What is the best way to find the optimum load for a given rifle?

  • 973 Views
  • Last Post 29 May 2022
John Alexander posted this 08 May 2022

A while back we had a fairly long thread going on Tony Boyer's (and a large share of the JB benchresters) method for finding the optimum load for benchrest shooting (tuning). Although the discussion centered on whether the Boyer method was a valid approach, little was said about other procedures for finding the optimum load for a rifle. We constantly hear about "working up a load" What does that mean to you? I think it would be interesting to see how we each go about this chore.

Anybody interested in cast bullet precision in rifles including match shooters and others afflicted with the compulsion to be able to shoot the smallest average groups must try to solve this problem.  How do you go about finding this load ? How many groups for each comparison? How many shots per group?  Once you have what seems to produce the smallest groups, when should the load be tested later to confirm that it is till the best?

Let's limit the discussion to only after a bullet and powder have been selected to avoid getting into the pros and cons of bullet design and the trials and tribulations of selecting a powder.

How do you do it?

John

 

Attached Files

Order By: Standard | Newest | Votes
Lee Guthrie posted this 10 May 2022

Being normally VERY short on time (no, I am NOT lazy cool), in attempting to wring maximum accuracy from a rifle I look at match results from the FS: shooters using my equipment that did well are copied, bullet choice and alloy, powder choice and amount, etc.  I have a few times made a cerrosafe casting of the chamber and throat, but due to past clumsiness the bullet diameter is frequently just enlarged until it will not chamber properly or accuracy drops drastically.  From their data I will bracket loads looking for a drastic increase/decrease in accuracy.  I seat the bullet to engage the rifling: sometimes just a kiss, and others a jam fit.

For hunting applications it frequently becomes a matter of "such and such" mould/grain weight/design has traditionally worked well in that model rifle in "X" chambering.  From that point a manual suggested load is tried, and if it seems both fast enough and reasonably accurate then I bracket the loading looking for a significant change - good or bad.  Shooting for "minute of deer" is a lot different than match shooting as extreme accuracy may have to be sacrificed for impact velocity, and a good bullet design and alloy for match may be completely wrong for hunting.  (This might be a good place to advise against letting "friends" try their loads over your chronograph:  scoped rifles tend to shoot a lot lower at 10-15 feet from the muzzle.)  A hunting rifle and cartridge that want a hard bullet at the velocity needed may have to be experimented with using cast soft nose bullets, which will somewhat degrade accuracy just by that method of casting.  Bullets have to cycle easily even if accuracy suffers, so COL may not be the best for max accuracy.

My method is pretty much lacking in the precise finesse of others.  A lot of trial and error.   I trust my magic wand to make up the difference.  Sometimes nothing works.  (Case in point:  Winchester '94 in 38-55.  NOTHING would work, until I learned of Starline brass made just for this rifle.)

Attached Files

joeb33050 posted this 13 May 2022

Attached Files

RicinYakima posted this 09 May 2022

First, I research Mattern, Whelen, Sharp and Harrison. Then start looking into reloading manuals from 1945 through 1980. Looking for precision, rather than velocity, little research in the gun rags. 

Second, I find out the rifling "form": size, throat and do a chamber cast.  

(Having researched for the 1903 Springfield rifles for over 30 years, I don't do five shot groups, as they only predict the next five shot group.)

I measure a fired case neck with a plug gauge, and never size bullets smaller that the fired case neck. I brush case neck every reload and reject any rounds that have a different seating pressure.  

Pick a load out of the research and load 25 to shoot one target, no called shots and no excuses. Load 25 at 0.3 grains above and 25 at 0.3 grains below. Shoot each of  them for one 25 shot target. I'm looking for a pressure level that works with my alloy and is OK with my barrel vibrations. 

 

Picking the closest 20 shots, I square the target to find the real center of group and measure the group size. With the 22 Hornet, 250/3000 and 7.62X39 I have found my best load within 10% of the starting load. 

My best loads have always shot best with the front band tapered, or lightly pushed, into the throat, FWIW. I no longer worry about seating depth with lubed cast bullets. Powder coated is an other animal. 

