Velocity plateaus in load testing: Why?

[maximus otter]

I’m an experienced shooter and reloader. I’ve read widely concerning load development, and l’m soon going to be starting the process for a new (to me) Tikka Super Varmint in .223

l’ve read often that skilled reloaders will perform a test known as the Satterlee test or method, i.e. they’ll load twenty cartridges, two each of loads precisely 0.2 grains (or .3 or .5, as appropriate) apart, shoot them and chart the velocity results. This will result in a chart which is essentially linear as one would expect (more powder = more velocity). However, typically there will be one or more “flat spots” on the line where increasing the powder charge produces little or no increase in muzzle velocity. The object is to pick a load in the middle of this flat spot as your accuracy load, allowing one a little “sneeze room” for slight differences in powder temperature, load dispensing, etc.

My question: Why does this happen? Why do differing powder weights not affect muzzle velocity? Everyone seems to accept the phenomenon, yet no-one l’ve read can explain why it happens.

Am l missing something obvious?

maximus otter

PS: l’m aware of and understand ideas like nodes, barrel harmonics, optimum barrel time and so on.

 

 

[Popsbengo]

26 minutes ago, maximus otter said:

I’m an experienced shooter and reloader. I’ve read widely concerning load development, and l’m soon going to be starting the process for a new (to me) Tikka Super Varmint in .223

l’ve read often that skilled reloaders will perform a test known as the Satterlee test or method, i.e. they’ll load twenty cartridges, two each of loads precisely 0.2 grains (or .3 or .5, as appropriate) apart, shoot them and chart the velocity results. This will result in a chart which is essentially linear as one would expect (more powder = more velocity). However, typically there will be one or more “flat spots” on the line where increasing the powder charge produces little or no increase in muzzle velocity. The object is to pick a load in the middle of this flat spot as your accuracy load, allowing one a little “sneeze room” for slight differences in powder temperature, load dispensing, etc.

My question: Why does this happen? Why do differing powder weights not affect muzzle velocity? Everyone seems to accept the phenomenon, yet no-one l’ve read can explain why it happens.

Am l missing something obvious?

maximus otter

PS: l’m aware of and understand ideas like nodes, barrel harmonics, optimum barrel time and so on.

 

 

Maybe an incomplete or disrupted combustion affect? Or could it be a measurement aberration caused by quantisation error in the chrono?

I'm also dubious of claims to be exactly 0.2 grains etc being weighed when, for example,  there's a +/- 0.1 grain error in a RCBS Charge Master.

[TJC]

Do the test 5 times with the same loads and then see if you have the same nodes or plateaus. Evidence from those who have tried it says no which basically suggests the 10 round process is not repeatable and is therefore statistically irrelevant. 

[Popsbengo]

16 minutes ago, TJC said:

Do the test 5 times with the same loads and then see if you have the same nodes or plateaus. Evidence from those who have tried it says no which basically suggests the 10 round process is not repeatable and is therefore statistically irrelevant. 

I agree,  having read the original article I'm pretty sure the tests are irrelevant due to statistical errors.  Claiming to measure velocity at a resolution better than 0.3% (8 in 3000) with a Magnetospeed is suspect: Magnetospeed quote 0.5% error in their spec.  Also measuring powder weights to better than 0.4%. (0.2 in 50 with likely measuring tolerance of +/- .1) without resorting to mega expensive lab scales in a stabilised environment is suspect.

[Andrew]

 You really need a pressure gun. If you look at a tracing of a pressure curve you will see that (generally) the peak pressure eventually takes a steep upward turn while adding to little to velocity. I believe this was the reasoning behind the idea. I have used it myself in developing loads. It won't pin point a narrow load band but you can track at which point the powder added gives smaller velocity gains indicating that you're probably hitting a peak with that particular powder.~Andrew

[Popsbengo]

15 minutes ago, Andrew said:

 You really need a pressure gun. If you look at a tracing of a pressure curve you will see that (generally) the peak pressure eventually takes a steep upward turn while adding to little to velocity. I believe this was the reasoning behind the idea. I have used it myself in developing loads. It won't pin point a narrow load band but you can track at which point the powder added gives smaller velocity gains indicating that you're probably hitting a peak with that particular powder.~Andrew

Andrew, this is interesting and explainable but is it not a different effect to the supposed narrow band plateau(s) being reported in the so called Satterlee test?  After all, the suggestion is that the velocity increases again after the so called plateau effect.

