Something to note: There are safe that have panels of breakable material on the inside to prevent drilling If they break, the safe is permanently closed, requiring destructive entry So an auto dialer is still a needed utility. And having an open source one is a boon to the community
yeah would love to see this open-sourced, so people can improve the design. Also there are combinations locks without the handle, that open when you turn the dial the other way, so if that takes more force to turn, it means its open.
Oh man, you had it the whole time! My favorite part was when you guys opened the safe. The excitement was so genuine. Those borescopes are super useful. Great work as always!
oh man that was so exciting to see the bore cam work. Just like in the movies! I was bummed that your robot wasn't the final solution. I know what it's like to put so much work into something only to have it not pan out in the end. I think you almost needs to purchase a small safe that WILL work with your robot just so you can gave it on the shelf and show folks how awesome it is, but it's AMAZING! Great job, Zach!
I think after I get settled in after I move I may try to find a free or cheap locked safe on Facebook marketplace or Craigslist to keep working on this project
Ah-Ha, drilling and scoping works, as a 35 yr safetech that's how I opened 90 % of the dial locks, I did have some success manipulating antique safes where there were greater tolerances, modern locks are very hard, great job, great vids
I mentioned on your last video to measure the amount the handle moves, as I suggested i use a laser pen taped to the handle when I open these. Glad you finally got it open.
Man what a journey! Loved following along because it encompassed a real life engineering journey with all the highs and lows, heartbreak and successes. Even though you had so much time invested in the auto dialer, it was great that you pivoted and found another working solution and at the end, found the reason why the auto dialer didn't work initially. Looking forward to the next one!
I enjoy microcontroller projects, I also enjoy look picking etc. My thought was rather than brute force the combo you could teach the controller to manipulate it. Depending on whether the motor can report a high sensitivity regarding torque I think it would be the most efficient option.
It’s so awesome to see this project have a good ending. You got so wrapped up in the auto dialing you were committed to see it through. With the deadline approaching you were gonna crack that safe one way or another. Fun series
I know there is probably a few comments on this - but with all the set backs and such - i really enjoyed the journey and the mathematics / programming it takes to accomplish such a feat - the time invested, if the device just did every single combination ever - might of done so already hehe
I found my brothers combo using a camera, he locked his manual and receipt inside when he brought it home, I used the camera to turn the pages and found the page with the three numbers in big red ink. We used the holes for bolting it to the floor for camera access.
Ack! Missed it by that much! I love that through each setback your lessons learned and the next iteration got you closer a winning solution. Looking forward to seeing the next shop.
Love it dude way to get the job done one way or the other. Its good that you got to discover your design issue too without having that plague any future iterations. Well done!
If you had an analog deflection readout on the handle servo instead of the binary limit switch, you would have not only found the combo, but also could use it to probe the gate locations automatically and shrink the search space.
I had thought about doing something like that, but I did not have enough time. If I pick this project back up in the future I will certainly try it! Thanks for the suggestion
This has been quite the rollercoaster and I've enjoyed being on this journey. Great video series and your persistence knows no bounds. Looking forward to more of your videos
Yeah, measure the distance the handle moves with each attempt and save the combo associated with the top 100 handle distances. After the program completes sort by distance the handle moved and start checking by hand from the largest movement.
Thanks! I appreciate that! I've been watching Bob since he started and I listen to making it every week! He has definitely been an inspiration for me and so many other people.
So the auto-dialer needed more of an analog sensor on the handle and not a limit switch. You could then record and review all the combinations with the longest stroke to find the winner.
It's so sad that you had a solution that would have worked if you'd had a better way to work the lever. I'm glad you were able to laugh at the fact that for less than $20 and well less than an hour of time you could have solved the problem without spending 1/2 a year. But we all know the tinkering is the fun part. Hope the move is going smoothly!
Also...did you try to figure out what went wrong with trying to read the false gates? If it's cheap enough they probably made the false gate spacing identical on all the safes so reading the false gates in that case should allow you to determine exactly where the real gate is.
