Back off 1/4 or no on head retorque

[Dogma]

Some other data from the textbook (ISBN 0-07-056899-5, 'cuz I'm too lazy for another bibliographical notation): After some algebra grinding on the geometry, some representative torque factor values (K) are given for calculating the torque required (T) to induce a tension (F) on a bolt of diameter (d) according to T=KFd (independent of pitch!).

K Condition
0.30 nonplated, black finish
0.20 zinc plated
0.18 lubricated
0.16 cadmium plated
0.12 with Bowman anti-seize
0.09 with Bowma-Grip nuts (whatever those are)

This would seem to contradict the study the textbook quotes on the previous page. Presumably, the tested bolts were either zinc or cadmium plated, probably cadmium, to be consistent with the torque factor chart. But this chart is about bolts in varying conditions while the study I talked about before was focused only on the effects of lubrication.

I wish I knew what lubricant they were talking about in this chart. The textbook has a source citation, but nobody is going ot track down a copy of Fastener Facts by the Barnes Group, from 1980 to look on page 90, right? Anyhow, if I had to guess they were probably talking about the common practice of using engine oil. Anyhow, anti-seize on a fastener can yield almost twice the induced tension of a lubricated bolt at the same torque! Something to keep in mind... Check the service manual to see how they want you to prepare your fasteners.

Edit: Without tracking that book down, we can't be sure of whether those torque factors are calculated theoretical values, or values taken from actual bolt tests.

[posplayr]

I may have been ambiguous in my choice of words. I wasn't trying to make a new theoretical statement. I was only trying to report that, in tests, bolts tightened to a specific torque reading, did not show any significant change in average tension induced when lubricated. What they found was less spread in the data for tension actually induced at a given torque wrench reading.

The article quoted in my textbook is in a back issue of Machine Design that I'm having trouble finding on-line, though I thought for sure they had back issues available (J.C Blake and H. J . Kurtz "The Uncertainties of Measuring Fastener Preload," Machine Design, vol. 37 Sept. 30, 1965, pp. 128-131).

Here's the pertinent data though: 1/2-20 UNF bolts tightened to 800in*lb, dry and lubricated. Mean tension induced for dry bolts was 34.3 kN with a standard deviation of 4.91 kN. Mean tension induced for lubricated bolts was 34.18 kN with a standard deviation of 2.88 kN. The dry bolts ranged from 23.6-42.7kN, and the lubed bolts ranged from 30.3-40.5kN.

I double-checked the quoted figures, and yes, the lubed bolts induced less tension on average, but by an amount lost in the variability of the test. The principal conclusion of the study was that lubricated bolts are imperative for maximizing consistency. Even so, the standard deviation for lubed bolts is almost 10% of the mean. For really critical applications (none of those on a GS), you need a method of measuring bolt tension directly.

Well this does clarify a few things and refutes some of my arm chair theorizing. We know of course that these results are sensitive to the type of materials so steel bolts in aluminum of steel nuts on steel bolts against aluminum heads?

So at the risk of generalizing these results, at least in this case the average toque is essentially the same, whereas the variation is almost twice as much dry to lubricated. This suggests that the culomic friction is the same or very small dry and lubed. Also the wide variation suggests something pretty random. I doubt it is the variation in the torque wrench as even my cheap one is probably good to within 10% (using a weight and getting a feel for the consistency).

So where is the randomness coming from,? It must be what amounts to the stiction capture phenomena. The initial grabbing which eventually grabs and holds. So the torque wrench is in a sort of chicken and the egg proposition. If I'm in stiction, I can't measure torque. But I have to turn the wrench till it stops which is when it is in stiction and it grabs.

So the lubrication is primarily required to reduce the stiction effect I would suppose.

Bottom line:
· lube it if you got it and
· stay away from stiction, back off the bolts.
Thanks for the references. I'm sure the physical phenomena is even more complicated but these rules should still hold.

[tejasmud]

It made me curious. I assume the means of fastening the nut to the flange creates a reduction in surface area whereas when torque is applied the K is slightly less.

Grip Nut http://www.emhart.com/products/grip/assembly.asp


Interesting thread.

I torque fasteners on sealed and non-sealed joints everyday, conducted torque analysis on products, and have done numerous capability studies on assembly lines as well. I usually lubricate all threads, and apply/measure installation torque dynamically. Always loosen a fastener to re-torque. I apply these methods sub-consciously in anything I bolt together now I think about it.

