Why might my rifle have lost zero

[Formidilosus]

Are you putting the paint/polish on the treads? I’m only asking because I’ve seen pictures of yours with the paint on top of the screw head and surrounding it

Yes. Used just like blue loctite.

 

[stan_wa]

Gun disassembled all fasteners and scope tube degreased with break clean

Action screws installed with 55 in-lb
Ring to base 55 in lbs
Scope ring clamp 20 in- lbs.

Will report back once I get a few range trips in

 

[stan_wa]

Gun disassembled all fasteners and scope tube degreased with break clean

Action screws installed with 55 in-lb
Ring to base 55 in lbs
Scope ring clamp 20 in- lbs.

Will report back once I get a few range trips in

range update I went to the range with a fresh install as mentioned in my previous post. Here's a quick update on how it went:

• Baseline Shooting- group 1: I started with a 5-shot baseline and registered a Mean Radius (MR) of .43.
• Drop Tests- group2 : I did some drop tests, as much as I could tolerate, dropping the rifle 12-18 inches onto a foam mat over concrete. shot a group. shots 6,7 touching shots 8,9 touching but sepreated form 6,7 by 1.25". MR of .61 (yikes)
• Group 3: shot 3 more MR of . 38

Throughout this, I noticed the group centers were moving. For simplicity, I calculated a 16-shot overlayed group, which had an MR of .58 with a Standard Deviation (SD) of .29.

Before realizing the issue, I had 23 shots of combined data with an MR of .25 and an SD of .13.

I wasn't able to "clearly" excite zero by any specific action (e.g., running dials, drops, or magnification changes). The biggest jump I saw was between two shots where nothing changed: shots 6 and 7 were touching, and shots 7 and 8 were touching but 1.25" apart.

It's clear to me that something is off as accuracy data is by far the worse i have seen over 580 shots on this rifle and 3 scopes. My guess is the scope, so next time I head to the range, I'll bring a backup scope that I know is good, tested by @solarshooter

Group	# of shots	zero shift (in)	MR		collective DATA
1	5.00	N/a	0.43		MR	0.58​
2	8.00	0.40	0.61		sd	0.29​
3	3.00	0.84	0.38

 

[ResearchinStuff]

Small sample size noise man.

 

[Reburn]

Small sample size noise man.

I have to agree here. I would start with 20 shots. Then drop it. If a 0.1 mil or less then rock that B****

 

[stan_wa]

The combined group samples size is statistical significant and the data that matters. Its best to Think of all the groups as one which is why I was able to confidently conclude this scope is “grouping “ poorly

 

[Macintosh]

For simplicity, I calculated a 16-shot overlayed group, which had an MR of .58 with a Standard Deviation (SD) of .29.

Before realizing the issue, I had 23 shots of combined data with an MR of .25 and an SD of .13.

@stan_wa maybe I missed it—apologies if so.

What was your original baseline group size, and what size group was it measured from? Is that the MR of .25 and SD of .13? Was this measured from one 23-round group or 2 smaller groups, or many groups? “Combined data” makes it sound like multiple smaller groups.

The most recent groups, its not clear if the “aggregate” larger group included breaking and building a new position for each of the 3 smaller groups. If so that is noise that will cause your groups to open up for the vast majority of shooters—that translates to possible inconsistency with your baseline that would potentially reduce the ability to draw any conclusions from a comparison with your baseline.

Did your extreme spread (of dispersion) change from before to now?

In the pics of groups posted I saw several groups with different round counts, so i couldnt see for certain what your previous baseline was and how that was arrived at and what consistency you had in your periodic checks.
It seems like the problem was originally a change in zero, and now its a change in precision. Is it only one or both?

Initially I would have guessed the problem was a loose fastener or scope slipped in the rings from insufficient torque, action slipped in stock, etc. After re-assembly it seems you are talking about a different issue (group size, not change in zero) so Im trying to figure out if you are seeing a real change in precision versus other variables (position changes, changes in group sizes, etc) just creating noise.

