Reloading: Two-Step Sizing and Concentricity

Two-Step Sizing and Case Neck Concentricity
by: Germán A. Salazar

Those of us who handload rifle ammunition with accuracy in mind are well aware of the importance of neck to body concentricity in our cartridge case.  While there may be some dispute over the degree of importance of obtaining perfect concentricity and its effect on accuracy, we can all agree that a more concentric case is better than a less concentric case.  The two principal factors affecting neck to body concentricity are: the quality of the brass itself (i.e. neck thickness variance, case wall thickness variance) and the quality of the reloading dies used.  It is mostly the sizing die which affects concentricity, while the seating die can have a small effect (see September 2009 article on that topic).  It is ultimately the sizing die that either maximizes concentricity within the limitations presented by brass quality, or fails to do so.

This article's title refers to two-step sizing, a term that may not be familiar to all, so an explanation is in order.  When sizing with bushing dies, such as the Redding Competition and Type S dies, it is well known that a neck diameter reduction of more than 0.005" should be accomplished in two steps, with each step sizing the neck down no more than 0.005" and preferably less (see October 2009 article about neck tension).  This two-step procedure not only minimizes sizing die induced concentricity deterioration, but it also yields a finished neck diameter equivalent to the bushing's nominal size.  It is often found that when using a single bushing to size more than 0.005", the finished neck diameter is smaller than the bushing's nominal size.  Those are the basic reasons for two-step sizing.

Two-step sizing can be accomplished by using a full-length sizing die twice or by using a full-length die and a neck die in sequence.  I prefer the full-length and neck die combination because of the potential to have varying headspace on cases run through a full-length die twice.  I haven't tested that theory, but it is a concern for me so I use two dies.  The very narrow question which we seek to explore in this article is: Which die should be used first in a two-step sizing process in order to minimize sizing die induced runout?

At first blush one might suppose that the order in which the dies are used is of no consequence.  The Redding competition neck die is a chamber type die which should support the case well during sizing and the Type S full length die is a high quality die that one would not expect to induce much runout.  Nonetheless, as you will see below, there can be a difference in the concentricity depending on the order in which the dies are used.

The brass used in the initial test was Norma .30-06 with unturned case necks.  These vary a bit more in thickness than Winchester or Lapua, with 0.002" to 0.003" neck thickness variance not being uncommon.  However, this is older brass from the 1970's; current Norma brass checks out better.  The brass was checked on the Audette tool ( see the July 2009 article) for case thickness variance and index marked on the thin side; all cases were under 0.003" in case wall thickness variance which is an acceptable level.  The cases were all sized with the index mark in the same location relative to the sizing die on both the full length die and the neck die.  Because one die might tend to increase or decrease the concentricity variance created by the other to a greater or lesser extent based on the case's orientation relative to each die, this indexing process ensured that we were working each case in exactly the same manner.  Anyone duplicating this project without access to an Audette tool can simply make an index mark anywhere on the case, the point is simply to keep the case-to-die orientation consistent throughout the process.

The cases were fired in a .30-06 chamber with a 0.340" neck diameter with loads that I've pressure tested and which I know generate approximately 58,000 psi which is just a bit under the SAAMI maximum average pressure of 60,000 psi for the .30-06.  As-fired, the case necks measured 0.3395" and the object was to size them to 0.331" by using the two dies, first with a 0.334" bushing, then with the 0.331" bushing.  The full length die was set to push the shoulder back 0.001" (see the August 2009 article) and I checked brass before and after the neck sizing operation to make sure that the neck die was not affecting headspace.  The dies were both Redding, a Type S full length die and a Competition Series neck die.  The bushings used were a Wilson 0.334" and a carbide 0.331" formerly made by Butch Lambert.  Neither die was used with an expander ball as these can have a detrimental effect on concentricity.

Since this article is at its core about measurements, a word about that process is in order.  All neck diameters were measured with a Mitutoyo electronic caliper and are reported here to the nearest half a thousandth (0.0005") and even that level of precision is a bit suspect for calipers on a relatively soft object, but diameters aren't our primary concern.  Concentricity was measured with a Sinclair concentricity tool using an MSC dial indicator and I read it to half a thousandth, rounding up or down as needed.  That is all the precision we can expect from this type of tool and the results should be read accordingly.  Readings were taken 0.150" from the case mouth in order to avoid the slight irregularities which can occur at the mouth.

The brass was divided into two sets of 20 pieces each.  Set 1 would be sized first with the full length die using the 0.334" bushing, then with the neck die using the 0.331" bushing.  Set 2 would be sized first with the neck die and 0.334" bushing, then with the full length die and the 0.331" bushing.  The charts show the measurements taken on each case: first the as-fired concentricity, then after the first sizing operation and again after the second sizing operation.  The last column shows the total change induced by the two sizing operations from the as-fired state.  I also averaged the measurements at each stage and these are really the interesting results.

