Big Push
Posted by K
School is out! All of the stuff that gets in the way of robotics: classes,
homework tests, manditory after school sports, report cards and graduation
ceremonies are done!
Summer 2017
Directed Studies
Posted by K
One of the technology courses had a curve thrown at it. In September a
student's year's project was build and develop a semi-autonomous aerial drone.
Then there was an announcement from the Transport Canada, the federal
regulation body. Drone useage rules would be stricter, and linked to the
criminal code to enable police departments to enforce them. These new rules
made legal private drone use inside Vancouver city limits virtually impossible.
Until the school completes Transport Canada's requirements and recieves
academic exemption status, drones will remain grounded.
The student involved with this aerial drone would switch to working on an AUV.
This one individual would become an inverse tasmanian devil leaving a trail of
creation in his whirling path. Items to his credit were the 3D design and
physical creation of the current tray system, the kill and arm switch controls,
thruster junction box / gimbal arm connector and an independent pitch zeroing
system (mechanics, electronics and programming)!
Spring 2017
Schizophrenia
Posted by K
Spring time, warming weather, summer vacation looms and student activity
bedlam. Organizing meaningful meetings was an exercise in futility due to the
number of students coming and going on school tours or extracurricular
activities. Somewhere in there, the one that was supposed to be the webmaster
was disappeared into a model UN group and the programmers were comitted to a
school tour to Waterloo University for half a month. We're still waiting for
the webmaster to resurface somewhere.
Spring 2017
Get wet already...
Posted by K
Lots of time expended chasing phantoms with the sensors. Time to splash. With
the prototype tray system and basic electronics taped to it, pool time was
booked to check / calibrate the depth sensor. The test would commence without
the propulsion system attached. It took over 28 pounds of scuba weight belt to
convince the electronics hull with Raspberry Pi, OSD-5000 and MEAS transducer
sink. Gingerly the hull was submerged at increasing depths and visually checked
for leaks at each plateau. The collected data looked good from the sensor.
Finally something straight forward!
Winter 2017
Sensor Rework II
Posted by K
The real OceanServer OS5000-USD Compass/Depth Option sensor was tested on the
faculty advisor's ancient decade old netbook running the the manufacturer's
demonstration software. The data stream came through five by five.
The sensor was handed off to the team, and the report, "this thing is
still sending wonky depth data!"
Suprise, fustation... cause?!? What changed? It worked on a Windows XP netbook,
it partially worked on a Linux Raspberry Pi 3. The programmmers fingered Linux
drivers and moved to find and install the latest FTDI Linux drivers. Done,
retest... same result. "Were these drivers the right ones? Perhaps the
second latest drivers?", asked the programmers.
Install those ones. Done, test... SAME result. Revert back to the original
drivers... installation error... unstable operation... oh what a mess. ...
debug the OS or just reinstall? The law of diminishing returns dictates:
reinstall.
In the mean time the question,"Is it a Pi thing?", arose.
New test try running a tty terminal on a MacBook Air. Result: same.
Stumped.
Call Tech Support again. Result: "We're sorry we only support MS Windows
environments." Frown. Walk away for a couple of days.
Thinking about something different, "how long can our batteries run a
Raspberry Pi3?", was the segueway to the solution of sensor conundrum.
The Pi3 has a current consumptions close to 2A. The battery port is rated at
2A... very tight current tolerances. But wait the sensor is also powered by the
Pi. Check out the USB ports on the faculty advisor's ancient netbook. It
supplied 5.2V. Compare to the Pi and MacBook Air: 4.8V. Could the 0.4V drop be
screwing up the 24bit transducer AD converter on the the sensor module?
Experiment attach a powered USB hub to the Pi, attach the OS-5000USD to the
hub and run the code. Wow every thing read perfectly: compass bearing and
depth. Question answered. Problem solution attainable.
Time chewed up: five weeks.
Winter 2017
Sensor Rework
Posted by K
Navigation for the AUV revolves heavily upon the the OceanServer OS-5000USD 3D
compass/Depth module installed in the AUV. This year's programming team wanted
to know why the PID used in RoboSub 2016 was only 80% ... in their opinion. The
compass module was collected from the parts cache for testing. Directional
information from the sensor was immediately available, but there were some
peculiarities in reading depth information. The depth information did not seem
to update. Backtrack to the sensor to the manufacturer demonstration software.
Same result... scratch your head... try it on a different computer... same
result... scratch your head somemore... ask a fresh set of eyes to check...
same result... what the h377?! Contact the manufacturer's technical support.
Email pictures to technical support. "That's not a OS5000USD Compass/Depth
sensor, its an OSD5000USD Compass (only)," was tech support's response.
Wow.
The chain of events. OSD5000USD Compass/Depth sensor removed from 2016 AUV,
test fit onto new tray, put in old box (mislabled), and collected first
OSD5000USD visible for this test: a comedy of errors choreographed by many with
a cost of three weeks.
Fall 2016
Tray
Posted by K
During the 2016 RoboSub competition, one of the largest sources of fustration
was the inability to pull cylindrical component sections out of the hull to
perform servicing and adjustment. Issues with the way the hull was constructed
prevented smooth tranport of the component sections. The change for correct
this was to build a tray-drawer system to hold electronics, batteries and
sensors. Over the next few weeks a prototype built from scrap wood and
furnature hardware. The electonics were removed from the AUV and a trial fit on
the new trays was made.
Fall 2016
Forward
Posted by K
Discussions were held with returning and new robotics team members covering
what worked, what skidded by and what failed in the AUV. That conversation was
followed with the question, "What now and what is possible with the
resources at hand?!"
The main takeaways were: the internal electronics tray system had to be redone,
software reliability on the sensors and control circuits had to be improved and
essentially make do with the minimal number of purchases.
Fall 2016
The Fall
Posted by K
Expenses for the AUV during San Diego tour were settled at the start of the
school year, because the fiscal year end happened April 30, 2016. Then came a
"funny" discovery courtesy of the school administration. The new
budget featured a 58% cut to the robotics program. Merde.
Fall 2016
Reboot 2017
Posted by K
So the return home from RoboSub 2016 was nice. The return flight was smooth.
The parents were happy the chaperones hadn't lost any students. The mid-August
land recovery of the AUV in its road case from the Washington State address was
uneventful. The road case would sit waiting until the start of the 2016-2017
school year.
Summer 2016