I have a bit of trouble understanding floating voltages. I thought about quite a few circuits and scenarios involving isolation transformers and floating voltages but something is still off. Here are some points which contain a contradiction to me:

  • An isolation transformer separates the primary from the secondary in terms of current path (such that the current cannot cross between the two sides)
  • In AC, a path is not required (e.g., capacitor plates)
  • Voltage means there is a path for electrons to move as this "pressure" is measured. It makes sense to me that voltage measurement implies that a circuit is actually closed and that electrons are allowed to flow even though it will happen at very low currents "work needed per unit of charge to move a test charge".

With these points in mind why can someone then measure the voltage between isolated circuits where one of them is supposed to be floating? The mere existence of this voltage means there's a path for it to flow. (Here are some examples.)

My own answer would be: even though the circuits are isolated, the voltage is measured because of an incredibly small current (which would also be the maximum current in case of short circuit). This current is caused by a very large resistance. If this is the case, would a sufficiently large voltage for the floating side jump across the air resistance or winding's isolation if close enough?


simulate this circuit – Schematic created using CircuitLab

  • 1
    \$\begingroup\$ Before lightning happens there is a voltage between a cloud and earth - look up electrostatics. \$\endgroup\$
    – Andy aka
    Feb 15, 2021 at 16:25
  • 1
    \$\begingroup\$ Apparently some concepts are missing in the exposure of the problem, in any case, in a transformer, for example, even if a winding is without reference to earth (floating) it will have stray capacitances to earth that will end up defining a voltage. \$\endgroup\$ Feb 15, 2021 at 16:51
  • 1
    \$\begingroup\$ "voltage potential means there is a path for electrons to move". What is "voltage potential"? I would use either "voltage" (meaning the path integral of the electric field) or "potential difference" (meaning the same BUT in the special case where said integral only depends on the endpoints.) The fact that you need a geometric path to define a path integral does not mean that path needs to be a physical conductive path. Far from it . A good voltmeter tries to make the path along which it measures voltage as close to a path in a vacuum as possible. \$\endgroup\$ Feb 15, 2021 at 17:07
  • 2
    \$\begingroup\$ "Voltage means there is a path for electrons to move" - no, what makes you think that? \$\endgroup\$
    – Sim Son
    Feb 15, 2021 at 19:29
  • 1
    \$\begingroup\$ Through a path or with the help of some displacement currents (I think the term is capacitive coupling) for example in-between capacitor plates. Now if there is a potential difference it means it has...potential. By completing the path the electrons will travel. And here lies the contradiction for me: If there's truly a voltage between floating circuits, then by connecting the two there will be some current flowing (it would be absurd there not to be a current: imagine connecting the two circuit aka short circuit and the voltage to remain the same).But if a current exists why are they floatin? \$\endgroup\$ Feb 17, 2021 at 21:09

3 Answers 3


From a physics standpoint, I understand what you are saying. Potential difference is meaningless for two utterly isolated and closed systems, since there can be no forces between them, and one cannot do work on the other. The instant those systems appear in the same universe, though, naturally they do interact, and it becomes possible to compare potentials.

Then it becomes a measurement problem; any device used in an attempt to compare potentials will necessarily perturb the now-hybrid system. The measurement equipment becomes the very thing that defines their potential difference, by providing a path for charges to be exchanged, and work to be done.

In an argument from a practical engineering perspective, though, I will disagree that the presence of a potential difference means that charges are moving, in the same way I disagree that a water tap must be passing water if there's a difference in pressure either side. The tap can be closed, and still have a pressure difference.

Besides, Ohm's law alone, V=IR, is undefined if R=∞, and any potential difference must therefore be a function of other influences, such as the measurement equipment itself or parasitic coupling.

Also, voltage being "work needed per unit of charge to move a test charge" may be true, but it is also "work that a unit of charge could do if permitted to move", even if it's not moving. This is why it's called "potential".

I concede that whatever voltage source or coupling is responsible for any potential difference between the two isolated windings, it is clearly a possible path for current, and for an exchange of energy, but I argue that from a practical perspective, as long as the coupling is loose enough (high enough impedance), systems on either side may be considered "floating" with respect to each other. They are like two independent closed systems in separate universes, a situation which will remain true until the impedance between them becomes appreciably small.

I think it's fair to say that all the work MarcoReps did was to keep inter-winding impedance high enough to be negligible.

  • \$\begingroup\$ Yes thank you. You understood exactly what I meant. It seemed weird to me that isolated circuits could have a defined voltage between them. I was trying to imagine what would happen if in the case of such potential difference someone were to connect the two ends together. Now I get it. \$\endgroup\$ May 9, 2023 at 5:13

When I make medical power supplies we add shielding inside the transformer. Wind the primary first, them a sheet of copper. Maybe a second sheet of copper, then the secondary. The shielding reduces the P to S capacitance current. enter image description here


I recently found a very good answer to my dilemma on the Marco Reps youtube channel. The idea is that between the primary and secondary sides there's capacitance and thus current can actually "leak" into the secondary and then through ground and thus completing a circuit.

The voltage he measured was a wopping 60V between sides. Also current does actually flow through.

Here's the video: https://www.youtube.com/watch?v=9JinSfCKuNQ


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.