From measurements over the years, breakaway torque (static torque) is always slightly higher than dynamic. Depends on the surface the fastener is turning against also. Breakaway torque is used during studies on compression set here, but two readings are taken. Initial breakaway then Actual breakaway, which initial is usually higher on some surfaces due to adhesion. Powder paint is a good example.


None of which is relevent to the thread, but are just observations in measuring torque 5 days a week for the last 13 years.

All this makes me think what is the range of tolerance is given in the service manual for torque? If big enough, probably could just torque it down to spec and go. I'm admit I am a little retent with torque.



As for leaks, surface condition can make a big difference. Even the design can be an issue. There are two spots the manual mentions to apply Suzuki Bond on the clutch cover on the 700. Clamping force between the two bolts may not be suficient, who knows, but the Engineers obviously made note. A bandaid for a design issue is my thought.

Applying more torque is usually not a good thing on gasket leaks. I would bet the leak will cease once the torque of the head fasteners is brought back to specification.

[chef1366]

Rocket science I tell ya

[Dogma]

Pos, I was wondering about the variation too. I couldn't find the article the numbers came from, but I would imagine that the testers were interested in the variations present during assembly with "good practices" so the bolts were probably new and clean. Nothing obvious like grime, corrosion, or damage in the threads. I can only think of two things to explain it. One is variation in the thread forms themselves. I've looked into the tolerances allowed by ANSI standards, and there's a fair amount of variation allowed. I would imagine JIS threads have similar tolerances (and DIN, and ISO...). The other is maybe misalignment of bolt and nut and whatever else was involved in the test assembly.

I guess one other possibility (or maybe just my first guess again) is the unpredictability of friction. We like to talk about Coulombic friction, and we used it to death in mechanics and physics classes, but it really is highly idealized. Coulombic friction is between smooth, well-matched surfaces, and threaded fasteners are anything but. Very finely ground (or maybe lapped) threads might approach Coulombic friction, but real, practical parts are uneven and tend to bind on each other, gall a bit, etc. Real friction in many cases is so insanely complex, the only way to deal with it is to test with real materials that are representative of what you're interested in, and if the tested behavior is linear enough you can gloss over the details with a measured coefficient of friction.

Well, I typed all that, and I didn't say much different than what you did about stiction. Lubrication can reduce or eliminate stiction, in Coulombic friction and in the more complicated scenarios.

[Dogma]

...
All this makes me think what is the range of tolerance is given in the service manual for torque? If big enough, probably could just torque it down to spec and go. I'm admit I am a little retent with torque.
...

I'm thinking, with as much variation as seems to be possible, I guess we just have to eliminate as much variation as we can and shoot for the middle of the specified torque range. But that can't be right. The designers should be taking the expected variation into account when the torque values are specified. I guess that's why they sometimes have to point out that lube is necessary, because the variability of the dry fastener is more than the application can tolerate. If I'm right about that, then anywhere within the specified torque range should be fine, as long as your threads are all a reasonable approximation of factory condition. That is, clean and undamaged.

But I'm speculating here. I don't know any of the guys doing this kind of engineering. The closest I ever got was a guy in charge of designing engine covers for Harley. (Those engines are made of covers, I tell ya!) there wasn't any chance of me gleaning anything from that guy though. I couldn't stand him. That, and he was about three levels of customer away.

Going completely off topic here, he did admit that Harleys had an oil leaking problem for a long time. Apparently, they actually used tapered pipe threads for their plugs! They now use SAE straight thread o-ring seals and the leaking problems are supposed to be gone. But really, thread-sealing pipe plugs? yeah, way off topic.

[Dogma]

Off topic, but since we have the engineer brain trust in residence...

I never thought about this until yesterday. I've been a subscriber to the "slappy valve is a happy valve" school. Thinking nothing about setting one at .1 rather than leave it at .04. When I initially adjusted my valves on assemble I set some WAY loose (in the .11 to .12 range) just to get it together (shim shortage). I adjusted them all back to within spec yesterday (and didn't spit a shim in the 200 miles of break in ride thank god)..

My thought (which as I always say is dangerous) is, that would change the effective lift of the cams right? a .310 lift cam on a valve that was at the loose end would be effectively lifting less that a valve at the tighter end of the range.

Am I thinking correctly?

