One lock, fully interrogated

[rohare]

It had occurred to me that posting a journal of manipulations in the style of Oldfast would be a fun thing to do here. After having read through most of his thread I now realize he's already got that covered for the community. Instead, I thought it might be a useful resource to post a series of different manipulations of the same lock, with the same combination. The manipulations will consist of an AWL approach, an AWR approach, a 3-2-1 approach, and anything else that strikes my fancy after that.

The idea is to extract every single bit of knowledge possible from a single lock by approaching it with a variety of different methods. I am very new to manipulation, but I'm going to put myself out here and risk making silly mistakes in a public forum because I think the project will be fun and I always learn faster when I approach a new subject methodically and in great detail. For each new manipulation I will do entirely new graphs and I will fully document any additional tests such as High/Low or amplification. I will be calling this the “Full Interrogation” of the lock (term borrowed from a Mark Bates interview). So without further ado, here is our lock:

Ilco Unican 673000
Group 2, 3 wheels

[center]

Dial and Mount.jpg

Internals.jpg

This lock was purchased on eBay and was advertised as having been a store display. It certainly looks like one because it was already on a plastic mount complete with a slightly damaged change key. I doubt this lock has ever been used in any other capacity. It’s nice and clean on the inside and there are no obvious signs of wear. In case anybody notices and wonders about it, I have bent the relocking spring out of the way to disable the relocking mechanism.

Left Contact Point: 98¼
Right Contact Point: 7
Width of Contact Area: 8¾
(Measurements taken with all wheels parked at L50)

Manipulation #1
All Wheels Left Method

In this manipulation I discovered the previously unknown combination. In future manipulations of this lock in the “Total Interrogation” series I will continue using the same combination so that various approaches to the lock can be directly compared under the same circumstances.

The first manipulation of the lock will be a traditional all wheels left approach measured in increments of 2½ and measuring to ¼ of a dial mark. Here's the graph.

AWL Graph.jpg

I found two good looking indications at 20 and 72½ with some suspicious activity between 45 and 55. The area between 45 and 55 didn’t look like enough for me to bother with right away because I had two good indications to investigate, but I kept it in mind for the future just in case.

The next step I took was to amplify the two indications so as to zero in on the correct numbers. I went back to the indications and tested each number individually instead of just checking every 2½ numbers. Here are the results:

Amplifying L20:
17.0 – LCP 98.00 – RCP 7.00 – CA 9.00
18.0 – LCP 98.25 – RCP 6.75 – CA 8.50
19.0 – LCP 98.50 – RCP 6.75 – CA 8.25
19.5 – LCP 98.50 – RCP 6.75 – CA 8.25
20.0 – LCP 98.50 – RCP 6.75 – CA 8.25
20.5 – LCP 98.50 – RCP 6.75 – CA 8.25
21.0 – LCP 98.50 – RCP 6.75 – CA 8.25
21.5 – LCP 98.00 – RCP 7.00 – CA 9.00

As you can see I got a little paranoid part of the way through and started checking every ½ number. I've been burned before because I wasn't sufficiently careful while amplifying an indication. I found a 3 dial-mark wide area from 19-21 with a clearly low reading. Looks like a gate to me!

Amplifying L72.5
70.0 – LCP 98.00 – RCP 7.25 – CA 9.25
71.0 – LCP 98.50 – RCP 6.75 – CA 8.25
71.5 – LCP 98.50 – RCP 6.75 – CA 8.25
72.0 – LCP 98.50 – RCP 6.75 – CA 8.25
72.5 – LCP 98.50 – RCP 6.75 – CA 8.25
73.0 – LCP 98.50 – RCP 6.75 – CA 8.25
73.5 – LCP 98.50 – RCP 6.75 – CA 8.25
74.0 – LCP 98.00 – RCP 7.25 – CA 9.25

The indication showed up as 3.5 dial-marks wide (from 71-73.5) so I split the difference, rounded up to the nearest half and called it L72.5. If this turns out to be wheel 2 I’ll translate that into a right oriented number.

The next task was to determine which gates relate to which wheels. Following are the results of my tests. I decided to use a High/Low test first since that’s what I usually do. The test was done on L20 and L72.5 using +/- 10 numbers.

