Twin Tweaks ...
AND WE HAVE OUR WINNER!
72 SERVO 05.2012
cosmic wedgie occurs when the EEPROM doesn’t have
enough time to reset at the beginning of a program.
When this happens, the robot wanders off to never never
land and engages in odd behaviors like we were seeing.
Fortunately, there is an easy fix: All you need to do is reset
This was an easier fix for us than it is for a lot of
folks. All of the chips on our Mark III were socketed
because it was always built as a platform for expansion.
Some folks, however, insist on soldering the OOPic brain
directly into the board. We personally think that sockets
as a general policy are a good thing, but if the brain is
directly soldered to the board it can be removed if you
are a fan of desoldering. The whimsical novelty of solder
suckers and solder braid wear off quickly though, and
thankfully there is another solution for those that
want to reset the EEPROM without having to remove
the chip itself. The Mark III board includes two jumpers,
and when the jumpers are shorted together before
resetting the board and trying to download a new
program, it will be treated the same as if you had removed
the EEPROM. We tried both methods and they both
worked like a charm.
Resetting the EEPROM, however, was only enough to
bring the bot back from never never land temporarily. To
keep the cosmic wedgie from becoming a chronic
condition, we had to treat the cause. The cause,
fortunately, is simple. The treatment for a cosmic wedgie is
to give the OOPic enough time to reset at the beginning of
the program. The consensus is that five seconds is
sufficient, so one of the most important lines of code in an
OOPic program is oopic.delay = 500 at the beginning of the
program. The offending line in our program was easy to
find, and it read oopic.delay = 50. Just as with the IRS or
SEC, being one zero off is indeed a big deal. We added a
zero, reprogrammed the bot, and finally resumed the
We like to include these experiences because we think
dealing with unexpected problems is a part of the process
experienced by all tinkerers. Working with a kit is never as
smooth as following the step-by-step instructions. There will
always be things left unmentioned by the manual; there will
always be unexpected speed bumps on the road to your
Once we were able to test the Mark III with our
ephemeral obstacle, we were surprised that once again
it gave a result markedly different from the Scribbler.
The Mark III consistently spotted the obstacle from a
distance of 36 cm and evaded it accordingly. On one
hand, this was an impressive result for the Mark III.
The sensors were sensitive enough so that they even
spotted our feathery simulacrum of the bush that foiled
TerraMax. On the other hand, too much sensitivity to
ephemeral obstacles could leave the bot feeling trapped
when, in fact, it could plow obliviously through like
We decided that the Mark III’s result was the better