 

[Andrew]

Can't explain it as I have been a bit more coarse in my testing, looking for the end mark in loading a given powder. This kept me from incrementally loading a powder that was never going to reach a (needed) velocity level.~Andrew

[brown dog]

To the original question: I've only ever heard of OCW in the context of harmonics etc, and that makes total ballistic sense to me.  The multiple 'velocity plateaux' concept is a new one on me; which makes no sense.  My kneejerk would be to wonder if it's a process defined/developed by someone who didn't quite understand OCW?

[MJR]

Two completely different processes. OCW to find a charge that promotes accuracy by tuning the load to the barrel harmonic. 

Saterlee load development looks for a charge weight spread that effectively reduces velocity variation to promote long range accuracy and ease reloading by giving a 'window' charge weight of several tenths of a grain.

I've used both and at the moment favour the saterlee method. Interestingly I haven't yet found a load using this method that is not accurate. I have also found very accurate 100yd loads using the OCW method that gave high velocity SD and went to pieces at longer ranges. 

[No i deer]

Do these plateaus occur on the Quickload predictions...?

If no why not...?

[MJR]

No idea😁

[Popsbengo]

8 minutes ago, No i deer said:

Do these plateaus occur on the Quickload predictions...?

If no why not...?

I suspect they don't.

[MJR]

But quickload, whilst being an extremely useful tool is a prediction, not fact. 

I am currently looking to improve charge weight accuracy to re run previous tests and further investigate the charge weight/velocity match increments.

[brown dog]

7 hours ago, MJR said:

Two completely different processes. OCW to find a charge that promotes accuracy by tuning the load to the barrel harmonic. 

Saterlee load development looks for a charge weight spread that effectively reduces velocity variation to promote long range accuracy and ease reloading by giving a 'window' charge weight of several tenths of a grain.

I've used both and at the moment favour the saterlee method. Interestingly I haven't yet found a load using this method that is not accurate. I have also found very accurate 100yd loads using the OCW method that gave high velocity SD and went to pieces at longer ranges. 

Yes, and OCW makes physical and chemical sense: The link between barrel time and harmonics is a cause of nodes at the target that's easy to account for.

I haven't read about 'saterlee' beyond this thread; but, as per the OP's question, I'd be interested in hearing why the postulated vel plateaux are thought to be happening.   (I acknowledge Andrew's point about diminishing returns at the upper ends, but I believe this Saterlee is postulating mid curve vel nodes, akin to OCW nodes...and these make no physical/chemical sense.)

For example, artillery and mortar ammunition generally has charge increments. Nowhere do the velocities 'plateaux' as the increments are increased.  More chemical power means greater vel. More grunt, more speed.

What's the vel plateau theory?

 

[maximus otter]

3 minutes ago, brown dog said:

...as per the OP's question, I'd be interested in hearing why the postulated vel plateaux are thought to be happening.

What's the theory?

 

That’s the question...

maximus otter

[baldie]

7 hours ago, No i deer said:

Do these plateaus occur on the Quickload predictions...?

If no why not...?

Because Quickload is just a computer programme, that people rely upon far too much.

Real world shooting, shows it up for what it is...a very basic guide.

[brown dog]

The reason quickload doesn't show mid curve velocities plateauing with increasing charge weight, will be because, excepting empirical small-sample experimental and statistical error, it won't be happening.

This is like a school physics experiment where students measure how far a toy car rolls off a ramp and draw a graph of the results.  As they raise the end of the ramp one book at a time, the graphed outcome will (should) show that the car will go further for each height raise

- more energy in, more car velocity. End of.

- but not all the graphs produced will show that. A sneeze, a push, aruler error, whatever will mean some get the expected graph, and some don't. But the science doesn't change.

