Experiment for quantifying lot to lot variations of powders

[Michael Courtney]

Hodgdon claims their Extreme line of powders has smaller lot to lot and temperature variations than other brands. We've got some ideas on the back burner to test their claims regarding temperature variation, but we're starting to get much more practical about quantifying the lot to lot variations in velocity for some of their Extreme powders.

The basic idea would be to pick a powder and cartridge, say H4831 in the 25-06. Keep everything else constant (powder charge, lot of brass, lot of bullets, lot of primers, etc.) and shoot 10 shots each with a carefully measured powder charge from a given lot. Repeat for several lots and then compare the average velocities between the lots. (We think we might need to interleave the shots from different lots to eliminate the possibility of changing environmental factors, changing friction with barrel fouling, etc. from confounding the results.) Repeat with different powders and cartridges, H1000 in 6.5x.284; H4350 in 30-06, Varget in .223 etc.

What do you think? Is this a convincing experimental design to detect lot to lot powder variations? What might you suggest we do differently?

 

[Joe King]

RSI's shooting lab, I mean the whole $760 enchilada would probably have the equipment needed to get meaningful results

 

[Michael Courtney]

RSI's shooting lab, I mean the whole $760 enchilada would probably have the equipment needed to get meaningful results

We are ballistics professionals. We have strain based systems, including the PressureTraceII. We also have PZT-based pressure measurement systems. Pressure measurements would be more appropriate if the intent of the study was focused on the safety aspects of lot-to-lot variations. We're expecting the average velocity and pressure variations to typically be from 1 to 3% from lot to lot. With careful reloading procedures, lot to lot variations would be unlikely to present safety issues unless they were 5-10%. Also, our experience with the PressureTrace II system suggests that typical accuracy expectations are on the order of 2-3%. Experiments are especially challenging when the measurement uncertainty is about the same size as the expected effect being studied, maybe larger.

The focus of the study is on velocity variations that would be of concern to precision long range shooting. Our LED-based chronographs are capable of 0.1% accuracy when calibrated immediately prior to use. It's hard to see what would be added by pressure measurements only accurate to 2-3%, when the precision of our chronographs allows us to see the variations under study, but the pressure measurements do not.

Chronograph based studies also cost a lot less in terms of prep and are much faster in terms of range time. 50 shots per hour is typical. Pressure-based studies are much more expensive and time consuming, and while some shooters are wowed by the graphs of pressure curves, I'll take reliable velocity data over pressure data of quality that is marginal for the intended purpose.

Don't get me wrong, PressureTrace II is great for its intended purposes, many of which are discussed at the web site:

RSI - PressureTrace

I just don't see what it would add to this study.

 

[rscott5028]

Michael,

I look forward to reading about your results.

A few considerations for this or future experiments...

10 is a pretty small sample.

Perhaps test in parallel with 3 different rifles?

Shoulder fired weapons are said to show different MV depending on individual shooter technique. Perhaps use a lead sled or machine rest?

Certainly you would use cartridges for which the powder is intended. But h4831sc is known to work for many cartridges. Are the results the same regardless of which cartridge (capacity and/or case fill ratio within reason) is tested?

Are the results consistent with minimum, moderate, and maximum loads?

Are the results consistent with light, medium, or heavy bullets?

Does it matter if the load has been optimized for consistent MV?

Isolate temp, atmospheric and other external conditions.

Does chamber pressure track consistent with MV for this experiment?

Are the initial fouling shots statistically different than the rest?

Are all canisters of the same lot # the same?

Are all Hodgdon Extreme powders equally consistent?

Are Hodgdon Extreme powders more consistent than other powders that do or don't make similar claims?

I suppose you need to narrow down the number variables in order to have a manageable experiment. But, my concern would be whether or not your conclusions would be the consistent for all conditions rather than just assuming that to be the case.

-- richard

 

[Joe King]

We are ballistics professionals. We have strain based systems, including the PressureTraceII. We also have PZT-based pressure measurement systems. Pressure measurements would be more appropriate if the intent of the study was focused on the safety aspects of lot-to-lot variations. We're expecting the average velocity and pressure variations to typically be from 1 to 3% from lot to lot. With careful reloading procedures, lot to lot variations would be unlikely to present safety issues unless they were 5-10%. Also, our experience with the PressureTrace II system suggests that typical accuracy expectations are on the order of 2-3%. Experiments are especially challenging when the measurement uncertainty is about the same size as the expected effect being studied, maybe larger.

The focus of the study is on velocity variations that would be of concern to precision long range shooting. Our LED-based chronographs are capable of 0.1% accuracy when calibrated immediately prior to use. It's hard to see what would be added by pressure measurements only accurate to 2-3%, when the precision of our chronographs allows us to see the variations under study, but the pressure measurements do not.

