This post was printed in the December 2010 issue of Doors and Hardware
[Click here to download the reprint of this article.]
Someone brought this to my attention yesterday and I thought there had to be a mistake. I was handed a slide from a recent presentation on NFPA 72 – National Fire Alarm and Signaling Code, which said that all exit doors with access control had to unlock upon actuation of the fire alarm. I checked the 2007 edition and sure enough, there it was:
NFPA 72 – 2007:
“6.16.7.1 Any device or system intended to actuate the locking or unlocking of exits shall be connected to the fire alarm system serving the protected premises.
6.16.7.2 All exits connected in accordance with 6.16.7.1 shall unlock upon receipt of any fire alarm signal by means of the fire alarm system serving the protected premises.”
This could easily be interpreted to mean that every exit door with any type of access control (card reader, keypad, etc.) would have to unlock upon actuation of the fire alarm. Obviously, this would present a problem from a security standpoint if “unlock” was interpreted as unlocking for ingress as well as egress. Exit doors with mechanical locks (operated by a key) aren’t required to unlock upon fire alarm and allow ingress, and most exit doors have to allow egress all the time – not just during a fire alarm. Not to mention that many locks used in access control systems allow free egress all the time.
I checked with NFPA and I was told that “under this provision every exit that has access control would require unlocking by the fire alarm system, unless the AHJ or other codes state otherwise.”
That’s not good.
Luckily, a code change proposal was submitted, and the 2010 edition of NFPA 72 contains revised language which clarifies the intent:
NFPA 72 – 2010:
“21.9 Electrically Locked Doors.
21.9.1 Any device or system intended to electrically lock a required means of egress door in the direction of egress shall be connected to the fire alarm system serving the protected premises.
21.9.2* Electrically locked doors in a required means of egress shall unlock in the direction of egress as prescribed by other laws, codes, and governing standards.”
There’s additional information in Annex A of NFPA 72 – 2010:
“A21.9.2 Doors are commonly locked for various security reasons. Though doors are permitted to be locked to prevent ingress, doors are generally not permitted to be locked to restrict egress unless specifically permitted by governing laws, codes, and standards. Examples of special locking arrangements include delayed-egress locking and access control locking. Approved locking requirements by governing laws, codes, and standards can vary extensively. For example, some might require all fire alarm initiating devices to immediately unlock electrically locked egress doors, while others might permit such doors to remain locked when a single manual fire alarm box is activated. Some codes might also permit electrically locked doors to remain locked when a single smoke detector has activated. These allowances are typically permitted only in sprinklered buildings and are generally used as additional safeguards to counter efforts to breach security, without compromising occupant safety.”
The 2010 edition also clarifies the use of battery back-up on electrically-locked doors:
“21.9.3* For all means of egress doors connected in accordance with 21.9.1, and where batteries are used in accordance with 10.5.6.1.1.(1) as the secondary power supply, the batteries shall not be utilized to maintain these doors in the locked condition, unless the fire alarm control unit is arranged with circuitry and sufficient secondary power to ensure the exits will unlock within 10 minutes of loss of primary power.
21.9.4 Locks powered by independent power supplies dedicated to lock power and access control functions, and that unlock upon loss of power, shall not be required to comply with 21.9.3.
21.9.5 If means of egress doors are unlocked by the fire alarm system, the unlocking function shall occur prior to, or concurrent with, activation of any public-mode notification appliances in the area(s) served by the normally locked means of egress doors.
21.9.6 All doors that are required to be unlocked by the fire alarm system in accordance with 21.9.1 shall remain unlocked until the fire alarm condition is manually reset.”
And from Annex A:
“A.21.9.3 A problem could exist when batteries are used as a secondary power source if a fire alarm control unit having 24 hours of standby operating power were to lose primary power and be operated for more than 24 hours from the secondary power source (batteries). It is possible that sufficient voltage would be available to keep the doors locked, but not enough voltage would be available to operate the fire alarm control unit to release the locks.”
In summary, the language in the 2010 edition of NFPA 72 helps to clarify the requirements for the interaction of the fire alarm system and doors with access control. The revised language is specific to unlocking “in the direction of egress,” so this section does not require the doors to allow free access/ingress through the door. This does NOT mean that it’s acceptable to lock doors in the direction of egress and release them upon fire alarm – it’s simply addressing locking scenarios allowed by other codes. For example, delayed egress locks are allowed to prevent egress in certain occupancies for 15 seconds as long as all of the requirements are met. NFPA 72-2010 states that those devices have to be connected to the fire alarm system and shall unlock in the direction of egress. That’s much clearer than the previous edition.
There will still be lingering questions about the use of battery back-up, but NFPA 72-2010 does help explain the intent. When a secondary power supply uses batteries, the batteries can’t control the locks unless the fire alarm will unlock the locks within 10 minutes after the loss of primary power. If we think about delayed egress as an example again, in theory the delayed egress lock would still function correctly during that 10-minute period, so egress is not compromised. Although the 2010 edition of NFPA 72 has not yet been adopted in many jurisdictions, these changes may help to illustrate the code’s intent to an AHJ or others.
This post was originally created on 4/28/10 and was printed in the December 2010 edition of Doors & Hardware magazine.
You need to login or register to bookmark/favorite this content.
Lori,
Thank you for this article. I sent it to an architect of mine who was dealing with an AHJ that was enforcing the unlock on fire alarm rule at a large new hospital project. They were meeting with the AHJ the next day, took the article, and got the code official to change his mind and go by the “intent of the code”. It made me look like a star!
