When you convert data in a GPIB voltmeter, there are some interesting things that happen. In the process the data takes on many forms, and the way the data takes on those forms and is processed needs to be understood so that you can understand meter accuracy.
The first step in using a GPIB voltmeter is to do the conversion. The input from a voltage source being measured first goes to an A/D converter. That converter will produce an integer that is representative of the voltage being converted. In the figure below, the analog input voltage (shown as a narrow line) produces an integer (and since that could be 8 bits, or 12, bits, etc., we show that with a thick line). Click here if you want to see a simulator for an A/D converter.
The integer produced by the A/D depends upon the input voltage, but it is not usually the case that the integer is equal to the voltage input to the A/D. You need to compute the actual voltage - and that is going to be a floating point number. The voltmeter will need to have some sort of computational capability to perform the computation. That's going to be something like this:
Vcalc = Lower Limit + Count*(Higher Limit - Lower Limit)/(2#Bits)
If the voltmeter is really an oscilloscope with 8 bits, and the lower limit is -4v with an upper limit of +4 volts, you would get
Vcalc = -4 + Count*8/(256)
That is the computation
that would be done on an oscilloscope with an 8 bit A/D, set on 1v/Div,
when there are 8 divisions (4 up and 4 down from the center of the scope).
In many cases, GPIB instruments send data to a computer as an ASCII string, so there has to be some further data manipulation to complete the process and get the data to the computer. The entire operation would look something like that shown in the figure below.
When you work with a GPIB instrument, it is not immediately apparent the many forms the data takes in the process of moving from the input, through the instrument, to the computer. Note the following.