Crimp or Neck tension?

[Klem]

Guys,

If anyone's interested, here's an experiment attempting to shed more light on the old question of whether crimping or neck tension produces more accuracy.

A while back we had this discussion and everyone had a preference but no one seemed to have any empirical evidence supporting it.

I have done some testing over the past few months, enough to satisfy me at least.

Same load divided between two methods of seating;
1. .001" of neck tension finished by a 0.1mm taper crimp.
2. .004" of neck tension before seating, no crimp.

4rd groups at 100yards, bench rested.

The results are pretty close, however .004" neck tension wins the day. The added bonus being it saves time in reloading with one less step.



Notes:

• Load; 120NBT 28.5gns H4895, 2.25"OAL, CCI450.
• Bench rested on calm days.
• No rounds or groups discounted.
• Groups measured using On Target.
• Forster seating die, Redding taper crimp die.
• The gaps in the list indicate separate shooting sessions.

[kmon]

Thanks for running this test. I have been in the neck tension camp for years for bottle neck cartridges, of course a crimp is needed for straight walled cases. Standard Deviation numbers are telling as well, thanks for including.

[rabiddawg]

How do you achieve the desired neck tension?

Are you turning down the expander?

I crimp but would always prefer eliminating a step, saving time And get better results

Appreciate you sharing this

[Klem]

How do you achieve the desired neck tension?

Are you turning down the expander?

I crimp but would always prefer eliminating a step, saving time And get better results

Appreciate you sharing this

I use a Forster full-length sizing die without the expander ball on the stem.

With the expander ball the Forster imparts .001" of tension. Without it, .004".

Just sheer luck that the Forster turns out the preferred neck tension for a semi-auto. I tried the same with a Redding and the outcome does not change; its .001" with or without the ball.

[keystone183]

What weapon are these fired from?

[Jakal]

I chose the Redding "S" type die so I could use their neck tension bushings. They come in .001 increments to let you tailor your neck tension. I run .003 neck tension with the Grendel and no crimp at all.

[Joseph5]

I am using RCBS small base dies and I have them adjusted so that they apply a slight taper crimp when I am seating bullets. No extra steps and I couldn't tell any difference in group size either so I see it as just a little bit of insurance against bullet movement during the chambering cycle. When a good crimp did make a big difference for me was when I was shooting and loading for a 300WM. It made velocities more consistent and shrank the SD so I got tighter groups.

[Klem]

What weapon are these fired from?

The original Lilja Group Buy.

[SG4247]

Umm,

Would the numbers change if you just counted 80 rounds or 5 groups from the .004" neck tension column?

Seems like the shot count and group size count should both be equal for a fair comparision.

May not change much, just curious.

[458MinMag]

Well done Klem! Good experiment, and very good shooting! Thanks for sharing.

[Klem]

Umm,

Would the numbers change if you just counted 80 rounds or 5 groups from the .004" neck tension column?

Seems like the shot count and group size count should both be equal for a fair comparision.

May not change much, just curious.

You can compare two independent populations with different sample sizes in statistics. There are conventions which inform the process; for example are the two samples independent, are they normal ('normal' as in a Normal distribution) and are the variances significantly different (in this case they are not)? The traditional test when comparing two means is the t-test. The test makes no assumptions about the population sizes being equal so the fact that they are different is irrelevant. Plus, we don't want to discount any groups to make those two lists equal because the larger a sample is the more reliably it reflects the whole population. The whole population in this case is all the bullets that could ever be fired out of the test barrel with a .004" neck, or a crimped neck.

For a simple question like which one is better then we can simply compare the means and use the STD Deviations to inform the comparison. As for if the two sample sizes were the same (i.e. 80 shots each) would the answer be different...No. The right hand column numbers would be different but the answer would still be the same.

Feel free to manipulate the data as you see fit.

'Lies, damned lies and statistics.' (Mark Twain)

[nimitzman]

good info Klem the results are closer than i would have expected. i never crimped my 7.62 handloads when i was shooting M1A

[JASmith]

Hot Dawg! I didn't know there are folks around who can understand and correctly use statistical inference.

For me, using it is a continuing struggle with mixed results!

[SG4247]

Im not that curious, but shouldnt group size n=30 to be significant?

Thanks
SG4247

[Klem]

Im not that curious, but shouldnt group size n=30 to be significant?

Thanks
SG4247

Significance is whatever the popular consensus is in he field or discipline you are researching. Statistical Significance is usually set prior to data being collected.

The n or 'frequency'=80 rounds for Group 1, and 108 rounds for Group 2. Or you can define n to be the number of groups fired; 20 for group A, and 27 for Group B.

You can interpret the findings however you see fit. For example, you can look at the variance between means and decide that the difference is not significant enough and therefore the null hypothesis reigns (in this case the null hypothesis is that there is no difference in accuracy between crimping and neck tension). Or you can decide that the sample sizes are not a true reflection of the wider population. That you need to shoot more groups.

If you are not trying to publish the findings or convince anyone else then you get to decide the parameters.