visa to visadev issue (visadev not getting all data sent)

Grabbing data from a R&S ZNB network analyzer. I have code that works well for grabbing data when the netan is defined as a visa object, however, when trying to rewrite as a visadev object, I cannot grab all of the trace data. I realize a lot of the formatting data, etc, can be done a lot cleaner, forgive me for that right now (wrote this function a LONG time ago).

Function that works (visa object):

function data_to_save = rhode_netan_save_trace_data(ip_addr, channel, trace_name, file_name)
    %rhode_netan_save_trace_data('192.168.10.103',2,'Trc2','clever_name_for_file');
    
    dev_name = [ 'TCPIP::' ip_addr '::INSTR'];
    netan = visa('ni', dev_name);
    
    netan.InputBufferSize = 1000000;
    fopen(netan);
    
    fprintf(netan,'sense:sweep:step?');
    freq_step  = str2num(fscanf(netan));
    
    fprintf(netan,'sense:freq:start?');
    start_freq = str2num(fscanf(netan));
    
    fprintf(netan,'sense:freq:stop?');
    stop_freq = str2num(fscanf(netan));
    
    fprintf(netan,'sense:sweep:step?');
    freq_step  = str2num(fscanf(netan));
    
    freqs = transpose(start_freq:freq_step:stop_freq);
    
    command = [ 'calc'  num2str(channel) ':parameter:select ' '''' trace_name ''''];
    fprintf(netan,command);
    
    %formatted data
    fprintf(netan,':form:dexp:sour fdat');
    
    %get data from the selected trace
    %fprintf(netan,'trac? ch1data');
    fprintf(netan,['trac? ch' num2str(channel) 'data']);
    
    raw_data = fscanf(netan);
    formatted_dB_only_data = sscanf(blah, '%f,');
    
    fclose(netan);

raw_data in the above function returns as a 1x22248 char. formatted_dB_only_data is a 1601x1 double (which is what I expect as I have the network analyzer set to 1601 points).

Now, let's try this with the netan as a visadev object:

function data_to_save = rhode_netan_save_trace_data(ip_addr, channel, trace_name, file_name)
%rhode_netan_save_trace_data('192.168.10.103',2,'Trc2','clever_name_for_file');

dev_name = [ 'TCPIP::' ip_addr '::hislip0'];
netan = visadev(dev_name);
% configureTerminator(netan,"LF")

freq_step = str2double(writeread(netan, "sense:sweep:step?"));
start_freq = str2double(writeread(netan, "sense:freq:start?"));
stop_freq = str2double(writeread(netan, "sense:freq:stop?"));
freqs = transpose(start_freq:freq_step:stop_freq);

%select the trace you want to grab data from
command = [ 'calc'  num2str(channel) ':parameter:select ' '''' trace_name ''''];
writeline(netan, command);

%formatted data
writeline(netan, ":form:dexp:sour fdat");

%get data from the selected trace
data = writeread(netan, ['trac? ch' num2str(channel) 'data']);
% write(netan, ['trac? ch' num2str(channel) 'data']);
% data = read(netan, 1601, "double");

formatted_dB_only_data = str2double(strsplit(data,','));

When using 'writeread' to populate data, it comes back as a string of length (according to strlength(data)) 10898. If I run that command a second time, it comes back with garbage binary data of strlength() == 7, and every time I run it (breakpoint set at 'data = ...'), it will keep returning garbage data of various length. I should also note that if I stop the script and run again, the size of data according to strlength() changes every time (3 runs come back with a strlength of 22247, 22224, and 5355). Obviously this should be returning trace data for 1601 points every time. If I use the write and then read (1601, "double"), I do get a 1x1601 matrix back, but the data is garbage (6.22e-38, 6.54e-43).

Anyone have any thoughts on this? I was really hoping that converting a lot of my functions from visa objects to visadev objects would be more trivial, but I suppose as long as I understand how this works in this example, I can figure it out for other scripts / devices.

Thanks.

[Andrei]

What MATLAB release are you using? A number of issues with visadev were resolved in MATLAB R2022a. If you are not able to upgrade to this release, contact MathWorks technical support for a workaround.

[Scott]

R2022a.

I should note, as shown in the code above, I am not doing "periodic polling" of the instrument that could interupt a concurrent data transfer as mentioned in that bug report. I am simply trying to do reads once I issue a command.

Answer 1

[Scott]

Closing the loop on this:

It seems the bug mentioned by @Andrei Ursache was a good working theory, however this bug was not resolved in the original release of MATLAB R2022a, it was fixed in R2022a_Update_3. After filing a service ticket, I was instructed to download update_3 manually since it is not officially released as of the writing of this post. This is the process to download and install the update. Once I ran the update, MATLAB sucessfully grabbed all of the data from the ZNB.