Vector timing - from binary (.VEC)

We would like to derive the timestamps for each individual sample for the fast-velocity data (Vector velocity data, pg. 39 System integrator guide) via parsing the binary (.VEC) file, the information given on timing, is not completely clear in the post: http://www.nortek-as.com/en/knowledge-center/forum/velocimeters/30181049#63704243

We would like to confirm if we have the correct formula to calculate the 1st time stamp of the Vector velocity data:

T = time, 1st time stamp for Vector velocity data

T_timestamp[VV-DH] = time-stamp derived from "Vector Velocity Data Header" (pg38 System integrator guide)

delay = 1, the second second of the burst, during the second second, no velocity data is collected.

sampling rate = sampling frequency, Hz

formula:

T = T_timestamp[VV-DH] + delay + (1 / (2 * (sampling rate) )

best regards Ed.

Thanks. All clear now. Cheers Ed.

Hi Mary,

 Burst interval is the time between the start of data collection cycles, not the amount of time the instrument is asleep. If you wanted three minutes of data and a ten minute off period, burst interval would need to be 780 seconds. 

P.J.

Hi P. J.

I would like to compare the measurements of the Vector with the Aquadopp HR (results from the bin at the same height where the volume for Vector is), and get the correlation.

I had my vector in bursting mode, with the sampling rate of 32 Hz, burst interval of 600 seconds and sample per burst of 5760, My Aquadopp burst interval was 7 second with the average interval of 2 sec. 

In order to make the outputs the same form, first, I averaged 32 measurements for each second of vector and put the result on the starting second to have the output on 1 Hz. Next, I averaged the data for two seconds  of the vector every 7 seconds and put the result on the starting second (i.e. avg the result of sec1 and sec 2, put it on second 1....)  and repeat this process again  after 7 seconds  to make it similar to the aquadopp output.  I was wondering if this is the right way to make them the same form in order to be able to compare the results. 

Thanks a lot,

Mary

P. J,

I tried to assign the time stamp to the vector similar to the aquadopp as I explain in my last post but that is exactly what I am not so sure about. I would appreciate any more explanations in this regard. I could not find the time stamp in the menual.

What do you mean by averaging the time stamps alongside the velocity data to keep things lined up, don't you make new time series this way that does not exist in the aquadopp?

Also, attached please find a sample of my east velocities over top of one another. I have averaged the 32 Hz of Vector to have one data for each second  to make this plot. I can see almost the same structure but with different averages which might be due to the fact that they are just at the same elevation but not in one point (although close).. 

Thanks for the helps..

Mary

Hi PJ,

Thanks for the comments. I guess the reason I was confused is because I did not get the time stamp for aquadopp. In one post I have found this formulae:

T=T_timestamp+meas load %/(2*sampling rate*100) as the time stamp of aquadopp.

  --- In my setting I had not used burst sampling option, which I guess lead to a diffrent time stamp.Therefore, in my hdr file (which I have attached, there is no sampling rate (just N/A) and the measurement load shows 74% although I specified 100% in the setting)

My thought is, since my meas/burst interval is 7 s and the average interval is 2 s, say we tell it to start at 12:00:00, that means ADP collects the data for 2 s and average that and output it for  time 12:00:00 and then average data starting from 12:00:07 until the end of 12:00:08 and out put it at 12:00:07 and.... Would you correct my thought process?

Also, 

> For such slow velocities, I would make sure to pay a lot of attention to the Vector data quality and see how much screening the velocities changes the statistics.

For checking the data quality in vector:  we define horizontal/ vertical velocity range in the planning. In the manual it mentioned that sampling at 16 Hz will provide data with an uncertainty of 1% of the velocity range. so I think the lower I specify the velocity range the lower error I will get. But where can I find the uncertainty associated with 32 Hz?-- and for Aquadopp, I am not sure how I can check the quality?

Hopefully my questions are not too confusing,

Thanks a lot,

Maryam