302 wind calculation

On Mar 29, 10:55*am, mattm wrote:
On Mar 29, 12:31*pm, T8 wrote:



On Mar 29, 11:29 am, Darryl Ramm *wrote:

For the 180 degree wrong problem, any system that uses track and TAS/
GS difference calculations risks being 180 degrees wrong it it just
has that difference on two reciprocal tracks. If you run straight down
a ridge and do a rapid turn back the other way and fly the same track
back the other way there are two perfectly equivalent trigonometric
solutions which will give the wind from either direction. The flight
computer has absolutely no idea what direction you are crabbing the
glider into the wind.

I understand the problem of very limited information. *This is a
different problem :-).

In this case picture wind from 360, trip out on ridge with track 270,
303 shows wind from 360 @ 12, relative wind 90 degrees, arrow points
left (i.e. "wind blowing on glider" from right). *On the reverse trip,
having failed to get a wind update, the
computer now shows track 90, wind from 360 @ 12, relative wind 90
degrees, arrow points left. *See the issue? *Even in the absence of
updated wind, the relative wind should be showing 270 based on old
wind, new track. *I verified this behavior again yesterday, turning
through 45 degrees (slowly), watching track change, seeing relative
wind stay constant despite changing track until the device updated the
wind. [edit: I managed to mung the tracks up the first time].

Ignoring the 180 degree wrong wind issue, if you do do +/- 30 degree
or more track changes (not track reversals) do you get wind updates
that agree with a PDA software. What if you reset that PDA calculated
wind before each test does the wind calculated then agree?

No. *On "typical" thermal soaring days with winds under 15, but with a
shear I *know* is there, I won't get a zigzag update on the 303 unless
I alter track +/- 45, and sometimes not even then. *A single 360 turn
always works, but who wants to make such drastic maneuvers? *+/- 30 is
vastly more acceptable and that would do the trick on the older CAI
system. *My old PDA software was WP Pro 9.11. *In general, the wind on
the PDA was preferable to the 303, but less useful because I do my
final glides on the 303 (too many things to go wrong in WP, too hard
to read the PDA at high speed). *XCSoar has the nifty feature that you
can do wind calc by either circling, zigzag or both. *In zigzag only,
it out performed the 303 in yesterday's somewhat unusual conditions
(repeatedly transiting a known shear layer).

"relative wind" that is relative to track (not heading).

Understood, yes.

-Evan Ludeman / T8

Hum, interesting. *A number of flight computers have a mag compass
attachment which allows computing the wind vector in straight flight.
I had
thought the 302 had one of these, too, but now I see that it doesn't!
That
would solve this whole issue. *Did the LNav have that?

-- Matt

A good heading sensor would indeed allow computation of highly
accurate real-time vector winds since most glider computers already
have ground speed, track and true airspeed. 3-axis MEMS gyro
stabilized magnetometers MAY work but multi-antenna GPS derived
heading data would be better if such a device were available. (Google
"GPS heading sensor")

On Mar 29, 9:31*am, T8 wrote:
On Mar 29, 11:29 am, Darryl Ramm *wrote:

For the 180 degree wrong problem, any system that uses track and TAS/
GS difference calculations risks being 180 degrees wrong it it just
has that difference on two reciprocal tracks. If you run straight down
a ridge and do a rapid turn back the other way and fly the same track
back the other way there are two perfectly equivalent trigonometric
solutions which will give the wind from either direction. The flight
computer has absolutely no idea what direction you are crabbing the
glider into the wind.

I understand the problem of very limited information. *This is a
different problem :-).

In this case picture wind from 360, trip out on ridge with track 270,
303 shows wind from 360 @ 12, relative wind 90 degrees, arrow points
left (i.e. "wind blowing on glider" from right). *On the reverse trip,
having failed to get a wind update, the
computer now shows track 90, wind from 360 @ 12, relative wind 90
degrees, arrow points left. *See the issue? *Even in the absence of
updated wind, the relative wind should be showing 270 based on old
wind, new track. *I verified this behavior again yesterday, turning
through 45 degrees (slowly), watching track change, seeing relative
wind stay constant despite changing track until the device updated the
wind. [edit: I managed to mung the tracks up the first time].

I have never noticed behavior like that. I'll look for that next time
I play around. Seems like a simple firmware bug that maybe might be
reproducible. Did Dave Ellis have any comments on this behavior?

Darryl

On Mar 29, 9:45*pm, Darryl Ramm wrote:
On Mar 29, 9:31*am, T8 wrote:



On Mar 29, 11:29 am, Darryl Ramm *wrote:

For the 180 degree wrong problem, any system that uses track and TAS/
GS difference calculations risks being 180 degrees wrong it it just
has that difference on two reciprocal tracks. If you run straight down
a ridge and do a rapid turn back the other way and fly the same track
back the other way there are two perfectly equivalent trigonometric
solutions which will give the wind from either direction. The flight
computer has absolutely no idea what direction you are crabbing the
glider into the wind.

I understand the problem of very limited information. *This is a
different problem :-).

In this case picture wind from 360, trip out on ridge with track 270,
303 shows wind from 360 @ 12, relative wind 90 degrees, arrow points
left (i.e. "wind blowing on glider" from right). *On the reverse trip,
having failed to get a wind update, the
computer now shows track 90, wind from 360 @ 12, relative wind 90
degrees, arrow points left. *See the issue? *Even in the absence of
updated wind, the relative wind should be showing 270 based on old
wind, new track. *I verified this behavior again yesterday, turning
through 45 degrees (slowly), watching track change, seeing relative
wind stay constant despite changing track until the device updated the
wind. [edit: I managed to mung the tracks up the first time].

I have never noticed behavior like that. I'll look for that next time
I play around. Seems like a simple firmware bug that maybe might be
reproducible. Did Dave Ellis have any comments on this behavior?

Darryl

I described the behavior much as I have here. Dave was puzzled, but
concluded that my 302 was not broken.

I am able to reproduce this bug at will.

*All* of the issues I see with the 302 have the look and feel of
software problems.

-T8

On Mar 30, 6:56*am, T8 wrote:

I am able to reproduce this bug at will.

And here it is http://tinyurl.com/ygtwbvf

Ridge on the left, correct wind showing in XCSoar in PDA (12 knots
from 090 relative), 303 showing 4 kts from 333 relative.

The wind calculation performance of XCSoar, using data from the 302,
is absolutely outstanding. It's fast, it's accurate, it doesn't cough
up garbage. The 302/303 is *buggy*.

What's missing in all these schemes is a display of the age of the
vector wind calculation on the main screen.

-Evan Ludeman / T8