MATLAB: How to convert a double to a data with the following features

how can i convert a double

The data consists of two bytes per point with no characters between bytes or points. The point data is sent high byte first. The data block has a header which consists of the # character followed by several ascii coded numeric characters. The first of these defines the number of ascii characters to follow and these following characters define the length of the binary data in bytes.

Best Answer

This would get you started:

bindatalength = sprintf('%d', numel(waveform)*2);   %length of binary data (2*numbers of points) encoded in ASCII.
lengthlength = sprintf('%d', numel(bindatatlength); %length of the length string, encoded in ascii , must be only one character
assert(numel(lengthlength) == 1, 'wavelength has too many points');
commandheader  = sprintf('ARB %c%s', lengthlength, bindatalength);

You then need to encode the waveform itself as two byte integer in big endian form. As far as I can tell there is no information provided on how to do that. The way I'd find out what to do would be to send the command ARB1? to read a known waveform and work out the format.

Possibly, the [-1, +1] range is normalised to uint16 [0 65535], if that is the case:

waveformbinary = uint16((waveformArray + 1)*65535)

Or maybe it's converted to signed 16-bit, int16 [-32768, 32767]:

waveformbinary = int16(waveform*32767);

Or maybe it's something else. As said, ARB1? could help you figure it out.

Once you've got the above worked out:

fprintf(yourdevice, commandheader);   %write as ASCII
fwrite(yourdevice, waveformbinary, 0, 'ieee-be');  %write binary data in BIG ENDIAN format (high byte first)
fprintf(yourdevice, '\n');  %terminate command

Related Solutions

MATLAB: Does FREAD read data incorrectly from the instrument when using Instrument Control Toolbox 2.2 (R14SP2)

It is likely that your instrument returns data in binblock format. This means that it comes back with a header that is as follows:

 #XYYY<data>>

where:

the pound sign is literal
the X is an ASCII digit representing the number of bytes in YYY
YYY is the number of bytes in the waveform.
 <data> is the data in binary format, 1 or 2 bytes per value depending on the FORMAT setting.

Note that X is the number of bytes, not values, so if you are transferring in WORD format, the number of bytes will be twice the number of values to read.

For example, with 2500 data points, a header would look like the following:

 #42500<data in binary format>

If you are using a direct FREAD command to read everything (including this header information), then all of your waveform will still be read (except the last few bytes). You will probably find that the first few values of your data are not what you expect and that they are the same every time.

To solve the problem, read the data in as a binblock. There are two ways to do this. One is with the BINBLOCKREAD function in the Instrument Control Toolbox. Simply change

%get data

fprintf(s, 'waveform:data?');
out = fread(s, points, 'int16');

to:

%get data
fprintf(s, 'waveform:data?');
out = binblockread(s, 'int16');

Note: The use of the BINBLOCKREAD function requires Instrument Control Toolbox 1.1 (R12.1) or later.

If you want more control over the read, you can simply do the parts yourself. You can either go through the whole process of interpreting the header, or if you know how many points are going to be returned, precalculate the length of the header and discard those bytes.

For example, for 500000 bytes and a header of:

#6500000

you could use the following code:

fprintf(s, 'waveform:data?');
fread(s, 8, 'int8'); % discard header
fread(s, points, 'int16'); % get the waveform data of known length>

Additional information on binblocks and reading data from your instrument should be available in your hardware documentation.

MATLAB: Reading from Keysight scope to MATLAB and the integer/byte value limit

In case anyone ever encounters a similar issue here is how I solved it: In the end the issue was offset voltage on the scope coupled with v/div setings. For some reason zeroed offset was a major problem. I raised the offset voltage (which basically raises the waveform on the scope) until I started capturing the correct waveform, as well as played a bit with the volt/div settings (to a lesser extent). Weird issue.

Best Answer

Related Solutions

MATLAB: Does FREAD read data incorrectly from the instrument when using Instrument Control Toolbox 2.2 (R14SP2)

MATLAB: Reading from Keysight scope to MATLAB and the integer/byte value limit

Related Question