High test on L20:
w1&2 @ L20 / w3 @ R30
LCP 97.75 – RCP 7.25 – CA 9.50
w1&3 @ L20 / w2 @ R30
LCP 98.25 – RCP 6.75 – CA 8.50
w2&3 at L20 / w1 @ R30
LCP 98.50 – RCP 6.75 – CA 8.25

Low test on L20:
w1&2 @ L20 / w3 @ R10
LCP 98.50 – RCP 6.75 – CA 8.25
w1&3 @ L20 / w2 @ R10
LCP 98.25 – RCP 6.75 – CA 8.50
w2&3 @ L20 / w1 @ R10
LCP 98.25 – RCP 6.75 – CA 8.50

The High/Low test on L20 was inconclusive. The high test indicates that L20 belongs to w3, but the low test shows that w3 is the least likely to own L20. The two tests directly contradicted each other, even after double and triple-checking for errors. It just goes to show that High/Low tests are fraught with peril.

High test on L72.5:
w1&2 @ L72.5 / w3 @ R82.5
LCP 98.50 – RCP 6.75 – CA 8.25
w1&3 @ L72.5 / w2 @ R82.5
LCP 98.50 – RCP 6.75 – CA 8.25
w2&3 @ L72.5 / w1 @ R82.5
LCP 98.50 – RCP 6.75 – CA 8.25

Low test on L72.5:
w1&2 @ L72.5 / w3 @ R62.5
LCP 98.00 – RCP 7.00 – CA 9.00
w1&3 @ L72.5 / w2 @ R62.5
LCP 98.75 – RCP 6.50 – CA 7.75
w2&3 @ L72.5 / w1 @ R62.5
LCP 98.75 – RCP 6.50 – CA 7.75

Here, the high test showed nothing of use, but the low test couldn't be clearer. L72.5 clearly indicates for wheel 3. It's easily seen here that when wheel 3 was at L72.5 there was a smaller contact area and when it was not at L72.5 there was a larger contact area. At the same time, there is nothing in the high test that contradicts that conclusion.

I decided to do a little more investigation since this is a “Full Interrogation”. This time I parked all three wheels at a low spot on the graph (R35) and did another test using that low area as a baseline. Then I brought wheel 3 over to the two indicating area’s to compare them against the baseline. Here are the results.

All wheels @ R35 (baseline test – low area of graph)
LCP 98.25 – LCP 6.75 – CA 8.50
w1&2 @ R35 / w3 @ L20
LCP 98.28 – LCP 7.00 – CA 8.75
w1&2 @ R35 / w3 @ L72.5
LCP 98.25 – RCP 6.75 – CA 8.50

The indication got worse when w3 was on L20, but remained the same when it was on L72.5. This gave more credence to the conclusion that L20 was not the combination number for w3 but it still wasn't very convincing.

In the spirit of a Full Interrogation I tried yet another approach to investigate the different indications on wheel 3. I had already tried comparing the two indications against a baseline in a low area of the graph, but what would happen if I compared them against a baseline in a high area of the graph? Here are the results.

All wheels @ R60 (baseline test – high area of graph)
LCP 98.00 – RCP 7.00 – CA 9.00
w1&2 @ R60 / w3 @ L20
LCP 98.25 – RCP 7.00 – CA 8.75
w1&2 @ R60 / w3 @ L72.5
LCP 98.75 – RCP 6.50 – CA 7.75

Voila! That made three tests assigning L72.5 to wheel 3 and one test that was inconclusive but could have been interpreted to mean that L20 belonged to wheel 3. Considering that one weird test was the first one I did, I'm glad I continued to investigate in greater detail.

X - X - L72.5

The next thing I considered was what to do about the indication at L20. The high/low test on L20 was weird so I disregarded it. I decided to do the same "baseline" tests for w2 as I did for w3 except with w3 on its gate. Here are the results.

