I suggest you pressure your cooling system to 30 psi to check for any leaks around the studs.
ARP states prior to Tensile Strength Overload the "bolt will neck down prior to rupture. That makes sense as if you stretch a rod in length it will have to acquire a thinner diameter/cross section as the material to lengthen will have to come from somewhere...
i measured 5 tension bolts in two places to the hundred thousandth of an inch as to diameter. the average was .36890 or 9.37 MM. I don't have a new tension bolt to measure but thought i had heard they are 9.5 mm. If that is the case we do have a bit of stretch going on. i will research this further just for my own curiosity...
One final material consideration as to 10 mm aftermarket studs is to the extent they offer a larger area they increase actual clamp. If we use my 9.37 mm diameter and compare to a 10 mm stud the additional clamp is 13.8%. Significant!
CHIPS MOTORSPORTS 10 MM 4340 STUD KIT: ANALYSIS AND INSTALLATION
Prior to getting started just a bit of background. i recommend and, based on testing, would run without qualification only two stud kits. Turblown and Chips Motorsports. I worked a bit with Chip on the design of his studs but do not sell them and did pay retail for the set i installed in my motor 2 days ago. If you want to buy a set contact Chip, not me. The reason for this addendum to my Stud Section is i was impressed with some of the details of the studs and wanted to share them with the community.
Just to review, I recommend ONLY stud kits that contain 16 studs. While the rotary has a power section (Southern half of the block) and a cooler induction section (Northern half) I want equal clamp around the entire perimeter.
And just exactly what is preventing this potential engine rebuild? 16 tension bolts that are 26 years of age. If the motor has been previously opened it is a good bet that the 16 tension bolts that were specified by Mazda (what do they know) to be torqued to between 24 and 28 foot pounds were overtorqued to the point where they could have gone into "YIELD" and have lost much of their clamp. (see beginning of section for more on yield).
I have no proof of my conclusion but it is that you are less likely to have a coolant seal failure if you have more clamp. A very good reason to include a proper stud kit on your rebuild.
Another reason to consider more clamp is the dreaded front iron crack (FIC).
As you see, the Mazda bolts are around 150,000 psi which places them at Metric 10.9/Grade 8. First class. Why then the 24-28 torque spec? Proper torque on a 10.9 oiled is 40! I had a suspicion but contacted my top metal guy and confirmed it. Our 10+ inch tension bolts clamp 6.3 inches of Aluminum together along with the cast iron side irons. Aluminum is second only to Plutonium as to thermal expansion.
Translation: it expands lots when it gets hot.
If we were to torque the tension bolts to the normal rated spec, around 40, when the motor warmed and the aluminum expanded the bolt would stretch into the Yield or perhaps Tensile area.
The first spec on my 6 page CPR Spec Sheet is rotor housing parallel. If the housing is off parallel by .0024 Mazda says no go. You typically develop a combustion seal leak to the coolant... the motor pushes coolant into the overflow tank until it overheats. Rebuild time. If any of the 16 tension bolts has been stretched to Yield it will only be a matter of time until you have a coolant problem.
Bottom line: if you torque the tension bolts beyond the Mazda spec... 28... you will be risking bolt stretch and clamp loss resulting in coolant leakage, especially if the motor runs hot at some point.
As mentioned, i tested 3 aftermarket 10 mm 16 stud aftermarket offerings. Two exceeded the Mazda tension bolts and posted very nice numbers. The other one was significantly less than OE.
Chipsmotorsports and Turblown are your only aftermarket stud options.
"It was always a PITA to install a stud kit because you had to thoroughly clean the rear iron mating surface, then attempt to fill the 16 wells with Hondabond... generally a mess but no leaks. Oh well. Note how the ChipsMotorsports stud has no threads at the mating point. No threads, no silicone, no mess. No Leak. "
"Going to have to edit this as i recently found a slight leak on my motor on startup. Of course there is no such thing as a "slight" coolant leak so the motor had to come apart. I am back to Hondabond around the stud prior to adding the O ring.
Here's the better looking back end of my motor. Thanks Chip for a really nice new product for our 26 year old cars. I also recommend Chip's oilpan brace (picture in the build section) and rate it best in class. As to studs, I recommend both Turblown and Chips.
TENSION BOLT TECH
You are looking at three springs.
