Weak spots

Up to Velocimeters - single point

"Weak spots"

Posted by Atle Lohrmann at March 22. 2002
For down-looking Vector probes, the presence of a boundary close to the Vector sampling volume may give rise to problems.  This is especially the case if the boundary is hard (rocks, concrete, glass, etc.) and/or the water echo is weak.  

For each velocity range, there are one or two distances that give rise to problems. The existence of these “weak spots” can be identified in the data record by a decrease in the correlation and an increase in the velocity variance. The problem is mostly seen in flumes with a hard bottom but has also been observed in the field, especially at the higher velocity ranges.

For a standard Vector, the distances from the sampling volume to the boundary that should be avoided are:

Range     "Weak spots"
7 m/s:    2 cm  and 4 cm
4 m/s:    3 cm  and 6 cm
2 m/s:    5 cm  and 9 cm
1 m/s:    8 cm  and 20 cm
0.3 m/s   20 cm
0.1 m/s   46 cm
0.01 m/s  312 cm

The distances are approximate and has a vertical extent of about 1 cm


The problem is less acute at the lower ranges and this - by itself - is a good reason to avoid the higher ranges unless needed.  Another way out of the problem is to use side-looking probes; the issue goes away if there are no boundaries in the path of the transmit pulse.

"Weak spots"

Posted by Atle Lohrmann at August 29. 2002
Can you explain how the heights for the weak spots are calculated and why the numbers you provided seem to differ from those given in the nortek manual (pg 18)?

"Weak spots"

Posted by Atle Lohrmann at August 30. 2002
Hi

The weak spots are related to the spatial separation between the pulse pairs transmitted by your velocimeter.  To be more precise, a weak spot occurs when the first pulse hits the bottom as the second pulse goes through the sampling volume. The position is thus deterministic and can be calculated.

The numbers in the manual refer to the NDV whereas the previous posting refers to the Vector.  That is why they are different.

- Atle Lohrmann

"Weak spots"

Posted by Atle Lohrmann at July 28. 2005
For the Vectrino, the corresponding numbers are:

Range        "Weak spots"
4 m/s:       2 cm  and 5 cm
2.5 m/s:    3 cm  and 10 cm
1 m/s:       5 cm  and 12 cm
0.3 m/s:    10 cm  and 23 cm
0.1 m/s:    23 cm and 45 cm
0.03 m/s:  38 cm and 75 cm

The distances are approximate. The vertical extent depends on the bottom composition and is about 0.5 cm for a flat bottom.

Please note that the Vectrino has some degree of self-adjustment.  Also, the numbers are subject to change since are configured with the Vectrino software.  The above values refer to Vectrino software version 1.05

"Weak spots"

Posted by Atle Lohrmann at January 26. 2006
As point of clarification. For lab work (ie in a flume or tank) are the weak spots restrained to a 1 cm vertical extent.

ie if I was using a Vector in a flume, mounted such that there was 5 cm between the sampling volume and the flume bottom, for a velocity range of 0.1 m/s, would this suffer from boundary echo effects? From my reading of the manual and earlier posting, this would only be a problem if the sampling volume was 45.5 - 46.5 cm from the bottom, ie above where I would like to mount the Vector. Is this correct?

Thanks, R Waters

"Weak spots"

Posted by Atle Lohrmann at January 27. 2006
Hi

The vertical extent of the weak spots is a function of the bottom composition.  If the bottom is well defined (e.g. sand) the extent is no bigger than the transmit pulse or about 1 cm.  If the bottom is rough, the vertical extent can be larger.  It is also a matter of the relative strength between the water scattering and the bottom echo - if the water scattering is high the whole issue goes away.


Quote
From my reading of the manual and earlier posting, this would only be a problem if the sampling volume was 45.5 - 46.5 cm from the bottom, ie above where I would like to mount the Vector. Is this correct?

Yes, this is correct.

- Atle Lohrmann

"Weak spots"

Posted by Atle Lohrmann at June 01. 2006
Quote (Atle Lohrmann @ Aug. 30 2002,08:42)
To be more precise, a weak spot occurs when the first pulse hits the bottom as the second pulse goes through the sampling volume.

... because the reflection of the first pulse will interfere with the third pulse, because they both reach the sampling volume @ the same time?

"Weak spots"

Posted by Atle Lohrmann at June 13. 2006
The first ping hits the bottom and the reflected signal of the first ping reaches the sampling volume at the same time as the second pulse goes through the sampling volume.

- Sven

"Weak spots"

Posted by Atle Lohrmann at October 08. 2007
I've become familiar with the problem of weak spots with the traditional downward looking probe.  We're currently looking into testing a sideways looking probe but it could take some time before we actually get it.  I was wondering, in the meantime, would tilting the downward looking probe so that it's 15, 30 or 45 degrees from vertical do much to change the interference of weak spots?  We're working with a gravel bed (D50= 7.1mm).  Would it make a difference if the sample volume is pointed upstream or downstream of the probe?  Thanks.

