A Detention Lock Can't Tell You It's Unlocked: Exploring a Southern Steel 10120AM Down to Its Two Micro Switches

A field story from a jamb-mounted motor deadlatch with its factory plug cut off at the wires: two SPDT micro switches, a cam that only counts rotations, a datasheet that reads like it's hiding something until you realize it's the most honest document in the box, and a door that settles back locked every time we tried to convince it otherwise.

The setup

The job: a detention-grade deadlatch, jamb mounted, on a door that needs remote unlock from a controller. The lock is a Southern Steel 10120AM Motor-Operated Electro-Mechanical Deadlatch. Southern Steel is Southern Folger now, and the 10120AM is their medium/max-security swinging-door electric lock: 10 gauge steel case and cover, 19 pounds of it, 1 inch of stainless latchbolt throw, UL listed for three-hour fire doors, ASTM F1577 Grade 1 impact rated, fail secure. You order it keyed one side (10120AM-1) or both sides (10120AM-2), and you order the motor as either 115 VAC or 24 VDC. This one is the 24 VDC version.

The want list was simple: a relay from the door controller pulses it, the bolt retracts, the door gets used, and when the door closes again the lock locks itself. No door contact wired back to the controller, no relock logic on the controller side, none of that. The lock does its own relocking, locally, off a door position switch.

The problem: the factory field connector, a 6-pin Molex (part number 03-09-1064), had been cut off the pigtail at some point in this lock's life. No plug, no pigtail labels, no wiring diagram anywhere. The official Southern Steel datasheet is a spec sheet, and a good one, but it carries exactly zero wiring. "For more information please call 210-533-1231."

The way out was the service manual for a different brand. The Airteq 9924-SSA (24 VDC) and 9912M-SSA (120 VAC) are drop-in replacements for the 10120AM with the same internal topology, and the 9900-series service manual includes real wiring diagrams. Between that, the datasheet's function description, and a meter, everything in this article got figured out on a bench and then confirmed on the door.

If you're doing the same job, the part numbers that matter: 76763785 is the factory wire harness, 76763791 is the 24 VDC motor wire harness, 76763787 is the 115 VAC one, and 96905075/76905076 are the internal micro switches. Craftmaster Hardware (800-221-3212) sells Southern Folger detention parts, and honestly, ordering the harness is cheaper than the evening you'll spend on cut wires. We did it the hard way on purpose, because the hard way is where the understanding is, and the understanding is what this article is actually about.

First: what's actually in the box

Open the cover (10 gauge, so bring a real screwdriver) and the electrical side of this thing is almost insulting in its simplicity. Two SPDT micro switches, three terminals each. One two-wire motor. That's the entire electric lock.

The two switches are not interchangeable, and figuring out which is which is job one:

Switch one sits on the motor cam. A cam on the motor's rotation path actuates its lever, so this switch reports where the motor mechanism is in its rotation. This is the switch that stops the motor.

Switch two moves with the latch mechanism. It's actuated by the latchbolt/latchback linkage, so it reports whether the latch has cycled: bolt extended (door has been opened and the bolt went back out) versus bolt retracted (the motor pulled it in). This is the switch that decides whether a door close is allowed to re-arm the lock.

Now the trap, and it's a good one. Both switches are stamped NO and NC like every micro switch you've ever met. Those stamps are defined by the switch's own actuator at rest, and the switch's actuator has no idea what state the lock is in. The NO contact on the cam switch might be the one that's closed when the lock is locked, or when it's unlocked, depending on how the cam sits at rest. Guessing off the stamps is how you wire a lock that instantly cancels its own unlock or goes deaf after one cycle.

So: meter before you wire. With the lock at locked rest, find which throw is made on each switch, and label by function, not by the stamp. On the cam switch, the throw that's made at locked rest is the LOCKED throw, the other one is the UNLOCKED throw. On the latch-cycle switch, the throw that's made with the bolt extended is the CYCLED throw. Three labels, ten minutes with a meter, and every connection after that is obvious.