Low baseline test:
w1&2 @ R35 / w3 @ L72.5
LCP 98.50 – RCP 6.25 – CA 7.75
w1 @ R35 / w2 @ L20 / w3 @ L72.5
LCP 99.00 – RCP 6.25 – CA 7.25

High baseline test:
w1&2 @ R60 / w3 @ L72.5
LCP 98.75 – RCP 6.50 – CA 7.75
w1 @ R60 / w2 @ L20 / w3 @ L72.5
LCP 98.75 – RCP 6.25 – CA 7.50

It looked like I had another winner! On both tests I got a clear improvement when placing w2 on L20 so I concluded that this was the gate for wheel 2. On this lock changing from a left oriented number to a right oriented number adds .5 so L20 became R20.5.

X - R20.5 - L72.5

Now that I had two numbers, I got dialing.

Open!

L50 - R20.5 - L72.5

What did I learn?

I learned to really dig in and investigate each indication. Remember that “suspicious activity between 45 and 55” that I supposedly kept in mind for future use? Well I forgot all about it. This was a failure to fully interrogate the lock. I should have investigated it when I was done with looking into L20 and L72.5. It turns out I had every single gate showing up clearly in my first graph.

I learned that if you take readings at enough points on the dial, you can actually paint a picture of the drive cam gate. If you look at the amplification of L20 above, I was taking readings pretty close together and captured the gradual descent of the nose against the left contact point. It read 98, then 98.25, then 98.5.

I am more and more wary of High/Low tests in general. They seem to introduce an element of randomness to the process by arbitrarily parking wheels at some set distance from the indication being investigated. Why 10 dial marks? Why not 5, or 12, or 57? The state of the wheel pack 10 numbers away from the indication is going to be totally different from lock to lock, and from combination to combination in the same lock, so why use it as a standard? What seems to be happening is that you are comparing an unknown datum (the indication you are investigating) to an unknown datum (the state of the wheel pack at an arbitrarily assigned number of steps away from your indication) and hoping to arrive at useable data.

I just came up with the "high baseline" and "low baseline" tests on the fly and I don't know if that technique already exists somewhere under a different name. I have seen people on the forums “park wheels in a low area” so maybe this is the same thing. I will continue using it to determine if it's actually consistent but I cannot believe that it is less valid than a High/Low test. I imagine it must be more consistent since you are comparing an unknown datum against a known datum. That’s a big step up from comparing one thing you don’t know about to another thing you don’t know about. You already know from the graph what the state of the wheel pack is at the points you are comparing your indications to, so you can see if there are departures from that state and to what degree.

I apologize if I'm coming off as arrogant here or if I'm stepping on anyone’s favorite technique. I'm more than willing to be proven wrong on my distrust of the High/Low test or on the validity of the tests I just came up with and I will happily eat crow if it results in my having a better understanding of these locks and the techniques used to manipulate them. I am really new to this so I'm sure I'll embarrass myself somewhere along the line

Questions I have

Am I doing the high/low tests wrong? If there's more to it or I'm applying the technique incorrectly I really want to know.

Do the “low baseline” and “high baseline” tests I threw together already exist under another name? I’m willing to bet there are people out there who already came up with this idea and either use it or have discarded it for valid reasons.

Is the fact that I found all three gates on the first graph unusual? I thought it was almost a certainty that you would only find one gate at a time and that exceptions to this were very rare.

[Oldfast]

....Instead, I thought it might be a useful resource to post a series of different manipulations of the same lock, with the same combination.

....The idea is to extract every single bit of knowledge possible from a single lock by approaching it with a variety of different methods.

GREAT idea rohare! I look forward to the progression of this thread. From it, I'll give what I can and take what I can.
I still have only a small number of locks to work with, but it's been more than enough to keep me busy and learning
for many months now. I'm costantly working new combinations through each one while varying my techniques.
Each lock is a worthwhile investment that brings a looooooong time of enjoyment and learning.

I am very new to manipulation, but I'm going to put myself out here and risk making silly mistakes in a public forum because I think the project will be fun....

Don't short yourself. In all honesty, I can already see that your line of thinking is more advanced than mine was when starting out.
After a read through last night & this morning-- the conclusions you're reaching as well as the methods you're questioning seem
quite sound to me. I have a very good feeling that I'll be learning some good stuff from you.... and I look forward to it.

