Manipulation of 1920's Mosler vault

[ericrwnage]

I am attempting to open a vault built into a house constructed in the mid-1920s in Washington State, USA.

The door is a nondescript olive green but the handle gives it away as a Mosler safe; according to the stamped serial number it's from about 1910. I found a nearly identical vault door from a former US Post Office (1883-1917) in Ohio, which clearly identifies it as a Mosler safe door. (The below image has been flipped horizontally with a perspective correction applied.)



The dial is of S&G make similar to the D017-007 (but larger). I have been assuming the safe was drilled open at some point in the past 100 years and the dial, if not the lock, was replaced as part of the repair. I couldn't put a date range on the replacement.

I built an autodialer that assumed 3 wheels, a tolerance of +/- 0.5, and a final clockwise rotation to 90 to engage the deadbolt, as with S&G locks. This ran through all combinations for 2 days but did not find a solution.

I also spent about an hour attempting to manipulate it, with no prior experience, but I couldn't easily identify the contact points. I thought I found one at 31, which seems uncommon (though I note it is 1/4 turn after 6, a contact point on my S&G, as if the spline key were in the wrong slot?).

I have been assuming the lock mechanism was replaced with an S&G along with the dial, but given the design of the wheel pack in a Mosler safe, this might not actually be possible. Is it likely that the lock mechanism is original, just with a newer S&G dial? If so, what would the Mosler lock model be?

I have about a week left to get this open before the owner has it drilled, and I have only occasional access to it. I can tweak my autodialer algorithm (e.g., to assume 4 positions and multiples of 5, to try a pattern other than LRL, or to change the bolt retraction behavior), or continue to try to manipulate it. I'm hoping for advice on the best use of time and any tips from others who have encountered similar vaults before. Thanks!

[billdeserthills]

Have you determined there to only be three wheels in your safe's lock?

Many vaults use four wheel safe locks to add a bit more security

Last vault door I found in a house was easily opened. I noticed someone had marked the dial in three places & dialed the few possibilities until it opened

[98G]

Spin the dial counterclockwise at least 5 full rotations and stop on 75. Now turn it clockwise and see how many wheels you can feel picking up.

Every time you get to 75 going clockwise it should get just a little harder to turn until you've picked up all the wheels.

On your autodialer program, you may want to have it turn to open until it reaches a hard stop instead of stopping where you think it should be.

[MartinHewitt]

Hi,

The door is not a high security bank vault door. It falls into the plate vault / fire vault door category. So your assumption of 3 wheels is probably correct, but counting and knowing it is better.

The two door look similar (the photos are rather small), but the distance between dial and handle are different pointing to a different bolt work. Also most S&G dials do not work on most Mosler locks. My guess is, that this is the genuine dial of a more vintage S&G lock, e.g. an M6730 which was used on HHM vault doors.

Maybe your dialler did dial the correct number, but the lever was to lazy to fall due to petrified grease. You might want to try applying a rubber mallet over the dial, esp. with the dial positioned around 0 to loosen up the lever. Maybe you can then feel contact points. This may then be enough for manipulation with possibly later more mallet work to get the lever to drop. If you get clear contact points and clear sounds of the lever moving when wiggling the dial it might be also ok to re-try the autodialler. But be careful. An autodialler puts a lot of stress on the lock.

Of course 90 might have not been far enough to get to stop it.

[ericrwnage]

To which position should I turn the dial to retract the deadbolt? On my S&G (of modern vintage) the deadbolt retracts at 95 I think; my autodialer dials to 90 each time and tries to detect a stall.

If this is a 1910-1920s lock mechanism, would trying every other number or every third number or something be sufficient? What is a reasonable RPM for an autodialer like the ITL-2000, anyway?

I did find a recent Reddit post with a photo of an HHM safe installed by Purcell Safe Co. in Washington. It appears to have the same S&G dial and nearly the same hinges. I am reasonably sure this is a Mosler handle, though?

[Squelchtone]

Yes, it looks like a Mosler handle, and your dial is not original to that fire door. A better picture would help though as the one provided is too distant to see detail.

Does the safe dial say Rochester or Nicholasville on it? It looks like a 1960's to 1980's S&G dial. Mosler would not put an S&G lock on their doors, they would use a Mosler lock. Your dial could be newer with a Mosler lock inside, or the entire lock was changed out for an S&G 6700 and the S&G dial.

every 2 numbers in medium speed mode would be fine on an ITL dialer.

