Action Screw Torque Accuracy Test.

[diderr]

So I finally broke down and bought a Wheeler Engineering FAT torque wrench. I decided to do some testing to find the ideal inch pounds for my action screws. I went from 45" lbs. to 65" lbs. in 5 lb increments. I shot a three shot group at a distance of 100 yards with each setting here are my results.
Rifle:
FN Patrol action
McM A5 Bedded in Devcon w/ pillars installed
Bartlein 1-10 5r @ 23" chambered in 308 W.
Premier Heritage 3-15

Handload:
190 SMK
40.5 IMR4064
Lapua case
BR2 primer
2.8355" COAL

Group Sizes:
45" lbs.: 0.460
50" lbs.: 0.374
55" lbs.: 0.289
60" lbs.: 0.573
65" lbs.: 0.874

 

[Elite_KG]

It's almost like you knew 55 would be the best....

 

[Bear24]

That is very interesting, that is quite a difference. Maybe I need to try a little lighter than 65 with my pillar bedded FN SPR

 

[Blowby]

Good info! I went from 40 on Saturday to 45 on Sunday with a noticeable improvement. Torqued the action bolts to 50 last night so I'll see if my groups improve like yours did. HS, pillars, bedded.

 

[Archer762]

There must be other factors at play. Did you go front to back, or back to front each time? How much time did you leave between groups? A barrel heating up will open up the groups. I can't believe a 10lb difference in 2 screws makes the difference between a .3 moa rifle and a .9 moa rifle, unless the high torque values were actually causing some kind of twisting of the action.

 

[Elite_KG]

I've read in several places that the higher torque values indeed distort the action. I've seen a video test that corroborates this by measuring the action flex as torque is increased. Some quick searches will likely reveal a great deal of info on the matter.

 

[diderr]

There must be other factors at play. Did you go front to back, or back to front each time? How much time did you leave between groups? A barrel heating up will open up the groups. I can't believe a 10lb difference in 2 screws makes the difference between a .3 moa rifle and a .9 moa rifle, unless the high torque values were actually causing some kind of twisting of the action.

I went front to back each time, I let the barrel cool after each string. 65 was rushed a little and didn't get as long to cool as a storm was moving in which probably explains why it jumped from .573 to .874. Still; I'm leaving it at 55.

 

[paul2atlo8]

Nice info, will try the same torque sequence.

 

[billyburl2]

Now how are you going to check the accuracy of the torque wrench?

 

[turbo54]

Interesting, but in all objectivity, you simply did not do enough shooting to prove anything

Not to poo poo your test...

Perhaps the results would be repeatable, but with only a few shots at each torque setting, the results could be skewed by a whole host of reasons.

If you've ammo to burn, repeat the test, with 5 round groups, at least 2 groups per setting, in random order, loosening between each setting...

...and have someone else do the torqueing, so the shooter is blind to what they're testing.

Even better if performed so the shooter can't see their results from one shot to the next.

 

[3wide]

Maybe try more groups at each torque setting, to confirm.

One three shot group is a pretty small sample size.

 

[Wildwillalaska]

We need some industrious type to verify the calibration of each torque setting, the do a ladder or OCW tests of sorts, say 12-120 groups at each torque setting from 38-72 inch pounds in 2lb increments. Should also likely test long vs short action for any distinction as well as different test btw Rem 700, FN SBR, Defiance, Surgeon, etc., to see what distinctions exist between manufacturers.

I set this out in jest, but admit real interest in any results. Found the original post overly interesting.

 

[Temp9]

Interesting, but in all objectivity, you simply did not do enough shooting to prove anything

Not to poo poo your test...

Perhaps the results would be repeatable, but with only a few shots at each torque setting, the results could be skewed by a whole host of reasons.

If you've ammo to burn, repeat the test, with 5 round groups, at least 2 groups per setting, in random order, loosening between each setting...

...and have someone else do the torqueing, so the shooter is blind to what they're testing.

Even better if performed so the shooter can't see their results from one shot to the next.

+1. This has been discussed endlessly (and controversially) on these forums. Single 3 shot groups don't prove anything. Good on you for getting out and testing it in the first place, though. Now try again with more ammo.

