Pouring Cold Water on the Fail-Safe Fallacy

Replacing dogmatism with rationalism in locking

Today I am going to go out of my way to alienate fire engineers.

Don’t worry, though. I’m also going to be contradicting some widely cherished traditions of today’s security industry - just in case anybody imagined I’d had enough of panning my peers…

Let’s get this out of the way for a start. Magnetic locks do not lock doors, they just make them a little more difficult to open.

And I’m not just playing with definitions here. Exceed the holding force of a magnetic lock and the door opens, right? As long as the door furniture and fixings you’ve used to attach the lock and armature plate are all reasonably robust, nothing gets broken when you do that. Close the door after you’ve opened it and we’re back to step one again, just like nothing happened.

The holding force of the electromagnet is dependent upon a bunch of factors, one of which (perhaps the most important one) is the amount of electrical energy it is being fed, which in turn is going to depend upon the power supply you’re using, the distance between the door and that supply, and the cabling you’ve used to make the connection. You might not take a whole lot of notice of little things like the way you make the connections between the cores of that cable and each switching or connecting element along the way, or how carefully you strip the wire to avoid damaging or removing strands, but these are also factors when you’re attempting to remain within the limits set by the lock manufacturer when they specified what the holding force of the magnet is going to be.

There are a lot of mechanical factors too that can significantly reduce or even disable the magnet’s ability to cling to the door. Dirt and corrosion between the magnet and the armature plate make a difference. Misalignments of the armature plate make a difference.

I’ve had a situations where somebody applied a little piece of clear adhesive tape to the magnet. Everything looks fine, and the door even feels as if it’s locked, but give it a tug and it just pops open - in the same sort of way that a lock that’s got a volt drop problem will, even though this is not the case with the sticky-tape and the poor engineer who’s trying to fix the puzzlingly frustrating problem.

Conversely, there was an incident (on a Friday night, of course) when some bright spark decided to squirt some superglue on the armature plate of the main entrance door to a block of apartments. The mag lock was nicely installed - there were actually two locks, flush mounted into a timber door frame - and the door was heavy and stiff. Under normal circumstances, they worked really well, because you didn’t get the sort of leverage you often do when the lock is at the transom and you’re pulling on a handle a meter away on a door that flexes…but when somebody permanently fixed the plate to the magnet using the strongest glue known to mankind (or at least the population of Southend-on-Sea) that door was not coming open.

Another time I was given a great practical demonstration of what force means in the F=mA equation by a client, who showed me how he was able to take advantage of leverage and oscillation on a flimsy metal door to easily exceed the 600lb of holding force that was supposed to be preventing people from wandering into the server room in the headquarters of Europe’s most popular MMORPGs at the time - also doing no physical damage to the lock or the door.

Take any ‘real’ lockset - something that features a mechanical element that is preventing some other mechanical element from moving unless you actually break something - and the story is quite different.

Whether it’s a solenoid lock, an electromechanical mortice lock with a little solenoid inside or a motorized lock with a moving latch, or even an electric strike (also geld secure by a solenoid), these things present a uniform level of holding force all the time that the solenoid (or the latch) are in their locked position. It doesn’t matter if the voltage drops a little, they have an operating voltage range, and all the way from the lowest voltage - where you’re barely able to get the thing to switch - all the way through to the point where you’re melting the solenoid coil with way too much power, they are locked.

Push, pull, stamp your foot and do a little dance - it makes no difference to the holding force of the lock, and the only thing you can do to get the door open is to either use a valid means of authentication and authorisation to release the thing…or break it.

By the way, did I mention that architects are also idiots when it comes to doors and locks? Today I’m scatter gunning in the alienation department.

Okay, I do understand what the issue is on this topic, and I do get how the coordination piece between one discipline’s needs to fill a hole with a nice looking piece of swingy wood, another discipline’s needs to take control of who can swing that wood, and another discipline’s needs to reduce the chances of everyone in the building burning to death if that swingy piece of wood either does or does not do the right job in a fire can be hard. But haven’t we built enough buildings yet to work out that this is a fundamental necessity when we build something?

It isn’t difficult. Just give it some priority nearer the front of a project and put somebody in charge of doors - you know somebody is going to have to do it eventually, so why not learn the lesson you’ve all failed to pay attention to on your last gazzillion construction projects, perhaps?

I know. Nobody in the archi/structural world gives a crap, and they never actually will. So if they’re going to alienate me, then I don’t see what I have to lose by alienating them.

Just saying…

Okay, so a lock is a lock. A magnet is not.

They can sometimes make it easy to ignore the coordination issue - just forget about the ironmongery that’s already on the door and slap a mag on it. How any self respecting interior designer can say ‘okay’ to a magnet on a door just boggles my mind, especially if it ends up with a Z/L bracket, and especially especially if that’s the case and somebody fails to include a cover! Shudder…

Of course, nobody (and I do mean nobody, including me) wants to see anyone going anywhere near a door on site with a chisel, but unless somebody has really f*cked up, there should rarely need to be.