Your door may have originally had a Mosler model 5-H lock on it which does come in 4 number combinations, but some spinning of the dial to count wheels as others have mentioned is a better idea than us guessing.

Squelchtone

[MartinHewitt]

Do you hear the rumbling of gears when you turn the dial?

[MartinHewitt]

Regarding the RPMs: If the dialler rotates to fast the wheels can overshoot. How much to fast is depends on the lock. To be exactly it is mostly the ramp down where an to abrupt halt is the issue and also the clashing of the pin at wheel pickup, which might snap them. A flat ramp will also limit the maximum speed, because there are lot of short movements that never get to the end of the ramp. The turning resistance you feel when you turn the dial is helpful information. My home safe is very very light there and I have to slow down manual dialling, to get it open. If it is more like a modern lock, then I would target the ITL2000 speed with a max. runtime of ... 20 hours?

[ericrwnage]

Here's a closeup of the dial, which says Nicholasville, KY so I believe that dates the dial to after 1974.



I'll be onsite tonight and will try to determine the number of wheels and presence of gears. It does have much higher resistance than my practice lock which spins very freely.

The point about the dialer deceleration is interesting. I've just been using a constant RPM with very abrupt starting and stopping (or direction switching). The positioning is highly accurate but I suppose a wheel might be flung a little when picking it up to move it just a small step, or continue turning even after the motor switches direction. But slowing down before making contact with the pin would greatly expand the runtime.

I'd love to study the ITL's speed curves if someone has one. I'm not sure the right method to do so, maybe a high speed camera and some traceable points on the chuck?

[ericrwnage]

It's pretty hard to tell but I think I could count 4 wheels. The dial is a little askew so there is some extra friction.

Since I cannot clearly identify the internal sounds, nor could I identify the contact points, I decided to move ahead with the guess that this is a Mosler 5-H lock. (There's a nearby house built by the same architect who could answer this for me, but alas I don't think that will be possible.)

I've got the autodialer testing 4 number combinations of multiples of 5, at a relatively low speed, and pushing the status to me over the Internet so I can keep tabs on it. It turns left to 20 at the end of the sequence to try to engage the deadbolt; this is cutting it a little close with the motor's stall detection setting so it may get the correct combination but not notice.

[ericrwnage]

No dice with round two of the dialer, either. I think unfortunately this is probably getting drilled open this week.

I'm still curious to study a commercial dialer if someone has one and wants to send me a PM.

[Squelchtone]

Have you tried dialing some test combinations? Last one I opened that was 4 digits was a Mosler 5-H with a combo of like 19 20 19 80, just dialing 20 20 20 80 opened it as well. 10 20 30 40, 20 40 60 80, 50 50 50 50, and other similar combinations may lead to an easy open. when was the town incorporated? when was the school building built? there are some memorable historical or pop culture years that people may choose such as 1911, 1945, 1969, 1984

Dialing order RL-4R-3L-2R-1L to stop at 16

In case the lock inside is an S&G 6700 3 wheel, 10 20 30, 25 50 75, 20 40 60, 20 60 40, 20 80 40, 50 50 50, 50 25 50, 25 50 25 are all good combinations to try.

Good luck,
Squelchtone

[ericrwnage]

The vault was successfully manipulated open by Kai Wolk of Pro Safe Services Inc., sparing it from the drill.

The lock was indeed a 4-wheel Mosler 5-H lock with a retrofit S&G dial. Due to the sloppiness of the dial, Kai says it was very difficult to get accurate feedback from the mechanism, but after a few hours he was confident he had the contact point pinned down, and not too long after he had it open. Two of the numbers were the same, reducing the difficulty to solving 3 wheels.

Why didn't the autodialer solve it? To retract the bolt the lock needed to be dialed to 80, not 20; the retrofit dial caused everything to be offset from where it "should've" been. There was also the possibility that the "petrified grease" issue was causing the wheels to stick to one another even when the dialing direction was reversed.

There were some concerns the autodialer might have worn down the fence, or the pins, or caused the spline key to shear off. None of those things appeared to be the case.

[Squelchtone]

That's excellent news! Very happy it got opened without having to drill it!

Thanks for coming back to share the update

Squelchtone

[Kaesekopf]

Excellent! I wondered what happened