SpruceJedi

This is probably a can of worms here. I am also sure its been asked before.

Some people say to reduce the amount of torque you put on a bolt when anti-seize is being used due to the reduced friction on the threads. I am going through and torquing all of my suspension bolts. I have anti-seized everything. How much should I reduce? Some places say reduce by half. That being said, my leaf spring bolts say to torque to 100, that would mean I would only torque to 50. Thats not very tight for an M14 bolt. Whats the consensus?

dhoge88

i always torque the specified amount regardless of antiseize. if anything i feel it wouldn't hurt to over torque a few lbs with anitseize because it reduces the friction need to pull it back apart. but ya, if it says 100 ftlbs then do 100 ftlbs IMO

SpruceJedi

that was sort of my take. Especially on suspension stuff. I just don't feel right about reducing the torque spec on that.

JJCIII

Always torque to spec. The base of antiseize is graphite powder. When a lug, bolt etc is torqued the graphite base becomes a solid within the threads so the only way for it to break free from that state is if you were to physically break it with a breaker bar or whatever. Therefore it cannot come loose by itself

5-90

The critical value for installing a threaded fastener is "tension" or "tensile preload," which is a present stretch in the body of the fastener.

There are three ways to measure this:
- Directly through elongation. This is commonly done with high-performance connecting rod bolts, but it is only doable when you have free access to both ends of the bolt.

- Directly through compression. SPS/Unbrako makes "preload washers" - they are composite parts, and there is an "indicator" (essentially a plastic bit in the middle that "squidges out" when a certain amount of compression exists under the head. This compressive force is essentially equal to the tensile strain in the screw boty.

- Indirectly through applied torque. As tension increases in the screw, friction between the mating threads increases (read that bit one more time, you will see this material again) and the screw becomes more difficult to turn. By measuring how much force is used to turn the screw, the tensile preload may be estimated. ("tension-by-torque," even with a perfectly accurate wrench, still has a wide range of actual values. I think the variation is +/-20% of spec, assuming you follow directions exactly.)

A modification of this last is the "torque angle" method, where a very low initial torque value is used (typically 25 lb-ft or less,) so the initial error is much lower; and then the screw is turned a predetermined amount - which is given as an angle. This is usually more accurate - since the amount of turning of the screw has no dependence on the torque between mating threads. This was initially done for cylinder head screws on Diesel engines, and has been gaining ground on critical screws elsewhere (and, it should be noted, it is the de facto standard for hydraulic and pneumatic fittings in industry, with the torque spec given as "flats past finger tight" - meaning you turn the fitting in using just your fingers/hands until it stops readily - don't crank it! - then turn it in increments of 60* - one flat - until you get where you need to be.)

The advantage of torque angle is that, since it isn't dependent on the friction of mating threads, it doesn't matter what lubricant you use.

However, let's go back to the "applied torque" method -

Since "applied torque" varies heavily dependent on the friction of the mating threads, the cleanliness of the threads will matter. Defects in the thread will matter. The amount of mesh (typically 75% of thread depth) can make a difference. The materials used makes a difference (Cf is different, say, between an aluminum-iron interface than a steel-iron interface.) Essentially, anything that can change the friction from the value in a nominal "clean, dry" thread will alter applied tensile preload with the same applied torque.

Therefore, when you apply a lubricant, applying the same torque (turning force) will result in increased tensile preload! This can be bad - since you can actually end up pulling threads out of mating cast parts, overdo preload on roller/ball bearings, crush bushings & gaskets - you get the idea.

Interestingly, the variation in tensile preload is reduced as a lubricant is applied - the greater the lubricating effect, the lower the variation. I did an experiment on this for Materials & Processes a few years ago in college (second time around.)

CONDITIONS:
10L20 steel block for female threads, using a ground-thread tap. Holes were drilled & tapped for 1/4"-20UNC Class II thread using a cutting tap, then finished with a "fluteless" tap and lubrication (to debur the threads.) Female thread blocks were soaked in solvent overnight, then cleaned the following day to remove all traces of lubricant. Solvent was blown dry with shop air. Blocks were 3/4" thick, a minimum of 1/2" of material surrounded the holes in all directions. Holes were drilled & tapped through
SAE5 1/4"-20x3/4" SHCS used for male threads. Screws were cleaned in a similar manner to the blocks, they were black oxide coated, and arrived with some "storage oil" on them. All traces of this were removed.
Lubricants used were: none, engine oil, chassis grease, paraffin wax, engine assembly lubricant, honing oil, LocTite, RTV, PTFE paste, and never-seez (overkill, I'd have to find the report to get figures for them all.)
Screws were torqued to 5 pound-feet (SAE spec limit for 1/4"-20 Grade 5 is 95 lb-ins - just under 8 lb-ft - in general use.) Torque wrench used had its calibration checked before each day's use, and had not varied by more than 0.2% (the auto shop, across the way, had a measuring/calibrating device.)
Ten data points were used to test each case (not a huge statistical universe, but enough to get pointed in the right direction. I wasn't getting paid to do this...)

