Cable Vector - Horizontal tilt sensor - Orientation

Hi

After some somewhat confusing data was received, I want to confirm that I am right with respect to the current flow and the orientation of the probe.

I have been mounting the instrument with the mark on the case (Presumably the "Z" in Drawing N2008-137) facing up, and the probe facing down. This is different from the cases shown in N2008-137.  I am also assuming the arrows represent flow in the direction of the arrow.

Could you please have a look at the attached figure, which shows the configuration I use and my interpretation of the orientations, and let me know if I am correct (assuming the arrows represent flow in the direction of the arrow).

I am concerned because the figure N2008-137 seems different to the figure in the manual for the fixed head for the Z direction (where positive Z is towards the head for the fixed head and away from the head for the cable probe, and I had not realised that.

My main concern is with the Y direction.

Thanks for your assistance

KP

Hi Kevin,

Could you please provide me the probe number for your instrument? Then I can check the velocity calibration file to be 100% sure.

Best regards

Jonas Røstad

Thanks Jonas

I have two instrumnets.  The probe numbers are N4415 and N4449.

KP

Hi Kevin,

The XYZ coordinates are relative to the probe and independent of whether the Vector points up or down. It is also independent of how the canister is mounted. The definitions are attached. The arrows do represent flow in the direction of the arrow.

The mounting drawing N2008-137 is important when you choose to measure in ENU coordinates. The ENU velocities is a calculation of XYZ velocities, pitch, roll and heading. The pitch, roll and heading sensors are inside the canister.

The standard mounting in drawing N2008-137 was established after the production of your instruments. To get right ENU coordinates mounted the the way N2008-137 says you need new head files. I am happy to update your head files to meet the "new" standard.

Best regards

Jonas

Thank you Jonas.  The diagram that you provided XYZVector.jpg (from the manual) is the interpretation that I have been using.  I will now have to think through other reasons for my rather strange data. 

Is there any advantage in updating the head files, and how is this done? I presume that would mean that the diagram XYZVector wuld no longer be valid.

Hi again Jonas.  I have been thinking more about this problem, and I did a few simple tests and am now convinced that my instruments do not follow the conventions in the diagram XYZVector.jpg (which is the same diagram as in the manual I have). I am attaching a pdf file with a description of what I did, and the results. The X axis is OK.  However, the Y axis is reversed. The output data show +ve values for Y, when the definition diagram shows they should be negative, with the probe pointing down and being towed from right to left.  I have asked 2 colleagues to check my logic, and they agree with my interpretation. Obviously this is of concern to me.  Could you please have a look again and let me know what is going on here.  I will also send a separate attachment with the .vec file which was collected in the simple test. Many thanks.  Kevin.

The .vec file

Hi Kevin,

I do agree that it seems like your y-axis has wrong direction. Could you please check that the probe has got the right beam order.

Do both of your probes behave this way? 

The XYZ definations will change if you choose to go with N2008-137.

-Jonas-

Hi Jonas

Both probes behave in the same way.

I am not sure how to check that the probe has got the correct beam order, except that I am assuming that Probe 1 is the one with the red sleeve (directly opposite the flat area with the probe number on it (see attachment), and the other beams are as described in the definition diagram.  If there is another way to check, please let me know how to do it.

This is obviously a major concern as it impacts on other data that I have collected using the original definition diagram.

Kevin.

Hi again!

Before I go back and check previous data to see what affect this has had on my interpretation, I want to check all possible explanations. I have updated firmware at various times (I think).  Is it possible that the Y axis reversal problem occurred at some point due to a firmware update, and if it is possible, could you please let me know when this might have happened, so that I know when to start re-checking my data?

Kevin.

Dear Kevin,

Thank you for your patience - I have finally figured this out. Sorry that it took it's time. 

It turns out that the instrument is configured to meet N2008-137. As you can see from the attached the y-axis is defined opposite than the definiton in the manual which is confirmed by your bucket test.

The "z" on the canister orientation is not important before using the pitch/roll and heading data.

Best regards

Jonas

Thanks Jonas.  But I really need to know when this happened.  Has the N2008-137 definitions being in place since the instruments were new? I assume the N2008 is a date, so  did the change come to my instruments during a firmware upgrade, and if so, what is the earliest date this could have happened. This information is very important as I (and others who  have used these instruments, and very possibly other users) will need to go back over the data to check our interpretations.  This is not a trivial task as you might imagine!   

