A data presentation method for LoLo/RAM/Alti data using LabVIEW analysis software
1.Introduction
LabVIEW is a high end data acquisition package which offers an ‘object’ based coding method. Data is manipulated using a set of library based icons arranged to make array handling, data smoothing, graphing and calculation easy to arrange, store and recall. Visit the National Instruments site for a detailed explanation. As such it is easier to use than the standard recording altimeter packages and much easier to use and modify than Excel. These notes look at a dataset shared by Randy Brust from a set of flights by a buddy with an ‘Encore’ (Figure1) ….. sample rate set to 10s/s.
Figure 1.…….Encore screen dump
2. Objectives
‘Rough’ out a LabVIEW method for reading Ascii tab/space/comma delimited data.
Split the dataset to give individual launches.
Analyse the data for meaning.
Post the result to ‘old hands’ for audit.
3. Methodology
Open LabVIEW routine Altitude2.vi, see Figure 2.
Edit the file path to where you have stored the Ascii file (this is all you have to do!)
The software runs a 5 point moving average smoothing filter (the data looked fairly digital) as the pressure transducer only resolves to 1m
A moving average (aka Olympian averaging) is a standard technique for under sampled or poorly resolved data.
A 5 pt moving average, adds 5 consecutive numbers, finds the average and assigns the result to the time midpoint.
The data is split into individual launches based on a start time T_start where height >0.5m and end time T_end where height< 0.5m.
Each launch is given a consecutive flight number F0, F!,F2.……and saved as Ascii in a ‘derived_data/yearmonthday/ directory
The routine displays the data to show that is has been read correctly.
Open LabVIEW routine Altitude3.vi, see Figure 3.
Edit the ‘start path’ to where the individual launches have been stored.
Edit the ‘Dir’ path to the date you have run the data
Edit the launch files to the ones you wish to look at.
The user may analyse single launches, multiple launches from the same date or run launch comparisons from days 1,2,3.
Run the routine.
The graph shows a single launch height trace.
The software calculates flight duration T_dur from T_start and T_end and finds the maximum launch height and the time for maximum launch height T_max. Time for launch is calculated from (T_max-T_start).
The height data is differentiated once to give height gain/loss as a function of time (instantaneous sink rate) and a 7 point moving average is run as the differentiated data is still fairly digital.
Mean sink rate is determined by cutting the array from T_max to T_end and averaging the result.
Precis data is written to a flight log file.
Open Flight_log.vi, figure 4, for the date in question , print and file
The routine graphs launch time/launch height, histograms the height and Tdur variables.
The whole exercise would take some 15-20mins.
Observations
Apologies to Randall Brusts colleague for picking his day to pieces. He only wanted to get out if the house for a fly, his data ends up in the UK and gets scrutinised in minute detail.
Altitude2
Note that this dataset is compromised by a ‘spike’ shortly after launch. This may be due to air inflow into the fuz, the transducer would be measuring a ‘dynamic pressure head’ or could be an organ pipe resonant condition at that particular velocity. The latter is the favoured explanation as the effect occurs a second or so after launch and not when the launch velocity is at a maximum. The reader should ignore this anomaly as it is the software under evaluation and not the data. (Put the altimeter in a balsa box with a pinhole to fix the issue).
Altitude3
Figure 3.1 graphs a single launch F2, the data show:-
Stall after reaching maximum height at T=7s
Sink rates of 4-6 times mean sink rate at t=28/35/44s, symptomatic of too much down EL on the downwind leg followed by a balloon into wind.
The height loss at T=66s is probably an approach and landing issue.
Little in the way of thermal activity with sink rates negative for 95% of the time
The positive sink rates are likely to be due to elevator pumping and not thermal activity
Suggestion
Cure the stall at maximum height and run some turning trials to perfect the turn and minimise height loss.
Figure 3.2 graphs 5 consecutive launches and the graph zoom functions looks at the first 10 seconds only. The data show:-
All 5 launches show excellent consistency with respect to roll out to climb in the first 1 second of the launch
Climb angle is consistent
All launches are in the range 39-40.5 metres
Suggestion
No suggestions necessary, the model trim is good and the launcher has good climb technique.
Figure 3.3 looks at the same 5 launches but expands the X scale to see the complete flight
In all launches , the model stalls from maximum height, losing 2-3 metres at worst. When does the pilot change his camber trim ?
Launches F2, F3, F4 all have the same mean sink rate and it is tempting to assume that all these flights were without thermal assistance. Should this assumption be correct, then -0.53 m/s is the still air sink rate at his trim, wing loading and wind strength……the following notes assume this is correct.
Launch F5 was much the same as F2,F3, F4 but encountered lift at T=66 s at 6m height to extend the flight some 10s. Pilot did not spot it or didn’t bother to work it.
Launch F6 hit ‘helpful ‘air at T=20s , stayed with it until T= 50s , then ran close to -0.53 m’s mean sink rate thereafter.
There is still evidence in all the flights of an imperfect turn strategy.
Figure 3.4 retains F2 as the reference condition and plots best two flights of the day. The data show:-
Thermal conditions for F9 and F10 were still of the ‘helpful’ and not ‘boomer’ variety.
F9 was abandoned for a re-launch………..strategy worked with a 2.40 in difficult conditions.
I still see evidence of EL pumping in the turns.
Figure 4 shows the flight log in précis form. This is a file which is written by altitude3.vi. The data show :-
There is no correlation between launch time and launch height.
The launcher needs 10 launches to loosen his arm.
Lift was sporadic and conditions difficult
Conclusions
I am not going to make any, but this is the way I shall evaluate my DLG data
Yours is the judgement as to whether this is far too nerdy/obsessive to adopt. (I need all the help I can get)
The software is yours for the asking but you will need LabVIEW to run it…….not cheap.