"Gary Schafer" wrote
|For the pitch and roll, I don't think an accelerometer is not going to
| do it. It will give you rate of change. You could do some math and get
| degrees but it would not be accurate if the rate was different the
| next time you had a pitch or roll.
|

This is NOT correct. The ADXL is NOT measuring rate of change. It's output
measures the gravitational acceleration directly. I suggest that you read
AMD's data sheet on this product to understand how it works. I have used the
ADXL to measure the inclination of a telescope tube with some success.

Regards,

Vic
--
__________________________________________________ ______

Victor Fraenckel - The Windman
vfraenc1 ATSIGN nycap DOT rr DOTcom
KC2GUI

Home of the WindReader Electronic Theodolite
Read the WIND

"Victory at all costs, victory in spite of all terror, victory however long
and hard the road may be; for without victory there is no survival."
- Winston [Leonard Spencer] Churchill (1874 - 1965)
|

| Have you considered integrating the signal? I'd suspect that on average,
| "up" would be the 'average' of the roll and pitch over a period of time.
| Interesting engineering problem.

The ADXL202 is a two axis device. It measures the gravitional acceleration
vector on mutually perpendicular axes defined by the top face of the IC.
Check the data sheets at Advanced Micro Devices' web site to see how this
works.

I routinely integrate the measurements on both axes by making 10 readings on
each. This does nothing to compensate for the skewing of the ADXL's axes
relative to the axes of the device that it is mounted on. I believe this to
be a geometry problem.

HTH

Vic

--
__________________________________________________ ______

Victor Fraenckel - The Windman
vfraenc1 ATSIGN nycap DOT rr DOTcom
KC2GUI

Home of the WindReader Electronic Theodolite
Read the WIND

"Victory at all costs, victory in spite of all terror, victory however long
and hard the road may be; for without victory there is no survival."
- Winston [Leonard Spencer] Churchill (1874 - 1965)

Dost thou not know, my son, with how little wisdom the world is governed?
-Count Oxenstierna (ca 1620) to the young King Gustavus Adolphus

|

Well, it IS pretty close to correct. An accelerometer measures exactly what
its name implies - acceleration. And 99% of the time the name implies
linear acceleration. So if you integrate that once, you have velocity. But
nowhere in this process do you get pitch, roll, azimuth, or any other
angular measurement. In order to get angular measurement, you need either a
pendulum, or some sort of gyro-stabilized platform.

In all cases with which I am familiar, the problem involves instruments
which are fairly expensive. The least expensive would be the sort of
instruments which are used to display attitude in aircraft. Start with the
needle and ball and all the implementations of artificial horizons that have
evolved over the years. The most accurate and stable over long term would
be some sort of inertial navigator. There may be newer, solid state gyros
which have evoved over the last few years, but I doubt that cost has
decreased significantly.

"Vic Fraenckel" wrote in message
...
"Gary Schafer" wrote
|For the pitch and roll, I don't think an accelerometer is not going to
| do it. It will give you rate of change. You could do some math and get
| degrees but it would not be accurate if the rate was different the
| next time you had a pitch or roll.
|

This is NOT correct. The ADXL is NOT measuring rate of change. It's output
measures the gravitational acceleration directly. I suggest that you read
AMD's data sheet on this product to understand how it works. I have used
the
ADXL to measure the inclination of a telescope tube with some success.

Regards,

Vic
--
__________________________________________________ ______

Victor Fraenckel - The Windman
vfraenc1 ATSIGN nycap DOT rr DOTcom
KC2GUI

Home of the WindReader Electronic Theodolite
Read the WIND

"Victory at all costs, victory in spite of all terror, victory however
long
and hard the road may be; for without victory there is no survival."
- Winston [Leonard Spencer] Churchill (1874 - 1965)
|

On Wed, 07 Jan 2004 02:34:27 GMT, "Vic Fraenckel" wrote:


| Have you considered integrating the signal? I'd suspect that on average,
| "up" would be the 'average' of the roll and pitch over a period of time.
| Interesting engineering problem.

The ADXL202 is a two axis device. It measures the gravitional acceleration
vector on mutually perpendicular axes defined by the top face of the IC.
Check the data sheets at Advanced Micro Devices' web site to see how this
works.

I routinely integrate the measurements on both axes by making 10 readings on
each. This does nothing to compensate for the skewing of the ADXL's axes
relative to the axes of the device that it is mounted on. I believe this to
be a geometry problem.

HTH

Vic

I took a look at the usdigital web site, and their inclinometers seem to fit the bill. For two of
each, ( $700), I wouldn't hesitate for a second if it were a customer job. If a 'home' project,
your appraoch is probably the correct one. Either way, the problem of mounting the device(s) so that
'up' is really up is ongoing. Any ideas how to 'calibrate' up in the field....er, on the boat?
Norm

"maxlynn" wrote in message
news:VpLKb.102178$pY.42460@fed1read04...
Well, it IS pretty close to correct. An accelerometer measures exactly
what
its name implies - acceleration. And 99% of the time the name implies
linear acceleration. So if you integrate that once, you have velocity.
But
nowhere in this process do you get pitch, roll, azimuth, or any other
angular measurement.