"Weak spots"

Posted by Atle Lohrmann at October 09. 2007
Hi there

The problem of boundary influence is felt for each individual receiver element.  If you reduce the effect in one beam (and remember they are all tilted about 30 degree relative to the vertical), then it is likely to show up stronger in one of the others.  And for downlooking probes you need all three receivers to correctly measure the velocity.

The answer, as you suggest, to problem with gravel beds (and high velocities) are side-looking probes.  With these you only have to worry about the distance to the side-wall and this is generally much more manageable than looking down into a gravel bed.

Best regards, Atle Lohrmann

Re: Weak spots

Posted by Jonas Vermander at December 10. 2009

Hi

I'm using the Vectrino to measure orbital velocities caused by waves in a wave flume. The probe is put downwards so he measures horizontally. I'm getting proper results, but near the bottom (from 5cm) things go wrong. I've measured with ranges 2.5 and 1, and my closest boundary is about 40cm away. This distance are not in the list of weak spots...?

Is there perhaps a minimum distance i should keep from the bottom in order to maintain correct measurements?

"For a standard Vector, the distances from the sampling volume to the boundary that should be avoided are:"

What exactly is ment here by "distances from the sampling volume", are these distances only in the horizontal direction? (i'm holding my Vectrino Vertical) or also vertical?

I'm hoping you can clear things up.

Thanks in advance, Jonas.

Re: Weak spots

Posted by P.J. Rusello at December 10. 2009

Hi Jonas,

The Vector and Vectrino measure velocities in a volume located approximately 15 cm and 5 cm from the central transducer respectively. Thus the distance measured from the central transducer to the boundary needs to be over 5 cm for a Vectrino to measure velocities since the Sample Volume has a finite height (default is 7 mm, so 3.5 mm on either side of the center of the Sample Volume). Please look in the Vectrino manual for a more thorough description of the Sample Volume and its location. It's also discussed quite a bit on these forums.

The "distances from the Sample Volume" refers to the distance from the center of the Sample Volume to the boundary as measured along a line from the central transducer aligned with the probe head's principal axis. For your situation this would be the vertical distance.

Keep in mind weak spot positions are only a general location and they will vary a bit based on speed of sound and can extend over a range of about 1 cm.

 

P.J.

Re: Weak spots

Posted by Jonas Vermander at December 11. 2009

Hi,

first of all thanks for your quick reply!

Some things still aren't clear though. You're saying that the weakspot distances will be vertical in my case, i would think they are horizontally, since this is the direction the beam is sent out? (My X- and Y-axis are horizontally) Discussing these directions gets a bit tricky, so i made a drawing as attatchment where i've shown my axes. The drawing is a cross section of the wave flume where there is a seperation on the bottom in the middle. What do you mean by the way by principal axis?

The situation drawn gives at certain ranges very high standard deviations and unrealistic velocities.

The horizontal distance from the probe to the boundary is approximately 37cm. The distance from the centre of the sampling volume thus becomes 32. The weak spots are these:

2.5 m/s:    3 cm  and 10 cm
1 m/s:       5 cm  and 12 cm
0.3 m/s:    10 cm  and 23 cm

These values differ quite a lot from 32cm, so i wouldn't expect problems.. Or is it (also?) the vertical distance that matters? This would make sense since this is about 5cm from the bottom.

Thanks in advance,

Jonas.

 

Attachments

Re: Weak spots

Posted by P.J. Rusello at December 11. 2009

Hi Jonas,

In your drawing, what I'm referring to as the probe's principal axis is the vertical or Z axis. This is because everything is built around this axis and it is the dominant dimension of the probe head and in this case probe stem.

The following discussion is all referenced to your drawing so there should be minimal confusion of directions! ;)

The central transducer sends out a pair of pulses which is directed along the Z axis and will encounter the bottom in your drawing. If the first pulse hits the bottom and then passes through the Sample Volume when the second pulse is passing through it, there will be pulse interference. This is what we call a weak spot.

Only the vertical distance matters and I think you are simply too close to the boundary. If you have the Vectrino+ firmware you can start a distance check to measure this distance. If not (and actually even if you do have the + firmware) take a look at the Probe Check output. This will show the return signal strength as a function of distance from the central transducer. The back of the Vectrino manual has examples of what the Probe Check should and should not look like. The one you should check for is to see if the bottom is located too close to the Sample Volume.

P.J.

Re: Weak spots

Posted by Natalie Youngblood at August 20. 2010

Can you clarify the difference between weak spots and pulse-to-pulse interference?  From the research I have done I was under the impression that they were different, however in some the Nortek documentation the terms seemed to be used interchangeably.

