AR Sponsor
Posted: 2/15/2026 12:46:21 AM EDT
[Last Edit: RD20][Edited]
|
Have you ever had a castle nut that you didn't stake come loose after it was properly torqued to spec, 38-42 ft lbs? This is not a question of if staking is necessary or not, and no one needs to know that yours has never come loose because you stake all of your castle nuts either. I'm just curious if anyone has had one come loose after proper torque was applied and NO staking was used. |
Joined:
May 2023
Posts:
660
EE: 0% (0)
WI, USA
|
Nope While I'm sure you'll find someone who has witnessed it, because it is possible, you should find very few because it is very highly improbable. When I'm pretty sure I'm not gonna change a lower's buffer tube, end plate, or rear takedown pin, I do apply some blue locktite between the buffer tube and steel castle nut. I don't know if using locktite changes the value of my "no" response. |
|
I see it all the time. Castle Nut torque value is inadequate. Staking is hit and miss. I crank down on the nut, and when it feels right, say 'click' in my head, and give it back to the shooter. Never had one come back to me. When you've put enough lowers together as I have, you just know. I know that sounds all Bravado, but there it is. Jay |
The Right tends not to riot. They mostly have three speeds:
Uninvolved, Peaceful but passionate protesting, or Genocide
Uninvolved, Peaceful but passionate protesting, or Genocide
|
@AZCOP How many ft lbs would you guess your mind-click is calibrated to? I also thought 38-42' lbs was considerably low. If we feel the need to stake this component, perhaps the torque spec isn't high enough. There's plenty of other torqued items we don't stake. |
|
Originally Posted By RD20: @AZCOP How many ft lbs would you guess your mind-click is calibrated to? I also thought 38-42' lbs was considerably low. If we feel the need to stake this component, perhaps the torque spec isn't high enough. There's plenty of other torqued items we don't stake. It's just one of those things you get a feel for when you do it enough. I did torque one once, then tightened it down further, and it didn't tighten up just a little bit. It's the same for gas key torque specs. If your gas key was torqued to 'spec', you better have a better than a damn good stake on it. Jay |
The Right tends not to riot. They mostly have three speeds:
Uninvolved, Peaceful but passionate protesting, or Genocide
Uninvolved, Peaceful but passionate protesting, or Genocide
Member
Joined:
Jun 2022
Posts:
12858
EE: 100% (30)
TX, USA
|
Originally Posted By Aimless: I had a castle nut come loose on my Anderson Forced reset AR. I can't swear what it was torqued to though. Pretty much the same. But the light and optic also loosened on the same trip to the range. I fired a LOT of rounds in rapid succession that day 
|
“That’s not how it works. You buy more tools and power than you need and as you fuck everything up you gain the knowledge to become a better carpenter.”
Tveddy
Tveddy
ARFCOM’s CHO
Joined:
Aug 2022
Posts:
4359
EE: 100% (1)
FL, USA
I've seen one come loose while shooting, but I don't think it was properly torqued to begin with. The guy just tightened it back down by hand and kept shooting at the trash pile 
|
"I do not love the bright sword for its sharpness, nor the arrow for its swiftness, nor the warrior for his glory. I love only that which they defend."