Im wondering all this because my impression is that you are using slightly different, or even very different, methodology (# shots, split into different groups, possible position breaks, etc) for each of your larger “groups”, which could be an apples : oranges comparison that will certainly gum-up your data. You have to use precisely the same methodology for your baseline and your comparison so as much as possible the gun itself is the only relevant variable. Unless you are extremely careful controlling those variables such as position, you just cant make the conclusions you need in this case by comparing groups that use significantly different group sizes, arrived at in different ways (ie position breaks, one is one group the other 3 aggregated groups, etc). If I understand correctly you’re talking about a +\- 1/4” change in mean radius (.25 to .58)—less than 1 scope click—and a change in position or methodology could easily create a much larger difference than that.

Personally, what I would do is zero, then shoot one 20-round or 30-round group from one position without breaking position, and record mr, sd and extreme spread of dispersion. Id take a pic for reference. That’ll be your baseline. The group will be bigger than your smaller round count groups. As far as zero shift, if your future small-round count groups (3,5,whatever) fall inside the extreme spread-sized circle of your baseline group, your zero didnt shift, even if the individual 3 or 5 round groups centers move a bit. Even for MR, I think you really need to compare similar round count groups using same methodology—there is going to be more variation in mr in small round count groups when one or two rounds from the edge of the Normal distribution winds up in that group, so you have to account for that and expect that one will show up once in a while.

Apologies if my impression is wrong, I didnt see some of this specified in the various posts, but I may have just missed it.

 

[stan_wa]

@stan_wa maybe I missed it—apologies if so.

What was your original baseline group size, and what size group was it measured from? Is that the MR of .25 and SD of .13? Was this measured from one 23-round group or 2 smaller groups, or many groups? “Combined data” makes it sound like multiple smaller groups.

The most recent groups, its not clear if the “aggregate” larger group included breaking and building a new position for each of the 3 smaller groups. If so that is noise that will cause your groups to open up for the vast majority of shooters—that translates to possible inconsistency with your baseline that would potentially reduce the ability to draw any conclusions from a comparison with your baseline.

Did your extreme spread (of dispersion) change from before to now?

In the pics of groups posted I saw several groups with different round counts, so i couldnt see for certain what your previous baseline was and how that was arrived at and what consistency you had in your periodic checks.
It seems like the problem was originally a change in zero, and now its a change in precision. Is it only one or both?

Initially I would have guessed the problem was a loose fastener or scope slipped in the rings from insufficient torque, action slipped in stock, etc. After re-assembly it seems you are talking about a different issue (group size, not change in zero) so Im trying to figure out if you are seeing a real change in precision versus other variables (position changes, changes in group sizes, etc) just creating noise.

Im wondering all this because my impression is that you are using slightly different, or even very different, methodology (# shots, split into different groups, possible position breaks, etc) for each of your larger “groups”, which could be an apples : oranges comparison that will certainly gum-up your data. You have to use precisely the same methodology for your baseline and your comparison so as much as possible the gun itself is the only relevant variable. Unless you are extremely careful controlling those variables such as position, you just cant make the conclusions you need in this case by comparing groups that use significantly different group sizes, arrived at in different ways (ie position breaks, one is one group the other 3 aggregated groups, etc). If I understand correctly you’re talking about a +\- 1/4” change in mean radius (.25 to .58)—less than 1 scope click—and a change in position or methodology could easily create a much larger difference than that.

Personally, what I would do is zero, then shoot one 20-round or 30-round group from one position without breaking position, and record mr, sd and extreme spread of dispersion. Id take a pic for reference. That’ll be your baseline. The group will be bigger than your smaller round count groups. As far as zero shift, if your future small-round count groups (3,5,whatever) fall inside the extreme spread-sized circle of your baseline group, your zero didnt shift, even if the individual 3 or 5 round groups centers move a bit. Even for MR, I think you really need to compare similar round count groups using same methodology—there is going to be more variation in mr in small round count groups when one or two rounds from the edge of the Normal distribution winds up in that group, so you have to account for that and expect that one will show up once in a while.

