June 19, 2020, 10:26 AM | #26 |
Senior Member
Join Date: October 21, 2010
Posts: 1,024
|
Ewa Thoughts, your last post (#25) referred to using fired case volume for the 185 gr, .45 ACP example. QL advises using the sized volume for cartridges having less than 30,000 PSI max chamber pressure. Does GRT not agree? If not, why not?
|
June 19, 2020, 11:49 AM | #27 |
Junior Member
Join Date: June 18, 2020
Posts: 9
|
Off the wall results.
I took the 185 gr, charge from Shooting Times (6.1 grs.) and using the measurements from some Win 45 ACP cases fired in my Para 45/14, I came up with a case length of .891", H2O volume of 27.8 Grs, and used an overall cartridge of 1.24". I then took your submitted Berry's 185 bullet (flat based dimensions) and placed it in my GRT simulation. I then adjusted the Initial Pressure (IP) using the Yellow Star wizard to Pistol and lead (as plated bullets act more like lead than hard ball). That made the IP 1160 psi. I set the Gas Seal at Soft Lead. I got a simulated velocity of 935 fps and peak pressure of 13584 psi. Subtracting the Shooting Times velocity from the GRT sim. velocity (935 - 812) = 123 fps difference, and dividing the difference by 812 shows a 15% difference.
I then looked at the picture of the bullet base on Shooting Times and measured the approximate hollow base dimensions. This was a guess, but I cam up with L = .3, A = .296, and B = .2". This is very rough, you may want to drop your actual measurements here. Adding those hollow base measurements into the simulation produced a velocity of 838 fps, and a pressure of 10617 psi. So taking the Shooting Times velocity of 812 fps and subtracting that from the GRT sim. of 838 fps, I get a difference of 26 fps. Dividing (from your view point) 26/812, I get an difference of 3.2%. Inspection of the effect of the hollow base dimensions on the simulation is not minimal, it is 935 - 835 = 100/835 = 11.9%. 11.9%/15% = .79, or 79% of the difference. So now you may understand that the details can severely affect the outcome of the simulation. At least for this combination of components no warning of high or over pressure were displayed. You may also understand that comparing published data to any other published data is very difficult because the firearms and loading techniques vary so much. It much better to question results from your own fired results. I am sure that you have found differences in your fired loads as compared to published load manuals. That is at least comparing one party to another, not to a third party that can not speak for itself. |
June 19, 2020, 11:53 AM | #28 |
Junior Member
Join Date: June 18, 2020
Posts: 9
|
Higgite
The GRT simulation is not a copy of QL. GRT bases its calculations on the calculated final chamber volume as the case as expanded to fill the chamber. If the case did not expand, then the correct volume would be the case as sized. I will double check with Gordon on this however, to be sure I have not answered incorrectly. Last edited by Ewa Thoughts; June 19, 2020 at 11:55 AM. Reason: Point to questioner |
June 19, 2020, 12:09 PM | #29 | |
Staff
Join Date: September 25, 2008
Location: CONUS
Posts: 18,433
|
Quote:
As a writer and editor, I completely disagree that what I wrote "begs for a reply." It doesn't. What it begs for is a revised version of GRT that will generate results closer to what the real world experiences. Since this exchange has gone from my intention of offering some feedback in the hope of improving GRT to a urination contest, I have gone back and tried playing with various aspects. Fixing the hollow-base bullet by inputting the dimensions of the cavity made a negligible difference. Just a few minutes ago I played with the initial pressure. Taking that all the way down to 500 psi (which is rather unrealistic) still left the velocities for the 185-grain bullet load about 100 to 150 fps faster than what the article reported as real world results. This seems to suggest that the error is generated by something in the powder data, and that's way over my pay grade. I have given you what I can. I obviously can't provide data that I don't have, and which I stated in my e-mail to Gordon that I don't have. Your continued harping on that appears more and more to be a defensive tactic rather than a sincere effort to foster a cooperative exchange.