The results were very surprising; I was expecting to see no real difference between the two methods, but, in fact, there was a substantial difference that is too large and too consistent to be a random occurence.  Sizing with the neck die first followed by the full length die produced cases with significantly better neck-to-body concentricity that using the full length die first and the neck die second.  As a purely numercial matter, Set 2 which was full length sized last, increased runout about half a thousandth on average whereas Set 1 which was neck sized last increased about two thousandths on average - essentially a 400% difference in sizing die induced rounout attributable to the order in which the dies were used!  I certainly consider that a significant difference and one worth noting when reloading.

Look at the average Total Change figure, that represents the difference between the as-fired average runout and the final runout after going through both dies.  Set 1 shows 2.1 thousandths which we will round to two thousandths due to the imprecision of dial indicator reading, and Set 2 shows 0.6 thousandths which we likewise round to half a thousandth.  If we look one column to the left, the actual runout after both sizing operations, we can see that Set 1 has about twice the avearge runout of Set 2.  Regardless of how we measure the effect, it is fairly clear that on these cases, neck sizing first, followed by full length sizing produced more concentric cases.  This was not at all what I expected to find and caused me to take on some additional work.

In the past, I have typically full length sized first, which the test of Sets 1 and 2 indicates is the inferior method.  However, my match brass which is normally neck-turned, doesn't have as much runout as that in Set 1 which was done in my usual sequence.  Therefore, while I hadn't initially planned on it, this became a four part test with Sets 3 and 4 done in the same manner as above but with neck turned brass (0.0125" thick necks) (see December 2009 article on that topic).  The cases for sets 3 and 4 were Lapua .30-06 which has been fired four times with loads similar to those used in the Norma brass. 

As with Sets 1 and 2, all cases in Set 3 and 4 were checked on the Audette tool, indexed and inserted into the dies in the same orientation each time.  All lube was wiped from the cases prior to each concentricity measurement.  I did everything possible to ensure that the results were not subject to outside influences or other variables.  The difference in results between Sets 3 and 4 versus Sets 1 and 2 shows the difference between sizing methods with neck-turned or unturned brass.  The fact that the set pairs were of different brands is not particularly significant as this test is about neck to body concentricity and the neck is the key element in the brass.

The data from Sets 3 and 4 does not indicate a significant difference between the concentricity of the two sets.  While Set 3 has a nominally lower average runout (0.8 vs. 1.1), the amount of the difference is within the range of variability and reading error for this type of instrumentation and the number of cases checked.  I consider both Sets 3 and 4 to have essentially one thousandth of added runout (over as-fired) from the sizing process and both to have a total runout of one and a half thousandths.  For all practical purposes, they are the same.

We come to a tentative conclusion then, that for unturned brass it is better to neck size first, then full length size when using a two-step sizing method; whereas for neck turned brass, there is no particularly significant advantage to one method over the other.  Of course, this is simply my sample of 80 pieces of brass and two dies, hardly a completely scientific test, but about as good as most of us reloaders do with our hobby time.  The real interesting questions are:  Why are the results so different?  What makes the unturned brass so sensitive to the order in which the dies are used?

The answer to that question can only be found in the neck thickness variance of the unturned cases.  I have no hard data to back up my thoughts on this, in other words, this is pure speculation.  I suspect that the thicker side of the case neck reacts to the reforming process in the die somewhat differently than the thin side.  The neck sizing die theoretically supports the case with its sliding sleeve, but necesarilly has some clearance around the case body.  This allows the case body centerline to shift  from the neck centerline as the neck is sized with the thick side of the neck "resisting" the sizing process more than the other.  Now we have an off-center neck.  Looking at the data from Set 2, we can see that the neck sizing operation (which was first) induced the largest concentricity loss of any of the sizing operations on any set of cases.  The full length sizing which followed actually tended to make the cases a tiny bit straighter, or at least not worse.  Obviously, there is no clearance inside the full length die and the body and neck alignment is subject only to the quality of the die (including how well centered the bushing is in the die).  The body can't "push away" from the centerline of the neck in the full length die and, in fact, may tend to straighten a slightly crooked case.

This all makes perfect sense, the only fly in the ointment is that the next-worst first resizing operation was in Set 3 which had neck turned brass and was full length sized first.  That should have been an almost perfect arrangement, yet the numbers aren't too good.  I can't explain that and I'll continue to look at the data and run more tests.  One of my plans is to use the full length die for both operations on some brass and see how it reacts and whether brass life is shortened; this will necessarily be a long-term project.

If you are one of those people who skips to the end of an article in order to get the result without understanding the process, neck size first on unturned brass and it doesn't matter on turned brass.  You can stop reading now.  If you read the article, and understood the process, and have an interest in learning more, I think it should be very simple for anyone to duplicate this experiment.  Like any project of this type, our data set is small and results can only be considered an indication of tendencies rather than an absolute rule. Nonetheless, I believe this has been a worthwhile and illuminating exercise, I hope you think so too.

All contents Copyright 2012 The Rifleman's Journal