Practicing manipulation with my new audio amplifier

[femurat]

After a few years, I'm trying to get my touch back. Not being able to open a vault door gave me the motivation to train again. So I'm spending at least one hour a day playing with my old safe locks.
I also ordered an audio amplifier and a contact microphone to play with.
Here are some manipulations that went smoothly









There are some more, that are not interesting enough to end up here.

And then, with one of these locks set on an arbitrary combo by a person not interested in locks, I hit a wall. The combo may be inserted badly, unusual or even "illegal" (forbidden zone, fly safety eccetera), but I should be able to open it anyway.
Here are the graphs I made till now. Reminds me my latest failure. Looks like there's something that I still don't get.



With graph #1 I got a clear gate signature around 83 and another one, small but still interesting, around 68.
Then, due to an unusual high low test result at 63, I waste time in graph #2. I only discover that it's useless to graph wheel #3 between 10 and 45 because it's clearly too high.
Time to get back in track with graph #3, but I got nothing more than a low area between 55 and 80.



At that point, I decided to test wheel #3 around that promising gate signature and chose 82 as its center. Unfortunately, it's not a gate, otherwise it wouldn't have masked the other two wheels between 22 and 70.
There's still a small interesting point at 83.



Any hint on how to continue?

Thanks

[Alexander Mundy]

Gut feeling.
I'd do 45, AR, 75.
If nothing shows, AL, 45, 75.
Neither 45 or 75 are based on gates but on the sine waves to try to uncover a gate on 1 or 2.

Is that a damn Lagard you are graphing?

[femurat]

The lock is a R6700 SERIES SARGENT GREENLEAF.

Your intuition is quite impressive. And almost correct...





Here are the new graphs I made. Graph #5 following Martin suggestions, that wrote me on another forum, made me find a clear gate signature on wheel #2 at L24. Then in graph #6 I tried to find a low spot on wheel #3 but didn't find anything important.
So I went on with graph #7. I should have tested my very wide gate at 83 but almost forgot about it. When I completed the part of the graph that could have a gate, from 70 going Right to 20, I examined the previous graph and noticed the interesting spot at 83. As soon as I tested the combo, I was still turning and the lock got stuck. Turns out the last number was 7. 83-24-7.

This manipulation taught me to always trust a gate when I see it.

Now it's time to get the combo changed and try again.

Cheers

[Alexander Mundy]

Based on graphs 6 & 7 I wasn't too far off of where the low spots are on 1 & 3. I totally missed on what lock but it had tale tell signs of being a Lagard to me.

[femurat]

In fact you almost guessed them! Impressive.

The lock was actually shadowing like a Lagard.

Cheers

[femurat]

Tonight I think I got my personal record: 34 minutes to open this same lock. This time, ironically, I didn't use the audio amplifier.

The main reason of this speedy opening is luck, because I found a gate right at the beginning of my graphs: The first one at 16, saving me the time and effort to test the rest of the numbers. And the second number almost halfway to the end, at 55.
Also, I decided to brute force wheel #2, without graphing, every 2 increments. This is a risky decision, because if one of the other wheels number is wrong, I don't get the gate position on the tested wheel. But it reduces the time needed drastically.

Another reason is that, due to the less work required, I decided to assign the gate found at 16 on graph #1 at wheel #1, and to test wheel #3. I would have tested #2 and #3 together, but followed the suggestion to isolate the wheel. Another lucky decision.
I haven't changed the combo yet, so I could test #2 and #3 together to see if the gate would have showed up anyway.
To avoid the risk of missing the gate due to its offset, I dialed all the wheels to the Left at 16, then moved only wheel #3 Right to 16, and then started the test going Left. This way I got the gate position in the same direction I have to dial it.

Another thing that sped up the process was the decision to write down the dots on the graph only when they changed, and not every time I tested a number. This prevented me to draw them slightly over or under the line, loosing the difference between shy and bold readings. But I don't need such precision because the gate reveal himself with more than 0.1 increment in this lock.

Feels good!

Cheers