__________________
NRA Life Member / Certified Instructor NRA Chief RSO / CMP RSO 1911 Certified Armorer Jeepaholic |
|
June 19, 2020, 12:21 PM | #30 |
Junior Member
Join Date: June 18, 2020
Posts: 9
|
Final
Since we are not able to communicate in a positive manner, I will discontinue any post here. I have tried to help, requesting a copy of your GRT .grtload file so I could vet the problem you are experiencing. However, all I get is I cannot give you any data.
Sorry I tried to help. |
June 20, 2020, 12:22 AM | #31 |
Senior Member
Join Date: October 23, 2005
Location: US
Posts: 3,649
|
Sigh... I’m afraid we may have just tarred, feathered, and ran off an opportunity to have some constructive input with GRT. It is a FREE tool. FREE. And, more importantly, still developing and willing to take inputs from users to help refine data. I see potential for great value in that. I’ve personally observed it to be in the ballpark of velocity (can’t speak to pressure because I can’t measure it) on a few known loads, plus their results in most common situations seem to jive with loads listed in published data. The load that is way off is of the same sort that has already been identified to have similar large variances in Quickload, which is an excellent product but is not free and does not utilize user data to refine the model (to my knowledge).
I guess what I’m trying to say is “lighten up a little” and give them half a chance.
__________________
Support the NRA-ILA Auction, ends 03/09/2018 https://thefiringline.com/forums/sho...d.php?t=593946 |
June 20, 2020, 01:46 AM | #32 |
Senior Member
Join Date: September 28, 2013
Posts: 3,769
|
It is indeed free and pretty good. I appreciate it, and continue to use and support it. It could be even better without this episode though.
-TL Sent from my SM-N960U using Tapatalk |
June 20, 2020, 02:09 PM | #33 | ||
Staff
Join Date: September 25, 2008
Location: CONUS
Posts: 18,433
|
One last thought:
Quote:
Quote:
In fact, when you use the Yellow Star wizard, it sets the gas seal/friction entry to Standard for electroplated projectiles.
__________________
NRA Life Member / Certified Instructor NRA Chief RSO / CMP RSO 1911 Certified Armorer Jeepaholic |
||
June 20, 2020, 02:45 PM | #34 | ||
Senior Member
Join Date: August 25, 2008
Location: In the valley above the plain
Posts: 13,401
|
Quote:
If you wish to have meaningful discussion, that's probably the best place to do it. You can have live interaction - whether via text or voice chat - and get through a topic or concept quickly and clearly, rather than dealing with the slow nature of discussions and misunderstandings on a forum. Quote:
QL wants to be as accurate as possible, by modeling with a small set of the best data reasonably obtainable. GRT seems to be okay with modeling based on a large sample size and a "close enough" average. Both approaches have their pros and cons. Right now, my issue with data submissions for GRT is the open invitation to submit real world data, but then there are a ton of hidden requirements, limitations, and exceptions. For starters, there is the requirement for the data to be "2x2x2x5." That is: 2 cartridges (of different caliber) 2 bullet weights (per cartridge) 2 charge weights. (Min and Max) 5 rounds of each combination. That works out to 20 rounds per cartridge, and 40 rounds total. And, from discussions that I've seen in the discord server, they don't really want your data unless the charge weights are starting charge and max charge, and velocities were recorded with a Magnetospeed or Labradar. So, you're basically just recreating published data, while dumping components into the dirt, and your data is "suspect" unless you shoot it over a Magnetospeed or Labradar. They also don't want data from revolvers. Because GRT's model automatically reduces 20-40% of the velocity for firearms with "vents or cylinder gaps", depending upon cartridge and pressure, they don't want neutered (revolver) data to begin with. (This is from Gordon, himself.) They also get picky about barrel length. For example, Gordon has stated that he doesn't want data for 700X, 800X, and "some other very fast propellants" unless testing was done in barrels at least 8" in length, with a preference for 14". And, then there are the exceptions to the exceptions, that are, yet again, not listed or talked about anywhere. For example: A user on their discord server asked if he could submit mid-range .300 AAC Blackout data from a bolt action rifle (since locked breech is what they want) to "fill in the gaps" and provide more data for the model. The response was that his data would not be useful, since the GRT model for .300 AAC would be aimed at gas-operated firearms (understandably). And then Charlie unceremoniously directed him to create his own powder file for "nonstandard use" by using techniques in the GRT Optimal Barrel Time video: https://www.youtube.com/watch?v=XxDJVROdyxc "We want your data!" "Wait... That's not good enough, that's the wrong gun, you didn't fire enough rounds, your equipment was inferior. Oh, and this powder was on the secret list of special powders that have special test requirements that we never talked about or published anywhere. Also, again, your gun was wrong ... but the other, other wrong."