Apologies if my impression is wrong, I didnt see some of this specified in the various posts, but I may have just missed it.

Thanks for giving me such a detailed reply. It’s very kind of you to spend that kind of time energy to help out a stranger and it’s a really appreciated. I understand what you’re saying with consistency, sorry if my presentation was confusing but I have about 200 rounds of logged shot data with a collective Mr of .32 and that was with changing load variables. My control “ baseline was 23 shots at Mr of .25
The most recent testing was 16 shot Mr of .58.

I agree, shooting always 10 round groups, for example, would make things easier to compare, which is why I calculate things as combined groups with MR so if I shoot 100 single shot groups or one 100 shot group the data would report the same way in terms of radius because I calculate the collective center of all the shots taken between Any groups which there should be no changes

In terms of extreme spread that’s a 1” group for 23 shots vs a 2.5” group for 16 shots

to illustrate baseline pic

Vs recent range trip



In the life of the rifle I have never seen anything close to this. I think it’s important to point out this small changes in Mr are much more meaningful than changes in group size
. Ie a quarter inch in MR is a huge change.

 

[Macintosh]

Are the new groups all stringing vertically like that? Seems relevant. Anyway, seems your next step of switching scopes to see if the problem isnt present with a different scope is the right one.

 

[stan_wa]

In all of the groups, the vertical is larger than the horizontal by about 2X

 

[solarshooter]

It seems like the problem was originally a change in zero, and now its a change in precision. Is it only one or both?

Yes, the issue discovered at long range was a shift in group center, but the group itself was still good.

After retorquing everything, it now seems to be an issue with group size. Zero hold issue could still be there, but testing has not been repeated. Or, the large group could be explained by a scope not holding zero.

It could still be a scope issue or a rifle issue. Testing with a different scope will hopefully answer this. For instance, if groups are shot with a different known good scope and they return to typical performance for this gun, then it is likely the scope, and the scope should go back to Trijicon under warranty. If groups do not return to typical, then it is likely the rifle, and the rifle should go back to Seekins under warranty.

 

[Macintosh]

In all of the groups, the vertical is larger than the horizontal by about 2X

I have seen several issues like this where it was a loose fastener, but in my limited experience that was more random, it didnt exhibit as linear stringing. Has anyone seen a loose action screw or scope attachment screw (rail or rings) exhibit pronounced stringing?

If that's consistent then something changed (or was exacerbated) when you re-torqued and is now moving significantly more in one direction than the other. Assuming the barrel is still torqued onto the action and there is zero contact with stock, and all screws are still torqued, its unlikely the action and barrel itself changed in a way that would overnight cause consistent stringing like that. My money says you'll see a difference with a different scope.

 

[solarshooter]

After listening to the latest @Formidilosus ep, which I think was spawned from this thread, I'm also a little skeptical of barrel torque. But I would wait to dive into that until the rifle vs scope portion of the fault tree has been isolated.

 

[wind gypsy]

After listening to the latest @Formidilosus ep, which I think was spawned from this thread, I'm also a little skeptical of barrel torque. But I would wait to dive into that until the rifle vs scope portion of the fault tree has been isolated.

Good thought. Per the other thread by @Duh (I think it was him?) about his seekins barrel coming off with the can, it seems seekins has more than one instance of not torquing the barrel very tight.

There’s also getting to be a fair number of rounds on that tube for a 28 nos.

 

[Duh]

Good thought. Per the other thread by @Duh (I think it was him?) about his seekins barrel coming off with the can, it seems seekins has more than one instance of not torquing the barrel very tight.

There’s also getting to be a fair number of rounds on that tube for a 28 nos.

Yep that’s me. Seekins said it’s not common and recommended 65 ft lbs to torque it back on.

 

[stan_wa]

d 65 inch lbs to torq

You mean ft lbs right

 

[Duh]

You mean ft lbs right

Oh yeah, ft lbs. good catch

 

[ResearchinStuff]

That's still not very much torque on something as big as barrel tenon threads