Data Editing

Is there any way to edit points in recorded data?

We frequently get spikes in data from contactor closures and the like, despite all our best efforts at screening, and need to edit them out before analysis.

We can do this in HBM Glyphworks (GED), but it's a terribly expensive package for just this function when Dewesoft can do practically everything else we need.

The function is to select a section of data with cursors, and either delete all the points completely, not maintaining the time series, set the value of the points to zero, or join the start and end points with a linear ramp or half sine.

This functionality would really shorten our analysis chain and greatly enhance Dewesoft.

Hi,

an option would be to filter out the spikes using an IIR filter or to export the datafile to a .txt file, edit it there and import it back to DEWESoft, but this option isn't very practical.

Is it possible that you send us an example data file and we try to find another solution on it?

Filters don't work for several reasons.

One is that occasionally the spike can lock up a DS-NET (it's something close to EMP), so the section for editing can be 10s or so.

Another is that the spike duration can be long enough to fall into the frequency band of interest.

Exporting and editing is what we do now, and would like to avoid.

I'll see if I can find a bit of data to send you next week.

Here is some sample data.

There is a section where the DS-NET has locked up, which is rather extreme.

More normally, there is just a big spike in the data,

Unfortunately, you can see that the data has a spiky appearance anyway.

It is normally filtered at 20Hz for analysis, this does not remove interference spikes.

We would greatly appreciate the ability to select the section where the DS-NET has locked up, and edit the data as described above.

We are also investigating whether Flexpro has this kind of facility.

Hi,

currently there is no option to select data with cursors and set the data to zero, but you are able to edit out the spikes using math formulas. You can find an example in the datafile "Sample pan data2_edited.dxd" on the ftp server. I attached instructions on how to access ftp server. On the ftp server you can find the file in "temp/Spikes" folder.

In the added formulas I used if functions to check, if the data is higher than a set number. If the number is higher, the output value is equal to the previous value (prev function), if not, than the output equals the input.

Another option is to use time channel as a condition for the if function. I used this option in the "G31_S1_Ch3_Pan_Height_V2" channel.

I hope this information helps you and if you have any more questions I will gladly help you.

Can't seem to access ftp server - maybe our security or traffic volume.

I get the concept though, and had considered a variation.

I was just going to set it to zero, didn't know about prev function, which could come in handy.

Using the time channel is a programmatic method of what I want to achieve - much more precise, but not very quick, and gets a bit lengthy with multiple spikes.

All good learning though, thanks for the help.