VISA (GPIB) connection - syntax correct but instrument not understanding commands

Hi All! A bit of a problem getting an instrument to understand what I am passing it, any advice would be very much appreciated!

Background:

• I am sending commands to a semiconductor analyzer from Agilent
• I am connecting via a USB-to-GPIB interface

Symptoms:

• A connection tool verifies that I have a successful connection
• a connection tool also provides a console in which I can enter texts commands (the same as the ones I pass through matlab), and this works perfectly.
• when I connect to the instrument via matlab with the Instrument Control Toolbox, using code like the below, the connection seems to be successful (see next bullet), but the commands that I deliver are not.
• When I send a command (via fprintf()), I get no immediate error in the matlab window, but using the console I can query the instrument for errors, and I get a "command not understood" error. So the instrument is receiving something, but it isnt understanding it
• When I query the instrument (via fscanf()), I get a VISA timeout error in the matlab window, and the same error as usual from the instrument.

An example of the code that I use is the following; I believe it is correct syntax because this code and others have been provided by the manufacturer, and because similar code that I have written based on these examples and those in the instrument control API have been verified to be correct syntax by the manufacturer:

%Test code to verify instrument and Matlab connectivity

%Close any old instrument connections
newobjs = instrfind;  %Find all previously connected instruments
if (length(newobjs)~=0)
    fclose(newobjs);
    delete(newobjs);
end

instrument = visa('agilent', 'GPIB0::17::INSTR'); %set variable to instrument
fopen(instrument);  %connect to the instrument 

fprintf(instrument,'*IDN?'); %query instrument
idn = fscanf(instrument) %read from instrument

fclose(instrument); %disconnect from instrument
delete(instrument); %remove instrument from memory
clear instrument idn newobjs A %remove variable from memory

Answer 1

[Darin]

Problem solved. By default (I don't know why), EOSMode is set to 'none', i.e. no terminator string is sent with commands to the instrument, so it just sees a long string. Setting EOSMode to 'read&write' worked perfectly.

code:

instrument=VISA('agilent','GPIB0::17::INSTR')
instrument.EOSMode='read&write'

or

set(instrument, 'EOSMode','read&write')

[Walter Roberson]

Huh, so I had the right basic cause but it was implemented a quite different way!

Answer 2

[Walter Roberson]

Before you fopen(instrument), get(instrument,'Terminator') and see what that says, and see if it matches the end-of-line sequence that the device is expecting.

[Darin]

hm, it says that VISA-GPIB objects dont have the property terminator

[Walter Roberson]

Okay, that turned out to be a false lead on my part.

fprintf() for devices turns out to use %s\n as the default format, with the \n meaning newline. This differs from fprintf() for serial ports, which also use %s\n as the default format, but for those the \n means the Terminator string of the device.

488.2 compatible instruments use \n (meaning newline) as their line terminator, I see now.

Sorry, I do not have any further ideas at this time.

[Walter Roberson]

Ah: the \n for fprintf() for an instrument turns out to mean send the EOS character, http://www.mathworks.com/help/toolbox/instrument/eoscharcode.html . Same basic issue as with serial ports. Except, as Darin located, there is an EOSMode that determines whether the EOS is sent or not; http://www.mathworks.com/help/toolbox/instrument/eosmode.html

Sigh.

Join my Facebook "One World, One Newline" campaign!