__________________
Don't even try it. It's even worse than the internet would lead you to believe. |
||
June 21, 2020, 01:44 PM | #35 |
Staff
Join Date: March 4, 2005
Location: Ohio
Posts: 21,021
|
I tried matching QuickLOAD's 231 model to pressure and velocity data on Hodgdon's site for a jacketed bullet by changing the powder burn rate to change velocity and changing the start pressure (aka, initial pressure) to change peak pressure without much effect on velocity. Since QL allows start pressure to be changed continuously above 145 psi rather than just allowing fixed category numbers by the bullet construction, these two factors may be fine-tuned switched back and forth to get a match. I was at first surprised to find the start pressure had to go to over 8000 psi to get a match, but then remembered something I've posted repeatedly was at play:
In pistol cartridges with small powder spaces, it is not uncommon for the primer to unseat the bullet before the powder burn gets completely underway. I expect this was happening to Hodgdon during their tests. The amount of jump to the throat then effectively expands the powder space before the pressure has built much, and completing the initial burning stage (until the bullet is engraved) in that larger volume may well explain the big velocity difference. I still had some difference, but more like 30 fps over for the lower loads and slightly under for the top load. So I am thinking a jump in powder space probably explains a good portion of the velocity difference. I made similar changes in GRT and got the velocities down, though not all the way. I do wish the start pressure could be varied directly in value and not just by the construction category (unless I am just missing the method of doing that).
__________________
Gunsite Orange Hat Family Member CMP Certified GSM Master Instructor NRA Certified Rifle Instructor NRA Benefactor Member and Golden Eagle |
June 23, 2020, 01:44 PM | #36 | |
Staff
Join Date: March 11, 2006
Location: Upper US
Posts: 28,675
|
Quote:
And, just for what its worth, if an American tells you "yeah, yeah" or "yeah, right..." he usually means the same thing that Berliner does. (kiss my …..) though that is not the literal translation of those words... I found the discussion somewhat interesting, in general, but not anything important to me as I don't use any computer simulations, only real world results. The discussion does prove that GIGO is still a valid thing, probably always will be. Garbage In, Garbage out. IF your data isn't exactlty what the program is built to "see" your results won't be, either.
__________________
All else being equal (and it almost never is) bigger bullets tend to work better. |
|
June 23, 2020, 07:08 PM | #37 | |
Staff
Join Date: September 25, 2008
Location: CONUS
Posts: 18,433
|
Quote:
__________________
NRA Life Member / Certified Instructor NRA Chief RSO / CMP RSO 1911 Certified Armorer Jeepaholic |
|
June 26, 2020, 10:50 PM | #38 | |
Senior Member
Join Date: August 25, 2008
Location: In the valley above the plain
Posts: 13,401
|
I've been following the discussions on GRT discord. Thought a minor update was in order.
Apparently, GRT just absolutely cannot compensate for anything outside of the pressure window the powder model was intended for. This is one of several replies over the last week, in regards to people posting real world velocities that were more than 10% different than GRT's predictions, even with good data. In this case, GRT predicted 2,101 fps, while recorded velocity averaged 1,803, for 7.5x55 Swiss GP90/03. Response from Charlie: Quote:
__________________
Don't even try it. It's even worse than the internet would lead you to believe. |
|
June 27, 2020, 12:57 PM | #39 |
Senior Member
Join Date: September 28, 2013
Posts: 3,769
|
We have seen this in super light loads. It is quite alright for me. Software models always have limits. Good thing is the sim tends to be on the conservative side.
-TL Sent from my SM-N960U using Tapatalk |
June 27, 2020, 02:22 PM | #40 |
Senior Member
Join Date: September 28, 2013
Posts: 3,769
|
Test fired another load of 7mm Mauser yesterday.