??????
If I mount an ADXL with one axis vertical and the other horzontal, I measure
a constant gravitational acceleration on the vert. axis and 0 on the hor.
axis. If I now rotate the sensor, I get a nice sine/cosine relation between
the signals on both axes. So there's my angular measurement!

Meindert

On Tue, 06 Jan 2004 20:27:08 GMT, "Vic Fraenckel"
wrote:

You might also explore U.S. Digital http://www.usdigital.com/ as they have
an inclinometer that might do the trick. Check out the T6 unit. I have used
their shaft encoder products and am seriously considering the T6.

Hey, nice products. The A2T looks like it might fit the bill for me, and the
serial interface they use means I can hook a few of them into the 1 serial
port!

Only question is the dynamic performance - 250mSec settling time is a bit
slow.

If we have 30° roll with a period of 10 seconds (pretty severe, but it might
happen), then that is 3 degrees per second. I guess that's still better than
1 degree accuracy - just. At that price I might grab a couple of them & have
a play regardless, while waiting for the Hippy price to come back to me!

Thanks for all the help guys - I'll let you know how it works in the real
world as we are doing these tests for a project with DNV, so the results
should be public.

Dave

Norm,

| I took a look at the usdigital web site, and their inclinometers seem to
fit the bill. For two of
| each, ( $700), I wouldn't hesitate for a second if it were a customer
job. If a 'home' project,

I suggest that you check out the U.S. Digital T6 incremental inclinometer
($82). I believe it would do the job for you, especially since it has an
indexing feature. The absolute models is way to expensive. Also check out
their ICs for interfacing the encoder to a micro - perhaps the 7166. This is
the way I will probably go if and when!

Vic
--
__________________________________________________ ______

Victor Fraenckel - The Windman
vfraenc1 ATSIGN nycap DOT rr DOTcom
KC2GUI

Home of the WindReader Electronic Theodolite
Read the WIND

"Victory at all costs, victory in spite of all terror, victory however long
and hard the road may be; for without victory there is no survival."
- Winston [Leonard Spencer] Churchill (1874 - 1965)

Dost thou not know, my son, with how little wisdom the world is governed?
-Count Oxenstierna (ca 1620) to the young King Gustavus Adolphus

maxlynn wrote:
| Well, it IS pretty close to correct. An accelerometer measures exactly
what
| its name implies - acceleration. And 99% of the time the name implies
| linear acceleration. So if you integrate that once, you have velocity.
But
| nowhere in this process do you get pitch, roll, azimuth, or any other
| angular measurement. In order to get angular measurement, you need either
a
| pendulum, or some sort of gyro-stabilized platform.
|

All I can suggest is to go to the AMD web site and get the data sheets and
app notes and find out how the ADXL line of accelerometers work!

HTH

Vic

--
__________________________________________________ ______

Victor Fraenckel - The Windman
vfraenc1 ATSIGN nycap DOT rr DOTcom
KC2GUI

Home of the WindReader Electronic Theodolite
Read the WIND

"Victory at all costs, victory in spite of all terror, victory however long
and hard the road may be; for without victory there is no survival."
- Winston [Leonard Spencer] Churchill (1874 - 1965)

Dost thou not know, my son, with how little wisdom the world is governed?
-Count Oxenstierna (ca 1620) to the young King Gustavus Adolphus

What you say is true in a fixed one-g environment. You can derive a crude
measurement of angular rotation in any fixed, uni-directional acceleration
environment in the manner that you suggest. But as I understand the
proposed application, the accelerometer(s)/instruments are to be mounted in
a dynamic environment. The accelerometers mentioned are designed to measure
linear acceleration. That is what the manufacturer designed them to do.
They are NOT designed to measure rotation or rotational rates. That is
generally a domain reserved for gyroscopes and related insruments or
systems.

As an exercise, consider how you would perform your suggested computation in
an orbiting spacecraft(i.e., zero g).

Max Lynn

"Meindert Sprang" wrote in message
...
"maxlynn" wrote in message
news:VpLKb.102178$pY.42460@fed1read04...
Well, it IS pretty close to correct. An accelerometer measures exactly
what
its name implies - acceleration. And 99% of the time the name implies
linear acceleration. So if you integrate that once, you have velocity.
But
nowhere in this process do you get pitch, roll, azimuth, or any other
angular measurement.

??????
If I mount an ADXL with one axis vertical and the other horzontal, I
measure
a constant gravitational acceleration on the vert. axis and 0 on the hor.
axis. If I now rotate the sensor, I get a nice sine/cosine relation
between
the signals on both axes. So there's my angular measurement!

Meindert

Dave Baker wrote:

Anyone know of any pitch/roll sensors that are designed specifically for
measuring pitch & roll in vessels? The units that I have found apparently are
more for static applications, and are affected by acceleration, so are not
accurate in dynamic situations.

Try this for a start. There are several manufacturers of electrolytic
sensors and I am sure you can get what you need in just about any output
configuration.

http://www.microstrain.com/FAS-E.htm

You might also try and contact BP Oil Shipping, they installed strain
monitoring on their Alaska tankers a few years ago.

Rick