Matlab import fails when measuring range is exceeded

Hi,

we use the picoscope in a triggered environment. Sometimes our measurement runs ot of the range we set up. When this happens, we cannot open the exported matlab files anymore. Also ASCII is not usable because it creates "infinity" instead of a number in the columns. Is there a way to get the last measured value instead of infinity?
Currently we use Picoscope 6.6.?? in any newer version we cannot open the exportet matlab files at all with our signal analysis software.

Thank you for your Help.
Jones

The last "working" version is 6.6.23.2.

Hi Jones,

Could you please e-mail a psdata file to support@picotech.com so that we can attempt to recreate the issue?

Are you loading the files directly into MATLAB or another software application?

Regards,

Hi Hitesh,

I sent a sample file to you.
we load the files into Famos a signal processing software. Currently i am trying out different things.

Regards
Jones

Hi,

Additionaly:
With the new versions the import fails with "Invalid Matlab4 or 5 format". With the old version the import succeed but i need to use a band pass so i don't have to zoom in a lot until i can see the signal.

Regards
Jones

Ok one more result of my tests...
i analysed the mat-files in a hex-editor. Difference seems to bee a lot of zero's after each channel. If i delete the last channel by hand than i can load the mat-file.

Hi,

tested the files with octave. It seems that there are additional points added to the curve (zero values) so that the Matlab-file created with new versions add's zeros to 3 of the 4 curves.
In my case:
old file -> 703299 values
new file -> 980398 values
zeroes in the new file: 980398 - 703299 ~ 280000
(the file is getting bigger around 30%)

In detail if I save more than one channel into a matlab4 file, than every channel except the last one gets zeroes added. But the last channel gets the same size as the other channels. Thats why errors occur when loading the last saved channel. I veryfied this in a hexeditor looking into the headers of each matrix. The channel-matrices have the same length. But only n-1 channels are filled up with zeroes...

I think thats all i can do and test for my part, hope you can fix this issue.

Regards

Hi Jones,

Thank you for the update and additional information.

I've passed on this information via our bug reporting system. Once a resolution is found, it should be incorporated into a future release of the software.

Regards,

I am having the same trouble with saving my files.
I've tried around a little bit with older Versions of Picoscope.
I am now using 6.5.42.0, which I had stored somewhere on my harddisk.
This one works!
The older oldest available Version on picotech.com is as far as I've seen 6.6.57. This doesn't work.

[I used a Picoscope 4227 for the recording of the psdata-Files]

I have the same problem, with Pico scope 2204A version 6.9.16.17. when exporting to *.cvs file.
I missed the export to MATLAB file

EDIT:
I have the same problem with *.MAT files as well.
The psdata file was too heavy to upload, and the .mat file is ignored by the forum and wont upload...
any help about it?

Hi alonohev,

Could you please e-mail support@picotech.com and we will be able to arrange for you to upload files to our filestore site.

Many thanks,

Hi
While sending the file I have noticed that the problematic file had 0 files size. I have regenerated the *.mat files again, but got the same error with the same file.
All other data was uploaded with no problem
I'll work with *.mat files from now on.
Alon.

Hi,

This is being handled via an e-mail support ticket.

There is a bug relating to missing data from a psdata file which will be investigated.

Regards,