Lastly, I am very picky about the dies and press I use. If the dies and press will not load cartridges within 0.003" run out from the web of the case to the driving band, they go away. Nothing loads straighter cartridges than LE Wilson hand dies. 

Attached Files

  • Liked by
  • Bud Hyett
Bud Hyett posted this 13 May 2022

The first step is research; checking the Fouling Shot equipment listing for rifle and caliber as a reference. The second step is looking at alloy, the sizing on the body and nose as shot. The third step is counting how many shooters are using which loads and their results. 

If possible, I cast the chamber and leade to see what these dimensions are and to check the progression of wear on the leade. 

I have prejudices;

  • Reloder #7 powder.
  • VV N-335 powder.
  • Lapua brass where available.
  • Federal or W-W Large Rifle primers in .30-'06 length cases.
  • R-P 7 1/2 primers in .222 Remington head size cases.
  • AA 4100, #9, or Ramshot Enforcer powder for .32-20 or .25-20 loads in Plain Base. 
  • Annealing three hundred cases per rifle each Spring, then outside neck-turning to .015 wall thickness.

 Seating depth is shooting trial loads. I start with a witness mark of .050, then back off to zero witness mark before getting aggressive.

I have my own testing approach that takes a full day at the range to confirm two prospective loads.

Farm boy from Illinois, living in the magical Pacific Northwest

Attached Files

  • Liked by
  • RicinYakima
RicinYakima posted this 13 May 2022

Attached is a 25 shot group. It shows that the sight is 1/8th low and 1/4" left. That is not observable with 5 shot groups. 

Attached Files

  • Liked by
  • Bud Hyett
joeb33050 posted this 13 May 2022

Joe sez: "Nope. Zero the rifle 1.234" low, 2.345" right., or anywhere, the stdev of measures is independent of POI/POA. 

==============

Joe, now my BS detector has rung so hard that the bell is broken and fallen on the floor.  Maybe I should think about it some more before answering.

 

How about this. Set sights to hit anywhere. Fire 25 ex shots at 25 bullseyes. Measure x and y of each shot.  This is the data, all there is.

There are 12 2-shot groups, shots 1-2, 3-4, 5-6, 7-8...etc.

Put the x,y coordinates of shots 1 and 2 on a piece of graph paper and measure 2-shot group size. 

there are 5 5-shot groups. put the x,y coordinates of shots 1-5 on a piece of graph paper and measure 5-shot group size.

Or, calc stdev of x and y.

Calculate whatever you wish, the 25 x,y pairs are all the info there is.

The precision of each shot is in the location of the bullet hole, ALL the information in a shot is in the xy coordinates of the hole. There is no other information. We can calc 2-shot or 5-shot or 7- shot group sizes, or whatever, we can argue  about mean radius, but we're ALWAYS dealing with bullet hole location.

I spent half an hour and made an EXCEL worksheet that calculates 2, 3, 4, and 5 shot group size from xy coordinates.The computer calculates it all, makes calcs easy.

The x,y coordinates of the bullet hole are ALL the info in the shot; and 1 shot per bullseye is the "best" way to get and see that info.

joe b.    

 

Attached Files

  • Liked by
  • John Alexander
Ken Campbell Iowa posted this 17 May 2022

i have been waiting for someone to mention ::

 

first setting the goal.

**************

as a plinker, my goal is pretty blah ...  bean can groups at 30 yards are acceptable ... bean can groups at 100 yards is worth entering in my log book while downtown in big smiles city. ...

******

but even lowly standards are not always easy to meet ...  often just looking in the lyman book and then loading a 1700 fps amount of fast to medium powder gets me a 4 inch group at 50 yards ...  hmmm ... up and down on the powder or different powder doesn't help .. i don't even go through that exercise anymore, as long as the bullet holes are round.

it is always the lonely ( there aren't any other rules ) Rule One :: ... it is always bullet fit.

in the old days i would make a squish die and get an occasional 1 moa screamer group from a good rifle ... now i just order another Lee $26 mold ..  ... anybody else have molds for rifles long traded off and for rifles not traded for yet ?? ...