....As you can see I got a little paranoid part of the way through and started checking every ½ number.

I find myself doing that sometimes in order to choose a more accurate center.
.5 inc readings on either side of the gate is sufficient to find a more exact edge.

....On this lock changing from a left oriented number to a right oriented number adds .5 so L20 became R20.5.

If you found the pickup difference for w2 to be 1/2 an inc, then L20 will convert to R19.5
Regardless of the rotational direction or the the wheel you're working with... you always
want to dial beyond/passed the original # by the appropriate amount (.5 inc in this case).
Outta all this, that's the only correction I could find?! That's pretty damn good! LOL

I am more and more wary of High/Low tests in general. They seem to introduce an element of randomness to the process by arbitrarily parking wheels at some set distance from the indication being investigated. Why 10 dial marks? Why not 5, or 12, or 57? The state of the wheel pack 10 numbers away from the indication is going to be totally different from lock to lock, and from combination to combination in the same lock, so why use it as a standard? What seems to be happening is that you are comparing an unknown datum (the indication you are investigating) to an unknown datum (the state of the wheel pack at an arbitrarily assigned number of steps away from your indication) and hoping to arrive at useable data.

I agree. Early on I posed the same question (why 10? why not choose a more suitable test position based on your graph?)...
no takers. Too much left to chance in my opinion. Just as you mentioned- especially when we've gathered enough info
to make a more sound decision. I've not discounted it entirely... it certainly has it's uses. But the longer I manipulate,
the less I find myself hi/low testing, and the more I tend to use isolation as a means to tag a gate to a wheel.

Take your manipulation for example. My recent manipulation of a Mosler I'll be posting soon is actually QUITE
similar to the way yours has played out here. This's how I would approach it... something you can try anyway?

The most likely scenario here is that 72.5 is w3 and 20 is w2. Obviously, we cannot assume this
is the case... but we can at least use this info to start our search in the most likely of places.

I would first see if the most predominant gate (found @ 72.5) is on wheel 3 by isolating it.
This just means I will graph ONLY w3 every inc through the area in question. If the same
drop occurs, I know without a doubt it's on w3 because it's the only wheel I was moving.

Park all wheels at your other good indication of L20. Now switch directions, going right one revolution
and you'll pick up w3 at 20. Bring it down to 78, then take a reading. Right again to pick up w3, bring it
to 77, take a reading. You're graphing only w3 through the area (78-68) with the other 2 wheels at 20.
Technically, you wouldn't have to take a reading at every increment.... but you get the idea.

If the drop occurs you can tag the gate to wheel 3. The next order of business is to isolate w2,
running it through the area of 20. Bare in mind, this gets a little more complicated, as there's
more rotations involved. You're going to park wheels 1 & 3 @ L72.5 and run ONLY wheel 2.

W1 can be parked and not touched again. W3 however will have to be repositioned at it's known
gate EACH time after moving w2, prior to taking each reading. Graph from approximately 23-17.

Hope this helps. Or... at the very least... give you some new ideas to run with

Great stuff! I'm very much looking forward to more.

[rohare]

I still have only a small number of locks to work with, but it's been more than enough to keep me busy and learning
for many months now. I'm costantly working new combinations through each one while varying my techniques.
Each lock is a worthwhile investment that brings a looooooong time of enjoyment and learning.

Oh good! My wife is already dubious about the alarming number of locks I've collected since I started picking....

I find myself doing that sometimes in order to choose a more accurate center.
.5 inc readings on either side of the gate is sufficient to find a more exact edge.

Heh, I actually stole the idea from you. When I saw you using final numbers like "R37.5" or some such, I knew this must be what you were doing.

If you found the pickup difference for w2 to be 1/2 an inc, then L20 will convert to R19.5
Regardless of the rotational direction or the the wheel you're working with... you always
want to dial beyond/passed the original # by the appropriate amount (.5 inc in this case).

Ohhhhhhh. No wonder I occasionally have to re-dial the numbers to open it. Duh.

I would first see if the most predominant gate (found @ 72.5) is on wheel 3 by isolating it.
This just means I will graph ONLY w3 every inc through the area in question. If the same
drop occurs, I know without a doubt it's on w3 because it's the only wheel I was moving.

