As for impulse buys, could be the IP from Air-Avionics for their butterfly is available at a reasonable cost for a younger person so inclined.
On Thursday, May 6, 2021 at 10:31:16 AM UTC-7, jfitch wrote:
Air Avionics have been equally circumspect on exactly what they do, and how much each of the 9 axis MEMs sensors contribute to that. Using the magnetometer has some advantages, as the limiting case of straight line flight in steady flow can still produce a wind calculation, also it makes inherent an AHRS which is not possible without it. I'd agree that the result is the most important, I'd pop one in the panel and find out but they are a bit pricey as an impulse purchase - costing somewhat more than an Air Avionics did when they were still sold.
On Thursday, May 6, 2021 at 10:04:42 AM UTC-7, wrote:
On Thursday, 6 May 2021 at 16:54:14 UTC+1, jfitch wrote:
The paper is fairly circumspect on the details of what they are doing.. Their claim that this hasn't been done before seems disingenuous, given that the Air Vario had been out for a number of years producing (or claiming to produce) a similar result. Air's efforts are described in a university published paper from some years ago. Their methods appear to be similar. But I am glad that someone has taken up the challenge that Air Avionics has now dropped. Whether it is better or worse will require installing both in the panel and flying for awhile. The advantages they expound of knowing the true wind in real time are already well known to anyone with an Air Display S who has paid attention to it.
On Thursday, May 6, 2021 at 1:27:24 AM UTC-7, wrote:
On Thursday, 6 May 2021 at 06:48:03 UTC+1, krasw wrote:
On Thursday, 6 May 2021 at 00:15:44 UTC+3, Eric Greenwell wrote:
"The mathematical system theory also answers the question under what conditions
a wind estimation is possible. In the case of a borderline case of an exact
straight flight movement and completely calm air, this is not possible. All
triangles are identical. We (Huang and Meyr) could show, however, that the
random changes in the air movement are sufficient to make the system
"observable.” "
--
Eric Greenwell - USA
- "A Guide to Self-launching Sailplane Operation"
https://sites.google.com/site/motorg...ad-the-guide-1
For wind vector you need two TAS-GS difference/track vectors. If they are identical (same direction of flight), you have no wind solution (as stated in paper). Optimum would be two vectors with 90 degrees difference, and measured in extremely short duration (this is how Oudie "live wind" tries to do it, though I haven't got it working). This would give an ok solution in circling flight, but at the same time the angular speed while circling is not very high, so sensing wind vector at different parts of thermal is rather slow (on steep turn with AIR Glide S gives you basically 20-25 separate wind measurements with 1 sec average). In straight flight Hawk is relying on miniscule differences in flight track. Instead of two wind vector components with 90 degree difference in direction they have vectors that have only few degrees difference on average. I would love to see someone comparing wind indications of Hawk and Glide S someday.
Krasw, if you study the paper you will see that your end point is the author's start question i.e. how can you possibly estimate 3D "wind" at high frequency without heading data input from a magnetometer?
The Air Avionics wind and blue ball netto calculations are dependent on data from an accurately reading magnetometer - and installing one of those free from EM fields is its Achilles' heel, and also the reason I never got one when they were still on sale So, to my mind the Hawk method is definitely new - although I am more interesting in finding out whether it works rather than who did what first.

I thought the paper contained a lot detail on the concept and implementation including the statement:

"Mathematical system theory also answers the question under which conditions wind estimation is possible. In the limiting case of an exactly straight-line flight motion and perfectly calm air, this is not possible. All triangles are identical. However, we could show (Huang und Meyr) that the random changes in the airmass movement are sufficient to make the system "observable"."

From this, if true, one might infer that what would be to the pilot "miniscule differences in flight track" would actually yield plenty data for the system to calculate wind.