Pages

Feb 28, 2013

LEGO's magic number is 37,112

Have you ever asked yourself this question: "How many times can I assemble LEGO bricks before they wear out?"

Well... probably never but I did (on LEGO SE).  The story would end there if didn't had the knack to build a machine to test it.  Here it is:



The first version was using a metal tube attaching the top servo to the LEGO brick.  It was too flimsy and died during initial testing.  So I took an old CPU heat sink...


And using my trusty angle cutter, I made this new arm:


A successful test proved that the LEGO was no match for this chunk of aluminium.  I then made an infra-red proximity sensor to detect if the presence of the LEGO and rigged the whole thing to an Arduino.  This was shot before the test started:



and this was shot about two thirds into the testing that lasted a whopping 10 days!!!



So here are the 2 dead LEGO bricks after 10 days and 37,112 assembling and disassembling.



Don't try this at home. It's long, noisy and cruel to LEGO bricks.

In the end, both pieces failed at 50%. The studs of the bottom brick and the inside walls of the top brick are visibly worn. Both bricks can still hold on (not strongly) to normal bricks but, when put together, they can't hold. In a way, you could say that they are still in working conditions as long as they don't meet again.

Here are some more answers:


This story was covered in the follow media:

25 comments:

  1. Thanks for all the visitors today.

    ReplyDelete
  2. Did either of the lego bricks still work when combined with fresh legos or were they both completely dead? It looks like most of the damage was to the the receiving lego?

    ReplyDelete
    Replies
    1. Good question. Together they don't work when the top one is on top but work almost normally when the top one is at the bottom (no jokes intended). It performs the same way with other bricks.

      Delete
  3. Could screwing the arm onto the top Lego brick have weakened the brick?
    Could you repeat using some other fixative? (I KID!)

    great project!

    ReplyDelete
  4. 1) You should test if the number of PIN (yours is 6 pin lego) improuve or not the "torq" state.

    2) You should definitly work for Ikea as a stress tests engineer ! ;)

    ReplyDelete
    Replies
    1. I settled on 2x3 for the test. I thought it was a good ratio of studs/piece. If I get I little financing from LEGO, I'll do the mega tests. :)

      Delete
  5. Did you count the few times you attached them together during building/testing? Or did you start the long experiment with fresh ones?
    Point being is it 37112 or is it 'about 37120'...
    Great project nevertheless.

    ReplyDelete
    Replies
    1. Good point. I used old LEGO bricks for this first test so they were clearly used before. I'm going to make a follow-up vid about the details of this test.

      Delete
  6. Your pictures are broken.. I suggest using imgur or SE's imgur account to upload and then link them here.

    ReplyDelete
    Replies
    1. Mmmm... they look OK for me. Maybe it's something with how blogger handles massive traffic.

      Delete
  7. Hi, nicely done!

    Some numbers don't make sense to me, perhaps you could explain... You say: "two thirds into the testing that lasted a whopping 10 days!!"; but the number of iterations at that point was 36720, while it ended up at 37112... Did it get a lot slower afterwards, or is one of those numbers not correct?

    Thanks!

    ReplyDelete
    Replies
    1. Sorry about that. During the tests, I had to guess the current count based on the estimate that one full iteration was about 10 seconds. This was not a good estimate and it made me overshot the actual number. When the test was finally over I was able to see on the SD-card the final 'real' total of 37,112. I'm making a new video explaining some of this.

      Delete
  8. See the final results and some answers in this other video: http://www.youtube.com/watch?v=VLt8HkIOAuY

    ReplyDelete
  9. >Cruel to LEGO bricks

    LEGO bricks had it coming. Too long have my bare feet suffered at the hands of their pegs and damn near impossibly sharp corners. I aim to build an army of your machines, just to see bricks suffer in the cruellest way possible.

    ReplyDelete
  10. I'm curious if it was the pip side or the receiving side that failed (male or female connectors). Can the tested piece still work with new legos?

    ReplyDelete
    Replies
    1. In the end, both pieces failed at 50%. The studs of the bottom brick and the inside walls of the top brick are visibly worn. Both bricks can still hold (not strongly) on to normal bricks but, when put together, they can't hold. In a way, you could say that they are still in working conditions as long as they don't meet again :)

      Delete
  11. Hi Phillipe,
    Very interesting work. Always wondered about that question. IfF you're ever curious about getting a close look at the wear patterns I have an electron microscope and a bit of personal project time. y lab is https://sites.google.com/site/probelab2/

    ReplyDelete
    Replies
    1. @Glenn I will pick you up on that offer. Send me and email directly at ( phillipe at cantin dot com )

      Delete
  12. Greetings Phillipe,

    First let me say this was an interesting idea. Your rig was unusual but got the job done for sure. I hope when you build the new test rig that you redo your test for the 70's legos again as well.

    My concern is your math was pretty solid for the 4 days. I calculate though that it takes roughly 12 seconds per iteration and not 10. If this did in fact go a full 10 days that should have been closer to 72,000 and not 37,112 as stated in your blog.

    60 seconds in a minute.
    60 minutes in an hour.
    24 hours in a day.
    10 days in the experiment.

    60 * 60 * 24 * 10 = 864,000

    864,000 seconds in a 10 day period.
    12 seconds per iteration of the test.

    864,000 / 12 = 72,000

    So your original count was correct. In the video you said you were off by a "little" or "small" amount if that were the case you would not have been off by half.

    Is it possible your code was counting incorrectly? If it were counting every other iteration it is still easy to calculate the real total which I would imagine would be closer to: 74,224 if it went for a little over 10 days? If not it would have only went on for approx 5 days (half the time).

    Keep up the great experiments! I hope you solve the discrepancy in your numbers by a simple explanation. Even if you find a flaw in my math would make me happy!

    ReplyDelete
    Replies
    1. You got me thinking there for a moment and I even double checked my log file and it's OK. You have a point though, this is bugging me. Like I mentioned to Nathan from ArsTechnica it was disturbing my daughter so toward the end of the test I had to pause it for the nights. I wish I had more control points than just the count and that I could have ran it in one clean pass.
      Believe me when I say that I will totally overkill the next rig so that I have not only the count but timestamps and other data points.

      Delete
  13. Depending on how much detail you want to get in, you should compare bricks of different colors and opacities. My experience is that transparent bricks are more fragile than the opaque ones. Given what little I know about plastics I wouldn't be surprised if the dyes used to make different colors affected their durability, even among different opacities.

    ReplyDelete
  14. I'm actually impressed that the servo's lasted to the end of the test!

    ReplyDelete
  15. I'd be interested in seeing some close up macro photos of the Lego blocks to see the wear.

    Nice work BTW.

    ReplyDelete
  16. Thank you for your very interesting test. I used to play Lego. If you put Legos together vertically: not rotary motion, the result might be different. I believe that it might be more than 40000. Very good work. Thank you again. (Tokyo) 

    ReplyDelete
    Replies
    1. The next testing machine will use a linear motion, at least when putting them together. On the other hand, I think that the natural manipulation for taking them apart is with a rotating movement. Thanks for your interest and comment. (Quebec)

      Delete