Prior to a bit of stretching the bottom two were the same length. As you know, a spring's function is to provide tension. The upper spring is in tension as all the coils are tight against each other. The bottom spring has no tension as it was stretched to the point of "Yield."
Yield is the technical term for the point where the material under stretch is no longer elastic, that is, it no longer exerts a clamping force..
All bolts contain elasticity. It is this elasticity that holds the bolted parts together. As you apply stretch to the bolt by tightening the nut the clamp force rises. If you over-tighten the bolt you can reach a point where the material gives up it's elasticity and clamp force disappears... just like the spring pictured above.
It is this elasticity dynamic that drives all the torque charts. For instance Mazda directs us to torque the stationary gear bolts 12-16 foot pounds. I replace these important bolts with 10.9 Grade and have to torque them to 25 foot pounds to input the proper tension. This increases clamp force 43%.
Likewise Mazda stipulates the "tension bolts", after coating the threads with motor oil, to be torqued to between 24 and 28 foot pounds. Here we get back into a central question... given we are going to be making a multiple of the original horsepower... do we need to upgrade?
After receiving in excess of 130 13B-REWs for rebuild to higher power specs and carefully examining all of the incoming motors i say YES. I see galled flange surfaces on the aluminum rotor housings from interplate movement. Given we are often doubling the output and we have five plates being held together with 16 bolts some additional help is in order.
There is another very important reason to consider tension bolt replacement... it relates to the opening picture. Do your tension bolts resemble the first or second spring? There is no way to know.
If your motor has been previously rebuilt and the builder decided that since the power was going up he ought to add a bit more torque to the tension bolts, say 30 to 35, you might have inelastic tension bolts.
i experienced such a situation on one of my builds. The motor was in for a clutch replacement and the shop happened to put a wrench on one of the tension bolts and found it was loose. I do 8 cycles around the tension bolts and all are on spec. No doubt what had happened was the bolt had been previously stretched and had lost elasticity. You may have similar bolts holding your motor together. If your motor has been previously rebuilt you are speculating on whether the builder didn't overly crank on the bolts unless you replace them.
i have always been intrigued that Mazda calls for 24-28 foot pounds on a 9.5 mm bolt. A look at the torque charts shows the bolt quality at that torque to be Metric 8.8/Grade 5. Pretty close to hardware store quality. Why such a low Grade for such an important function?
Finally my curiosity led me to get some answers. I decided to send a couple of the OE tension bolts to a Certified testing lab. Since the aftermarket offers various stud kits as replacement i decided to check them as well. The results proved worth the investment and answered my questions.
It turns out the factory tension bolts were of much higher quality than indicated by the low torque spec. I had all the items tested for Yield, the point where any further stretch results in loss of elasticity/clamp and Tensile which is actual breakage. Since i haven't seen a broken tension bolt I think Yield is the key metric. And the Mazda bolts did better than I expected... Here's the Certified results:
It appears that a weak point in our engine's architecture is on knock or pre-ignition pressure pushes the rotor housing outward and with it the tool steel dowel pin that anchors only about a half inch into the cast iron front iron. The dowel pin is moved outward and cracks the iron. Result, a rebuild and shortage of BREW used (uncracked) front irons. Again, i am speculating here but if more clamp existed between the plates MAYBE this might be avoided.
We have coolant seal failures and broken front irons and our defense is a 26 year old 9.5 mm bolt that may have been compromised by some knucklehead torquing them to 35 pounds to offset a higher output configuration.
I switched to studs a number of years ago on the suggestion of Jose Le Duc. He suggested the CX Racing titanium studs which i ran for a number of years. They were amazingly light being titanium but eventually I learned they were made in China and i was concerned about their metallurgical pedigree as well as the nuts seemed to be not too happy on the threads. I swapped to another brand, ran them for a couple of years. Meanwhile I became more and more convinced that studs were so important that i decided to test them. Sort of like oil, we know it is important but just how to make the right decision? Studs are easy... test them for Tensile strength (breakage point) and more importantly Yield... where they are stretched to the point they lose clamp strength.
It is a shame to cover up my new studs but if it has to be done let's do it in style... my Mcleod double disc light as a feather take up that holds 800 ponies. Just had them take 6 pounds out of the flywheel.
As many know, in a really serious connection situation such as the bolts holding the lower cap on the connecting rod the amount of bolt stretch matters. Too much torque, too much stretch results in a broken motor. Bolts are torqued to actual stretch as measured by a rod bolt stretch gauge.