There was a period in security history when the choice of electromechanical lock in the market was both limited and price prohibitive - particularly if you needed to buy a half day of carpenter’s time to install each one of the damned things, but this is not the case anymore. If you take a tiny amount of time to get to understand how locks actually work then you’re going to be able to significantly reduce the amount of coordination that’s needed to create meaningful and useful door schedules during the design phase of a construction project, whilst also not disappointing the interior designer, the fire consultant or the security person.

But people seem to be just too lazy to learn anything new when it comes to architectural accepted practices.

This is where we come to the issue with fire people.

There are countries where enough people have burned to death in buildings, enough things have been stolen from buildings and people have wasted enough money on reworking bad installation work that the sector has grown up and established some sensible practices to try to avoid or at least minimize all of this stupid negligence.

There are also countries - many, many more of them, by the looks of things - where (for whatever reason) people have not learned the appropriate lessons.

Of course, I am not trying to advocate more needless cremations, criminality or (indis)criminate rework of buildings, but it’s incredibly tiresome to have to wait while people (and I am specifically looking in the direction of the fire engineering community, because it is predominantly them that get in the way here) creep (at a tectonically snail-like pace) towards enlightenment.

But the security industry has not helped.

We’ve (stupidly) allowed the terms “fail safe” and “fail secure” to establish themselves in the general vocabulary of locks. Just the kind of thing that the code-obsessed idiots love to latch on to when they’re just looking to ease through to retirement without their own little version of Grenfell getting in the way of a lifetime achievement award at the Annual Conference of People Who Deal With Fire Safety Stuff Association.

Electrically controlled locks are power to lock or power to unlock. Whether they’re safe or whether they’ve failed has very little to do with it.

Electro-mechanical locks that are power to unlock are one of the safest things you can put on a door - even if you want to have electronic access control.

Fit a push bar or a paddle on the inside and they’re absolutely free to exit - even if the access control system fails, the power fails, the fire alarm fails and if your arms fall off. Just stumble against the inside of the door and it’s going to open - no matter what.

But no. That’s not allowed, because the lock doesn’t say “Fail Safe” on the datasheet, and the fire engineer doesn’t know (and isn’t interested in) how locks work.

“So put a power to lock (what often gets wrongly called a “fail-safe”) electromechanical lock on the door and get the ‘added bonus’ that the door will always automatically pop open when the power is off.”

And who actually thinks that’s a good idea? Wouldn’t it be nice to arrive back at the office on a Monday morning to discover that the doors have all been open since the random power outage on Friday night… Well done batteries! You clung on to things for eight entire hours before the local crack addicts came in and stole all the tea bags…

“Include a fire interface that automatically disconnects lock power when the fire detection system is triggered.”

This requirement seems to appear in just about every fire consultant’s list of demands, but I’m almost certain that I’ve never had a customer where this has been something they actually wanted. Hit any fire call point and open your bank / museum / jewellery store (delete as appropriate) to anyone who wants to walk in off the street. Hmm. Yeah.

Sure, there are workarounds for all of these - but they’re ways of working around a problem you’ve created with a regulation that doesn’t make sense if you choose an appropriate locking and escape hardware solution.

Sure, there are situations where you need re-entry or you need an access route for the fire services, but (a) there are much better ways to do that, and (b) they’re fire services…they have a big red truck full of brawny individuals with axes…they’re gonna get in!

It would be wonderful if this year I didn’t have to hear the words - “Put a mag lock on. That’s fail safe.”

Sigh.

Magnetic locking devices wouldn’t exist if there were not sets of circumstances where they’re actually needed. I most definitely do not deny this, and I am certain that in the next twelve months I will be left in situations where I have no other option than to allow the words “well I guess we’ll have to use a maglock” crawl reluctantly between my grimacing lips, but when that happens I’m also going to be making sure that we’re not allowing security to be thrown out of the window at the same time.

And I am also going to do my darndest to make sure that what gets implemented is actually convenient for the people who are going to use it. None of this standing at a door trying to work out where the heck the button is - or which button I am supposed to press…or how our less able bodied compatriots are meant to find the button, press the button, open the door and then actually get through the door in one piece…for instance.

There are plenty of ways you can screw up an electromechanical lock install too, but it’s really hard. If you bother to take the time, you’re going to be able to find a drop-in replacement electromechanical version of every mechanical lock there is, so there’s very little excuse not to go that route. Getting cables to the lock remains the biggest challenge - one that is best solved in the factory, unless you’re working with a door that (through chance or design) already has an internal cavity through which the cable can be passed. But that problem is not unsurpassable.

But don’t forget that there are an ever increasing range of wirefree online options that we now have, that do not require you to connect physically to the lock, and are also entirely compatible with whatever your chosen access control and escape ironmongery option might have been.

There are plenty of examples of bad locking hardware choices all over Linkedin (and illegal or non-compliant ones where laws and regulations exist) but it’s not going to change anything, and our fire colleagues are going to keep on ignoring logic in favour of dogma, but I just wish people would stop calling things that are neither failed nor safe “Fail Safe”.

---