I always fly with automatic rescue, which almost always works. However, I had a crash due to a control error after a candle maneuver. Automatic rescue was set to about 10m and tested before. The helicopter came down vertically and the automatic rescue did not work at all.
I then did a test with another heli with a rescue height of 30 m. Several attempts showed that the rescue worked. After triggering rescue automatically, the helicopter came down to 28 - 23 m, but that was okay. During one flight, however, I was able to reconstruct the failure of the rescue, namely a perfectly vertical nosedive. Here, the rescue did not trigger at all. I had to catch the helicopter manually just above the ground. All altitude data refers to the Spirit GTR altitude sensor. I can rule out incorrect altitude measurement because I can read the altitude values in the Jeti log file. The log file proves that the rescue system triggered normally before, but not during the vertical flight to the ground. I suspect a firmware error. Perhaps the system cannot decide whether to rescue in a normal position or upside down.
I can provide the log file from Jeti. However, I assure you that I have carefully analyzed the situation. The automatic rescue can be recognized in the log data by the fact that the motor current suddenly rises sharply, even though I did not make any control inputs. Normally, the automatic rescue kicks in very, very quickly. In one case, however, it did not kick in at all.
Konrad
Firmware bug in automatic rescue?
Hello, I am not the manufacturer and cannot answer the question of why the system did not activate despite the height being recorded correctly. Nevertheless, here's a little advice: I would never rely on an automatic rescue system – I always have my finger on the rescue switch – yes, it takes a little practice, but it's absolutely reliable even at low heights – I've never had another crash due to pilot error since then – especially with candle figures, you can reach high speeds in the downward phase – that's exactly when I would never rely on an automatic system.
Yes, that's absolutely right. You should never rely on the automatic rescue system. That's why I have a button on the control stick that allows me to trigger the manual rescue system very quickly. I've been doing this since 15 years ago, when I was still flying with Skookum FBL. The automatic rescue system is basically a second line of defense if I react too late.
However, flying advice is not the topic here. Thank you for your input, but please start your own thread if you want to give further advice.
However, flying advice is not the topic here. Thank you for your input, but please start your own thread if you want to give further advice.
Hello,
we are sorry for the inconconvience. Generally it does not matter for Spirit unit in which posiiton model is. The problem is measurement of altitude.
When flying vertically for longer time, resulting pressure is always completely off - either because main blades are at 0° and/or different air circulation inside canopy. Difference is similar as between motor on and off.
Spirit unit has to compensate inverted and non inverted measurement, which is still technically possible. But during free fall maneuver pressure is suddenly very, very different.
Difference could be even 20 meters.
The only thing we could do to improve is adding another compensation algorithm for this kind of maneuvers. In next update we will focus on it. However in any case it is not bug, but physical limitation.
we are sorry for the inconconvience. Generally it does not matter for Spirit unit in which posiiton model is. The problem is measurement of altitude.
When flying vertically for longer time, resulting pressure is always completely off - either because main blades are at 0° and/or different air circulation inside canopy. Difference is similar as between motor on and off.
Spirit unit has to compensate inverted and non inverted measurement, which is still technically possible. But during free fall maneuver pressure is suddenly very, very different.
Difference could be even 20 meters.
The only thing we could do to improve is adding another compensation algorithm for this kind of maneuvers. In next update we will focus on it. However in any case it is not bug, but physical limitation.
Spirit System developer
Hello Tomas,
Your arguments contain a logical error. I am not referring to the actual altitude, but to the altitude measured by Spirit, which I can see in the log file. Differences between actual altitude and measured altitude can depend on aerodynamics. However, if the measured altitude falls below the set rescue altitude, it should react immediately and, after a short delay caused by the servo mechanism (in my case, fast MKS servos), a rescue action should be noticeable. Normally, this is the case. In the case described, nothing happened for several tenths of a second until I performed a manual rescue. I find it dramatic that even a (measured) rescue altitude of 30 m is not sufficient. I kindly ask you to check whether there is a firmware error in the rescue algorithm.
Konrad
Your arguments contain a logical error. I am not referring to the actual altitude, but to the altitude measured by Spirit, which I can see in the log file. Differences between actual altitude and measured altitude can depend on aerodynamics. However, if the measured altitude falls below the set rescue altitude, it should react immediately and, after a short delay caused by the servo mechanism (in my case, fast MKS servos), a rescue action should be noticeable. Normally, this is the case. In the case described, nothing happened for several tenths of a second until I performed a manual rescue. I find it dramatic that even a (measured) rescue altitude of 30 m is not sufficient. I kindly ask you to check whether there is a firmware error in the rescue algorithm.
Konrad
The firmware in that area is identical for many years and was very thoroughly tested already by thousands, possibly already milions of models around the world.
If you are able to provide Spirit realtime log, it would be very helpful. There we can see real circumstances and what is happening.
Also your Spirit settings and photo including Spirit unit can be helpful too.
If you are able to provide Spirit realtime log, it would be very helpful. There we can see real circumstances and what is happening.
Also your Spirit settings and photo including Spirit unit can be helpful too.
Spirit System developer