I get why you decided to go for the drilling. And I see that it turned out that it was the correct choice. But a part of me is disappointed that you didn't manage to do it the "pure" way: brute forcing through all possible combinations. With the safe having false gates on all wheels, you could have done the following. A: Find a single combination that lines up with a gate on each wheel. B: Map all gate positions on all wheels (rotating one wheel while keeping the other wheels in the positions found above). Make sure to map the "start" and "end" of each gate, to a precision of 1/2 or 1/4 of a position. C: Have your auto-dialing bot visit only combinations of those gate positions. D: If B doesn't work, try going to 2 below and 2 above your mapped gate positions, in addition to the gate positions themselves. For step A, you have gates at 8 positions out of 100 (i.e. every 12.5 positions). To find the first combination where 3 gates line up would thus take you about 12.5*12.5*12.5=1953 tries. (This is assuming that you don't take steps of 2 positions at a time, which would narrow it by a factor of 8, for 244 combinations to try). Step B: To get all the gates with 1/4th positions accuracy, you have to try 400 (sub-) positions per wheel. Since you work through the gates sequentially, that's 1200 positions total. (600 if you can get away with 1/2 position accuracy.) Step C would try only 8*8*8=512 combinations (out of a million). Step D would still only try 24*24*24=13824 combinations (out of a million). This sums to 244+600+512=1356 combinations *(max)* if step C opens it (just under 2 hours if you are manually dialing a combination every 5 seconds). Or 1953+1200+512+13824=17489 combinations *(worst case)* if it turns out that the gates are narrow enough that you need 1/4th position precision and step D (19 hours if you dial a combination every 4 seconds). I'm surprised how this is viable in a break-in. At least theoretically; no burglar would take that risk, unless you had the nuclear codes in that safe. And I wonder /why/ they went with 8 gates per wheel. The *only* number of gates that results in this attack taking a lower total number of combinations (assuming the safe opens after step C) is 7 gates per wheel. Even 6 gates per wheel would have again been slightly harder than 8 (1395 combinations total).
If you look back at the numbers he came up with in the first part, the first two numbers aren't on the lists, so brute forcing using those numbers wouldn't have worked even though it would reduce the number of combinations.
so the robot worked, but it was just not strong enough to actuate the safe handle far enough to trigger the correct code? wow, that kinda sucks, but good to know :)
Wow, can't believe how easy that was using a simple borescope camera. I think the moral of the story is: buy a crappy designed cheap safe because the expensive nice ones make it easier to crack? 🙂
Ok. You failed at building a safe cracking robot. Kind of a shame, since I really wanted to see the safe cracking robot do it, I've seen safes drilled open plenty of times. As safes go, this one is as basic as they come, with no intrusion countermeasures whatsoever, barely safer than a high school locker. You could probably open that one by feel alone. Just tension it and then go from there. It basically has just 512 combinations
I would call that a huge success. He engineered the thing and it just needs a small change in how the correct gate is detected. He pushed it further with the correct combination, it's just that the limit switch is binary, and the threshold was too far. So to avoid this problem it just needs to measure distances instead of a binary yes or no. That could even be used to automatically measure the false gates also which would make the thing even faster.
This series has been a joy to watch, but please finish the auto-dialing project. You were so close!
It is finished, just need a servo update 🤓😉
Something to note: There are safe that have panels of breakable material on the inside to prevent drilling
If they break, the safe is permanently closed, requiring destructive entry
So an auto dialer is still a needed utility. And having an open source one is a boon to the community
yeah would love to see this open-sourced, so people can improve the design. Also there are combinations locks without the handle, that open when you turn the dial the other way, so if that takes more force to turn, it means its open.
it's called external re-lockers and they are a pain but can be opened either way!
It's normally glass
Best line I'm going to blame the safe cheap design for the fact that I couldn't crack it
Oh man, you had it the whole time! My favorite part was when you guys opened the safe. The excitement was so genuine. Those borescopes are super useful. Great work as always!
This has been a really fun ride!
3d ago??? Are you a time-traveller?
@@Alex-oh5rt Video was probably unlisted and listed 6 mintues ago. But yeah, I was confused too XD
oh man that was so exciting to see the bore cam work. Just like in the movies! I was bummed that your robot wasn't the final solution. I know what it's like to put so much work into something only to have it not pan out in the end. I think you almost needs to purchase a small safe that WILL work with your robot just so you can gave it on the shelf and show folks how awesome it is, but it's AMAZING! Great job, Zach!
I think after I get settled in after I move I may try to find a free or cheap locked safe on Facebook marketplace or Craigslist to keep working on this project
Ah-Ha, drilling and scoping works, as a 35 yr safetech that's how I opened 90 % of the dial locks, I did have some success manipulating antique safes where there were greater tolerances, modern locks are very hard, great job, great vids
I mentioned on your last video to measure the amount the handle moves, as I suggested i use a laser pen taped to the handle when I open these. Glad you finally got it open.
Man what a journey! Loved following along because it encompassed a real life engineering journey with all the highs and lows, heartbreak and successes. Even though you had so much time invested in the auto dialer, it was great that you pivoted and found another working solution and at the end, found the reason why the auto dialer didn't work initially. Looking forward to the next one!
