long term accelerometer data

[nssid]

Hi,

I need to store data from an accelerometer (sampled at 1000 Hz) for a period of 7 days. However I am not sure how I can store such huge amounts of data. Is there any function in lab view that I could use to filter out data (acceleration) below a certain value and just store a limited data with the associated time stamps ?

I am open to any other suggestions, tips that could help me to store such large data sets.

Thanks,

Sid

[odessy27]

With 1000Hz you are looking at saving about 2.5 GB of data over 7 days. This article should help you. TDMS file format is easier, convenient and efficient. Do you absolutely have to save all the data? If you just want the data just above certain value, just add a small 'grater than or equal to' case and you will save only part of data above certain value. Also, instead of making single big file, I would create new files soon as file size exceeds certain value/ every 12 hour or so. 

  Bottom line is try to avoid storing huge chunks of data unless you absolutely have to. Decimate/ filter the original data to reduce the filesize.

-Nilesh

[GregFreeman]

---

@odessy27 wrote:

With 1000Hz you are looking at saving about 2.5 GB of data over 7 days. This article should help you. TDMS file format is easier, convenient and efficient. Do you absolutely have to save all the data? If you just want the data just above certain value, just add a small 'grater than or equal to' case and you will save only part of data above certain value. Also, instead of making single big file, I would create new files soon as file size exceeds certain value/ every 12 hour or so. 

  Bottom line is try to avoid storing huge chunks of data unless you absolutely have to. Decimate/ filter the original data to reduce the filesize.

-Nilesh

---

We recently did something along these lines but it was 102.4 khz! However, it wasn't over a 7 day period. The large files are fairly managable for tdms but I agree with breaking them up, especially if you are going to FTP them over to the host. The method of checking file size worked fine for us and we just kept a counter of the number of files that had been open. Any time you create a new one because one got too big, we kept the same name but appended the counter (1,2, 3 etc) to the end of the file name. You can do this in whichever what suits your test best, but Nilesh's advice is definitely what I'd go with.

[nssid]

Nilesh, Thanks for the reply. I was thinking of saving the files using different file names (I saw that the write to meas file express vi has an option for this). My problem however is I need to get data from a variety of sensors (4 accelerometers, 2 strain gages and 2 pot's) which brings by data size to be more than 10 GB. That being said, I think the 'grater than or equal to' case structure should help me in filtering out the unwanted data. I will give it a try.