measuring very small velocities

Hi. 

We recently deployed a pulse-coherent aquadopp near a lakebed. 

Velocities were very small (a couple of cm/s), so we cranked the pulse distance way up (to 6.5m).  Mean-flow profiles look beautiful.  But we'd like to measure turbulence stats, which requires measuring velocities of about 1mm/s.  Since the ambiguity velocity is only a couple of cm/s, and correlations are very high, I thought the current meter should be able to do this.  But the data file only spits out velocities to the nearest mm/s (we recorded along-beam velocities), which is pretty heavy rounding for this application (leading digits are just zeros).  Is there a way to not round to the nearest mm/s?

Thanks

Steve H.

Have you looked at beam velocity spectra? I've measured similar flows with the HR Profiler and seen a -5/3 slope in temporal spectra, meaning ultimately the velocity scaling might not matter too much.

I'm not aware of a way to change the output resolution from the data conversion utility, but my guess is someone in Norway can answer more definitively. You might ultimately need to write your own conversion utility for this depending on the resolution data is stored at in the binary data file.

P.J.

Hi Steve,

The data stored in the binary file should be in a resolution of 0.1 mm/s, it is probably our conversion utility that rounds the data. We have made new beta version that will  convert data with full resolution, I will send you the link on email.

Other users facing the same problem should first check our Software Download section for HR-Profiler software version 1.07 or later. If version 1.07 has not yet been released send us an email and we will provide you with a beta version of the software via FTP.

Best regards,

Sven Nylund

Thanks very much Peter and Sven!

The new software worked great, reporting velocity to the nearest 0.1mm/s. 

I estimated the rms velocity noise (from the standard deviation of high time and space derivatives) as 0.5-0.7mm/s, so while recording to the nearest mm/s wouldn't have been too bad, there was some small advantage to using the new software. 

Steve H.

Hi Mary,

  This is the ideal application for the HR Profiler. The uncertainty for the standard Aquadopp is much higher and would require significant averaging to measure very slow velocities.

  I don't believe we can report velocities with any finer resolution than the 0.1 mm/s listed above, which is a reflection of the accuracy we expect for this instrument. Please see the following conference paper for information on measuring very slow flows with the HR Profiler: http://www.nortek-as.com/lib/bibliography/internal-wave-generation-in-lakes-with-very-slow

  As a general note, our stated accuracy for pulse coherent instruments is 0.5% of the measured value +/- 1 mm/s.

P.J.

Thanks a lot P.J. 

Actually the resolution I got before was 1 mm/s which was due to using HR-Profiler software version 1.06 so I downloaded version 1.09 from the website and I am getting  0.1 mm/s resolution now!

>>>   As a general note, our stated accuracy for pulse coherent instruments is 0.5% of the measured value +/- 1 mm/s.

>>>   As a general note, our stated accuracy for pulse coherent instruments is 0.5% of the measured value +/- 1 mm/s.

I made a mistake. The *accuracy* of the HR Profiler is as stated in the brochure, 1% of the measured value, with the offset ideally a function of the ambiguity velocity but as listed +/- 5 mm/s. We'll hopefully be updating the specifications for this in next few months to better reflect our expectations.

So do you mean if the measured velocity is 2 mm/s,  accuracy would be 0.01+/-1 mm/s? 

 

how about the 0.1 mm/s accuracy that we expect?