The datasheet quote that explains everything

Here's the 10120AM's electric function, verbatim:

"Remote switch activates a motor which retracts the latchbolt. Latchbolt remains retracted until door is opened approximately 2", then it releases, automatically latches and deadlocks when the door is closed."

Read that slowly, because everything else follows from it. The motor retracts the latchbolt. The latchbolt remains retracted (that's the latchback, a mechanical hold) until the door opens about 2 inches. Then it releases, and when the door closes, the lock latches and deadlocks. Mechanically. On its own.

Nowhere in that sentence is there a state called "unlocked." The remote switch doesn't unlock the door. It starts a motor that pulls a bolt in and holds it out of the way. The lock is "unlocked" only in the sense that the bolt is out of the way, and the datasheet tells you outright that this condition ends when the door cycles.

The datasheet's own application note says the same thing sideways: it recommends a door position indicator switch, a door closer, and a heavy duty door pull. Why would an electric lock need a door closer? Because the unlock only means something while the door actually moves. Hold that thought.

What the cam switch actually does (the wrong turn worth taking)

Our first wiring design assumed the cam switch detected an unlocked position: unlock stroke runs, cam flips at the unlocked end, motor stops, and the lock sits there holding the bolt retracted, ready to be used. We built a two-leg scheme around that idea, one motor path gated by the locked-state contact for the unlock stroke, one gated by the unlocked-state contact for the relock stroke, self-stopping at both ends.

Then we put it on the door and watched it work, and the door corrected us.

The cam switch is a full-rotation stop, not a position sensor. It knows when the turn is done. It never knows where the bolt is.

Trigger the lock and the motor runs one complete cycle and stops at the end of the rotation. That's all the cam switch does: it opens the circuit when the rotation completes. There is no unlocked position to detect, no dwell point where the lock waits with the bolt held back waiting for somebody. The window where the door can be pulled open is the window while the mechanism is in the retracted part of its cycle, and when that window closes, the door is a door again.

The practical consequence, learned on the door with a hand on the knob: you have to pull the door right as it cycles. Or you give the door spring tension (a closer, or spring hinges) so the door does the pulling for you, which is exactly why the datasheet recommends the closer. Nobody stands at a detention door holding a trigger relay and hoping.

The good news, found on the same test: the motor turns one direction through the whole thing. Same polarity for the unlock stroke and the relock stroke, because the "relock" is just the mechanism completing its rotation. No reversing relay, no polarity tricks, no DPDT anything. Two current paths, both through the cam switch's common, both ending at the same two motor wires. The wiring design we'd built survived intact, it just means something different than we thought: the "self-stop at both ends" isn't two parked positions, it's one full rotation with the cam opening the circuit at the end.

Run the timeline with me

Four situations worth being able to answer before you put a lock like this on a door, because they're the ones people ask about at 7 PM:

Pulse fires, nobody ever opens the door. The motor runs its cycle, the bolt retracts, and with the latchback holding it, the datasheet says the door stays openable until somebody cycles it once. What we watched on this unit, though, is the lock spending itself: with the relay held longer than the stroke, the lock runs again, a couple of full cycles, and settles back locked. An unused unlock doesn't leave the door unlocked forever, it evaporates. (More on why that matters in the latchback section.)

Door closes before the pulse. Nothing happens. The cam switch has the relock path open at the locked end, so a door close with the lock already locked is a no-op. This is the case that makes the cam switch earn its keep.

Pulse fires while the door is already open. After the door position switch went into the common feed, this one stopped being possible: the lock can't trigger at all unless the door is shut. Before that wiring went in, it could, and an unlock stroke with the door standing open is a weird thing to watch with nobody touching the door.

Door closes after the latch has cycled. The normal case. Door was unlocked, opened at least 2 inches, bolt went back out, latch-cycle switch flipped, door closes, door position switch closes, and the relock path is made. The motor finishes its rotation, the cam opens the circuit, the lock is latched and deadlocked. The controller never heard about any of it.

