The Published Test Says Your Lock Is Not the Weak Point - The Frame Is

Last episode I said I wanted somebody's published test of what actually holds a front door shut. I went looking. It exists, it has a number, and what it grades is not the thing you have been sold. ASTM F476, Standard Test Methods for Security of Swinging Door Assemblies. It grades an entry door assembly 10, 20, 30, 40. Climbing those grades, almost everything gets harder - and one number never moves at all. IMPORTANT: EVERY FIGURE BELOW IS FROM THE 1984 EDITION The current edition is F476-23, approved January 2023. ASTM sells it and I have not read it. The 1984 edition is public because it sits in the Internet Archive's collection of standards incorporated into law, and that is the one I read. If the 2023 edition changed these numbers, I would not know. WHAT THE STANDARD SAYS IT IS FOR "These test methods cover door assemblies of various materials and types of construction for use in wall openings to deter unwanted intruders." It applies primarily to entry doors in single and multi-family residential housing. THE NUMBER THAT NEVER MOVES Static load on the bolt: 670 N (150 lbf). At Grade 10. And at Grade 20. And at Grade 30. And at Grade 40, the top grade in the standard. It is the same figure every time. WHAT DOES GET GRADED UP Jamb and wall stiffness: 6,000 N at Grade 10, 8,000 N at Grade 20, 16,000 N at Grade 30, and 22,000 N at Grade 40. At the top grade the frame is tested at roughly 33 times the load on the bolt. THE ONLY FASTENER RULE IT STATES "Door assemblies and components shall incorporate no screw, bolt, nail, staple, or other mechanical fastener that is accessible from the outside, and whose removal would permit entry by disassembly." Nothing about screw length. It is about whether the door can be taken apart from the wrong side. THE THREE-INCH SCREW, AND THE QUALIFIER NOBODY GIVES YOU It is real code text: "The strike plate shall be attached to the door jamb with four screws not less than three inches (3") in length, and penetrating at least one inch (1") into the nearest stud." That is Kansas City Building Code 2018, Section 430.4.3 - a local amendment. It is not in the unamended national model code. True in Kansas City. Unknown in your city until you look. WHAT I COULD NOT FIND A published test measuring how much difference the long screw makes. I wanted two identical door assemblies, short screws in one and long screws in the other, pushed until they failed, with both numbers written down. If it exists I could not get to it. So this episode states no multiple, no percentage and no pounds figure for that upgrade, and neither should anyone else. THE STANDARD'S OWN LIMIT "These test methods are not intended to provide a measure of resistance for a door assembly subjected to attacks by skilled burglars on high payoff targets." It is a test for unskilled and low-skilled break-ins and it says so itself. AND ONE ODD THING F476 was withdrawn in 2011 and only reinstated with the 2023 edition. For roughly a decade there was no current American consensus test for residential door security to find. I could not find a stated reason for the withdrawal, so I am not offering one. WHAT THIS EPISODE DOES NOT COVER Confrontation, weapons, or what force the law allows - that varies by state. No method of attack is described. If you rent, a strike plate is your landlord's decision. If your city has a security ordinance, read yours, not Kansas City's.