Joined:
Jan 2025
Posts:
22
EE: 95% (19)
TX, USA
The world's most average shooter, or worse
Joined:
Nov 2010
Posts:
44997
EE: 100% (1)
SC, USA
Joined:
Feb 2023
Posts:
745
EE: 0% (0)
NJ, USA
|
Attached File |
---------------
They who can give up essential liberty to obtain a little temporary safety, deserve neither liberty nor safety. - Benjamin Franklin, 1775
They who can give up essential liberty to obtain a little temporary safety, deserve neither liberty nor safety. - Benjamin Franklin, 1775
|
Since the governments torque number is based off their tube material, design, and rating does anyone know what the 6061 tubes, which there are many of out there, castle nut torque value should be. Anyone know the formula or method used to calculate the values of the nut tightened on an aluminum tube? I have only seen one that came loose more than once on a friend's full auto. But it was long ago and it could have just been coming loose from stretch or deformation of the cheap aluminum some old tubes were made of. |
Joined:
May 2023
Posts:
683
EE: 0% (0)
WI, USA
|
Originally Posted By j3_: Since the governments torque number is based off their tube material, design, and rating does anyone know what the 6061 tubes, which there are many of out there, castle nut torque value should be. Anyone know the formula or method used to calculate the values of the nut tightened on an aluminum tube? I'll break these parts down into a basic mechanics problem to explore part of that. The locking castle nut is a nut. The end plate is a rotation limited washer that doesn't rotate. The lower receiver (in specific, the buffer tower) is the second nut. At this point, it's an engineering/physics problem. The following formula makes the assumption of using a standard bolt with the standard coarse threads. The true formula for a buffer tube would need to include modifications to take into account the lighter threads and hollow rod, but the relative change from altering the material should be adequate, as any modifications to the formula would end up cancelling out when comparing two tubes of different materials but identical structure. T = DFK Torque = Diameter * Forceaxial * K In our case, our two buffer tubes are two cases of the same equation and they can be described as: Torque7075 = Diameter * Forceaxial7075 * K and Torque6061 = Diameter * Forceaxial6061 * K where K is 0.2 for dry nuts and from 0.15 to 0.18 for lubricated nuts. This means lubrication REDUCES the necessary torque by 10% ((0.18-0.2)/0.2) to 25% ((0.15-0.2)/0.2) or, looking at it the other way, a dry nut needs 11% (.2-.18)/.18) to 33% (.2-.15)/0.15) GREATER torque than a lubricated nut. So, that leaves us needing to know how Forceaxial is determined, because the diameter and the K value stay constant across different buffer tubes. It so happens that the Force Axial is also called the clamping force, and for tightening bolts/nuts, generally the proof force is 75% of the yield strength for the bolt. This 75% value is to give adequate leeway for forces added during operation or from the environment so as to avoid the bolt breaking. *A VERY IMPORTANT NOTE: The yield strength in these formulas is technically that of a BOLT with SPECIFIC GEOMETRY made of a SPECIFIC ALLOY with SPECIFIC TREATMENTS. The functional value for a "bolt" here is only derived from the yield strength of the material alloy. While I will show that you can use the relative yield strengths of the alloys to determine the RATIO of buffer tubes' yield strengths, a buffer tube's geometry is not that of a standard threaded rod or bolt. In our case, the buffer tube is the bolt in these equations. As the yield strength is based on the material and design, and since the design (threads, OD, ID) are nominally identical, the change Force Axial is directly related to the ratio of the yield strength new material (6061 to the original 7075) Forceaxial6061 = ForceAxial7075 * (Yield_Strength6061/ Yield_Strength7075) And we substitute our new value for Forceaxial6061 into the previous equation for Torque6061: Torque6061 = Diameter * ForceAxial7075 * (Yield_Strength6061 / Yield_Strength7075) * K Given the 7075 buffer tube's torque equation Torque7075 = Diameter * Forceaxial7075 * K can be re-arranged to solve for Forceaxial7075 Torque7075 / ( Diameter * K ) = Torque7075 / ( Diameter * K ) = Forceaxial7075 we get Forceaxial7075 = Torque7075 / ( Diameter * K ) Now we substitute our new value for Forceaxial7075 into our earlier equation Torque6061 = Diameter * ForceAxial7075 * (Yield_Strength6061/ Yield_Strength7075) * K Torque6061 = Diameter * (Torque7075 / ( Diameter * K ) ) * (Yield_Strength6061/ Yield_Strength7075) * K then we re-arrange it and reduce Torque6061 = Torque7075 * (Yield_Strength6061 / Yield_Strength7075) * Torque6061 = Torque7075 * (Yield_Strength6061 / Yield_Strength7075) It all comes down to the yield strength of 6061 v 7075. 