Hi,
i have use PicoScope6404B to make measurements with a sampling rate of 312,5MS/s during 0,5s. so that i get about 156MSamples. When i load the converted data in .mat file, i get the Warning: "Unable to read some of the variables due to unknown MAT-file error.". I have try the release 6.6.50 and the latest 6.6.57.
in .csv i get just the first 1MSamples o f my data.
What i did wrong ?
thanks, Bassem
high sampling rate
Re: high sampling rate
Hello Bassem,
We are aware of an issue relating to saving large files to a .mat format.
If you have any items such as filtering turned on, please turn them off and try saving the file again. You can also try the latest version of the PicoScope Beta software from
http://labs.picotech.com/
Are you loading the files into MATLAB or another application?
With csv files there is a limit of 1.048 million samples per channel owing to a limitation in MS Excel.
I hope this helps.
We are aware of an issue relating to saving large files to a .mat format.
If you have any items such as filtering turned on, please turn them off and try saving the file again. You can also try the latest version of the PicoScope Beta software from
http://labs.picotech.com/
Are you loading the files into MATLAB or another application?
With csv files there is a limit of 1.048 million samples per channel owing to a limitation in MS Excel.
I hope this helps.
Hitesh
Technical Support Engineer
Technical Support Engineer
Re: high sampling rate
Hello Hitesh,
also with the latest version i get an error message when loading the converted file in MATLAB.
Please help !
the PicoScope software allows spectrum bins up to 1048576, so that only 2MSamples of my data (about 156MSamples) will be used and the rest will be ignored. Therefore i want to calculate the FFT in Matlab. Furthermore in the spectrum view the frequency axis is unfortunately not logarithmic.
Bassem
also with the latest version i get an error message when loading the converted file in MATLAB.
Please help !
the PicoScope software allows spectrum bins up to 1048576, so that only 2MSamples of my data (about 156MSamples) will be used and the rest will be ignored. Therefore i want to calculate the FFT in Matlab. Furthermore in the spectrum view the frequency axis is unfortunately not logarithmic.
Bassem
Re: high sampling rate
Hello Bassem,
If the original .psdata file is less than 5MB, please e-mail it to support@picotech.com
Otherwise, please upload it to a filestore site or let me know and I can provide you with somewhere to upload it on our filestore site.
Thanks,
If the original .psdata file is less than 5MB, please e-mail it to support@picotech.com
Otherwise, please upload it to a filestore site or let me know and I can provide you with somewhere to upload it on our filestore site.
Thanks,
Hitesh
Technical Support Engineer
Technical Support Engineer
Re: high sampling rate
Hello Hitesh,
because of the high sampling rate the .psdata file is 158MB big. Please tell me, how can i upload it in your filestore site.
thanks for your support.
Bassem
because of the high sampling rate the .psdata file is 158MB big. Please tell me, how can i upload it in your filestore site.
thanks for your support.
Bassem
Re: high sampling rate
Hello Hitesh,
i have already uploaded a test-file with a half of the sampling rate (=156,3MS/s), so that the file is below 100MB
i get this error from Matlab (7.11.0, R2010b):
??? Error using ==> load
Unknown text on line number 1 of ASCII file
test.mat
"".
Error in ==> importdata at 195
out = load('-ascii',FileName);
Error in ==> uiimport>runImportdata at 432
[datastruct, OTextDelimiter, OHeaderLines] = ...
Thank you for your help, Bassem
i have already uploaded a test-file with a half of the sampling rate (=156,3MS/s), so that the file is below 100MB
i get this error from Matlab (7.11.0, R2010b):
??? Error using ==> load
Unknown text on line number 1 of ASCII file
test.mat
"".
Error in ==> importdata at 195
out = load('-ascii',FileName);
Error in ==> uiimport>runImportdata at 432
[datastruct, OTextDelimiter, OHeaderLines] = ...
Thank you for your help, Bassem
Re: high sampling rate
Hi Bassem,
I've picked up the file and the bug does not appear to have been fixed yet.
I will pass on a request for one of our Development Engineers to take a look.
Are you looking to post-process this data in MATLAB? It is possible to write MATLAB scripts that use the dlls in the Software Development Kit in order to capture data. Guidance is provided in the following area on the forum:
http://www.picotech.com/support/forum71.html
Best wishes,
I've picked up the file and the bug does not appear to have been fixed yet.
I will pass on a request for one of our Development Engineers to take a look.
Are you looking to post-process this data in MATLAB? It is possible to write MATLAB scripts that use the dlls in the Software Development Kit in order to capture data. Guidance is provided in the following area on the forum:
http://www.picotech.com/support/forum71.html
Best wishes,
Hitesh
Technical Support Engineer
Technical Support Engineer
Re: high sampling rate
Hello Hitesh,
i have try the export function of a reference waveform in Matlab and it works. Matlab can read the data without any error --> perhaps it is a tip, how your Development Engineers can fix the conversion error of a measured data to .mat file !!!!
i have another question:
as i duplicate my data (156MSamples) to a reference waveform, the .psreference file was just 3MB big and the .mat file contains only 4,2MSamples. The frequency spectrums of the original and the duplicated data in Picoscope are also different (i see peaks in the duplicated data every 370kHz) --> i think the duplicate data are undersampled, isn't it ??
i have try the export function of a reference waveform in Matlab and it works. Matlab can read the data without any error --> perhaps it is a tip, how your Development Engineers can fix the conversion error of a measured data to .mat file !!!!
i have another question:
as i duplicate my data (156MSamples) to a reference waveform, the .psreference file was just 3MB big and the .mat file contains only 4,2MSamples. The frequency spectrums of the original and the duplicated data in Picoscope are also different (i see peaks in the duplicated data every 370kHz) --> i think the duplicate data are undersampled, isn't it ??