Bullet: Sierra #1950 170gr Powder: 31.8gr IMR8208XBR Load ratio: 65%<70%, a light load. GRT sim muzzle velocity: 2235fps Labradar muzzle velocity: 2163fps # of shots: 9 So GRT is off by about 3%. Not bad at all. -TL PS. GRT indicates that the powder doesn't get 100% burnt. So there is unburnt powder coming out of muzzle. Last edited by tangolima; June 27, 2020 at 02:31 PM. |
June 27, 2020, 02:46 PM | #41 | |
Staff
Join Date: March 4, 2005
Location: Ohio
Posts: 21,021
|
Got GRT Working Pretty Well
Quote:
That window discovery leads to my second discovery, which is one of the differences between QuickLOAD and GRT: The drop-down menu that selects the gas sealing and bore friction DOES NOT affect the initial pressure (start pressure) as I had assumed they did. In QuickLOAD's model, bore friction cannot exceed start pressure. This is not very realistic, particularly in rifles, for which friction peaks when pressure peaks due to all the acceleration g-forces tending to upset the bullet outward harder against the bore. So friction actually is a function of pressure that is in addition to the start pressure friction to engrave the bullet. You can, however, compensate for gas sealing in QL by raising the start pressure. Its model for long barrel friction, which occurs after the pressure peaks, is then pretty good, and pressure-induced friction can be modeled by adding a few percentage points to the bullet weight. But GRT appears to keep all that separate and seems to have (I need to evaluate this more) a better friction model than QuickLOAD that keeps them independent variables. Once I found the start pressure was independent of the gas sealing model, a whole new range of possibilities opened up. Finally, going back to post #1, the 230-grain load with 5.3 grains of 231/HP38 was reported in the Shooting Times Magazine article to have a velocity of 724 fps, and that number seemed to match what Aguila Blanca had measured. If I had just thought back to personal experience, that would have seemed low. So, when in doubt, get a second opinion. I looked up Hodgdon's data for the same weight bullet in their 5" test barrel, and there 5.3 grains of W231/HP38 is listed as a maximum load and they report 834 fps and 832 fps for lead and jacketed bullets, respectively. That's 110 fps faster than the Shooting Times number and Aguila Blanca's similar numbers. That difference accounts for most of the apparent error first seen in the GRT numbers. It has to be born in mind that the QuickLOAD and GRT models both, unless you change them, assume test barrel-like dimensions, which are minimum in chamber size. Real chambers can be a LOT sloppier, as you can see from the slide breech face to barrel extension gap in a lot of production 45 Auto guns. When you include the tendency to see a lot of primers unseating bullets to increase the powder space, I think the discrepancies are pretty well explained. In GRT I put a lead bullet using the physical dimensions of the Sierra Match 230-grain RN FMJ bullet by setting the gas sealing and barrel friction to soft lead. I lowered the start pressure to 1160 psi. I lowered the default case capacity from 27.8 grains H₂O to 26.0 grains, which matches my measurements of unfired 45 Auto brass because I think QuickLOAD is correct that at low pressures where the cases are not pinned hard against the chamber wall you do not see so much expansion of the brass at the pressure peak. I gave it 5.3 grains of 231 without altering the model. I gave it a 5" barrel to match the Hodgdon test barrel. I gave it a 1.200" COL to match Hodgdon's COL. The result is as follows: Code:
GRT Pressure Velocity 96.5% MAP 829.6 fps 20270 psi/17374 CUP Hodgdon Pressure Velocity 93.9% MAP 834 fps 19716 psi/16900 CUP *psi/CUP ratio based on SAAMI MAPS ratio for the cartridge which are 21,000 psi/18,0000 CUP. Based on this, the 6.1 grain loads from the ST article would run about 26,145 psi in the Hodgdon test gun, so I added the load warning to the first post, as that's about 14% above the SAAMI +P MAP of 23,000 psi for 45 Auto. Not yet to the proof load range (31,000 to 33,000 psi), but enough to beat up a gun if you fired it much.
__________________
Gunsite Orange Hat Family Member CMP Certified GSM Master Instructor NRA Certified Rifle Instructor NRA Benefactor Member and Golden Eagle |
|
Thread Tools | Search this Thread |
|
|