6061 yield is 35k-40k psi. 7075 is around 70k-75k psi, making the above ratio about 1/2. So, by a standard approach 6061 should be tightened LESS than 7075 by almost 50%. Some caveats: This is how you figure yield strength for a solid rod, if I understand correctly. The yield strength of a tube is going to be much less, but the CONCEPT is the same. If you over-torque, deformation and damage to the threads can start to occur. If that occurs, it could make it EASIER for the nut to back off under additional forces. This formula assumes a bolt with coarse threading (which means deeper, thicker stronger threads) standard for its diameter, which would be 7 TPI with an AR castle nut diameter, but AR castle nuts are 16 TPI, much finer. So, this may lower the maximum torque for our buffer tube. Anodizing is a surface coating, so it likely alters the K value too, just as lubrication and galvanization do... but again, those values cancelled out. This is not a statement of how the AR was designed or how the AR's specs were determined. It's just the physics underlying the design. EDIT: Formatting and added a few more statements to make the math clearer. |
Joined:
May 2023
Posts:
684
EE: 0% (0)
WI, USA
|
Originally Posted By Tiribulus: Not disputing your detailed information, but wouldn't this be considered a bolt? I went back and forth about that. Once the tube is locked in place relative to the receiver (via endplate), the tube and the buffer tower together do become the bolt. The buffer tower, alone, can't be a bolt. It has internal threading, as a nut. Either way, after getting flustered over that nomenclature issue for a bit, I realized it was a semantic issue that didn't alter the math and just hit "Submit". |
|
Originally Posted By PreacherOfGadget: I'll break these parts down into a basic mechanics problem to explore part of that. The locking castle nut is a nut. The end plate is a rotation limited washer that doesn't rotate. The lower receiver (in specific, the buffer tower) is the second nut. At this point, it's an engineering/physics problem. The following formula makes the assumption of using a standard bolt with the standard coarse threads. The true formula for a buffer tube would need to include modifications to take into account the lighter threads and hollow rod, but the relative change from altering the material should be adequate, as any modifications to the formula would end up cancelling out when comparing two tubes of different materials but identical structure. T = DFK Torque = Diameter * Forceaxial * K where K is 0.2 for dry nuts and from 0.15 to 0.18 for lubricated nuts. This means lubrication REDUCES the necessary torque by 10% ((0.18-0.2)/0.2) to 25% ((0.15-0.2)/0.2) or, looking at it the other way, a dry nut needs 11% (.2-.18)/.18) to 33% (.2-.15)/0.15) GREATER torque than a lubricated nut. So, that leaves us needing to know how Forceaxial is determined, because the diameter and the K value stay constant across different buffer tubes. It so happens that the Force Axial is also called the clamping force, and for tightening bolts/nuts, generally the proof force is 75% of the yield strength for the bolt. This 75% value is to give adequate leeway for forces added during operation or from the environment so as to avoid the bolt breaking. *A VERY IMPORTANT NOTE: The yield strength in these formulas is technically that of a BOLT with SPECIFIC GEOMETRY made of a SPECIFIC ALLOY with SPECIFIC TREATMENTS. The functional value for a "bolt" here is only derived from the yield strength of the material alloy. While I will show that you can use the relative yield strengths of the alloys to determine the RATIO of buffer tubes' yield strengths, a buffer tube's geometry is not that of a standard threaded rod or bolt. In our case, the buffer tube is the bolt in these equations. As the yield strength is based on the material and design, and since the design (threads, OD, ID) are nominally identical, the change Force Axial is directly related to the ratio of the yield strength new material (6061 to the original 7075) Forceaxial6061 = ForceAxial7075 * (Yield_Strength6061/ Yield_Strength7075) Torque6061 = Diameter * Forceaxial6061 * K Torque6061 = Diameter * ForceAxial7075 * (Yield_Strength6061/ Yield_Strength7075) * K and given the 7075 buffer tube's torque equation is this: Torque7075 = Diameter * Forceaxial7075 * K re-arranging it to solve for Forceaxial7075 ... Torque7075 / ( Diameter * K ) = Torque7075 / ( Diameter * K ) = Forceaxial7075 Forceaxial7075 = Torque7075 / ( Diameter * K ) Now we take our earlier equation Torque6061 = Diameter * ForceAxial7075 * (Yield_Strength6061/ Yield_Strength7075) * K And we substitute our new value for Forceaxial7075 ... Torque6061 = Diameter * (Torque7075 / ( Diameter * K ) ) * (Yield_Strength6061/ Yield_Strength7075) * K Re arrange it Torque6061 = Torque7075 * (Yield_Strength6061 / Yield_Strength7075) * and reduce Torque6061 = Torque7075 * (Yield_Strength6061 / Yield_Strength7075) It all comes down to the yield strength of 6061 v 7075. 6061 yield is 35k-40k psi. 7075 is around 70k-75k psi, making the above ratio about 1/2. So, by a standard approach 6061 should be tightened LESS than 7075 by almost 50%. Some caveats: This is how you figure yield strength for a solid rod, if I understand correctly. The yield strength of a tube is going to be much less, but the CONCEPT is the same. If you over-torque, deformation and damage to the threads can start to occur. If that occurs, it could make it EASIER for the nut to back off under additional forces. This formula assumes a bolt with coarse threading standard for its diameter, which would be 7 TPI with an AR castle nut diameter, but AR castle nuts are 16 TPI, much finer. So, this would likely lower the maximum torque the threaded rod, our buffer tube. Anodizing is a surface coating, so it likely alters the K value too, just as lubrication and galvanization do... but again, those values cancelled out. This is not a statement of how the AR was designed or how the AR's specs were determined. It's just the physics underlying the design. I had always wondered how close to correct using the values for a 7075 bolt sized with similar torque value and thread engagement as the 7075 tube and then looking at that same size 6061 bolts value would be for the 6061 tube. It was a little over half the value using that worthless guess. |
| Absolutely! I used to just torque them to spec. I don’t have definitive numbers, but after I found the first loose one it seemed like I was finding them loose all the time. Mostly truck guns. I’m not sure if the constant vibration on logging roads is a contributing factor or not, but I had it happen on at least 4 rifles. I stake the hell out of them. |
|
Yep, seen several castle nuts come loose. Somewhere around 5 I would guess. In every single case, the castle nut was either unstaked or "staked" but obviously an inadequate job with no displaced metal. High frequency vibration does funny things to threaded fasteners. I RO/staff a lot of major matches. I have just under 50 majors in my log I keep. Each match has anywhere from 80-130+ shooters come through my stage. And easily 5x that number of smaller/local matches, but I don't really keep track of those. So it's no exaggeration when I say I've run somewhere between 10,000 - 30,000 shooters through a stage. I get to see a lot of shooters, a lot of guns, and a lot of malfunctions. Shooters come through with everything from Anderson/BCA-level home built entry level rifles that probably cost $500 or less to factory KAC/LMT/HK rifles where the handguard alone costs well over $500. Interestingly, KAC is the only manufacturer that I'm aware of that intentionally does NOT stake their castle nuts - they say it's not necessary on their integrated system. So far, I believe them. I've never seen a KAC/LMT/HK rifle have a castle nut come loose, but they've certainly had some amount of malfunctions/stoppages. No gun is immune. But I've seen a LOT more Anderson rifles go down than any of the "tier 1" brands. |
|
I have never had one come loose. Usually, on a new build, I won't stake the castle nut until I've fired several hundred rounds through it just to be sure that everything is good. Most have been staked after this test run but a few haven't and I've never had a problem. My method is fairly sophisticated. I put a big crescent wrench on the nut, then a 4 ft piece of 4 inch pipe over the handle. Starting with the pipe perfectly horizontal to the floor, I hang, two-handed, from the pipe until it hits my shop floor. (Not really but I'll bet a few of you were thinking "This guy's a lunatic!"