I enjoy microcontroller projects, I also enjoy look picking etc.
My thought was rather than brute force the combo you could teach the controller to manipulate it.
Depending on whether the motor can report a high sensitivity regarding torque I think it would be the most efficient option.
This have been awesome to follow!
To bad that the autodialer failed because of such minor problem - but it's a simple fix.
It’s so awesome to see this project have a good ending. You got so wrapped up in the auto dialing you were committed to see it through. With the deadline approaching you were gonna crack that safe one way or another. Fun series
Please continue the auto dial robot. See it through man.
I know there is probably a few comments on this - but with all the set backs and such - i really enjoyed the journey and the mathematics / programming it takes to accomplish such a feat - the time invested, if the device just did every single combination ever - might of done so already hehe
I found my brothers combo using a camera, he locked his manual and receipt inside when he brought it home, I used the camera to turn the pages and found the page with the three numbers in big red ink. We used the holes for bolting it to the floor for camera access.
Awesome man. Super nice work. You still made a good device.
Ack! Missed it by that much!
I love that through each setback your lessons learned and the next iteration got you closer a winning solution.
Looking forward to seeing the next shop.
Love it dude way to get the job done one way or the other. Its good that you got to discover your design issue too without having that plague any future iterations. Well done!
If you had an analog deflection readout on the handle servo instead of the binary limit switch, you would have not only found the combo, but also could use it to probe the gate locations automatically and shrink the search space.
I had thought about doing something like that, but I did not have enough time. If I pick this project back up in the future I will certainly try it! Thanks for the suggestion
Enjoyed the journey and glad your got it open.
At least you were able to find where refinement could be made and that you were on the right track!
I love these videos. I can't believe that camera is so cheap. That's awesome. The robot is cool. He had a thing to learn that's all
This has been quite the rollercoaster and I've enjoyed being on this journey. Great video series and your persistence knows no bounds. Looking forward to more of your videos
Seems like you just need to reposition the limit switch, or use a more analog sensor to measure the amount of throw. A torquier servo might also help.
Yeah, measure the distance the handle moves with each attempt and save the combo associated with the top 100 handle distances. After the program completes sort by distance the handle moved and start checking by hand from the largest movement.
if the handle servo detected what combination allowed the handle to turn the farthest, it didn't need to fully open the door
great journey and process, glad you finally got it!
Thanks Forby
Mitutoyo digital dial indicators have a serial interface to access measurements automatically
Was that a S&G lock? I think they'd be a bit tougher. Loved your determination and patience.
Wild. Great music cues, as always
Good job!
Awesome, congratulations on a job well done!
Great! where does he live?
Nice project. Fix the issues you found. But remember, solving a complex problem by adding more complexity is probably the wrong way.
I had to do the camera method but I had one that they use for two models one with the key unlock one without and they just cover over the keyhole
All in all very entertaining and a good lesson on Control System design : )
Great video. Always wondered how this worked. Best of luck moving. ❤😊👍
I think you should keep going, get your original safe cracking machine to work.
Now the all world knows the right combination, only need to find the right safe
Yes you did get the combination now can you repeat it on a different safe with the problem solved
Love your videos! Bob Clagett recommended checking out your channel on the "Making It" podcast. he was right, your content is fantastic!
Thanks! I appreciate that! I've been watching Bob since he started and I listen to making it every week! He has definitely been an inspiration for me and so many other people.
So the auto-dialer needed more of an analog sensor on the handle and not a limit switch. You could then record and review all the combinations with the longest stroke to find the winner.
So it turns out the safe was made out of butter the whole time! It's the thrill of the hunt we're after I guess.
The journey, indeed.
It's so sad that you had a solution that would have worked if you'd had a better way to work the lever. I'm glad you were able to laugh at the fact that for less than $20 and well less than an hour of time you could have solved the problem without spending 1/2 a year. But we all know the tinkering is the fun part. Hope the move is going smoothly!
So would it work if you take out the limit switch and put in something measuring the stop location?
Working blind and you got that close O U T S T A N D I N G
The contents of the safe might be secret but now the internet knows what the combination is
was
That was pretty cool but wish that your contraption would have worked it was pretty cool and genius! 👍🏼✌🏼
Also...did you try to figure out what went wrong with trying to read the false gates? If it's cheap enough they probably made the false gate spacing identical on all the safes so reading the false gates in that case should allow you to determine exactly where the real gate is.
FINISH IT!
the journey was fun for shure
The contents were two smaller safes that they don't have the combination for.
I get why you decided to go for the drilling. And I see that it turned out that it was the correct choice. But a part of me is disappointed that you didn't manage to do it the "pure" way: brute forcing through all possible combinations.
