[quote=“iskess, post:19, topic:18341, full:true”]
My experience is that an analog airspeed sensor is more reliable, but a digital is more accurate.[/quote]
In my experience this can not be generalized:
I had no I2C issues, but disturbances on the analog line. I2c and analog line issues depend both on cable routing, cable length, shielding and electromagnetical interference. If several sensors are connected to the i2c, the voltage of the inputs / outputs must be the same or you must have an active i2c-hub. The disturbing factors therefore depend on the individual frame structure.
The next source of error, which can seriously affect reliability, is sensitivity to sunlight while airspeed calibration on startup.The 4525Do is very sensitive(the pitot probe was covered):
If the battery is plugged in under other lighting conditions than in flight and the sun can shine brightly on the sensor, the measurement is completely useless, if airspeed calibration on startup is enabled.
On conventional planes I have therefore enabled ARSPD_SKIP_CAL.
At the tiltrotor VTOL we check before each flight ARSPD_OFFSET,
since in the case of incorrectly high values, a stall while forward transition could be the consequence.
The ARSPD_RATIO is, on the other hand, consistently stable.
Only my experiences with 2 analog and four digital sensors.
Alexander, I think you’ll find analog works fine, and it’s much cheaper. I have 5 of them on different airframes. The issues I have seen with the 4525 were on a stock 3DR Aero, not a custom build. I will go back and check since reading Rolf’s post about voltage mismatch, but I like to think that 3DR didn’t screw that up.
I also tried the 4525 on another airframe and found a higher level of noise when sitting on the ground, always giving 2-3 m/s. This is probably in spec but my analog always stays around 1 m/s. Graphing both sensors in flight I had to admit that the 4525 was less noisy, but I decided to stick with them and they have served me well on a variety of airframes for many years.
Now the 5525 is out and it is much quieter than the analog, so I’m going to give digital another try.
Yes, that price is certainly appealing and surprising. It even includes the pitot tube and hoses. I have one on the bench I have been testing. The signal is super clean. One thing that concerns me is the latency of the reported airspeed. When I remove the pressure from the tube, it takes a long time for the airspeed to wind down. It is much less responsive than the other 2 airspeed sensor types. It seems there it some smoothing algorithm in the driver perhaps? This makes me wonder if the noise is being masked by this averaging. This could be a problem in a very dynamic flight situation like a stall.
Rolf, can you corroborate my observation?
Rolf, can you explain more about AIRSPEED_OFFSET and how you use it. I have only ever set the ratio. You say you check it on each flight, does it change on its own, why?
Thank you.
ARSPD_OFFSET is calibrated on startup, unless ARSPD_SKIP_CAL = 0 (default = 1 !!!) The value moves the entire measurement up or down.
(Best, you’re playing on the bench with this parameter to form your own opinion or conclusion about this parameter)
@Rolf how has your experience been with play uavosd? Most importantly, how has your setup experience been (tedious, a nightmare or hopefully effortless)? Do you recommend it for pixhawk? I currently am using the pixhawk 2.1, do you know if it will work on it? Thanks in advance
It’s not in arduplane but with parrot disco (and I think same with ebee). Pitot sensor is connected with only speed tube, not pressure tube. Since it was possible to make the disco working with arduplane, how could it works with only one tube ?
If we replace common pitot tube with this : https://www.thingiverse.com/thing:2328663 it should theorically working. What do you think about it ?
The static pressure is measured at the inlet pin for the tube. If the air pressure there corresponds to the static pressure outside the aircraft, this is not a deterioration but a good simplification
Sensor is in a plastic box under foam.
So why we don’t make the same thing in arduplane assembly ?
(May be it seems a stupid question and sorry for my ignorance in that case but it’s a real question !)
Often the air pressure is lowered inside the fuselage due to the airflow accelerating around the body and Bernoulli’s principle. Sometimes the fuselage has cooling holes in the front allowing ram air in, and creates a higher pressure inside. It just depends on the airframe. To remove this variable we use a 2nd tube to sample the static pressure from the ports that are 90 degrees to the ram air in the pitot tube. If a manufacturer carefully designes a fuselage so the air pressure inside it accurately reflects static air, then no static tube is needed. I’ve installed airspeed sensors like this before, and it often work just fine, but not always.