Kevin

Hi Kevin,

I cannot find any correspondence of this being changed after shipping of these in August 2005. I am pretty sure this has been this way since they were new. The N2008 is an internal series of numbers saying that the drawing is of a Vector, not a year.

To be absolutely sure the beam to xyz transformation matrix is available in the .hdr in the data. This is from your data for VEH4449

Serial number                         VEH 4449
Transformation matrix            2.6707 -1.3108 -1.3591
                                                    -0.0298 -2.3059 2.3303
                                                   -0.3433 -0.3384 -0.3508

Your validation is the sign of element [2,2] [2,3] [3,1] [3,2] [3,3]

If you check this on your oldest data you should be safe.

Best regards

Jonas

Hi Jonas

This is data collected in 2006

Serial number                         VEH 4449
Transformation matrix                 10939 -5369 -5567
                                      -122 -9445 9545
                                      -1406 -1386 -1437

Although of a different order of magnitude, the signs of the elements are correct, so I guess things have not changed.

I will now go back and check all my interpretations, and this is not a trivial task.  I do wonder however, how many other users are out there who are unaware that the direction definition diagrams in the manual are not correct. It was only by chance that I stumbled upon the N2008-137 diagram, which I went looking for when I had some data that was clearly wrong. Perhaps Nortek should attempt to notify other users somehow.  I think this is quite a serious issue.  Thanks for sorting it out for me.

Regards

Kevin

Hi Kevin,

The difference in magnitude is a just a scaling by 4096.

This type of Vector is not a common one so the users are limited. I do agree that this is a serious issue and we will update our users.

Best regards

Jonas

Hello Jonas,

Maybe it's worth to mention that the drawing N2008-137 is not valid for all Cable probe systems but is only applicable for systems that do have a horizontal tilt.

The cable probe systems that do have a vertical tilt do follow the drawing in the manual.

Best regards

Jos Sneek

Dear  Jonas,

I read the following text and I am keen to verify my problem. I used a cable probe ADV to measure current choosing ENU coordinates but I　did not align my red sleeve probe with the canister. The canister is horizontally mounted and the header was upward looking. The red sleeve probe was about 60 degrees with the canister (see the attached picture. the canister is wrapped by orange tape.). Now, I check my data and I find almost all the heading, pitch and roll are out of range.

Is it because of the orientation of the probe? Can I fix this problem by converting coordinate system back to XYZ? Are my data still meaningful?

Thank you very much. Looking forward to your kind reply.

Best Regards

Li

Previously Jonas Røstad wrote:

Hi Kevin,

 

The XYZ coordinates are relative to the probe and independent of whether the Vector points up or down. It is also independent of how the canister is mounted. The definitions are attached. The arrows do represent flow in the direction of the arrow.

 

The mounting drawing N2008-137 is important when you choose to measure in ENU coordinates. The ENU velocities is a calculation of XYZ velocities, pitch, roll and heading. The pitch, roll and heading sensors are inside the canister.

 

The standard mounting in drawing N2008-137 was established after the production of your instruments. To get right ENU coordinates mounted the the way N2008-137 says you need new head files. I am happy to update your head files to meet the "new" standard.

 

Best regards

Jonas

 

 

The orientation of the Probes.

Dear Jonas,

I checked the .ssl file again just now. I found the "roll out of range" is probably due to strong currents. So now, the problem is simplified to:

Can I fix the impact of the wrong direction of the red sleeve probe? I think it will affect the ENU current velocity.

Thank you very much.

Regards

LI

Dear LI,

The misalignment of the probe is not a problem if you know the exact angle, which is measurable.
The problem is that the pitch and roll is out of range. The response of this sensor is not at all linear when it is out of range. Since the heading is a product of pitch, roll and compass, this will be corrupted as well.
What you can do is to convert this data back to XYZ. XYZ is related to the instrument only and has to be corrected for the 60 deg misalignment after the conversion. I am not sure if this will help you since you then would need to know the orientation of the instrument from other sources than from the Vector to get values related to fixed references as ENU.

Please tell me if there are questions.

