Reading from a GPIB device

[smercurio_fc]

The only reason why the Read would work is if the instrument is continuously sending data. Is it? Some older instruments did this. If so, then you need to perform the read operation a different way. You need to append what you read to a shift register until you've gotten all the characters for a number. So, if the instrument sends the measurements using, say, 12 bytes, you'd need to read at least that much (you will also have some termination character). However, the very first read may be in the middle of the number. I.e., if the value if 12.3456789, the first read may start at the "5" character.

What does the manual say?

[Dennis_Knutson]

I'd also like to know where in the manual it says that command string is supposed to tell the instrument to take a measurement and place it in it's buffer. Very unusual if it did all by itself. The standard for GPIB is to do nothing until it is specifically commanded to do so and that means sending the exact command for taking a measurement - not just configuring it.

There are basic GPIB tutorials on the web.

[LarsUlrich]

Those are 2 good suggestions (though dennis didn't suggest this, merely hinted at it).

I may have a wrong  read function.  I was looking at the driver for the HP 3456A and that is how they set up their read function, and I just assumed that they would work similarly.  I am not exactly sure how it reads data, I'll have to do some reading and get back to you.

Or I may have misinterpretted the GPIB read function.  The instrument takes measurements on its own (i believe it doesn't need a command string for this) but it may need a command to put the data in the buffer, which would then be read by the GPIB read function, which I don't currently have.

[LV_Pro]

If you have the National Instruments Measurement and Automation Explorer (MAX) installed, which is likely, you can experiment with your command strings that you send and the results you get back, before writing any LabVIEW. It sometimes helps, you are solving an equation with fewer variables that way.

[LarsUlrich]

Ok so I did a little fooling around with the VI.  The problem (I think) has to do with the way the data is formatted by the Keithley.

The "data" indicator is the raw, unformatted string as it is returned by the keithley.

N means no error, DCV means volts DC, and then the reading in exponential format.

The fract/exp string to number converter is unable to convert it correctly, because although I get a correct reading from the raw data, the converted number stays at 0 (as does the data written by the write to measurement file function).

My guess is that I need a way to drop the first 4 digits of the raw data string, but I am not sure how to do it.

[LV_Pro · Accepted solution]

Pass the string into the "Match pattern" function found on the "string" palette. It has an input to define what to match, produces three strings; before the match, the match and after the match. The later would be the number if you match on the NVDC or whatever. You can also use that for error checking, if the match isn't found the instrument probably sent an error message.