Glad I could help, Justin. There’s another post which will actually be published in next month’s magazine, that deals with hospital lock-down. It might help with your project: http://www.idighardware.com/2010/07/i-2-special-egress-locks/
Keep in mind that when a hospital wants to do something that’s allowed by a newer code but is not allowed by the 2000 edition of NFPA 101, they must apply for a traditional equivalency from the Joint Commission and a waiver from CMS.
Glad to see the additional clarifications being laid out. Here in the NW the general interpretation has always been “in the path of egress” for unlocking. We have had to use more delayed egress scenarios in the transitions from parking structures to buildings because there are non-alarm events that require egress as well. Most times a separate submittal and design narrative are required for these specific situations so the AHJ can understand the intent and execution. Stairwells create the most issues as you balance between security and safety, several times I have had to rely on the AHJ to help enforce my logic in regards to egress at all times not just alarm events. Building owners and contractors worried about cost just want a deadbolt to solve everything, which is never acceptable.
Keep up the good work, Travis! Sometimes we have to be the bad guy when we tell the owner what’s required by code, but they will appreciate you eventually. 🙂
Is there a ‘listing’ requirement, or a ‘code’ or ‘approval’ requirement for the system that controls the electric locks on the doors (for release during a life safety fire incident). In other words, can a non-life safety system (like an access control system) release the fire egress doors during a life safety emergency (assuming that it gets its signals from a life safety system that detects the fire and tells it to release the doors). if an access control system can actually be used to release the doors, then what ‘listing’, or ‘approvals’ does it have to have to allow it to be used that way?
Hi Bob –
The codes aren’t specific about whether the access control system can be used to release a lock during a life safety emergency – the only place I can think of that it’s mentioned is on an access-controlled egress door where the push button has to unlock the door for 30 seconds, “independent of the access control system.” If we take a stair door, for example, I think the intent of the code is that the stair door is unlocked directly from a signal switch at the fire command center (or via the fire alarm system depending on which code). I don’t think most AHJs would be comfortable with the access control system being involved, but I posted the question on their discussion forum so we can find out straight from the AHJs: http://www.thebuildingcodeforum.com/forum/door-hardware-forum-sponsored-ingersoll-rand-security-technologies/10044-access-control-system-life-safety.html#post93963
– Lori
Does NFPA 2010 then not require Fire Alarm tie in on Mag Locked doors if the REX button and PIR are providing free egress?
Hi Jason –
Per the IBC and NFPA 101, there are 2 ways of handling mag-locks. If it has a sensor and push button to release the mag-lock, it is an Access Controlled Egress Door, and it is also allowed to unlock upon fire alarm and power failure. If the mag-lock is released by an RX switch in door-mounted hardware, it is an Electromagnetically Locked Egress Door and it does not have to be unlocked by actuation of the fire alarm but does have to unlock upon power failure.
– Lori
Thanks Lori. Would a touch sense bar mounted on the door qualify as an RX switch in door-mounted hardware therefore not need to be tied in with the fire alarm?
Yes, as long as the door does not require panic hardware. A touch sense bar does not qualify as panic hardware because it can’t be certified to UL 305. You should also make sure that if power is lost to the touch sense bar, the lock will unlock as required by code.
We were recently told we will have to test that all magnetic control doors release upon activation of the fire alarm semi-annually even though they are not locked in the path of egress (because a sensor releases the mag when you walk up to it). Does this sound correct? That’s a lot of doors to test. These doors are used every day so one could say the sensors are tested everyday, right? Why or how would the doors ever be locked in the path of egress that we need to test?
Hi David –
Testing whether the doors unlock upon fire alarm activation would be different from the everyday operation using the sensor. It’s possible that the door could unlock properly via the sensor, but may not be correctly set up to unlock upon fire alarm. I wrote this post a while ago about testing for delayed egress locks, and the same would apply to testing of mag-locks: http://idighardware.com/2012/08/qa-testing-delayed-egress-locks/. In the post, I reported that NFPA 72-2010 required annual testing to make sure the locks would release upon fire alarm.
I checked the 2016 edition of NFPA 72 to see if there was a change from annual testing to semi-annual. Table 14.4.3.2 – Testing – Item 24, says emergency control functions must be tested upon acceptance, and annually. I’m not an expert on NFPA 72, so I will submit this question to NFPA and make sure I’m interpreting it correctly. You may have a local code requirement that exceeds the requirements of NFPA 72.
– Lori
Lori,
Thanks for the reply. My mistake, it was annual testing. Do you know if you are required to test 3 individual releasing mechanisms (fire alarm release, sensor, and push to exit button) or simply that the doors release upon activation of the fire alarm?
For NFPA 72, it appears to just be the fire alarm release, but NFPA 101 requires annual testing of doors with special locking arrangements in some types of facilities (http://idighardware.com/2014/02/decoded-annual-inspection-of-egress-doors/) which would require testing the normal operation (release by sensor, push button, door mounted hardware, etc.).
Lori,
What about battery backup on a fail secure egress door? I have doors that are the Von Duprin 98/99 QEL series. The spec calls for battery backup for these doors. Is this compliant?
Hi Michael –
Because the QEL devices provide free egress regardless of the power condition, I don’t see any problem with having battery backup to the QEL, as long as the AHJ is willing to take the time to understand how the system works.
– Lori