Yeah, that makes sense to me, geometrically. Tighter clearance would be more lift. Slappy is happy, yes, just as long as there is some clearance so the valve can close completely. The lower end of the clearance would be the amount that the clearance could close up between adjustments. In other words, a valve on the low side should be fine, so long as you check it again at the specified interval. No need to go to a larger clearance. No need to have 0.04 clearance at the end of the interval.

Wasn't this thread about how to re-torque?

[tejasmud]

Lots of variation.

The torque specifications are also assuming OEM gaskets, bolts, surface finishes, and calibrated tools are being used. It goes on and on, perfect world conditions as they say.

Calculation vs. Reality are two different animals is very true.

But, I don't believe the torque specifications are so that some variation is not factored.

In opinion, most cases it is what has worked in the past should work now engineering. When it doesn't work, notes are placed in the manual to correct the problem observed during testing. At most times it is too late and too expensive to just re-tool a casting once problems arise.

Similar to why we put Suzuki Bond under the half moons on valve covers.

Yes off topic, NPT, tapered threads have their own set of problems. I deal with that daily too. Chatter, scalloped, and flat threads leak.

O-ring seals, I don't always agree with the Parker handbook or SAE, but they do work if done with the correct materials and percent of compression needed for the application.


Yes, it is "Rocket Science."

Shoot, I get paid to burn, melt, and blow stuff up.

Well, sometimes.

It's bad when you can scrap more than you monthly salary in one second.

[KEITH KRAUSE]

Once a fastener is torqued and especially if it's in service awhile you can't accurately measure the torque. The fastener must be cracked loose first, then re-torqued. The cracking method also assumes there's no significant thread corrosion or other issue that would compromise the read.
Just as an example, I've seen a fastener torqued at 20 lb/ft, back off visibly (it was marked), then the wrench set again to 20 and the wrench "click" before any movement. There's no way that fastener is torqued to 20 at the moment but by simply being in service it's "stuck". The torque required to get a fixed/torqued fastener to move again is, in most cases, greater than the initial torque.
What's best regarding head gasket torquing raises a lot of opinions. It depends also on the material/quality of the gasket and what conditions it's used in. Some say (such as an independent shop owner for Vance and Hines when they re-built my motor) to leave the gasket alone. No 100 mile check or 500 mile or whatever. He has an excellent record of doing good work. When I called the actual maker of the gasket they said V&H was mistaken and they were sounding sarcastic of the shop owners method. Nice to be in the middle of that.
I can see where loosening first could create a possible leak now/later but I think it's not generally a problem. But if the fastener isn't cracked loose a little first there's no way you get an accurate read. You're wasting your time.
Now lets discuss what type of wrench to use. For old times sake?
No. I'm kidding!!

[chef1366]

What about a good old cheapo needle type torque wrench Keith??

[49er]

Geez, talk about over complicating a relatively simple process.

Most composite type head gaskets will not fail with a variation of up to 0.003" of head surface continuity, even with a 5 ft lbs variation in torque. Now if you are using a MLS gasket, your surfaces need to be truer, but IMO you will still get away with the same variation in torque.

The keys are: lube the treads, Use a recently calibrated torque wrench, increase the torque figure gradually ( someone said Ray does 10 ft lbs increments, I do too but reduce to 5 closer to the recommended torque setting), and always re-torque by first backing off 1/4 -1/2 turn, then torque to the recommended factory setting. All this torquing/re-torquing is done diagonally from the center bolts outwards. And, the engine must be cold.
The 850 Clymer figure is 27fl lbs, I round it up to 30 ft lbs

And yes, if you are looking for outright performance, you should set your valve clearances to the minimum figures. For road use, set to the max to increase the period between clearance re-checks.

IMO, mostly common sense, not rocket science.

[Dogma]

Geez, talk about over complicating a relatively simple process.

...

IMO, mostly common sense, not rocket science.

Heh. If it was common sense, we wouldn't have more opinions on hit than we have commenters.

[Dogma]

...
Now lets discuss what type of wrench to use. For old times sake?
No. I'm kidding!!

What about a good old cheapo needle type torque wrench Keith??

People argue about this? Seriously? Well, yeah, we argue about everything else around here. We probably couldn't agree on what color black is.

[posplayr]

Heh. If it was common sense, we wouldn't have more opinions on hit than we have commenters.

Dogma,
Thanks for providing that reference. It showed an interesting and surprising result. I will have to change my simplified pictorial now.
Jim

[Agemax]

People argue about this? Seriously? Well, yeah, we argue about everything else around here. We probably couldn't agree on what color black is.

its very dark white!