Velocity Drops with Powder Increase

[RDRFab_65]

I have been reading through the forum and found interesting arguments, but I am not finding anything to help me understand this velocity drop as my grain weight increases.

I performed a ladder test from 40 to 42.8 grains in +.2 increments.
As shown you can see the grain increases but drops a fair amount of fps. I find it hard to believe these are chamber issues and suspect these are ignition issues. I do have some concern that the scale i am using just might not be accurate enough. I do an initial throw then trickle up to final grain weight on a Hornady GS-1500 scale. I can drop anywhere from 3-5 powder grains before the scale registers an increase in weight and sometimes that increase shows .2 grain increase instead of .1 grain increase. So, there is some concern that the refresh rate or accuracy is too low on this scale.
Any thoughts on this topic would be greatly appreciated.

Load Information:
All of the rounds are once fired Hornady brass, full length resized to a .001-.0015 shoulder bump, and the necks are sized slightly under and pushed to final dimension using a -.002" expander mandrel. When loading to final depth, I have noticed a very consistent tension between rounds. Concentricity is between .001-.003" and OAL with comparator is +-.001" max from target depth. I use Lapua Scenar 139 gn bullets for 6.5 CM, Federal 210 primers, and the powder is RL17. I have a 29" Heavy Target barrel so thats why I use the slightly slower burn RL17.

 

[golong]

Some drop across nodes is normal, but I have only seen that large of a drop in a particular 308 barrel.

These are from most most recent 6.5 Creedmoor load development. Only a small drop.

41.32767
41.9 2839
42.5 2855
42.6 2858
42.7 2876
42.8 2872
42.9 2875
43.0 2889

 

[std7mag]

It's nothing your doing.
As soon as you said charge increase with velocity decrease i figured what powder you were using.

RL17 is known to do this.

It seems that once the optimal charge weight per case volume is achieved, then velocity will start to drop off.

Kinda like the fuel/air mixture in an engine.

 

[RDRFab_65]

Some drop across nodes is normal, but I have only seen that large of a drop in a particular 308 barrel.

These are from most most recent 6.5 Creedmoor load development. Only a small drop.

41.32767
41.9 2839
42.5 2855
42.6 2858
42.7 2876
42.8 2872
42.9 2875
43.0 2889

This is more what i was expecting to see, not a 30 FPS drop.

It's nothing your doing.
As soon as you said charge increase with velocity decrease i figured what powder you were using.

RL17 is known to do this.

It seems that once the optimal charge weight per case volume is achieved, then velocity will start to drop off.

Kinda like the fuel/air mixture in an engine.

This makes a lot of sense since i drag race on and off, and tune my own car. I still fear the scale i have is causing the FPS drop to be more drastic than the poster above mentioned. I will definitely keep this comment in mind when performing my second test.

 

[OKyote73]

If you are toying with the lands aka jump or jam, slight differences in bullets might cause this also

 

[T-shooter]

I've seen the same thing happen before.

 

[RDRFab_65]

If you are toying with the lands aka jump or jam, slight differences in bullets might cause this also

I typically find a velocity node and then reload a few rounds to verify the node(s). Then I check the jump at the end. I chose .020" jump for initial testing. Open to suggestions on this method here as well. I try to keep my oal with bushings no more than .001" variance. I use a Redding match die set with micrometer adjust.
I use a Rock Chucker Supreme with Hornady Quick Change Bushings. Removing the seater from the press and installing it for every round on a 10 round sample shows the same variance as if i were to leave the die installed, so i just roll with the quick change bushings for convenience.

 

[TAJ45]

Probably not erratic enough but firing pin spring can contribute to those results.
My first thought was the load being at top of max.

 

[dcali]

It’s just random noise. More powder means more velocity on average.

 

[CharlieNC]

So pick any charge weight and shoot ten shots; it will have an average, es, and sd. This degree of variability is present at every charge weight, and unless you take enough shots to provide a suitable average then trying to make inferences about an individual point as a "node" is fruitless.

 

[Fast14riot]

Alex, your scale appears to be about right, RL17 kernels are about 0.02-0.03gr each. So 3-5 kernels per 0.1gr on your scale makes sense.

If you do order the other scale I was telling you about, just beware that repeatability of a scale is more important than absolute value and two scales can cause some doubt. They say the man with two watches is never really sure of the time.