I just googled Satterlee's theory:

"The following quote from Scott summarizes a key observation with load development, “Im not sure it is optimal charge weight at all but optimal velocity for the bullet weight that determines barrel harmonics. The reason I say this is because I can get the exact same great results in a velocity node using seven different powders. As long as I keep the bullet going the same velocity from powder to powder I get the same result. Powder charge may be slightly different but as long as I’m in the velocity happy place for that bullet and my cartridge, powder is secondary.” "

Some confused concepts in there. He's describing a bullet velocity sweet spot. And relating it to harmonics. And saying that harmonic sweet spot can be hit with a number of different powders. And because of that, he's relabelling OCW as optimal bullet vel. Umm. I think? 

 

So, back to this:

 

[Shuggy]

Hi guys. Plateau and mesa effects in propellant burning are a real thing and have been known about for many decades by energetic chemists designing gun and rocket propellants. The generally accepted theory is that they are caused by the propellant additives (e.g stabilisers, deterrents) having different surface regression/pressure rates compared to the base propellant (nitrocellulose or nitrocellulose and nitroglycerine for small arms).

There are many academic papers about this in the literature if you care to look.

[No i deer]

3 hours ago, baldie said:

Because Quickload is just a computer programme, that people rely upon far too much.

Real world shooting, shows it up for what it is...a very basic guide.

Yes I agree Dave.

There alot of variables we can't seem too account for but some we can.

Most variables are all types of case prep including neck tension,primers and powder weighing.

I shoot for a group then fine tune from that. If i am happy with a group and I've got no pressures I then just play with the seating depth.

I check velocities last for figures to enter into my bal cal

[Popsbengo]

2 hours ago, Shuggy said:

Hi guys. Plateau and mesa effects in propellant burning are a real thing and have been known about for many decades by energetic chemists designing gun and rocket propellants. The generally accepted theory is that they are caused by the propellant additives (e.g stabilisers, deterrents) having different surface regression/pressure rates compared to the base propellant (nitrocellulose or nitrocellulose and nitroglycerine for small arms).

There are many academic papers about this in the literature if you care to look.

That's interesting. It would account for the effect.  However, I'm with Browndog,  this particular experimental result seems prone to statistical errors. Just too many variables that are only partially controlled.

[MJR]

Shuggy, that's interesting, could you give a suggestion as to where to find these papers.

Here are three saterlee tests done with three calibres and each clearly show the velocity/charge nodes. I intend to repeat these test with a more accurate powder weighing set up quite soon.

However in each case you can see the velocity and charge weight track together and then reach a point where they intersect.

243-105 Bergers, H4350.docx

129SST load development.docx

155TMK Velocity vs Charge weight Varget.docx

[Popsbengo]

56 minutes ago, MJR said:

Shuggy, that's interesting, could you give a suggestion as to where to find these papers.

Here are three saterlee tests done with three calibres and each clearly show the velocity/charge nodes. I intend to repeat these test with a more accurate powder weighing set up quite soon.

However in each case you can see the velocity and charge weight track together and then reach a point where they intersect.

243-105 Bergers, H4350.docx

129SST load development.docx

155TMK Velocity vs Charge weight Varget.docx

I don't understand what you are saying.  What I see in your graphs is two series showing a general increase in velocity with increase in charge but on axis scales that are very fine increments. I don't see a statistically significant effect.  In fact (if I understand your series) there's no repeatability between series.   Am I looking at the graphs wrong ?

[Shuggy]

@MJR Try Combustion of Solid Propellants by G. Lengellé, J. Duterque, J.F. Trubert for a good review. But be warned, it’s pretty technical stuff.

There are loads of others: Google Scholar is your friend!

[MJR]

Shaggy, thanks.

pops, you won't see repeatability, they are three different calibres. 

[Popsbengo]

13 minutes ago, Shuggy said:

@MJR Try Combustion of Solid Propellants by G. Lengellé, J. Duterque, J.F. Trubert for a good review. But be warned, it’s pretty technical stuff.

There are loads of others: Google Scholar is your friend!

Hi Shuggy,  you seem to have a handle on this effect.  Do you actually think the tests done in the original Satterlee article is demonstrating this phenomena amongst the statistical noise?