With the safe having false gates on all wheels, you could have done the following.
A: Find a single combination that lines up with a gate on each wheel.
B: Map all gate positions on all wheels (rotating one wheel while keeping the other wheels in the positions found above). Make sure to map the "start" and "end" of each gate, to a precision of 1/2 or 1/4 of a position.
C: Have your auto-dialing bot visit only combinations of those gate positions.
D: If B doesn't work, try going to 2 below and 2 above your mapped gate positions, in addition to the gate positions themselves.
For step A, you have gates at 8 positions out of 100 (i.e. every 12.5 positions). To find the first combination where 3 gates line up would thus take you about 12.5*12.5*12.5=1953 tries. (This is assuming that you don't take steps of 2 positions at a time, which would narrow it by a factor of 8, for 244 combinations to try).
Step B: To get all the gates with 1/4th positions accuracy, you have to try 400 (sub-) positions per wheel. Since you work through the gates sequentially, that's 1200 positions total. (600 if you can get away with 1/2 position accuracy.)
Step C would try only 8*8*8=512 combinations (out of a million).
Step D would still only try 24*24*24=13824 combinations (out of a million).
This sums to 244+600+512=1356 combinations *(max)* if step C opens it (just under 2 hours if you are manually dialing a combination every 5 seconds). Or 1953+1200+512+13824=17489 combinations *(worst case)* if it turns out that the gates are narrow enough that you need 1/4th position precision and step D (19 hours if you dial a combination every 4 seconds).
I'm surprised how this is viable in a break-in. At least theoretically; no burglar would take that risk, unless you had the nuclear codes in that safe. And I wonder /why/ they went with 8 gates per wheel. The *only* number of gates that results in this attack taking a lower total number of combinations (assuming the safe opens after step C) is 7 gates per wheel. Even 6 gates per wheel would have again been slightly harder than 8 (1395 combinations total).
If you look back at the numbers he came up with in the first part, the first two numbers aren't on the lists, so brute forcing using those numbers wouldn't have worked even though it would reduce the number of combinations.
Do you have the printable files for download? Working on creating my own. Thanks :)
so the robot worked, but it was just not strong enough to actuate the safe handle far enough to trigger the correct code? wow, that kinda sucks, but good to know :)
Btw you can crack safes with that device. If you fine tune a bit. So you can make money out of it😀
Hi I have a sentry rifle safe I have the convanation but it won't open I don't know if I'm doing something wrong can you help, thanks
no hope for my safe then... What camera did u used?
Affiliate link: amzn.to/3pt3uRo
So the real question- Do you have a better quality safe to test your safe cracking robot on to show the fans it actually working start to finish?
It's possible I'll pick this project back up in the future
Dear sir lithium cell 18650 useful for ups or not
Admit it, you couldn't handle it 😉
Some of us already knew what was in the safe 😌
Wow, can't believe how easy that was using a simple borescope camera.
I think the moral of the story is: buy a crappy designed cheap safe because the expensive nice ones make it easier to crack? 🙂
Those false gates were huge you could have cracked it by hand in 5 minutes
I mean if you can see the wheels, not really that hard
So you're saying that your robot couldn't crack it BECAUSE it was cheap? That's the wrong way round, surely. 🤔
Can you use AI to crack a safe?
Get a safe guy to drill it.. .we would have been down and lock installed in 2hours....
Ok. You failed at building a safe cracking robot. Kind of a shame, since I really wanted to see the safe cracking robot do it, I've seen safes drilled open plenty of times. As safes go, this one is as basic as they come, with no intrusion countermeasures whatsoever, barely safer than a high school locker. You could probably open that one by feel alone. Just tension it and then go from there. It basically has just 512 combinations
I would call that a huge success. He engineered the thing and it just needs a small change in how the correct gate is detected. He pushed it further with the correct combination, it's just that the limit switch is binary, and the threshold was too far. So to avoid this problem it just needs to measure distances instead of a binary yes or no. That could even be used to automatically measure the false gates also which would make the thing even faster.
totally fake, there is no cover over the mechanism
no it wasn't fun. next time find a same locking mechanism to built 6 months of project on it.
I'm really disappointed. You wasted all of my time on a series, and in the end, you gave up.
Seriously? No one made you watch this
Engineering has failures, and those failures are learnt from.
Just as much as we hope you learn from this failure of a comment too.
This video series was a waste of my time.
Nobody made you watch this.
@viktortodosijevic3270 Engineering has failures, and those failures are learnt from.
Just as much as we hope you learn from this failure of a comment.