Take the same powder charge and weigh it 10 times on your scale to see if it changes. This can help alleviate any doubt crawling into your head.

 

[RDRFab_65]

All of these are great comments. Thank you all.

 

[Zero333]

So pick any charge weight and shoot ten shots; it will have an average, es, and sd. This degree of variability is present at every charge weight, and unless you take enough shots to provide a suitable average then trying to make inferences about an individual point as a "node" is fruitless.

Absolutely...

This to me is the same thing like some folk giving you single digit ES from 3 or 5 shots, which is not hard to do. But single digit ES with 10+ shots is not that easy.

 

[Brians356]

More powder means more velocity on average.

Right up until the case ruptures, then velocity could start falling off.
-

 

[RDRFab_65]

i use quickload to determine ladder test ranges and verify the ranges by manually inspecting the cases. I need to go back and increase the number of rounds per load as some of you mention i need to increase the population of data.
I likely just wasted my own time with this test but thats ok, we can always go back to the range and do more testing.

 

[dcali]

Right up until the case ruptures, then velocity could start falling off.
-

Something will fall off for sure.

 

[dcali]

i use quickload to determine ladder test ranges and verify the ranges by manually inspecting the cases. I need to go back and increase the number of rounds per load as some of you mention i need to increase the population of data.
I likely just wasted my own time with this test but thats ok, we can always go back to the range and do more testing.

You didn’t waste anything. Put a trend line through the red series in the chart and you will have a good idea of what charge vs velocity will look like. It’s pretty close to linear in this small a range of charge weights. Adding more data will just show you that what you have (the trend line) is probably a good representation. What it won’t show is a drop in velocity.

 

[Fast14riot]

A single round per charge can help you determine your pressure max for your rifle. Also, I would narrow down a new test based on this info, say one step either side of what you like and shoot it 3x, round robin.

Or, pick a charge in that upper flat spot at 42ish grains, and do a seating depth test, then fine tune the powder charge.

Not enough data does not mean wasted data!

 

[ronemus]

I can drop anywhere from 3-5 powder grains before the scale registers an increase in weight and sometimes that increase shows .2 grain increase instead of .1 grain increase

Most scales use a noise reduction scheme which will lock in a value until there is "sufficient" change, then it displays a new value; the programmers do not have trickling in mind - they want a stable value. Some scales will let you adjust a parameter to get the response you want; check the manual. You can also add a bit of powder and "jiggle" the setup (lift the pan, tap the scale housing, bump the table, etc.) enough to get a new reading; that really makes trickling a PITA, so you may want to switch to another scale if the adjustment isn't there.

 

[RDRFab_65]

You didn’t waste anything. Put a trend line through the red series in the chart and you will have a good idea of what charge vs velocity will look like. It’s pretty close to linear in this small a range of charge weights. Adding more data will just show you that what you have (the trend line) is probably a good representation. What it won’t show is a drop in velocity.

when zoomed out by adding another 100 fps on the low and high side of the fps axis, it damn near follows the quick load data. I like how its zoomed in and shows the randomness to the true conditions. I will analyze this trend line and include it in future tests to track what you mention.

A single round per charge can help you determine your pressure max for your rifle. Also, I would narrow down a new test based on this info, say one step either side of what you like and shoot it 3x, round robin.

Or, pick a charge in that upper flat spot at 42ish grains, and do a seating depth test, then fine tune the powder charge.

Not enough data does not mean wasted data!

There seems to me to be a lower node possibly developing around ~2860 fps and the real obvious spot at 42 grains. i think i will add to this data by making a few rounds in the weights around those two possible nodes. I did find that 42.6 is all she wrote so it was good for max pressure testing.

Most scales use a noise reduction scheme which will lock in a value until there is "sufficient" change, then it displays a new value; the programmers do not have trickling in mind - they want a stable value. Some scales will let you adjust a parameter to get the response you want; check the manual. You can also add a bit of powder and "jiggle" the setup (lift the pan, tap the scale housing, bump the table, etc.) enough to get a new reading; that really makes trickling a PITA, so you may want to switch to another scale if the adjustment isn't there.

I am so used to the high response rate of lab grade scales that i totally overlooked this programming option. I plan to test a few kernels on one scale to say .1 or .2 grains and then transfer it to the other and see if it recognizes the same measurement. Also plan on testing repeat-ability and record some small scale tests to track drift over time.