Chronograph based studies also cost a lot less in terms of prep and are much faster in terms of range time. 50 shots per hour is typical. Pressure-based studies are much more expensive and time consuming, and while some shooters are wowed by the graphs of pressure curves, I'll take reliable velocity data over pressure data of quality that is marginal for the intended purpose.

Don't get me wrong, PressureTrace II is great for its intended purposes, many of which are discussed at the web site:

RSI - PressureTrace

I just don't see what it would add to this study.

I take then you can graph out the pressure curve as well. Just seems to me that only looking at muzzle velocity is akin to using blinders. My thinking is generate as much data as you can, not just one type of output. Maybe I read it wrong but once set up doesn't that system stay set up until you take it down? so while your shooting you can get feed back on MV, and a visual of the pressure curve of each shot, thus giving you a much more complete picture of exactly the variation of lot to lot?

For example (only) wouldn't it enable you to identify when the pressure curve starts, ends and the duration. and running that data in conjunction with your chronograph enable you to see just exactly the difference?

What do you think? Is this a convincing experimental design to detect lot to lot powder variations? What might you suggest we do differently?

You asked I gave you a suggestion, if you feel the need to get butthurt over it then just disregard my reply pretty simple. Myself if I was going to put out the effort I would want as big of a return on my investment as possible.

 

[Michael Courtney]

You asked I gave you a suggestion, if you feel the need to get butthurt over it then just disregard my reply pretty simple. Myself if I was going to put out the effort I would want as big of a return on my investment as possible.

That's exactly my point. it's not about any offense. The point is that adding a pressure measurement approximately doubles the cost, time, effort, and expense of the experiment, but I don't see how it doubles the value of the results.

How do the (less accurate) pressure measurements make the determination of velocity variations more convincing? I can see how they might illuminate the underlying causes of any velocity variations that might be observed (pressure variations).

I take then you can graph out the pressure curve as well. Just seems to me that only looking at muzzle velocity is akin to using blinders. My thinking is generate as much data as you can, not just one type of output. Maybe I read it wrong but once set up doesn't that system stay set up until you take it down? so while your shooting you can get feed back on MV, and a visual of the pressure curve of each shot, thus giving you a much more complete picture of exactly the variation of lot to lot?

If the extra data added no time or expense, then I would agree, more data is better. But pressure measurements slow data acquisition by a factor of two or three, and the analysis is considerably more involved as well. You're talking about several hundred pressure curves. The system requires attaching a strain gage to a barrel with an adhesive. The calibration procedure is rather involved to ensure accurate results, and the whole procedure needs to be repeated for each rifle barrel in the study for even marginal accuracy.

For example (only) wouldn't it enable you to identify when the pressure curve starts, ends and the duration. and running that data in conjunction with your chronograph enable you to see just exactly the difference?

Not really. If the velocity variations are smaller than 1%, how do we expect to see the reasons for it in pressure measurements that are only accurate to 3%? Theoretically, the velocity is proportional to the area under the pressure curve, but the pressure measurements are not accurate enough to correlate with velocity.

 

[Mikecr]

Hodgdon claims their Extreme line of powders has smaller lot to lot and temperature variations than other brands.

Are you going to prove or debunk this?
If, so you'll need a couple other brands for comparison.
If not, your experiment will only show variances in given lots of one brand -w/resp to those conditions Richard mentioned.

 

[Michael Courtney]

Michael,

I look forward to reading about your results.

A few considerations for this or future experiments...

10 is a pretty small sample.

Perhaps test in parallel with 3 different rifles?

Shoulder fired weapons are said to show different MV depending on individual shooter technique. Perhaps use a lead sled or machine rest?

Certainly you would use cartridges for which the powder is intended. But h4831sc is known to work for many cartridges. Are the results the same regardless of which cartridge (capacity and/or case fill ratio within reason) is tested?

Are the results consistent with minimum, moderate, and maximum loads?

Are the results consistent with light, medium, or heavy bullets?

Does it matter if the load has been optimized for consistent MV?

Isolate temp, atmospheric and other external conditions.

Does chamber pressure track consistent with MV for this experiment?

Are the initial fouling shots statistically different than the rest?

Are all canisters of the same lot # the same?

Are all Hodgdon Extreme powders equally consistent?

Are Hodgdon Extreme powders more consistent than other powders that do or don't make similar claims?

I suppose you need to narrow down the number variables in order to have a manageable experiment. But, my concern would be whether or not your conclusions would be the consistent for all conditions rather than just assuming that to be the case.

-- richard

Wow, if this experiment was funded by a few hundred K and I had a graduate student, you have a whole PhD thesis written out there. You've got some interesting ideas, but really much more than the $1k or so and 100-300 shots that we like to see for a typical experiment. We like projects that can be written up in the space of a typical Precision Shooting article and digested in one sitting by a typical long range shooter with an interest in technical details, but not necessarily in possession of an engineering degree. We like to keep a project to one independent variable and save exploration of other variables for follow up studies.