 

Thanks,

Natalie

Re: Weak spots

Posted by P.J. Rusello at August 22. 2010

Hi Natalie,

Weak spots and pulse-to-pulse interference are different names for the same thing. Pulse-to-pulse interference seems to be used when discussing profilers, while at Nortek we use weak spots when discussing point velocimeters like the Vector and Vectrino. Regardless, the cause is the same.

To clarify a little further, here are a couple of quotes from earlier in this thread describing the problem:

 To be more precise, a weak spot occurs when the first pulse hits the bottom as the second pulse goes through the sampling volume. The position is thus deterministic and can be calculated.

and

Please note that the Vectrino has some degree of self-adjustment.  Also, the numbers are subject to change since are configured with the Vectrino software.  The above values refer to Vectrino software version 1.05


So, the two pulses passing through the same volume at the same time creates interference, reducing data quality and generally leading to unusable data. This will show up in profiler as a blank area of a one or two bins in the data. I usually see it as reduced return signal strength/amplitude.

Hope that helps clarify things.

P.J.

Re: Weak spots

Posted by Kristina McCann at February 26. 2013

I have data that shows we unfortunately had the vector positioned in a weak spot during once of our first deployments. We were sampling in the range 0.3 m/s (weak spot=20cm) and the probes show distance to bottom was between 19.2-22.4cm. It was a sandy bottom (ocean floor off Oregon) and there is a lot of turbidity in the water, so I was hoping it wouldn't be too much of an issue after reading these posts, but the velocity data is much more variable than the other six months that we have data. Is there anything I can do about this after the fact? Do you feel that with these settings and this environment it shouldn't be much of a problem, and maybe is just coincidental that the data is more variable? Thanks!

Re: Weak spots

Posted by P.J. Rusello at February 27. 2013

Hi Kritina,

  Take a look a the correlation values during the section in question. Weak spots in the Vector (and Vectrino) tend to show up most strongly as drops in correlation. Unfortunately, there's no way to recover from weak spot issues in post processing.

  Given the variability in range to bottom, I owuld imagine there are a few sections of data which are not affected. The weak spot shouldn't be as large in extent as the bottom range you mentinoed. A rough bottom could make the problem better or worse depending on how it wind sup affecting scattering.

P.J.

Re: Weak spots

Posted by Ainhoa MM at July 24. 2017

Hi,

I'm using a Vectrino+ side-looking probe in a flume at my lab. I just realized that many of my measured points are at a distance from the side-wall of the flume equal to the values given for the weak spots according to my velocity range setting (I simply didn't know about this problem and I'm now reading about it, that's why I measured at these points). However, I have good SNR (>15) and COR (>80) values, that's why I didn't suspect any problem when measuring.

The fact of being placed at these distances (for a given velocity range) means that we will always have for sure pulse interference or only that we might have (it's a critical zone) but not necessarily?

Could I have interferences (and so, bad quality data) even if my COR values are high and signal doesn't seem too noisy?

Thanks in advance,

Ainhoa

Re: "Weak spots"

Posted by Clare Reimers at January 06. 2018
Previously Atle Lohrmann wrote:
For down-looking Vector probes, the presence of a boundary close to the Vector sampling volume may give rise to problems.  This is especially the case if the boundary is hard (rocks, concrete, glass, etc.) and/or the water echo is weak.  

For each velocity range, there are one or two distances that give rise to problems. The existence of these “weak spots” can be identified in the data record by a decrease in the correlation and an increase in the velocity variance. The problem is mostly seen in flumes with a hard bottom but has also been observed in the field, especially at the higher velocity ranges.

For a standard Vector, the distances from the sampling volume to the boundary that should be avoided are:

Range     "Weak spots"
7 m/s:    2 cm  and 4 cm
4 m/s:    3 cm  and 6 cm
2 m/s:    5 cm  and 9 cm
1 m/s:    8 cm  and 20 cm
0.3 m/s   20 cm
0.1 m/s   46 cm
0.01 m/s  312 cm

The distances are approximate and has a vertical extent of about 1 cm


The problem is less acute at the lower ranges and this - by itself - is a good reason to avoid the higher ranges unless needed.  Another way out of the problem is to use side-looking probes; the issue goes away if there are no boundaries in the path of the transmit pulse.
We are using a down-ward looking Vector for eddy covariance measurements in a shelf environment over sand.  We attempt to set the sampling volume height to 15 cm above bottom but due to ripples often have a weak spot problem with a velocity range setting of 0.3 m/s.  If we lower the setting to 0.1 m/s, would the trade off be phase wrapping when velocities exceed 0.35 cm/s?  If these can be recognized and corrected for it would seem a better alternative than poor quality data because of the weak spot problem.  I do want to make sure that we are not cutting off the measured velocity range by using the lower setting. Please advise.
 
Clare Reimers
Powered by Ploneboard
Log in


Forgot your password?
New user?