Coordinates - Horizontal tilt sensor and coordinate transforms

(Posted by ehackett on Nov. 20 2007)

I have a Vector with a horizontal tilt sensor.  I have collected data in ENU and want to convert it back to xyz.  Can I use the Matlab script for coordinate transformations with the horizontal tilt sensor, and if not, how do I need to modify it?

Also, when the data is collected in xyz with the horizontal tilt sensor: the 'X' data is the vertical component, 'Z' data is along the instrument axis (horizontal in this case), and 'Y' data is 3rd orthogonal component, is this correct?  Because in the data, I would expect the vertical component to be the smallest but the 'Z' data is showing the smallest component, which is inconsistent with my expectations b/c the 'Z' data is actually in the horizontal plane (as far as I understand).  Are the X and Z data switched prior to being output to the data file in Vector's with a horizontal tilt sensor?

(Posted by Atle Lohrmann on Nov. 23 2007)

Hi there,

have you checked this drawing?  Vector cable probe

(Posted by ehackett on Nov. 29 2007)

I hadn't seen this, but it does not help me too much.  My Vector is not a cabled version, it is rigid.  The explanation of the cartesian coordinates is consistent with my current understanding that the output 'Z' velocity is acutally the X velocity, but it is not still not clear to me whether I am able to use the same coordinate transformation matrices with the horizontal tilt sensor as is used with the vertical tilt sensors.

(Posted by Atle Lohrmann on Dec. 03 2007)

Hello

The coordinate transform is independent of the orientation of the tilt sensor. So, you can use the same math whether the tilt sensor is vertical or horizontal.

Best regards, Atle Lohrmann

(Posted by ehackett on Dec. 06 2007)

Thanks!  

This is still a bit confusing.  I used a vector with a horizontal tilt sensor, but I needed to mount it upward looking.  (Beggars can't be choosers!)  Therefore my vertical axis was along the body of the instrument which the drawings define as Z.  My Z velocities (easily identifiable as the small ones) were reported in the X data.  This seems to indicate that the x and z axes are indeed switched.

By the way, the vector documentation shows Z positive towards the instrument where the drawing linked above shows z pointing away from the sensor head.  Is this an inconsistency?  Are they both correct?

Hello,

I am not sure if I understand your question correctly, but the drawing shows how to mount the probe in order to align it correctly to the housing if you select to use ENU coordinates as the instrument configuration. If you select to use XYZ coordinates the attachment shows how the coordinate system is defines. The XYZ coordinate system is always defined according to the probe like it is shown in the linked drawing. You may also freely place the probe any way you like if you are using XYZ coordinates but if you are using ENU coordinates you have to make sure that you align it correctly to the instrument housing. I hope this answers your question.

Regarding the sign of the Z axis the linked drawing is correct whereas there is unfortunately an error in the August 2005 version of the Vector manual.

Best regards,
Sven Nylund

Vector cable probe

Let me try to clarify my situation a bit.  I borrowed a vector with a horizontal tilt sensor.  The probe is attached directly to the instrument (no cable).  I deployed the instrument with the probe (as well as the rest of the instrument) looking upward.  I configured the instrument to report in XYZ coordinates.  While inspecting the raw data it is easy to identify the true vertical (assuming the instrument is truly vertical) axis by scale since the data are an order of magnitude smaller than the horizontal axes.  The data stream identified in the header file (*.hdr) as being for X values (the first of the three velocity values in *.dat) contains the data that are an order of magnitude less than the other two axes.

The prior paragraph contans facts.  What follows is speculation.  Since this instrument was intended to be mounted horizontally, if the X and Z axes were switched with respect to the drawing linked above then the X and Y axes would be horizontal axes defined by which direction the instrument is aimed (X direction) and Z would still be a vertical axis.  (This is presuming the instrument is mounted horizontally as intended, with zero pitch and roll.)  This is a "sensible" way for the axes to be oriented.  The question is: Is this true?  If so, this is an undocumented "feature" of the Vector when configured with the horizontal tilt sensor.

Does this clarify my question?

Tony

Dear Tony, sorry but I am afraid that the combination of OI in London, CMTC in Charleston and Easter vacation really slowed us down.

1) The coordinates systems we use are determined by the compass orientation and the need to get the correct transform from XYZ to ENU. As such, the XYZ coordinate systems are not arbitrary but determined by the question "What does the XYZ system have to look like if the compass/tilt sensor transform is to be correct?"