|
|
Castle nut is probably the only thing I dont torque to spec and hardly care about staking. I just give it a good ugga dugga or 2 with whatever wrench I find and call it a day. Hell, I've used vise grips and a rag wrapped around it to not scratch it before. Its not uncommon for me to do 1k rounds on a rifle in a single training weekend. Never seen one come loose. But I also sometimes loctite them |
|
Originally Posted By PreacherOfGadget: I'll break these parts down into a basic mechanics problem to explore part of that. The locking castle nut is a nut. The end plate is a rotation limited washer that doesn't rotate. The lower receiver (in specific, the buffer tower) is the second nut. At this point, it's an engineering/physics problem. The following formula makes the assumption of using a standard bolt with the standard coarse threads. The true formula for a buffer tube would need to include modifications to take into account the lighter threads and hollow rod, but the relative change from altering the material should be adequate, as any modifications to the formula would end up cancelling out when comparing two tubes of different materials but identical structure. T = DFK Torque = Diameter * Forceaxial * K In our case, our two buffer tubes are two cases of the same equation and they can be described as: Torque7075 = Diameter * Forceaxial7075 * K and Torque6061 = Diameter * Forceaxial6061 * K where K is 0.2 for dry nuts and from 0.15 to 0.18 for lubricated nuts. This means lubrication REDUCES the necessary torque by 10% ((0.18-0.2)/0.2) to 25% ((0.15-0.2)/0.2) or, looking at it the other way, a dry nut needs 11% (.2-.18)/.18) to 33% (.2-.15)/0.15) GREATER torque than a lubricated nut. So, that leaves us needing to know how Forceaxial is determined, because the diameter and the K value stay constant across different buffer tubes. It so happens that the Force Axial is also called the clamping force, and for tightening bolts/nuts, generally the proof force is 75% of the yield strength for the bolt. This 75% value is to give adequate leeway for forces added during operation or from the environment so as to avoid the bolt breaking. *A VERY IMPORTANT NOTE: The yield strength in these formulas is technically that of a BOLT with SPECIFIC GEOMETRY made of a SPECIFIC ALLOY with SPECIFIC TREATMENTS. The functional value for a "bolt" here is only derived from the yield strength of the material alloy. While I will show that you can use the relative yield strengths of the alloys to determine the RATIO of buffer tubes' yield strengths, a buffer tube's geometry is not that of a standard threaded rod or bolt. In our case, the buffer tube is the bolt in these equations. As the yield strength is based on the material and design, and since the design (threads, OD, ID) are nominally identical, the change Force Axial is directly related to the ratio of the yield strength new material (6061 to the original 7075) Forceaxial6061 = ForceAxial7075 * (Yield_Strength6061/ Yield_Strength7075) And we substitute our new value for Forceaxial6061 into the previous equation for Torque6061: Torque6061 = Diameter * ForceAxial7075 * (Yield_Strength6061 / Yield_Strength7075) * K Given the 7075 buffer tube's torque equation Torque7075 = Diameter * Forceaxial7075 * K can be re-arranged to solve for Forceaxial7075 Torque7075 / ( Diameter * K ) = Torque7075 / ( Diameter * K ) = Forceaxial7075 we get Forceaxial7075 = Torque7075 / ( Diameter * K ) Now we substitute our new value for Forceaxial7075 into our earlier equation Torque6061 = Diameter * ForceAxial7075 * (Yield_Strength6061/ Yield_Strength7075) * K Torque6061 = Diameter * (Torque7075 / ( Diameter * K ) ) * (Yield_Strength6061/ Yield_Strength7075) * K then we re-arrange it and reduce Torque6061 = Torque7075 * (Yield_Strength6061 / Yield_Strength7075) * Torque6061 = Torque7075 * (Yield_Strength6061 / Yield_Strength7075) It all comes down to the yield strength of 6061 v 7075. 