Re: high sampling rate
Hi Bassem,
I suspect a reference waveform would be downsampled by the software otherwise there would be too much data to display.
Regards,
I suspect a reference waveform would be downsampled by the software otherwise there would be too much data to display.
Regards,
Hitesh
Technical Support Engineer
Technical Support Engineer
Re: high sampling rate
Hi Hitesh,
i think, the down-sampling operation should be mentioned in your handbook / help !!
Bassem
i think, the down-sampling operation should be mentioned in your handbook / help !!
Bassem
Re: high sampling rate
Hi Bassem,
I have passed on that request and also requested for someone to investigate the bug.
Thanks,
I have passed on that request and also requested for someone to investigate the bug.
Thanks,
Hitesh
Technical Support Engineer
Technical Support Engineer
Re: high sampling rate
Update:
Calling the 'memory' function in MATLAB will display the maximum possible size for an array as well as for all arrays. E.g.:
>> memory
Maximum possible array: 879 MB (9.219e+08 bytes) *
Memory available for all arrays: 1319 MB (1.383e+09 bytes) **
Memory used by MATLAB: 348 MB (3.654e+08 bytes)
Physical Memory (RAM): 3017 MB (3.164e+09 bytes)
* Limited by contiguous virtual address space available.
** Limited by virtual address space available.
This is determined by the PCs resources.
The data is saved as 32-bit float values. What this means is that if the size of an array in the file exceeds the maximum possible size, it will not be loaded. If the total size of the arrays exceeds the memory available for all arrays then it may be possible to load some of the arrays but not all.
A 64-bit version of MATLAB may allow a larger maximum possible array.
Calling the 'memory' function in MATLAB will display the maximum possible size for an array as well as for all arrays. E.g.:
>> memory
Maximum possible array: 879 MB (9.219e+08 bytes) *
Memory available for all arrays: 1319 MB (1.383e+09 bytes) **
Memory used by MATLAB: 348 MB (3.654e+08 bytes)
Physical Memory (RAM): 3017 MB (3.164e+09 bytes)
* Limited by contiguous virtual address space available.
** Limited by virtual address space available.
This is determined by the PCs resources.
The data is saved as 32-bit float values. What this means is that if the size of an array in the file exceeds the maximum possible size, it will not be loaded. If the total size of the arrays exceeds the memory available for all arrays then it may be possible to load some of the arrays but not all.
A 64-bit version of MATLAB may allow a larger maximum possible array.
Hitesh
Technical Support Engineer
Technical Support Engineer
Re: high sampling rate
Hi Hitesh,
i have tried with the latest 64bit version of Matlab
>> memory
Maximum possible array: 5310 MB
Memory available for all arrays: 5310 MB
Memory used by MATLAB: 666 MB
Physical Memory (RAM): 3551 MB
i could load the generated .mat file in Matlab without any problem, but only if the number of samples is less than 100MSamples. If it is higher than 100MSamples and although enough free memory is available (100MSamples --> 800MB are needed), i get this error message
Error using load
Number of columns on line 3 of ASCII file D:\100MSamples.mat must be the same as previous lines.
That means, the file doesn't contain a rectangular table of numbers, with an equal number of elements in each row !
I think, this is an error in PicoScope, when converting the data in a .mat file
Bassem
i have tried with the latest 64bit version of Matlab
>> memory
Maximum possible array: 5310 MB
Memory available for all arrays: 5310 MB
Memory used by MATLAB: 666 MB
Physical Memory (RAM): 3551 MB
i could load the generated .mat file in Matlab without any problem, but only if the number of samples is less than 100MSamples. If it is higher than 100MSamples and although enough free memory is available (100MSamples --> 800MB are needed), i get this error message
Error using load
Number of columns on line 3 of ASCII file D:\100MSamples.mat must be the same as previous lines.
That means, the file doesn't contain a rectangular table of numbers, with an equal number of elements in each row !
I think, this is an error in PicoScope, when converting the data in a .mat file
Bassem
Re: high sampling rate
Hi Bassem,
Can you please try converting the psdata file to a .mat file using PicoScope 6 Beta please and then loading it into MATLAB?
This can be downloaded from http://labs.picotech.com/
Thanks,
Can you please try converting the psdata file to a .mat file using PicoScope 6 Beta please and then loading it into MATLAB?
This can be downloaded from http://labs.picotech.com/
Thanks,
Hitesh
Technical Support Engineer
Technical Support Engineer
Re: high sampling rate
Hi Hitesh,
unfortunately i get the same error.
100MSamples seems to be a magic limit !!! i assume, it is small error, like an additional number or a carriage return or a tabulator, that disturb the regularity of the ASCII file.
Bassem
unfortunately i get the same error.
100MSamples seems to be a magic limit !!! i assume, it is small error, like an additional number or a carriage return or a tabulator, that disturb the regularity of the ASCII file.
Bassem