Best regards
Jonas Røstad

Dear Jonas,

I am so appreciate for your kind reply. Yes, I had tried to rotate the probe around z axis by a angle. 

I checked the data again. I think the pitch and roll out of range is because of the equipment met strong currents in location 2. Then the equipment is not stable at the bottom. In location 1, the pitch and roll are OK. Please see the following information from .ssl file. In the beginning 1 day, the vector was put in location1. And after that it was pull out of the water and put in location 2 in the second morning.

-------------------------------begin of .ssl file-----------------------------------------

Date and Time             Ecode Scode Level      Description
26/10/2011 06:00:01    00    30    Info       First measurement
26/10/2011 06:50:01    00    31    Warning    Orientation changed to Down (probe pointing Up)
26/10/2011 07:00:01    00    30    Warning    Orientation changed to Up (probe pointing Down)
26/10/2011 18:50:31    00    38    Error      Roll out of range
26/10/2011 18:50:33    00    38    Error      Roll out of range
26/10/2011 19:10:26    00    38    Error      Roll out of range
26/10/2011 19:20:24    00    38    Error      Roll out of range
26/10/2011 19:20:55    00    38    Error      Roll out of range
27/10/2011 05:20:21    00    38    Error      Roll out of range
27/10/2011 05:20:22    00    38    Error      Roll out of range
27/10/2011 05:30:05    00    38    Error      Roll out of range
27/10/2011 05:30:06    00    38    Error      Roll out of range
27/10/2011 05:30:07    00    38    Error      Roll out of range
27/10/2011 05:30:08    00    38    Error      Roll out of range
27/10/2011 05:30:09    00    38    Error      Roll out of range
27/10/2011 05:30:10    00    38    Error      Roll out of range
27/10/2011 05:30:11    00    38    Error      Roll out of range
27/10/2011 05:30:12    00    38    Error      Roll out of range
27/10/2011 05:50:28    00    38    Error      Roll out of range
27/10/2011 07:10:01    00    31    Warning    Orientation changed to Down (probe pointing Up)
27/10/2011 08:10:01    00    30    Warning    Orientation changed to Up (probe pointing Down)
27/10/2011 15:20:01    00    31    Warning    Orientation changed to Down (probe pointing Up)
27/10/2011 17:00:01    00    39    Error      Roll out of range

...........

--------------------------------------end of .ssl file-------------------------------------------------------------------------------

But, in the .ssl file, it shows orientation and the probe has the opposite direction (please see the warnings in the 2 and 3 lines of the .ssl file.).

Is it because we put the probe up while put the direction sensor on canister (shown in N2008-137, there is a "z" on canister) down?(I did not remember how we put the "z" on canister.)

I think I made the mistake with Kevin that I put the probe facing up and the "z "point on canister facing down. In addition, I use ENU coordinates. I think in our ADV version, the direction of probe is defined by the "z" on canister and the orientation is defined by Z axis. My canister "z" is downward, so the probe is defined downward. Then according to the manual, the XYZ coordinates is defined like the right panel in the attached file. Therefore, the orientation is up. Is it?

Then, which  is the positive direction for vertical velocity, the probe direction or the orientation? Will the x and y direction be affected?Now, how could I fix this problem and make my data useful?

By the way, we have a Nortek ADCP, and its final test checklist was lost and I do not know its frequency. Can you please find out the exact frequency for us if I give you its serial number?

Thank you very much.

Regards

Yours

LI

Dear LI,

 

The misalignment of the probe is not a problem if you know the exact angle, which is measurable.
The problem is that the pitch and roll is out of range. The response of this sensor is not at all linear when it is out of range. Since the heading is a product of pitch, roll and compass, this will be corrupted as well.
What you can do is to convert this data back to XYZ. XYZ is related to the instrument only and has to be corrected for the 60 deg misalignment after the conversion. I am not sure if this will help you since you then would need to know the orientation of the instrument from other sources than from the Vector to get values related to fixed references as ENU.

Please tell me if there are questions.

Best regards
Jonas Røstad

 

Dear Li,

To be able to use your data you need to know how the instrument were mounted. The instrument was mounted outside of linearity of the sensors, and in addition moving quite a bit during the deployment. It will be very hard to post process the data into something useful with all the uncertainties with this setup.

Best regards

Jonas