[Stan Weiss]

I am old school and have never used a stretch gauge.

Has anyone used a torque wrench along with a stretch gauge? If so how much variance in torque did you see for the same stretch over a set of 8 rods?

Stan

[JSchmitz]

I've never used a stretch gauge either. It makes sense that it would be a more accurate way of evaluating bolt clamping force. It isn't affected by variation in thread friction. Interested to hear opinions however.

[Elarson]

Unbrako makes really high quality socket head cap screws and published some pretty definitive engineering data based on extensive testing (done way before wikipedia and keyboard crew chiefs existed). Here's their take on fastener tightening accuracy:

Operator's feel +/- 35%
Torque wrench +/- 25%
Fastener stretch +/- 5%

FWIW,
Eric

[Half-Inch Stud]

No, not used to rely upon. Came with bearing longk ago in the early 80's so I tried it to see what gives and found it to be both stoopid and confusing;

Not a uniform width.
Not easy to set in perfect alignment in the Journal.
Journal-Saddle not guaranteed to be perfectly concentric before or after torquing with Plastigauge loading 1-side.

Oil wetted vs dry ought affect plastic yield; oil has a thickness, film friction.
Oh, dont'a spin the Journal thinking it will help the plastigauge.

If we had to rely on Plastiguage: Crank and Rod example:
Set film on lower saddle dead-square and centered with saddle,
Lower crank onto Main Saddles ( no rear seal ),
place another film on Journal top ( 180* from lower film ), Tighten-TQ Caps.

Remove Caps, fetch both films, take the average reading of both films ( added/2 ).
Too much handling for me.

Clearance by using Dial-Gauges:
Mics and Caliper gauges with 0.0001" resolving are quite affordable these days.
Measure North-South and East-West with your "developed feel" and have that Skill-based Talent forever. Re-measure until you run out of thoughts, then use the plastigauge films. Room-temp is required, otherwise there is a thermal correction factor. Hand temperature can affect the gauges, so be warmed-up (cool-hands).

Edit: SPS ( Unbrako ) was the Jenkintown PA be-all end-all technology for Fasteners design-manufacture, and their Torque.
Is too bad SPS diverted away from Automotive aftermarket long ago.

All TQ-set methods will suffer from variation (error) by differing operators. I opine that bolt stretch has varaition from gauge droop from gravity, and not staying on the fastener divet center. To me, gravity affect upon the stretch gauge will affect "the feel" for getting a 0.0001" resolution.

[Elarson]

Years ago, I designed a test rig for an aircraft engine company. The hub of the rotating stuff had to take some really abusive loads, so the bolt preload was hyper-critical. The bolts were 7/8" diameter by about 10 inches long.

I specified that the bolts be coated with a special thread lubricant that the engine company loved and I specified that the bolts be tightened to a measured stretch. I also included a reference to what I thought the bolt torque might be, based on typical thread lubricant.

The test cell technician (a really sharp guy) called me up and reported that he had tightened the bolts to my stretch number but it only took about half of the predicted torque. After doing some research, I found that their gee-whiz thread lubricant had a history of being about half the friction of typical thread lube. Mystery solved.

The lesson there....if a manufacturer (such as ARP) specifies a torque with a particular thread lubricant...use EXACTLY that combo. Don't assume that all thread lubricants are the same.

Eric

[mgarblik]

In another thread I mentioned I am using Molnar rods in various builds. Tom Molnar does not spec. a torque number for tightening his rods at all. He specs. stretch and torque + angle only. Talking to him about this, he said there was entirely too much variation in torque readings, torque wrenches, and technique. He understands that stretch is cool in a lab but impractical in many engines. So I did my own experiments. I have a Snap-on electronic torque wrench. Certified +- 2%. Tom does not not like ARP lube and does not spec or use it in his products. Even though he does use them as a fastener supplier. So I used CMP #2 extreme pressure grease which is what he specs. I used the torque+ angle specs. specified for the fasteners. Set initial preload, then added the angle. Within 5 seconds of reaching the angle, you can switch modes and read the final torque at a given angle. Through the set of 16 fasteners, the torque varied by as much as 25 ft. lbs. to reach the specified angle. I was very careful and had the rods in a rod vice for these tests. I was surprised at the results. I then measured the stretch through the rod set. The stretch was all within specs. +- .001. So my results back up why Tom Molnar does not give a torque spec. for tightening his rod bolts. Torque + angle is a much more reliable way to get the recommended clamp load. The OE manufacturers and aircraft industry know this and have been using this method for decades.

