EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: SharpEars on October 19, 2015, 04:56:17 pm
-
I am trying to understand the usefulness of the 3456A's ratio mode for measuring. When is it useful? Assuming I am dealing with two voltage sources that are relatively stable and correlated (i.e., move in unison), is it not exactly the same as taking two voltage measurements and dividing them to find the ratio?
Or, is there more accuracy to be gleaned using ratio mode? It is not clear to me how the accuracy of the ratio mode relates to the accuracy of taking two readings and subsequently calculating their ratio using a pocket calculator (assuming of course no voltage drift between the readings as a given).
For example, say I am building a 10:1 voltage divider (10 V to 1 V). Assuming I have an extremely stable 10 V source, would I be better served using ratio mode than taking two readings at the 10 V DC range (10.00000 V and whatever my 1 V attempt produces at each iteration) and manually dividing them?
-
I can't think of a reason why it would be any more or less accurate. The specs say:
RATIO
...
Accuracy: Total % signal error + total % reference error (same as 0.1V, 1V, or 10V DC volts).
Ratio mode makes three separate measurements and then computes the ratio with the processor. The exact process is described in the service manual in section 8-171. (And not two measurements, as you might expect, because everything is referenced to the VOLTS LOW terminal.)
Ratio is just a convenience function, as far as I can tell. But because the three measurements are done close to eachother, there might be an advantage that there's less of an influence from any longer term external or internal drift, but you say that's not an issue with your setup.
-
As an example, if you were calibrating a power sensor (or thermistor mount), and you had a similar gold standard power sensor with a known correction factor, you could set the 3456A in ratio mode (using a third power sensor for the reference) to indicate the value of the correction factor of the standard power sensor, then replace the standard power sensor with the power sensor under test, and read its correction factor directly from the 3456A.
-
For example, say I am building a 10:1 voltage divider (10 V to 1 V). Assuming I have an extremely stable 10 V source, would I be better served using ratio mode than taking two readings at the 10 V DC range (10.00000 V and whatever my 1 V attempt produces at each iteration) and manually dividing them?
Depends on the spec or if the ratio mode check will do. You can normally use the 24 hour spec for an in-range ratio calculation (measuring 10 V then measuring 1 V, both on the 10 V range, then calculating the ratio).
-
Ratiometric measurements are very common, especially in sensor work. Nobody needs to care about the reference quality in such measurements, and so in general nobody does. If you want meaningfully accurate measurements from such a circuit you need to follow its imprecise reference, and not some high quality one in your instrument.