*****

and more on topic ... when serious accuracy testing, i do aim for about 1700 fps .. try 3 different charges of same powder and 2 or 3 powders ... one 5 shot group of each  ... ( hold on guys ) ...  

pick out a few good lookers and keep loading these forever and eventually eliminate the loads that let me down .  i will never find the very best FOR SURE load in my lifetime ... i do find one that meets my goals ... or i put the rifle back in the rack, it ain't the load.

*******

i suppose you could say that one good way to develop good loads is to get yourself a good rifle.

ken, guilty but smiling plinker .

 

 

 

 

Attached Files

  • Liked by
  • John Alexander
joeb33050 posted this 17 May 2022

Joe,

Your table above showing that the estimated mean radius estimated by applying a constant to the 10-shot group size produced a very good estimated value of the mean radius actually calculated by measuring the locations of the ten holes.

Mean radius and group size are measures of the same thing, the ratios are the ratios. It is like measuring the distance from Chicago to Denver in miles  and kilometers. Some advocate for kilometers, others staunchly defend miles; ad hominem attacks ensue, blows are exchanged, shots are fired. Nonsense reigns.  

The same is true of ES and STANDARD DEVIATION-measures of the same thing.  

Please stop me right here if I misunderstand the above.

So if I understand, using the information from two holes (extreme spread) in most cases provides, for practical purposes, almost as good a measure as diddling around with the data from all ten holes.

Five 2-shot groups is a better estimate of precision than two 5-shot or one 10-shot groups. BUT, only if statistical calculations follow.. 

Not a very good argument for going to more elaborate mathematical means, than simple group size, to estimate precision. The idea of using the information from each individual shot is intellectually pleasing and should be a tiny bit better but doesn't seem worth the trouble when the expected difference in two adjacent groups is much greater than the error in estimating shown in your chart. In other words it makes more sense to shoot two groups rather than computing the mean radius of one.

It is clear to me that shooters ain't statisticians, that any stats rules-of-thumb are ignored, and that life goes on. Accurate/precise guns and loads are developed, the hard way, with no well-defined plan in mind. This is a hobby, supposed to be fun. I have fun shooting, and doing statistics; ignore the stats, cast, shoot, have fun.

 

John 

Attached Files

  • Liked by
  • John Alexander
Paul Pollard posted this 19 May 2022

It is clear to me that shooters ain't statisticians, that any stats rules-of-thumb are ignored, and that life goes on. Accurate/precise guns and loads are developed, the hard way, with no well-defined plan in mind. This is a hobby, supposed to be fun. I have fun shooting, and doing statistics; ignore the stats, cast, shoot, have fun.

Statistics would be useful if we could follow a plan to use them. Whenever a discussion of statistics and how to use them comes up, it usually results in finger pointing, "my way is more precise than you way", etc. As a  person with little useful knowledge of statistics, my eyes glaze over and I dream of shooting thousands of rounds, collecting very little data of  value. 

At my work, we used a statistics program called MiniTab. It was an expensive program, but even a dummy like me would look lie a brilliant statistician. 

My approach to finding accuracy is to use process control. I preheat my mold on a hotplate and check the mold temperature before starting to cast, eventually establishing what temperature to set the hotplate at. I mix a 100 lb batch of alloy at a time and cast into ingots. The casting is done in a 20 lb Lee pot with PID control. In this way my expected weights are the same throughout. I find the temperature where the alloy becomes liquid and set my pot temperature for 100 to 125 degrees above that for casting. This allows for faster casting, usually 5 casts per minute with a single cavity mold.

This gives me a uniform bullet, but I still don't know what it takes for accuracy.

Attached Files

  • Liked by
  • JeffinNZ
John Carlson posted this 09 May 2022

The first thing I do is study.  I look at match results using equipment and components as close to mine as I can find.  I also look at reloading manuals but generally find the information there rather limited for cast.  With this information I will establish the minimum and maximum powder charges that I will test.  I then load 5 rounds each of 6 incremental charges.  All my load development is fired over a chronograph.  Analyzing the results of this session will tell me whether the bullet/powder combination is worth further testing.  Here's a recent example.