Torque doesn't measure clamp load and is an indirect measure of friction and so many other factors. Wrenching technique also can change results 10-15%. Most automotive fasteners are so oversized, especially on vintage Pontiac iron , the torque method is good enough. The safety margin is so great. With the exception of connecting rods. Pontiac was concerned with this in the early 70's with the SD-455 rods. Thus the famous valve cover sticker. Torque+angle was not common practice yet.

[mzbk2l]

Originally Posted by Elarson

Years ago, I designed a test rig for an aircraft engine company. The hub of the rotating stuff had to take some really abusive loads, so the bolt preload was hyper-critical. The bolts were 7/8" diameter by about 10 inches long.
Eric

Similar to the rod bolts in the engines on some of the ships I was on. We only used stretch gauges, so I can't correlate stretch to torque.

[JSchmitz]

Originally Posted by Elarson

Years ago, I designed a test rig for an aircraft engine company. The hub of the rotating stuff had to take some really abusive loads, so the bolt preload was hyper-critical. The bolts were 7/8" diameter by about 10 inches long.

I specified that the bolts be coated with a special thread lubricant that the engine company loved and I specified that the bolts be tightened to a measured stretch. I also included a reference to what I thought the bolt torque might be, based on typical thread lubricant.

The test cell technician (a really sharp guy) called me up and reported that he had tightened the bolts to my stretch number but it only took about half of the predicted torque. After doing some research, I found that their gee-whiz thread lubricant had a history of being about half the friction of typical thread lube. Mystery solved.

The lesson there....if a manufacturer (such as ARP) specifies a torque with a particular thread lubricant...use EXACTLY that combo. Don't assume that all thread lubricants are the same.

Eric

This is exactly in line with what I believe. Critical to use specified lube on specified surfaces. Make sure to coat the washers and threads. I think a lot of people coat the threads and forget the washers. The thrust surface under the nut is a major source of friction.

[Half-Inch Stud]

@Elarson; nice. As part of an engine upgrade (more HP), I had calculated the clamping friction needed to optimize The Bolt-Circle, Fastener size, count, and TQ-Set for wooden Propeller hubs ability to not-fret-burn from the Combustion engine's instantaneous TQ-Pulse adding to the Propeller air-drag TQ. That included the differing wood used for < 37" and >\= 37" dia propellers.

Somebody on the team thought i geeked-out too far, until they smoked their wooded Prop hub after performing runway "mad-dash" up & then back. Wooden hub charred badly and they came running back to my 3-Page paper.

See, TQ-Set can be theoretically -perfect and safety wired, but if the clamping material (wood ) yields then the clamp force decreases and we all fret when something cooks.

A "siezed, rounded" Flexplate bolt is the perfect example of an unyielding Fastener system the yield with one semi-hard smack on the bolt head middle, then you can use your fingers to turn it out. Powerful example right there, to re-apply to Front Damper, Mains, Rods. Wondering if Rear Diffs Caps and Ring gear bolts have the scenario.

[JSchmitz]

I was taught in A&P school about proper torque wrench technique. Where to place your hands and where not to. Pull slow and evenly on the wrench. How to calculate for using a "crows foot".

[steve25]

If you are tightening up the rod bolts for the third time, or God forbid the 5th time you had better use the stretch method and especially if your running the rods to near there stated Hp limit previously.

[Stan Weiss]

Thanks for the replies. Good information.

Has anyone seen any changes in the relaxed bolt length after multiple uses?

Stan

[Elarson]

Originally Posted by JSchmitz

I was taught in A&P school about proper torque wrench technique. Where to place your hands and where not to. Pull slow and evenly on the wrench. How to calculate for using a "crows foot".

To add to that....I used to have a SnapOn digital torque wrench that would buzz when the torque setting was hit but would also record maximum torque. It was surprising to see how easily the final torque would overshoot if you didn't react quick enough to the buzz and release pressure. Same effect happens with a click-type wrench but there's no telltale afterwards.

Eric

[61-63]

I'm not qualified to be commenting in this thread but the thing that occurs to me reading the posts in it is that the factory didn't go to any of the lengths mentioned here. They were slamming the engines together zip pop and the rod bolts were being tightened with air wrenches set up to go to a pre set torque and then stop, which testifies to mgarblik's statement that the hardware was oversized and enough leeway was designed in that the torque variances the assembly process produced didn't matter.

[JSchmitz]

Originally Posted by Elarson

To add to that....I used to have a SnapOn digital torque wrench that would buzz when the torque setting was hit but would also record maximum torque. It was surprising to see how easily the final torque would overshoot if you didn't react quick enough to the buzz and release pressure. Same effect happens with a click-type wrench but there's no telltale afterwards.