2) In case of the Vector with horizontal tilt, the X axis is in the direction of the canister (postive away from it), the Y axis is to the left in the horizontal plane if the instrument is lying down and the Z axis is up (right handed).  When you mount it vertical and upwards, X will point up and Y/Z will be in the horizontal plane.

Hope this is getting clearer.   Best regards, Atle Lohrmann

Hi Tony,

Sorry for the confusion, the drawing in the Vector manual is correct. It shows the XYZ coordinates for the standard system with fixed stem. This does not imply that the linked drawings in this thread is incorrect since it all comes down to the combination of compass orientation and the transformation matrix in the various configurations we use.

Best regards,
Sven Nylund

Related to this comment, I thought I'd inquire because we have a couple cabled vectors...  The drawing which Tony refers to shows the Z axis in the direction the prongs are pointing, but the manual shows the Z axis (for a fixed-stem system) in the other direction (towards the stem end).  As the m-file posted on this forum explains, the "UP" orientation for a vector will be when the vector is pointing down, but my concern now (after seeing this picture of a cabled probe) is that this (Z axis pointed towards sea surface when probe is pointed down) is only true for the fixed-stem vectors.  Indeed, I can use aquacoord.xls or the transformation m-files to convert the beam to xyz or enu, but to use the transformation m-files posted here, I need to know how to set the statusbit0. For example, the statusbit0 would a "1" if a fixed stem vector is pointed  towards the sea surface (which is the "DOWN" orientation... if I am understanding the script correctly).  Also, because the .xls file is called "Aquacoord", it leads me to believe the xyz axes must be consistent with an aquadopp which would not be the case if the posted drawing was correct .  

Contrary to said drawing (not the one in the manual but the one poster here of a cabled vector), I believe our cabled vectors have Z in the same direction as a fixed-stem vector (Z positive towards cable) because when I compare velocity data from aquacoord and the transformation matrix they agree perfectly (please ignore typos, etc.).  I realize that aquacoord.xls doesn't use the z information, but I believe the transformation would be impacted because T would no longer equal T_org in the script (note R depends on T not T_org).

It would be nice to confirm or reject my suspicion.

Thanks,

J.P.

Previously Sven Nylund wrote:

Hi Tony,

Sorry for the confusion, the drawing in the Vector manual is correct. It shows the XYZ coordinates for the standard system with fixed stem. This does not imply that the linked drawings in this thread is incorrect since it all comes down to the combination of compass orientation and the transformation matrix in the various configurations we use.

Best regards,
Sven Nylund

Hi J.P.

When it comes to the status bit you simply take that from the status field for bit 0 in the sensor data for the data set that you are converting. The instrument measures the up/down orientation at the beginning of each burst (and at the start of a continuous measurement) and sets the status bit accordingly. This will enable you to convert to XYZ coordinates.

The drawing of the cable probe in this thread shows the XYZ system for a horizontal cable probe. A cable probe with the instrument housing calibrated for vertical operation will have the same XYZ system as for the fixed stem. One trick to tell the direction of the Z-axis is that when the instrument indicates UP-orientation (in the sensor status field for bit 0), the Z-axis will point UP as well. Since the X-axis is marked on the probe you will then find the Y-axis from the right hand rule.

Best regards,
Sven Nylund

Sven,

Thanks for your response.  Our situation is a little more complicated because the cabled vectors which we have are being deployed horizontally although they are vertically configured.  We align our probes with the aquadopps being deployed on the tripod.  The real issue is that the pitch, roll, and compass information is really of modest use (i.e., by aligning the head with the pressure case) for the cabled systems (especially pitch and roll)... because they are cabled.  In the future, I think we will purchase fixed-stem systems. 

Anyway, the information you have provided is useful.  Thanks for your help.

Cheers,

J.P.

Previously Sven Nylund wrote:

Hi J.P.

When it comes to the status bit you simply take that from the status field for bit 0 in the sensor data for the data set that you are converting. The instrument measures the up/down orientation at the beginning of each burst (and at the start of a continuous measurement) and sets the status bit accordingly. This will enable you to convert to XYZ coordinates.

The drawing of the cable probe in this thread shows the XYZ system for a horizontal cable probe. A cable probe with the instrument housing calibrated for vertical operation will have the same XYZ system as for the fixed stem. One trick to tell the direction of the Z-axis is that when the instrument indicates UP-orientation (in the sensor status field for bit 0), the Z-axis will point UP as well. Since the X-axis is marked on the probe you will then find the Y-axis from the right hand rule.

Best regards,
Sven Nylund