6061 yield is 35k-40k psi. 7075 is around 70k-75k psi, making the above ratio about 1/2. So, by a standard approach 6061 should be tightened LESS than 7075 by almost 50%. Some caveats: This is how you figure yield strength for a solid rod, if I understand correctly. The yield strength of a tube is going to be much less, but the CONCEPT is the same. If you over-torque, deformation and damage to the threads can start to occur. If that occurs, it could make it EASIER for the nut to back off under additional forces. This formula assumes a bolt with coarse threading (which means deeper, thicker stronger threads) standard for its diameter, which would be 7 TPI with an AR castle nut diameter, but AR castle nuts are 16 TPI, much finer. So, this may lower the maximum torque for our buffer tube. Anodizing is a surface coating, so it likely alters the K value too, just as lubrication and galvanization do... but again, those values cancelled out. This is not a statement of how the AR was designed or how the AR's specs were determined. It's just the physics underlying the design. EDIT: Formatting and added a few more statements to make the math clearer. This is just copy paste from chat gpt and I doubt anyone is gona read it lmfao I love it when people copy paste chat gpt as if its their own knowledge |
|
Originally Posted By PreacherOfGadget: I went back and forth about that. Once the tube is locked in place relative to the receiver (via endplate), the tube and the buffer tower together do become the bolt. The buffer tower, alone, can't be a bolt. It has internal threading, as a nut. Either way, after getting flustered over that nomenclature issue for a bit, I realized it was a semantic issue that didn't alter the math and just hit "Submit". Originally Posted By PreacherOfGadget: Originally Posted By Tiribulus: Not disputing your detailed information, but wouldn't this be considered a bolt? I went back and forth about that. Once the tube is locked in place relative to the receiver (via endplate), the tube and the buffer tower together do become the bolt. The buffer tower, alone, can't be a bolt. It has internal threading, as a nut. Either way, after getting flustered over that nomenclature issue for a bit, I realized it was a semantic issue that didn't alter the math and just hit "Submit".  |
Joined:
Jul 2025
Posts:
32
EE: 100% (6)
MN, USA
|
Yeah I have seen them come loose numerous times, quite often resulting in someone eventually bringing it in for fixing, with their now rotating buffer tube with the threads all trashed from the steel end plate chewing into it, because why bring it in to me before it requires a new buffer tube?…. I would always clean them well and loctite it well, 9 times out of 10 you can skip the staking then… |
a loaded gun won’t set you free, so you say…
| Nope. And I have never staked a castle nut. Even if it does come loose it won't hurt anything unless it unthreads far enough for the plate to move. |
"Some people have issues. Sounds like he signed up for an entire subscription." ~Brohawk
Proud member of Team Ranstad.
Arfcom St Jude Mafia 3 years
Arfcom callsign: trenchfoot
Proud member of Team Ranstad.
Arfcom St Jude Mafia 3 years
Arfcom callsign: trenchfoot
|
Originally Posted By tac556: I would always clean them well and loctite it well, 9 times out of 10 you can skip the staking then… Not to hijack the thread, but before anyone takes that advice, think through whether there's ANY doubt in your mind you might want to remove that tube anytime soon. Overcoming a stake is a lot less hassle than loctite, before, during, or after application. Just following the TM is never a bad idea. |
|
This thread was designed to examine if proper torque alone is sufficient for a castlenut to keep from becomimg loose, or if other means are truly necessary. In addition to "not staking," I should have included "and didn't use threadlocker," but the answers are still interesting. |
Joined:
Sep 2022
Posts:
1
EE: 0% (0)
TX, USA
|
Yes, it has happened more than once. One of the first lowers I built around 20 years ago, I didn't have the correct punch and figured I could just torque and forget about it. I was wrong because at some point I noticed the stock twisting. Later I had it happen again due to laziness. It also happened on a pre built lower either from PSA or Anderson. In the past I also sourced parts from a lot of different vendors and found the quality of the castle nuts and the plate varied quite a bit. It's fast and easy to stake and build it right, and have the assurance rather than be lazy and complacent. If you cut a corner that is that easy then it makes me wonder what other things that are done wrong on the gun. |
Joined:
Mar 2026
Posts:
219
EE: 100% (1)
OH, USA
AR Sponsor