Eric

Max torque is a cool feature! I have an old Craftsman clicker. You should also back them off to zero for storage. I see people on TV cranking bolts down like they're in a race. They're absolutely overshooting their torque target.

[JSchmitz]

Originally Posted by 61-63

I'm not qualified to be commenting in this thread but the thing that occurs to me reading the posts in it is that the factory didn't go to any of the lengths mentioned here. They were slamming the engines together zip pop and the rod bolts were being tightened with air wrenches set up to go to a pre set torque and then stop, which testifies to mgarblik's statement that the hardware was oversized and enough leeway was designed in that the torque variances the assembly process produced didn't matter.

Oversizing would cover a strength failure. But too much torque can distort parts affecting clearances and straightness, causing cracks, etc.

[Shiny]

Good info in this thread, thanks all.

Stan - Elarson or others can correct me if I'm wrong... Unless intentionally "torqued to yield", I expect stretch specs will keep the bolt stress below yield, so all stretch will be elastic and therefore fully recoverable. At least until the engine is run. Since stress is "additive", operating stress could add to that initial "stretch" (locked in stress) and cause the bolt to yield. In that case, the bolt could permanently elongate. IMO the fatigue life would be short if that happened, so I'd say if you found a rod bolt growing in length, you'd be lucky it hadn't already failed in fatigue.

And not to annoy, but I'm curious, so forgive my rambling.. what is the typical "design stress" at installation (ie without operating stress) for a rod bolt? If this belongs in a different thread, let me know.

For example, if UTS is 180ksi, are rod bolts tensioned to reach 50% of UTS or some targeted fraction of yield stress?

Similarly, what is a typical "stress adder" or max operating stress for a rod bolt? What percent of yield and/or ultimate stress is considered "safe" for a performance engine? What fatigue life is "good enough" and how is this measured in practice? Do engine designers sacrifice engines by running to failure at max rpm or is it all done on paper with calculated stress?

I'm asking because of recent discussions regarding material quality for rods and effects of same on failure risk.

General topic is of interest to me, as the best way to avoid failure and reduce sensitivity to material and manufacturing quality issues is to have design margin. For high-cycle fatigue, that usually translates into a "rule of thumb" goal of max stress <50% UTS. Then you simply don't worry about fatigue life.

In a rod bolt, adding margin probably means higher strength. It could also mean larger diameter to reduce axial stress. Does this happen?

My impression is rod bolts can get stressed much higher than their fatigue limit or this topic wouldn't exist.

Mike

[Elarson]

Great questions. But you're flirting with getting deep into bolted joint design which could easily whirl out of control. But here's the overall rules:

1) the rule of thumb for a typical fastener is to tighten it to no more than 75% of yield. Note that if you use a torque wrench that is +/- 25%, you will come close to yield (about 94% of yield) but not exceed it. For example, a Grade 8 bolt is 150,000 psi ultimate, approx. 132,000 psi yield so the target preload would be 132,000 x .75 = 99,000 psi.

2) If the bolted joint does not separate, the bolt load will not go above preload. Think about bolt stretch vs. load; and the only way to stretch a bolt farther than it's installation stress is to open up the joint. For example, if the bolt is preloaded to .005" stretch and operating load makes the joint bounce open by .001", then the bolt stretch is now .006 and the bolt load is 20% higher than the preload. If it's in something like a connecting rod that has a cyclic load, then the extra 20% stretch (load) becomes an alternating stress and that's what kills a fastener in fatigue. So a properly designed bolted joint never separates and the bolt never sees higher stretch (load) than it's installation preload. Which leads to the next rule of thumb:

3) If practical, it's always safe to design the bolt preload to be double the anticipated operating load. Then the joint never comes close to separating. So in the previous Grade 8 example, the operating load should be equivalent to no more than 99,000/2 = 49,500 psi. Aircraft engines and nitro funny car engines don't typically have that luxury and the joint/fastener design becomes more critical.

Hopes this helps without getting too deep in the weeds.

Eric

[Elarson]

Adding to the discussion of point #2....let's say the careless engine builder snugs the rod bolts down but forgets to torque them all the way. During running, the operating load exceeds the preload and the joint (rod cap) is bouncing open and closed every revolution. That makes almost all of the stress become an alternating stress instead of a steady stress and the bolt fatigue is rapid and fatal to the engine.

Eric

[1968GTO421]

My thanks to all here. Much food for thought.