Results?
- The unlubricated screws showed a variation of +/-18% from calculated average value. No failures noted. The measured tensile force was used as a baseline for values in following series.
- RTV and LocTite showed a variation of +/-12% from calculated average value. No failures noted. No changes were necessary to achieve the same tensile preload.
- Engine oil & chassis grease showed a variation of +/-10% from calculated average. No failures noted. A reduction in applied torque of 20% was found to give the same tensile preload.
- Engine assembly lube showed a variation of +/-6-7% from calculated average. No failures noted. A reduction in applied torque of 25% was found to give the same tensile preload.
- PTFE paste showed a variation of +/-5% from calculated average. No failures noted. A reduction in applied torque of 10% was found to give the same tensile preload (this was rather less than expected - probably to do with the concentration of PTFE and the carrier paste used. I didn't do a chemical assay - I didn't have access to GCMS...)
- Never-seez showed a variation of +/-2% from calculated average initially, but half of the holes failed (threads pulled out) due to excessive tensile preload. A reduction in applied torque of 50% was found to give tensile preload similar to the "clean, dry" threads - at which time a variation of +/-1% was noted from calculated average tensile strain.

Draw from this what you will - but to me, this showed the validity of the thumb rules I've been using for thirty years:

- Clean, dry threads, LocTite, & RTV all get torqued to spec.
- PTFE paste torques to 90% of spec.
- Engine oil & chassis grease to 75% of spec.
- Never-seez to half of spec.

This last is also something I've used to good effect to 'extend' tools - for instance, the stub shafts on the D30 require "clean, dry" torque of 185 pound-feet. I coat the threads with never-seez and torque to 93 pound-feet, since my torque wrenches top out at 150 lb-ft. This effectively preloads the wheel bearing assemblies (they haven't failed me yet) and brings the measured applied torque within range of my tools.

The Pitman shaft nut on the steering box is treated the same way - with the addition of coating the splines with never-seez as well (makes my life easier to take it down again.)

Front end studs? Same deal - and I coat the tapers as well (so instead of needing to put a pickle fork in and bash Hell out of it with a BFH, I can brace the female end up and swat the end of the stud with a smaller hammer, and have it drop loose.)

This also explains the different value for the front left cylinder head screw - coat with PTFE paste, then torque to 100 pound-feet (instead of 110, as the rest of the screws.)

It's so nice to take conventional wisdom or thumb rules and confirm them experimentally.

ptreaster

Interesting ...

SpruceJedi

holy crap dude. I hope you had that pre-typed somewhere and just copy/pasted it. I would feel bad if you did a book report just to respond to my post. Very interesting information. I torqued to spec last night, all went well, no broken bolts and I slept well knowing my leaf springs wouldn't fly off due to not being torqued properly. Thanks for all the info!

Brown Dog

Excellent write-up 5-90. Most people have no idea what goes into determining the proper torque value for a given fastener. I'm not a smart man, so I rely on "Machinery's Handbook" for information on such matters.In the 28th Edition, hardcover, the discussion on "Torque and Tension in Fasteners" ensues on page 1428. But for some real entertainment, flip back two pages and read "Distinguishing Bolts From Screws". Just some light reading before bedtime.

Tim Shideler
Chief of Engineering and Production
Brown Dog Offroad
tim@browndogindustries.com

SpruceJedi

Good Idea Brown Dog. I didn't think of looking in there. I have that same book (26th edition I think) from my machinist days back in 2000/2001. I will check it out for sure. FYI, my motor mounts are still holding strong, those things are bada$$. Money well spent!

Brown Dog

Thank you for your kind words.

5-90

I like to teach, and I've finally learned to type - so it's easier than it used to be!

@Brown Dog - I hear you on Machinery's Handbook. I have four of them - 12th thru 26th ed - and two copies of Bosch Automotive Handbook - 6th & 8th Ed. (The latter is like MH, but targeted specifically to the automotive industry. Available through Bentley Publishers, you can sometimes find it on eBay as well. Expect to spend about $45-50, cover.)

5-90

The primary difference between a "bolt" and a "screw?" A bolt threads into a nut. A screw threads into a part (casting, bracket, whatever) - not a nut.

Which is why you see me talking about "Cylinder head screws," "manifold screws," "water pump screws," "oil sump screws," "engine mount screws" - you get the idea.

Brown Dog

The Bosch book is great.

Brown Dog

Yeah. I have looked up all of the references that are footnoted in that section. It should come as no surprise that to a minor degree, the courts were involved.

5-90

Yeah. Ridiculous. Bad enough we have accountants inserting themselves into engineering decisions (with no knowledge whatever of engineering,) but why do the damned lawyers have to get involved too?