If a case has 3% less internal volume then to get the same pressure with everything else the same you need to add about 3% more powder.
The problem you run into is that adding that additional powder will cause more velocity because there is a longer push on the bullet.
If it is important to keep the same velocity then use the same amount of powder. Even though the pressure is reduced the bullet will have very close to the same velocity.
For most shooters this problem is resolved by sorting cases by their volume. (after firing and uniforming them)
I don't think so. Adding
more powder to a
smaller case volume (pressure cell volume) will increase the pressure significantly, not decrease it. Keeping constant velocity/pressure with changing case volume requires charge weight to change in the same direction as case volume. Unfortunately, as I stated above, the responses are not linear. They may
seem to be linear if the change increment is very small, but as the increment gets larger, the values will move farther and farther away from a simple linear response. This is exactly why extending the freebore on a chamber, which facilitates seating a longer bullet farther out of the case so as to retain more useable case volume, allows us to achieve the same velocity at lower pressure.
Above is an example of some predictions I ran using Quickload. I started with a well-characterized 90 VLD/H4895 load I shoot in one of my .223 F-TR rifles.
The values for case volume (Lapua brass), charge weight, pressure, barrel time, and velocity are shown in the (
first) data column marked as "100% Case Vol.". These data have been calibrated so that predicted and actual velocity match exactly by adjusting the burn rate factor, Ba.
In the "97% Case Vol." (
second) data column, I reduced
both the case volume and the charge weight (
shown in red) values by 3%. You can readily see that the predicted values for pressure, barrel time, and velocity are lower than the parental values in the "100% Case Vol." column, suggesting that a 3% reduction in charge weight over-compensates for the 3% reduction in case volume.
In the (
third) column (also marked "97% Case Vol."), I kept the 3% reduced case volume, but increased charge weight until the barrel time (
23.18 gr, shown in blue) matched the value in the parental load "100% Case Vol." column. As expected with comparable barrel times, the predicted pressures in the two loads are also comparable (56,965 psi
versus 57026 psi). However, predicted velocity is almost 20 fps slower. Further, the predicted charge weight required to maintain the same barrel time with the 3% reduction in case volume was 23.38 gr, only a 2.2% reduction in charge weight.
It may seem like a difference in charge weight of 0.8% is trivial (2.2% decrease versus 3.0% decrease), but in reality it may not be trivial at all. In this specific example, a 0.8% difference amounts to only 0.2 gr. However, in a case with 2 to 3 times greater capacity than the small .223 Rem case, the actual difference in charge weight required to maintain a given pressure following a change in case volume (i.e when switching lots of brands of brass) could actually be much larger. Notice that a 3% decrease in charge weight concurrent with a 3% decrease in case volume reduced the predicted velocity by almost 40 fps, certainly more than enough in this particular load to move it significantly outside the optimal window.
The magnitude of the difference will depend a number of things, including type of powder, case volume,
case volume differential, and where the original load actually was on the pressure curve relative to MAX pressure. In the end, you're still going to have to determine empirically by actual testing where the new load will shoot. Dropping the charge weight by an equivalent percentage as the decrease in case volume is certainly a reasonable approximation, although to be absolutely sure, I might go even a bit further for safety's sake. In a small case like the .223 Rem, that extra little bit won't mean much extra effort when re-optimizing charge weight. It might be a little extra work in a much larger case, but still not a huge deal. The bottom line is that reloading parameters don't always behave in a linear fashion and a program like QuickLoad can be very useful in make prediction for non-linear functions. In cases such as the OP's where the difference in case volume is relatively small, you can probably get away with treating it as if it was linear. If the differential is larger, definitely err on the side of caution.