What the door position switch actually buys

Honest accounting, because the door position switch got wired through the lock and it's worth knowing exactly what it does and doesn't do.

It buys two things. One: with a longer relay hold time, it stops the double-cycling, the lock running its rotation over again because the trigger was still there when the cam re-armed. Two: it prevents triggering with the door already open. Both gains are real, and both are why it's in the wiring.

It does not buy a lingering "unlocked on demand" state. Nothing can, because that state doesn't exist in this mechanism. The unlock is a cycle, not a position, and the door position switch can't change what the cam switch fundamentally is.

And the knob that actually controls behavior is the relay output time. Held long: the lock spends a couple of cycles on an unused unlock. Trimmed to about a second, just enough to complete one rotation: one pulse, one cycle, done. That's not a defect, that's the whole personality of the lock in one number.

The no-latchback option, explained

The mechanical hold that keeps the bolt retracted until the door cycles is called the latchback, and it's the default behavior of the 10120AM: retract the bolt electrically, and the bolt stays out of the way until the door opens about 2 inches, then it releases and the lock re-latches on close. Practical meaning: a remote unlock lingers. Pulse the lock, and the door is openable until somebody actually uses it and a full door cycle happens. For "unlock the door for whoever's coming" over the next ten minutes, that's exactly what you want.

Southern Steel sells the opposite as a factory option: NL, No Latchback. Per the datasheet, it "allows latchbolt to extend without opening door," and it's "normally used with" the D function. With NL, the bolt goes back out on its own as soon as the motor stops. Practical meaning: a remote unlock self-cancels. The lock re-locks itself within moments, and the unlock pulse is only useful in the instant somebody's pulling the door. For a door that should spend every unlock immediately and never sit with the bolt held back, NL is the option code you want on the order form.

Two things to know about NL. One, it's an order-time factory option, not a field mod: if a job needs it, it gets specified when the lock is ordered, and the behavior should be confirmed with Southern Folger (210-533-1231) before anyone depends on it. Two, don't confuse it with the D function (Half Cycle Holdback), which sounds related and isn't: D is for a remote maintained two-position switch that holds the bolt retracted on purpose, and there's no UL fire rating with it. NL is about the latchback doing or not doing its thing after an ordinary momentary pulse. Different option, different job.

And a confession that belongs in the record: we never determined whether this particular unit carries NL from the factory. The behavior we watched with nobody pulling, the lock spending its unlock and cycling back to locked, looks like NL behavior, and the datasheet's standard-function text says the latchback should hold the bolt until a door cycle. One of those is not like the other, and the honest answer is that the option code on this unit's build sheet is unknown. If lingering unlocks matter to a design, that's the phone call to make before anything gets counted on. What's not unknown: with the relay trimmed to about a second, the lock does one cycle per pulse, and the door ends up exactly as locked as it started.

The wiring, as connected

No diagram for this one, just the instructions, in order. Every step assumes you've metered first and made the three functional labels from earlier (LOCKED, UNLOCKED, CYCLED).

1. Meter first. Lock at locked rest, door closed: find the made throw on the cam switch, tag it LOCKED, tag the other UNLOCKED. Bolt extended: find the made throw on the latch-cycle switch, tag it CYCLED. Ten minutes, saves a re-wire.
2. Check the motor label before anything gets energized. This unit is the 24 VDC build: about 3.5 amps inrush, 0.32 amps running. If you find 115 VAC on the label (2.4 amps inrush, 1.8 amps running), stop and redesign the feed, because 24 volts into 115 volt motor circuitry is a paperweight factory.
3. Always-on +24 VDC goes to the COM of the door position switch. The NO side of the door position switch, the contact that closes when the door is shut, is now the feed for both legs.
4. Unlock leg: door position switch NO goes to the unlock relay's dry contact, and the other side of that contact goes to the cam switch's LOCKED throw. The relay is the controller's momentary unlock output, and its contact carries motor current, so it's rated at 5 amps minimum.
5. Relock leg: door position switch NO also goes to the latch-cycle switch's CYCLED throw, and the latch-cycle switch's COM goes to the cam switch's UNLOCKED throw.
6. Cam switch COM goes to one motor lead. The other motor lead goes to 0V, the supply common, straight through the polyswitch if the harness has one in that lead.
7. Tape and label the unused throws. The NC side of the door position switch is where a monitoring loop would go on a fancier build, series with the lock's status contacts back to the controller, so a failed door position switch can't take the remote unlock down silently. This build's controller had no door input to feed, so that loop stayed unused, but the terminals are there.
8. Two wiring notes that came out of the bench: the relay and the door position switch contacts both carry motor current, so an alarm-grade door position switch with a 0.5 amp contact rating is marginal at 3.5 amps of inrush; interpose a relay if that's what's in the parts drawer. And the trim: the unlock relay output got set to about one second, enough to complete one rotation, not enough to re-trigger anything.

