Testing
Some notes from various tests...
Lab Night (6/7/2011)
Some mass measurements using scales rated to 180kg... initials used to protect the guilty ;)
Riders with helmet and gear
P 97.1kg (103.6kg with bag)
D 110.6kg
M 72.4kg
J 80.5kg
R 93.8kg
Vespas
P Vespa PX 200 with tools - rear wheel 83.9kg + front wheel 42kg (total 125.9kg)
D Vespa PX 200 with tools - rear wheel 78kg + front wheel 39.5kg (total 117.5kg)
M Vespa PX 200 racer - rear wheel 77.5kg + front wheel 33.8kg (total 111.3kg)
J Vespa GT 200 - rear wheel 84.5kg + front wheel 60kg (total 144.5kg)
Totals
P 97.1 (103.6kg with bag) + 83.9 + 42 = 223kg (229.5kg with bag)
D 110.6 + 78 + 39.5 = 228.1kg
M 72.4 + 77.5 + 33.8 = 183.7kg
J 80.5 + 84.5 + 60 = 225kg
Note: For Labjack + Tablet and another 1Kg (that's included in my bag weight)
M Test
M took the SCII for a whirl - the HP measurements directly from the SCII were still incorrect and the excel spreadsheet show hp readings could be gained but there were issues with resolution - hp readings jumped about and may need to be averaged and more samples taken. e.g. that last part of the graph is probably showing 12hp rather than 15 followed by 8 followed by 16 followed by 10...
Wheel circumference measurements
Sava MC31 = 1360mm
Heindau racer = 1314mm
Working HP Test (8/7/2011)
I finaly sorted out my maths issues and got correct Hp readings!
Now the issue is that I need to adapt the times that I test for speed. At low speed it skips a lot of measurements. I also still have the issue of a high reading followed by a low (suggesting the true reading is somewhere inbetween).
Some more math... currently sampling at 100ms which is 10Hz. 10Hz = 1.360 x 10 x 60 x 60 / 1000 = 48.96 km/hr. So anything lower than that I start missing readings.
Worked out the issue at low speed
The power measurement is only taking place on a change in acceleration, so if the last reading is the same as the previoius one (i.e. the wheel hasn't turned enough to trigger a new reading)- so exactly 0. See below the values to the right are calculated acceleration and power. So the power reading needs to be updated to check for the same value and hold off calculating horse power to the next change and then calculate it for the full range of time.
Fri 8. Jul 16:52:31 2011,3.07086,11.0551,-4.71902,-4.83348 Fri 8. Jul 16:52:31 2011,3.07086,11.0551,0,0 Fri 8. Jul 16:52:31 2011,3.07086,11.0551,0,0 Fri 8. Jul 16:52:32 2011,3.07086,11.0551,0,0 Fri 8. Jul 16:52:32 2011,2.46959,8.89052,-5.51623,-4.70301 Fri 8. Jul 16:52:32 2011,2.46959,8.89052,0,0 Fri 8. Jul 16:52:32 2011,2.46959,8.89052,0,0 Fri 8. Jul 16:52:32 2011,2.46959,8.89052,0,0 Fri 8. Jul 16:52:32 2011,2.46959,8.89052,0,0 Fri 8. Jul 16:52:32 2011,2.46959,8.89052,0,0 Fri 8. Jul 16:52:32 2011,2.46959,8.89052,0,0
Streaming Test (11/08/2011)
OK here are the first runs in Streaming Mode
Mostly good but the odd bad reading (6 bad out of 14,000), so I need to do sanity checking on reads and throw out bad ones
Although LabJack documentation says that timer measurements could be out by 1us, I am seeing more like 100us differences
Speed and hp is effected by these differences and averaging the readings over 5 might be the solution
I suspect though that issues are caused by poor debounced, possible the trigger edge is being caught a slightly different times
Next I will implement the better debounce circuit
Afterthought: This could also be due to the sensor vibrating by ~50um
Above: Blue actual readings for 100km/hr and green what a rolling average of 5 would look like
Above: Blue actual readings vs rolling average of 5 would look like for period
Debounced Streaming Test (23/08/2011)
I have a pretty nice signal now with the Reed Switch Debounce
Signal Generator (09/11)
Worked out the source of errors with a signal generator and help from LabJack support
http://forums.labjack.com/index.php?showtopic=5479&st=0
Update 30/09/11
LabJack support provided a firmware update to test (v140) and it worked!
