I've not found any improvement by sorting cases by weight.  But I am seeing significant reductions in ES and SDs by doing the above.  I had a 10 shot group this weekend with an ES of just 18 fps and SD of 6 fps.  That's the lowest I've ever recorded on my .303 handloads (for 10 shots).  Gone is the occasional very large ES values of 100 fps.  50 fps is now the highest I recorded since culling the brass. I've now set aside about 15% of the batch of 200 cases I was working with. 

Has anyone else tried this approach, and has it worked for you?

 

Maybe check volume by water?

I'm guessing that since they all have the same external dimensions (neck sized only for last 7 firings) and all very close in weight, may not be much difference in internal volume. 

I did this a while back when someone was going on about HXP I think being very variable. I weighed a bunch then wet tested for volume with tap water & a drop of dish detergent per quart to minimize the meniscus.

Let me see if I still have the records.

Yep!

Rating .303 British cases by manufacturer for critical specs. These are the average from 5 sample cases.

The tests produced a bit of interesting myth busting again (“Thinner than normal” American made cases & thinner rimmed cases from Rem. & Win.), & a couple of inconsistencies. I expected case weights to be the exact inverse of case volumes, but it wasn’t so. I expected so I wouldn’t be as concerned with the volume results as I would with the weight, rim thickness & variations in those 3 areas.

Case weight.

Remington lightest @ (161.04 GR).

Win 2^Nd. @ (168.0).

S&B 3^Rd. @ (180.62).

PPU 4^th. @ (182.38).

HXP Heaviest. @ (184.62).

Case volume of water.

Win smallest @ 53.28 Gr.

HXP 2^Nd. @ 53.78 Gr.

Rem 3^Rd. @ 54.6 Gr.

PPU 4^th. @ 55.54 Gr.

S&B largest @ 55.95 Gr.

Rim thickness.

S&B Thinnest with 0.0555”.

Rem 2^Nd. With 0.0586”.

HXP 3^Rd. with 0.0600”.

Win 4^Th. With 0.0607”.

PPU Thickest With 0.0608”.

Case to case variations.

Weight:

Win lowest @ 0.4 GR.

Rem next @ 1.5 Gr.

PPU 3^Rd with 4.1 GR.

HXP 4 Th. With 11.3 Gr.

S&B last with a whopping 18.3 Gr.

Volume:

HXP lowest with 1.2 Gr.

Rem 2^Nd at 1.8 Gr.

Win 3^Rd. at 2.7 Gr.

PPU 4^Th. At 3.1 Gr.

S&B highest @ 3.9 Gr.

Rim thickness variation.

PPU, Win & Rem in a dead heat for least variation @ 0.0030” between.

HXP 2^Nd at 0.0045”

S&B 3^Rd. with 0.0050”.

For grins, I'll check the case weight and volume on the 20 or so PPU cases I put aside (those that gave high and low bullet velocity).  They have been neck sized 7 times now (since the last P/L resize) so external dimensions should be very close to being the same.  I'll need to first trim them to the same length in order to measure volume. 

I will also reload the cases I put aside again to confirm that it is the cartridge case that caused the large bullet velocity spreads.  I can try and predict which cases will give low velocities and which cases will give high velocities based on weight and/or volume measurements.  If I can, then it would be possible to screen these cases out by taking these measurements. 

  

Yes, I agree its a small sample. But it was based on what "spare brass" from each brand I had available so I standardized it at 5 just to be consistent.

The big surprise was the volumes not matching the weight differences. Thicker case less volume thinner case more.

Thinking about how cases are made I wonder where the variations come from.

They all start out as a brass disk, then they're progressively drawn out, trimmed back ,drawn out again, primer hole punched & primer pocket swaged, necked down then trimmed to Max OAL.

I couldn't decide if progressive drawing die wear, or initial plug thickness/volume determined the variations.

The empty case weight spread (max - min) was 4.9 grains.  That’s a fair amount. 

I then weighed the same cases filled with water and subtracted the empty weight.  Max spread was just 0.8 grains of water. That seems quite good and considerably less than the spread of empty case weight. 

I don’t think we can tell anything by weighing cases. I plan to let the chrono and target decide which cases cause fliers. 

After resizing the brass, do you clean it to remove the lubricant?

I was thinking that a difference in residual lubricant in the neck, could affect the launch of the projectile, as would a change in neck tension. 

Also if the thickness of the brass around the neck is different, it will alter the tension. 

I think you could easily go nuts chasing every variable, but it's very interesting to see your experiments and results. 

Cases that get F/L or P/L resized are tumbled to remove the lube.  I've been using a wet tumbler for the last few years.  I neck size the .303 cases with a Lee Collet die, so no lube required.  After 8 to 10 neck sizing reloads, I then do a P/L resize to bump the shoulder back a few thou.  

My plan is to reload these 30 cases I've set aside and see what ES and SD they produce.  If I'm right, those 30 cases will have an ES of around 100 fps and confirm my suspicions.

Case weight empty            7.9 gr variation

Volume (weight of water)   1.3 gr variation

I’d guess that the variation in height of the meniscus might account for 0.3 gr, so really the volume variation was probably just 1 grain.  That is surprisingly very tight considering nearly a 8 grain variation in case weight, which must primarily come from rim thickness variations (hard to measure).  All cases had depth uniformed cut primer pockets and flash holes reamed and de-burred. 

So, it’s back to plan A, there is no practical way of determining which cases will cause large muzzle velocity variations except by loading them, firing and recording velocity.