开云体育

Hi Dennis / and Nielsen (Telecom).

@Dennis... Indeed the reference to floating is in the very subject. In fact, Nielsen wrote more about his question on the subject than in the body of the message.

@Nielsen... By "vs floating DUT" I assume you refer to float the DUT with the sole purpose of allowing the use of a single channel, non differential scope input, to take a differential measurement on a circuit (a measurement of the voltage between 2 different nodes on a circuit in which none of them is ground on the DUT).
What you suggest, using 2 channels of the scope, in ADD mode with Channel 2 inverted is, by all means, a much better approach to taking a differential measurement than floating the DUT and is the very reason why most 2 channel scopes have the option to invert channel 2... To allow for differential measurements.
But, when you float the DUT, although dreadful and although inspiring all the caution that Jeff underlined and more (and wait... even more), you accomplish more than just allowing a differential measurement to be taken so, you must be careful in assuming that you can use the differential approach (CH1 + CH2, CH2 inverted) in every case where someone would float a DUT.

For example, the 2 nodes from which you intend to take the differential measure are both, at a DC voltage level that's beyond the scope inputs' maximum (say 500V).
The difference signal that you intend to measure is small so, you won't be able to handle it using 2 x 100:1 probe. In that case you can't actually solve this problem by using alone CH1+(-CH2), because you are supposed to connect each channel's ground lead to the DUT's ground, and that will subject each of the scope's inputs to a voltage that exceeds their maximum.
And no, by no means you can think of not connecting the ground leads and assume the scope would be protected, because if you didn't float the DUT (which is the proposal of the topic to start with), then it's referred to earth's ground as much as the oscilloscope itself is, and even if you don't connect the ground leads, each input will still see 500V.
In that case, sometimes, floating (as dreadful as it is and as caution inspiring as it is) is the only solution *at hand* and using CH1+(-CH2) can't solve it.
I said "at hand" because there are sometimes other *technically-sound and yet safe* alternatives, such as galvanically isolated probes.
There are many other examples of situations that an differential input setup can't accommodate one (or more) of the reasons behind the need to float a DUT.

So, answering to your consideration that "it seems ok", the right answer is: It depends...
It's OK whenever the technical limitation of your differential setup can be respected, given the circumstances of the DUT, be it (CH1 + (-CH2)), or scope with differential input stage, or differential input probe, or galvanically isolated probes.
It's not OK whenever the circumstances of the DUT vs signal being measured would disrespect any of the input's limitations then, sometimes, Floating ***may*** be the only solution at hand.

Rgrds,

Fabio