Park all wheels at your other good indication of L20. Now switch directions, going right one revolution
and you'll pick up w3 at 20. Bring it down to 78, then take a reading. Right again to pick up w3, bring it
to 77, take a reading. You're graphing only w3 through the area (78-68) with the other 2 wheels at 20.
Technically, you wouldn't have to take a reading at every increment.... but you get the idea.

Cool! This is exactly what I was looking for; a concise description of a "wheel isolation" test. I sort of blundered into it during the manipulation but I intentionally avoided using the other indications and started from a low area on the graph that didn't hold an indication. I didn't have any reason for this particularly except a vague sense that I didn't want to mix test results but even at the time I thought I was being overly cautious.

Great stuff! I'm very much looking forward to more.

AWR coming soon... but that will be a shorter post.

[CPT1911]

Love to see all the discussion of Hi/Low testing and wheel isolation! I will be watching this closely as I have had nothing but trouble (seriously, like a 20% success rate) with the Hi/Low technique....

Thanks!

[rohare]

Ilco Unican 673000
Group 2, 3 wheels

Left Contact Point: 98¼
Right Contact Point: 7
Width of Contact Area: 8¾
(Measurements taken with all wheels parked at L50)
(Measurements also taken before I learned that L50 was a gate)

Combination:
L50 - R19.5 - L72.5

Manipulation #2
All Wheels Right Method

I’ll cut right to the chase on this one. In Manipulation #1 I ran all wheels left and did a bunch of tests to figure out which indications applied to which wheels. Now that that’s out of the way we can get straight down to examining how differing methods give us differing results. Hopefully, by comparing the various approaches, we can derive some useful knowledge about those approaches, about this lock in particular, or about locks in general. So, here’s the AWR graph.

AWR.jpg

For easy reference, here’s the original AWL graph.

AWL.jpg

Broadly, the results of the two graphs were very similar. The high and low regions roughly coincide, the w2 and w3 indications (R19.5 and L72.5 respectively) show up clearly on both graphs, and the w1 indication (L50) looks just suspicious enough to investigate on the AWL graph, but barely makes an appearance on the AWR graph.

One interesting thing I’ve noted is that the wheel 2 and 3 indications were very marginally more significant on the AWR graph but the wheel 1 indication was very marginally more significant on the AWL graph. That means that running right didn’t specifically improve the indications of right oriented or left oriented gates and the same goes for running left. Therefore (assuming this lock is representative of group 2 locks in general), there is no particular reason to prefer AWL or AWR except that running left means you could potentially avoid having to translate the orientation of one number. Big freaking deal. Nevertheless, AWL did work better on this particular lock. I just mean that statistically, I doubt there’s really a significant difference between the two.

Honestly I can’t deduce much more from this manipulation, but then again, I wasn’t really expecting to see a lot of difference between an AWL graph and an AWR graph. The only question that really comes to mind is why exactly there are any differences at all. Logically, all else being equal, they should be the same. Perhaps this is a consequence of the slight difference of position that the wheels will have due to their rotational offset? Or perhaps the way the wheels rest on the spindle? Is interaction with the spindle capable of affecting the wheel pack without the influence of the drive cam, drive pins, or fly's? Inquiring minds want to know...

Next up, a full interrogation of just wheel 3 comprised of four graphs.

[rohare]

Ilco Unican 673000
Group 2, 3 wheels

Left Contact Point: 98¼
Right Contact Point: 7
Width of Contact Area: 8¾
(Measurements taken with all wheels parked at L50)
(Measurements also taken before I learned that L50 was a gate)

Combination:
L50 - R19.5 - L72.5

Manipulation #3
Wheel 3 analysis

Manipulation #3 consists of four graphs detailing wheel 3 individually. Here they are:

Graph 1 Low Around Right.JPG

Graph 1: Wheels 1 and 2 parked at a low spot on the graph (L10) and wheel 3 dialed around right.

Graph 2 High Around Right.JPG

Graph 2: Wheels 1 and 2 parked at a high spot on the graph (L80) and wheel 3 dialed around right.