We've also noticed that Precision Shooting and Varmint Hunter and other typical publication venues are occasionally willing to publish two part series, but are not usually willing to endure a thesis level exploration of a subject. PS has published two of our friction articles, and also published a two part series on bullet stability by Don Miller and myself. Varmint Hunter has published several of our ballistic coefficient articles, but has not yet responded favorably to a recent submission. Accurate Shooter sent a positive response to a submission on bullet weight variations, but they have not actually posted the piece. A thorough exploration is more likely to get done if a venue is willing to publish it.

A sample size if 10 will typically give a uncertainty in the mean velocity of 5 feet per second with Hodgdon Extreme powders and our current brass prep and barrel cleaning techniques. That is good enough to see if lot-to-lot variations are consistent with the claim of 8 fps or closer to Hodgdon's results for other brands. Reducing the uncertainty by a factor of 2 requires four times as many shots. 40 shots for each lot makes this a much more expensive and time consuming experiment, close to $4k and over 1000 shots.

Would you rather see the same cartridge and powder in three different rifles, or different powders and cartridges in three different powders? I think the initial study would do better to study five or six different powders in different cartridges and rifles. Once these results are in hand, then they would better inform the most interesting avenues to proceed in future studies.

So maybe the question to ask is, of all the possible studies, what would be the best use of the first 200-300 shots, and why?

 

[Michael Courtney]

Are you going to prove or debunk this?
If, so you'll need a couple other brands for comparison.
If not, your experiment will only show variances in given lots of one brand -w/resp to those conditions Richard mentioned.

You're right of course, but testing three brands would cost three times the time and effort.

Hodgdon has published a graph claiming 8 fps lot to lot variation. Testing several powders from the Extreme line would at least determine if that result is typical or cherry picked.

 

[Joe King]

Ah ok now I have a better Idea of what your after and working with. Here's how I would proceed initially.

target 3 - 5 set temps, for example 30,60,90 deg f. load 10 rnds for each temp from each lot # of one and only one powder. 10 rnds will give somewhat crude data, but it will be usable without an undue amount of time involved in loading and shooting time. I believe that the greater number of different lots would be more revealing than more round of any one given lot. you could probably get away with 5 rounds of each sample but I personally would always be wondering if my results where skewed.

 

[Mikecr]

Credible investigations best leave credentials behind.
I don't know who's paying you to do whatever ya fancy, but things need to be done correctly for any human benefit.

What purpose or benefit rises in challenging a claim of a single powder provider?
If the results 'seem' favorable to one, then what of the others?
If the results sway otherwise, who would benefit or suffer given any libelous utterance of it?

I really think you're dead in the water on this.

 

[Joe King]

Credible investigations best leave credentials behind.
I don't know who's paying you to do whatever ya fancy, but things need to be done correctly for any human benefit.

What purpose or benefit rises in challenging a claim of a single powder provider?
If the results 'seem' favorable to one, then what of the others?
If the results sway otherwise, who would benefit or suffer given any libelous utterance of it?

I really think you're dead in the water on this.

Maybe it's just curiosity at play here, for that reason alone it's worth doing for the very same reasons that we develop the most accurate loads we're each capable of. It's really no different than say buying a box of Amaxs, a box of Berger VLD's and a box of some other custom LR bullet and taking every measurement possible of each and comparing them to the others. We all know a guy that did just that and more and we're buying a book he wrote based on that (and much more) and universally agreeing to the contribution.

In short why the hell not?gun)

 

[blipelt]

Try looking up Donovan Moran. He is a 1000yd. benchrest shooter and has done tremendous amounts of testing on powder, bullets, and cleaning supplies. He even wrote his own ballistic programs that quite frankly equal to or better than purchased programs in my opinion.

Brent

 

[Michael Courtney]

Ah ok now I have a better Idea of what your after and working with. Here's how I would proceed initially.

target 3 - 5 set temps, for example 30,60,90 deg f. load 10 rnds for each temp from each lot # of one and only one powder. 10 rnds will give somewhat crude data, but it will be usable without an undue amount of time involved in loading and shooting time. I believe that the greater number of different lots would be more revealing than more round of any one given lot. you could probably get away with 5 rounds of each sample but I personally would always be wondering if my results where skewed.

We've thought a lot about an experiment to test temperature variations. This is a much harder experiment, because it requires both going to the range when the temperature is the needed number, as well as waiting several minutes between shots for the barrel to cool. If you wait several minutes for the barrel to cool, then the experiment takes all day and the temperature changes through the day. We've decided it would be best to wait to attempt this experiment until we are able to actively control the temperature of the environment from 0 to 125 deg F and then do the experiment right.