| |
In lugnet.lego, Scott Arthur wrote:
| |
In lugnet.lego, Ross Crawford wrote:
| |
In lugnet.lego, Mark Bellis wrote:
| |
I believe that over the years the tolerance has been widened, with the aim
of reducing production cost.
|
OK so I just went to my box of childhood LEGO (circa 1970-80) to check this
out. Discounting the CA bricks which have known deterioration problems, I
put together 4 stacks of 10 ABS bricks, 2 red and 2 white. I made sure I
picked out the “best looking” ones to avoid damaged edges etc which may have
had an effect. I noted 2 things:
- Several of the bricks had significantly less “clutch power”, though obviously this could also be from deterioration;
- The stacks ranged from 95.9 to 96.6 mm. It would be interesting to know if the brick tolerance is +/- 0.1 or 0.3 mm, either way they were all inside it. But there was still significant difference even with these old bricks.
So I would invite others who think the tolerance has widened to do similar
tests on older bricks to back up their claim.
|
If exact brick size variation tends towards(1) a normal distribution there
will be equal numbers of oversized and undersized bricks. Measuring the
length of a “stack” of bricks will basically cancel out any error.
Scott A
(1) I say “tends towards” as I understand the ND is a theoretical construct.
|
Have you accounted for common mode error? Many bricks are produced by the same
mold. If 12 2x4s are produced per machine cycle, then 1 in 12 will be from the
same part of the mold.
The randomness of the distribution depends how many machines are producing that
shape in that color and how many are produced at a time in the mold.
I think it is likely that all the pieces from one mold will be closer in height
than two randomly chosen pieces, since the mold will be checked for consistent
depth during production. Therefore a thin batch is more likely as a mold
producing 12 thinner pieces will produce 1,200,000 pieces that thin in its
lifetime.
The height of pieces depends on how the machine closes on the mold, since the
machine pushes bricks off the mold as if they were being pulled upwards off
another brick, albeit sideways. See the movie on the Lego Club website features
area.
Therefore the height of a particular brick type is likely to vary slowly over
time, since a whole run in one color for a range of sets will be molded at once
if possible. That’s probably why a stack of plates from Yoda and SoL sets were
all on the thin side. By the time we consumers notice a problem, the mold in
question will have retired long ago - the time between the parts being produced
and us buying the sets. Therefore you have to integrate over a very long time
to find a Normal Distribution.
Mark
|
|
Message is in Reply To:
 | | Re: Quality issue - update
|
| (...) If exact brick size variation tends towards(1) a normal distribution there will be equal numbers of oversized and undersized bricks. Measuring the length of a “stack” of bricks will basically cancel out any error. Scott A (1) I say “tends (...) (20 years ago, 5-Jul-04, to lugnet.lego, lugnet.color, lugnet.general, FTX)
|
26 Messages in This Thread:
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
