Maker Pro
Maker Pro

peizo rate gyro drift

H

H.J.

Jan 1, 1970
0
Has anyone seen this amazing project where a small plane is supposed to fly
across the atlantic freaking ocean this week?

http://tam.plannet21.com/FAQs.htm#look

I noticed the guidance involves a Piezo Gyro. How do you think they keep
control of the gyro's drift due to temperature, natural gyro deficiencies
etc?

What I'm most interested in is the roll axis. I mean there is no dihedral
(wing bent-upward angle) to keep the plane from rolling, so a gyro must be
responsible for keeping the plane right-side-up in the roll axis for hours
and hours.

But dont those cheap piezo circuits have a slow drift? i.e. the plane is
level, but the circuit shows a slow rolling rotation at the rate of x
degrees per second? After a few hours (or even minutes) there would be
absolutely no accurate way of detirmining whether reported rotation rates
were drift or actual physical rotation.... how do you deal with this piezo
gyro drift in a system like this?
 
J

Jon Elson

Jan 1, 1970
0
H.J. said:
Has anyone seen this amazing project where a small plane is supposed to fly
across the atlantic freaking ocean this week?

http://tam.plannet21.com/FAQs.htm#look

I noticed the guidance involves a Piezo Gyro. How do you think they keep
control of the gyro's drift due to temperature, natural gyro deficiencies
etc?

What I'm most interested in is the roll axis. I mean there is no dihedral
(wing bent-upward angle) to keep the plane from rolling, so a gyro must be
responsible for keeping the plane right-side-up in the roll axis for hours
and hours.

But dont those cheap piezo circuits have a slow drift? i.e. the plane is
level, but the circuit shows a slow rolling rotation at the rate of x
degrees per second? After a few hours (or even minutes) there would be
absolutely no accurate way of detirmining whether reported rotation rates
were drift or actual physical rotation.... how do you deal with this piezo
gyro drift in a system like this?
This has already been done, by some university researchers, probably
under a NASA
contract or something. I read about it in Air & Space (Smithsonian)
last year.
This appears to be an amateur effort, though. (It took the pros 4 tries
to get a
plane all the way to their target.)

As for gyro drift, there are other instruments that can infer roll drift
and apply
small corrections over time to keep it corrected. If the plane starts
to drift off
course, you can assume it is either gyro drift or crosswind, either way
you apply
a very small correction to roll the other way, and adjust every minute
or so.
The gyro is to keep it stable in the 1 to 10 second intervals.

Jon
 
B

Brooke Clarke

Jan 1, 1970
0
Hi:

I read somewhere in the net that you can use a couple of IR sensing
diodes looking left and right in the daytime to sense roll in UAVs.
If the plane is headed straight an accelerometer would give you an
indication of down as would a 3-axis magnatometer. A combination of all
three of these would make a redundant system.
http://www.pacificsites.com/~brooke/Sensors.shtml

Have Fun,

Brooke Clarke, N6GCE
http://www.PRC68.com
 
O

onestone

Jan 1, 1970
0
H.J. said:
Has anyone seen this amazing project where a small plane is supposed to fly
across the atlantic freaking ocean this week?

http://tam.plannet21.com/FAQs.htm#look

I noticed the guidance involves a Piezo Gyro. How do you think they keep
control of the gyro's drift due to temperature, natural gyro deficiencies
etc?

What I'm most interested in is the roll axis. I mean there is no dihedral
(wing bent-upward angle) to keep the plane from rolling, so a gyro must be
responsible for keeping the plane right-side-up in the roll axis for hours
and hours.

But dont those cheap piezo circuits have a slow drift? i.e. the plane is
level, but the circuit shows a slow rolling rotation at the rate of x
degrees per second? After a few hours (or even minutes) there would be
absolutely no accurate way of detirmining whether reported rotation rates
were drift or actual physical rotation.... how do you deal with this piezo
gyro drift in a system like this?

Similar things have been done, maybe not the atlantic, but I've flown
and driven several automous models. mems gyros can be as stable as 70
degress drift per day, but you don't care about that. The gyro is for
short term corrections only. Tilt angle across the wings, ie 'roll' can
be accurately determined by an accelerometer as well, or by a 3D earth
field magnetometer. I utilise all three for redundancy. My system
combines 3 gyros, a 3D magnetometer, 6 axes of acceleration (up from 4
in the last version) and a GPS. It has proven to be incredibly precise.

Al
 
P

Pat Ford

Jan 1, 1970
0
onestone said:
H.J. wrote:


Similar things have been done, maybe not the atlantic, but I've flown
and driven several automous models. mems gyros can be as stable as 70
degress drift per day, but you don't care about that. The gyro is for
short term corrections only. Tilt angle across the wings, ie 'roll' can
be accurately determined by an accelerometer as well, or by a 3D earth
field magnetometer. I utilise all three for redundancy. My system
combines 3 gyros, a 3D magnetometer, 6 axes of acceleration (up from 4
in the last version) and a GPS. It has proven to be incredibly precise.

Al

Hey Al:
Do you have a website that I could look at? You've mentioned some very
interesting projects
Pat
 
J

Jon Elson

Jan 1, 1970
0
H.J. said:
No, this is much smaller than the 'aerosonde' you (probably) speak of, and a
much more impressive feat.
Yes, the earlier ones were about 3 x bigger, that's true.
which 'other instruments?'
and i think you are speaking of yaw, not roll, like i was. Yaw is easy to
fix, roll is not
No, I did, indeed mean roll. Yaw is not likely to change suddenly in a
conventional
design, powered craft. You might actually not need a yaw gyro, just use
the GPS
and detect the course over a minute or so. Roll and pitch are the axes
that need fast
correction, and a gyro would be pretty much mandatory there. The very
slow drift of
the roll axis would cause a course deviation, and when you think you are
holding the
wings level, due to roll gyro readings only, and you start to get a
course deviation,
then you tell the craft to roll the other way a hair to correct. It
doesn't really
matter whether it is wind or gyro drift, the gyro gives a short term
stable reference,
and you use the course to figure out where the zero is. The same can be
done with
pitch using altitude and airspeed readings.

Jon
 
Top