That's the whole circuit. Two SPDT switches, a two-wire motor, one always-hot supply, one door position switch, one momentary relay. Every path has a self-stop in it, the controller's only job is one clean pulse, and nothing anywhere depends on the controller knowing anything about the door.

How we left it

Here's the ending, and it's an OK one. The door got left in this configuration: door position switch in the common feed to both legs, so the lock only triggers with the door shut, and the unlock relay output trimmed to about a second, so a pulse is one cycle. Nobody standing by the door required, no double-cycling, no triggering on an open door, and every door cycle ends with the lock latched and deadlocked on its own.

And we're fine with where it landed, because the thing the lock "can't do" wasn't on this project's list anyway. A door controller that re-locks by sensing door input would need the door state wired back to it, and this build deliberately keeps the door position switch local to the lock. The controller on this project couldn't have done relock-by-door-input if we'd asked it to. The lock's limits cost this job nothing. It relocks on door close exactly as intended, and the remote side is fire and forget.

The real deliverable of the week wasn't the wiring, it was the map: what each of those two micro switches actually reports, where the cam switch's authority ends, what the door position switch adds, and where the one-second relay trim lives in all of it. If a future job needs a lingering unlock that auto-cancels, that's an NL on the order form, not a field fix, and now we know to ask before the lock ships.

What I'd tell someone else

• Meter before you wire, and label the contacts by what they do, not by the NO/NC stamp on the switch body. The stamps describe the switch's own rest position, which knows nothing about the lock.
• A cam switch is an end-of-travel stop, not a position sensor. Once you accept that, the wiring gets simple, and the two-leg scheme falls out naturally.
• The door position switch buys exactly two things: no double-cycle on a long relay hold, no trigger on an open door. Wire it for those, not for features the mechanism can't offer.
• The relay output time is the unlock's hold time. Long pulse, extra cycles. About a second, one cycle per pulse. Trim it and the behavior stops being a surprise.
• Latchback is the difference between an unlock that lingers until the door cycles and an unlock that self-cancels (the NL option). Know which one you have before you promise either behavior, and if it matters, make the call to Southern Folger before the lock ships.
• The 115 VAC build has the same wiring topology, but the ratings move with the motor, and the wrong-voltage mistake is fatal to the motor circuitry. Check the label first.
• And the boring one that saves the most time: order the factory harness (76763785, plus 76763791 for 24 VDC) even if you plan to do your own wiring. Having the pigtail with its own colors and pinout turns the whole exercise from forensics into plumbing.

Test environment: Southern Steel 10120AM Motor-Operated Electro-Mechanical Deadlatch, 24 VDC motor variant (Southern Folger / Southern Steel; Airteq 9924-SSA / 9912M-SSA used as drop-in service reference, 9900-series service manual "9924 2-15"), always-on 24 VDC supply, controller unlock dry contact, external door position switch mounted at the frame, metering done at locked rest and bolt-extended states, bench work October 6-9, 2026 with the door test on the 9th. Part numbers from the Craftmaster Folger Express parts list, pages 152-155. The lock stayed in service throughout: all metering at rest states with the motor circuit de-energized, and every wiring change verified before the door went back to work.