Graph 1a Low Around Left.JPG

Graph 1a: Wheels 1 and 2 parked at a low spot on the graph (R10) and wheel 3 dialed around left.

Graph 2a High Around Left.JPG

Graph 2a: Wheels 1 and 2 parked at a high spot on the graph (R80) and wheel 3 dialed around left.

When I compared the four graphs I was really startled to see the results side by side. I was expecting to see four different vaguely similar graphs with slightly more prominent indications on the graphs where I chose to park wheels 1 and 2 in low areas. Instead, I got a gigantic roller-coaster on two graphs, while the other two were fairly straightforward.

The "roller-coaster" graphs were both produced when running wheel 3 to the right. Wheel 3 is, naturally, normally run to the left. Parking the other wheels on high or low areas did not result in very much difference between the two right-run graphs. The most obvious feature of these graphs is the huge canyon in the left contact point that starts around 20 and ends around 65, but more importantly, there is no gate indication whatsoever. The gate (L72.5) is totally lost in an enormous low area that spans much of the wheel. These graphs would be entirely useless in attempting to open the lock.

The two useful graphs both came from running the wheel left, its intended direction. Whether wheels 1 and 2 were parked in high areas or low areas did not affect the appearance of the gate at L72.5 whatsoever. However, it is useful to note that parking w1&2 in the low area produced a much easier graph to follow. When they were parked in the high area, the graph showed lots of activity with the contact points deviating slightly every few numbers. On the graph where I parked them in a low area there are very few deviations from the general level of the wheel and the graph only gets active right near the gate itself, which makes the gate far more obvious even though the indications are identical between the graphs. I imagine that this degree of difference would make free spinning far easier. It certainly makes taking readings for a graph easier.

So my conclusion here is that careful analysis and conventional wisdom are fully in agreement. When running a single wheel for the purposes of graphing or for doing a wheel isolation test, don't waste your time running it in a direction it isn't intended to go, and definitely park the off wheels in a low area of the graph (or a gate if it's already known).

This also lends some more weight to the theory that AWL is a better method than AWR most of the time. If you are running left then you are more likely to get two good indications, w1 and w3, while obscuring only one indication, w2.

Yes, I know this is one of those things "everybody knows", but now I know why "everybody knows" it.

I also hasten to note that this may just be the lock I’m using rather than group 2 locks as a general rule. On the forum here, Daggers once asked about a weird indication problem where w3 of a lock gave an indication while running to the right, but never when running to the left. Farmerfreak responded that this can happen sometimes as a result of the wheel lifting up a bit when being turned one direction but pulling down away from the fence a bit when turned the other direction. It’s possible that’s what is happening in my lock.

Questions

Does the extreme difference I see between running w3 left and running it right com from the way the wheel sits on the spindle? I assume so, but wanted to see if anyone could confirm that.

[CPT1911]

This also lends some more weight to the theory that AWL is a better method than AWR most of the time. If you are running left then you are more likely to get two good indications, w1 and w3, while obscuring only one indication, w2.

Rohare, great point! I'm a newb and it never occurred to me!

I've recently experimented with parking wheels 1+2 and running wheel three in isolation, but I'm not sure about the different options for where/how to park them.

I see folks talk about parking 1+2 in the contact region. Why? And how does this better/worse than using the high point or low point?

Also, I assume that when we are talking about parking it in a "high" or "low" point, we are talking about a high or low point you found in a quick AWL run? Otherwise, where is the "high" point or "low" point coming from?

[rohare]

CPT1911,
Well, I'm a newbie too, but here's my 2 cents. I don't recall having seen anyone recommend parking w1&2 into the contact region. I have however seen people recommend parking them in the "forbidden zone" That is the area that w3 is not supposed to be in per the manufacturers directions. For example S&G instructs in their literature not to set w3 to any number between 95 - 0 and 0 - 20 (If I recall correctly). Honestly, I'm not sure why it should matter. Because you don't know where the gates on wheels 1 and 2 are, anywhere you park them is going to be a roll of the dice. If you already have done an AWL or AWR but didn't get any indications then by all means park w1 and w2 in a low area. This will give you the best possibility of getting an indication on w3.