Cartridge:  7.63x39

Bullet:  Lee 312155

Powder:  H4198

 Charge            Avg Vel         Vel Spread      Std Dev      Group

13.0                  1392               203                  88            2.277

14.0                  1464               26                     10            2.713

15.0                  1509               166                  55              2.598   

16.0                  1658               61                     25            .988

17.0                  1712               34                     13              1.284

18.0                  1785               38                     16            1.074

Given these results my next effort will be as follows:

1 group at 14.0 gn to determine if the exceptional chrono data is an anomaly and also to see if the group is twice the size of the 16-18gn groups

5 groups at 16, 16.5, 17, 17.5, and 18 gn

I will shoot these groups in reverse order, heaviest to lightest, to hopefully eliminate the possibility that the improved performance was the result of bore condition.

All sessions start with a clean bore and 5 fouling shots, same bullet and powder.

John Carlson. CBA Director of Military Competition.

Attached Files

John Alexander posted this 10 May 2022

Thanks Ric and John. Two different and interesting approaches, both by military rifle shooters. Thanks for taking the time to describe the details.  Hope others will confess how they approach this important aspect of CB shooting. How about some non-military shooters?

John

Attached Files

joeb33050 posted this 11 May 2022

There is currently a 2-shot-group thread on Rimfire Central.

Brent Danielson advocated the 2-shot group; I agree with the statistics,but objected because of POI change as POA changed.

I learned how to avoid that POA/POI change after a zillion 22RF shots.

If the shooter can conquer the POA/POI change, then 2 is the best shots/group number.

(If the POA/POI change is unconquered; then test results are questionable.)

Of course, the 1-shot-per-bull-22RF-score-target method, averaging bull-center-to-bullet-hole- center, is best of all.

Unfortunately, conventional statistical Z or t testing for means is wildly inaccurate, so there is no help in that direction.

Attached Files

John Alexander posted this 11 May 2022

It hAS always seemed to me that shooting 2-shot groups should be better that shooting 3-shot groups which are better -------------. And the mean radius method better yet, for efficiency in estimating precision with a given number of shots.

Mean radius method should be better that mean distance from aim point (22RF-score target method) for estimating precision of rifle but mean radius not as good for finding estimate of a true zero.

But how do you square my simpleminded thinking that 2 are obviously better that 3 and  4 better than ---, because you are measuring more shots, with the dope I dug up for the article in TFS #248 that indicated it was a settled question the most efficient number of shots per group was 7.  SEVEN??? 

John

Attached Files

Wm Cook posted this 12 May 2022

John, I’m assuming that the objective is to reach absolute match accuracy with one rifle, one bullet and one powder. That predetermines the requirement for match grade casting, reloading, bench equipment and practices.

If you only have one mold and one powder for an over the counter newly purchased production bolt gun it seems to me that the only variables remaining are alloy, lube, what size lube/sizer die, the size of the neck bushing you’ll use, the bullet seating depth, and finally the powder charge. I don’t think primers are critical so I’ll avoid that

As you know I’m new to cast accuracy but I’ll give this a shot (pun intended).

I’d first pound cast to get the chamber, free-bore, leade measurements. Then I’d use a small bore gage to get the bore diameter (I am incapable of measuring land tip to tip). I would verify the pound cast by tapping one of the cast bullets through the chamber and into the free-bore and into the lands (tapered design) or into the bore (bore rider design) removing the bullet by tapping from the muzzle. Obviously look for land marks on the bore-rider but I would focus on the free-bore length and diameter.

Either with a bore-rider, tapered or spitzer design the size of the die for the lube/sizer would be determined from the above work. At this point in my understanding cast accuracy I would size the bullets to a scrape fit into the free-bore for the reason that I want the gas check of the bullet to ride in the case neck of the brass so after chambering, the front driving band will be in both the case and in the free-bore so that the gas check is held by the case mouth.

Example: .308 Winchester has a case neck of .303. I would like the gas check to seat no further than .303 - .310 below the case mouth. My gas checks are about .015 in height.

With a Spitzer or tapered design you either live with what length you have available in the free-bore, leade and let the gas check protrude below the neck or you push out the lands. In the perfect world I would want the lands far enough out that the gas check is as I described and the bullet nose was just sticking in the lands.

I would like enough contact between the bullet and the lands that it sticks, but not enough contact that the round cannot be extracted without pulling the bullet. This “touchy feeling” thing determined the case neck bushing. Not a death grip on the bullet, just a modest grip and if the lands are broomed off a light or very light grip.

