

If we didn't use E85 on everything we build I could do that for you, we just never see those water temps, ever.

The car I spoke above is a drag car.

I have had the OEM bolts measured precisely( new, used, S4 and S6 ), and have all the data. This was taken into consideration when making our studs, and the torque specs etc.

I'm not going to share it as, as we have a lot of people copying us. I will see what I can do about replicating the above tests on a non alcohol based road course car.

BLUE TII

iTrader: (9)

Like Peejay says, Mazda tried close fit studs early in its rotary engine program and found they it causes cracking.
Pg 272 of "The Wankel Engine"
"A further cause of the cracking trouble was lack of flexibility in the trochoidal wall and conduction of heat from it to the cold end housings through the bolts holding them together. Improved flexibility and reduced heat transfer were simultaneously achieved by separating the side-bolt bosses from the trochoidal wall, and housings made in this way proved completely resistant to thermal fatigue."

All the studding and doweling, etc to make the engines "stronger" or "stiffer" might work to get it down the 1/4 mile in one piece, but it weakens the engine against other forces acting on it such as thermal and vibrational cycling that will destroy the engine in regular use or endurance racing.

TonyD89

iTrader: (10)

Hex head metric bolts and socket head cap screws in grade 10.9 can be had in almost any length. They will be custom, have to be ordered from Holokrome, Brighten Best, or another manufacturer. Possibly even through a supplier like Fastenal, (you wont find them on McMaster-Carr, though they may be able to order customs) but, they can be had. They might be expensive and will have a thicker unthreaded part than stock tension bolts.

fidelity101

iTrader: (11)

My voodoo magic method has always been chase all the threads with a tap/die for the iron/stud and use a light bit of WD40. I have often wondered if I used blue loctite what would happen but I have yet to test this theory since my engine builds have not suffered any tension bolt looseness that I am aware of.

JZG

On the 13B I had allot of issue with bad driveshaft harmonics that I eventually fixed but it took several events to get this addressed completely. So it got the crap shaken out of it. The 12A suffered through the normal 100MPH harmonic when it was raced.

You answered your own question. It's a balance issue, be it driveline or engine balance. I've seen this.

mustanghammer

iTrader: (3)

I would agree on the 13B as the driveline issues were really bad - way out of the ordinary. Still think my assembly process needs to change.

However, FBs have a harmonic vibration that occurs at around 98-100mph and goes away by 103-105mph. Its not real bad if pinion angle isn't weird and the drive shaft being used is in good shape. But it is always there and I have yet to talk to an FB racer that hasn't reported this issue.

There are ways to make this better with a multi-piece driveshaft and by moving the engine over about an inch towards the right side of the engine bay. But even the racers made these modifications still report a harmonic that comes and goes as the car passes through about 100mph. Also not every class allows these kinds of modifications.

peejay

I've noted that vibration at 80-85mph. And it goes right into normal highway speeds when you yank out the 3.91s and install 4.78s...

j9fd3s

iTrader: (3)

the miatae are infamous for doing this at 65mph.

peejay

I thought the Miata vibration was due to the factory tires? I spent a lot of time in a Miata a couple years ago and never noticed anything untoward.