 

[Sniper Uncle]

I agree on the more testing needed. But, I think you can short circuit the testing some by doing say three-five 5-shot groups at each of these three torque values: 45"#, 55"#, and 65"#. If the groups remain consistent with the original test, then I'd say you are definitely on to something. Then you could refine the test from there. I also agree with the poster who says you should loosen the screws between tests also. That way you are being consistent in more areas of the test.

 

[scudzuki]

Agree with Turbo, you need far more data points to reach a conclusion, at least 5 groups at each torque value.
The data you have collected seems to support the hypothesis that that action screw torque affects accuracy... but I'd be surprised to find it would have that much of an effect given that your action's been bedded.

Joe

 

[Lowlight]

Nice start, it is a beginning and gives people something to consider.

Especially when they get odd groups or erratic fliers they cannot account for.

 

[Pusher591]

55 is enough for you and your rifle and it obviously shoots. Who cares if it's not enough data, at least in your mind you'll feel more at ease if it's torqued to 55 because you've seen what it capable of at that setting. If you don't or can't replicate it then you know for the most part it's shooter error. Very interesting though.

 

[turbo54]

55 is enough for you and your rifle and it obviously shoots. Who cares if it's not enough data, at least in your mind you'll feel more at ease if it's torqued to 55 because you've seen what it capable of at that setting. If you don't or can't replicate it then you know for the most part it's shooter error. Very interesting though.

Note bold text.

Very true!

 

[bjay]

I think its worth the try.. ill try this

 

[SporterII]

Sometimes you got to take your findings and just apply them .
The torq wrench does not need calibration. It will likely repeat. Thats all that matters.
Front to back, me too.

 

[Twisted]

Note bold text.

Very true!

Couldn't agree with this more.

Sent from my Galaxy S3

 

[tomcat mv]

Has anyone tried this with an XLR chassis set up? I just recently got mine and torqued it to the recommended 65 in. lbs. and haven't shot it yet.

 

[haroldwilson]

Worked on my 3 screw Savage LRP.

Edit: I meant my F/TR. I haven't done this to my LRP. It shoots so good I am afraid to screw with it.

I only did the F/TR because I had to take the action out to adjust the trigger pull. I started at 10 in lbs and worked up to 35 in lbs in 5 in lb increments. It started opening back up at 30 and got worse at 35. I then loosened the rear screw back to 10 and repeated the test. I got the same results.

Savage Action Screw Torque Tuning within AccurateShooter.com

 

[jcann]

Just thinking out loud here,

I know 45 to 65 in/lb isn't a lot when compared to ft/lb but would you also need to possibly consider bolt fatigue, thread stretching and lubricant to bolt for proper torque values?

 

[Drifter™]

The torq wrench does not need calibration. It will likely repeat. Thats all that matters.

+1. I agree.

 

[swamp2]

Interesting, but in all objectivity, you simply did not do enough shooting to prove anything

Not to poo poo your test...

Perhaps the results would be repeatable, but with only a few shots at each torque setting, the results could be skewed by a whole host of reasons.

If you've ammo to burn, repeat the test, with 5 round groups, at least 2 groups per setting, in random order, loosening between each setting...

...and have someone else do the torqueing, so the shooter is blind to what they're testing.

Even better if performed so the shooter can't see their results from one shot to the next.

A big +1, this is right on the money. +1 to the other skeptics as well. Heck there is not even a control established here by shooting whatever strings/groups with no changes to torque. Once that is done some basic statistics on comparison of means is required. The proper procedure is called ANOVA (analysis of variance). An additional point is: if by any chance torques in these relatively small ranges do have some statistical effect on accuracy, there is absolutely no guarantee that the best setting will be the best for any other rifle.

Another suggestion would be to use some type of rigid shooting vise (not just a rest) to entirely eliminate the human factor. Not sure you can even buy one of these though... It is what Harold Vaughn did in his excellent book "Rifle Accuracy Facts". He obviously realized that the shooter had to be eliminated when talking about a scientific approach to what effects matter with the rifle/ammo/scope etc. and how much they matter.