Linotype for the alloy and the Carnauba Blue for the lube in my sizer should work with the pedestrian velocities I’m shooting.

With the restriction of having only one choice for powder I’d have three powder charges worked up that would give me about 1700, 1740 and 1780 fps. I’d shoot one 10 to 20 shot group at each velocity and my guess I’d have it close to being all it can be. There may be some jinking up and down with powder if what I started with looks ugly but with one powder you’re at a disadvantage.

I would greatly appreciate yours and others critical thoughts on what I sketched out. Thanks, Bill Cook.

A “Measured Response” is as effective as tongue lashing a stuck door.

Attached Files

RicinYakima posted this 12 May 2022

In a one in ten twist barrel, you would be better off looking at less the 1600 f/s. No magic, just history. 

Attached Files

joeb33050 posted this 12 May 2022

It hAS always seemed to me that shooting 2-shot groups should be better that shooting 3-shot groups which are better -------------. And the mean radius method better yet, for efficiency in estimating precision with a given number of shots.

Mean radius method should be better that mean distance from aim point (22RF-score target method) for estimating precision of rifle but mean radius not as good for finding estimate of a true zero.

 

A shot, a bullet hole in the target, contains a certain amount of information. A set of shots then contains a glob of information. If we put a piece of paper behind the target and change it every 2 shots, we get a set of 2-shot groups. Another paper changed every 3 shots, 3-shot groups. And so on. Measure group size or mean radius or group width or., and the bullet holes don't care-the holes are where they are.

If we shoot 25 shots, each at a bull, 1 shot per bull, and measure each bull center to bullet hole center; then the standard deviation of those measures is an estimate of the long-term variation in bullet location. From that standard deviation, one can estimate group size for any number of shots per group. Accuracy of the whole megillah varies as total shots, 25 in the above. Easy to measure each c-c distance. Easy to calc standard deviation with computer. Fairly easy to get any measure, such as 4-shot-group-size.

     But how do you square my simpleminded thinking that 2 are obviously better that 3 and  4 better than ---, because you are measuring more shots, with the dope I dug up for the article in TFS #248 that indicated it was a settled question the most efficient number of shots per group was 7.  SEVEN???  7? My BS detector is ringing. 

John

Attached Files

TRKakaCatWhisperer posted this 12 May 2022

Getting back to the original question of the "BEST" way... find the "OPTIMAL" load.

There are two words that are variables in themselves.

How I define best and how anyone else defines best will be different.  For example does it mean least number of shots or does it mean most reliable or repeatable?

Hmmmm.

Attached Files

Shuz posted this 12 May 2022

Would less than 1600fps apply to 1/10 twist in 25 caliber, as in a .250 Savage?😊

Attached Files

John Alexander posted this 12 May 2022

Bill asks: "John, I’m assuming that the objective is to reach absolute match accuracy with one rifle, one bullet and one powder. That predetermines the requirement for match grade casting, reloading, bench equipment and practices.

If you only have one mold and one powder "

-------------------------

Bill, Yes, that is exactly what I hoped people would limit the discussion to, not because the shooter has only one bullet and one powder but because he has progressed to the point where he thinks that bullet and that powder are hisr best bets.  Otherwise the discussion goes in all directions.  I was hoping for a focus on the shooting part.  We have already seen that JB bench resters think they can do this "final tuning" by reading performance of the load from single three or five shot groups. What do serious CB shooters do for this fine tuning, once they are concentrating on a single bullet and powder?

Short of time more later.

John

Attached Files

John Alexander posted this 12 May 2022

Joe sez: "If we shoot 25 shots, each at a bull, 1 shot per bull, and measure each bull center to bullet hole center; then the standard deviation of those measures is an estimate of the long-term variation in bullet location. From that standard deviation, one can estimate group size for any number of shots per group."

Joe, 

My point was that the method above will NOT give you group sizes representing the precision of that combination UNLESS the rifle has a perfect zero. That is why the mean radius method is needed to first determine the center of the group fired and then the distances from group center -- not from center of the bull. The further off the zero the bigger the error.

John

Attached Files

Show More Posts
Close