6
\$\begingroup\$

I am trying to better understand how ground loops affect a circuit and why it is important to do not have ground loop on a circuit. So I am trying to look at systems with ground loop to further understand and see the effects.

Here is a system with ground loops: enter image description here

As all the BNC connectors of the oscilloscope are connected between each other, there is a ground loop when we used two probes connected to the same ground of a board. Nevertheless, I never see problems on my measures due to ground loops or at least it was not impactful. So why this kind of ground loop has so little effect? And The ground loop with oscilloscope probe is pretty large! So it should have a large effect. So is there something to prevent the ground loop effect into the oscilloscope? Does anyone know when ground loop become relevant into a system?

\$\endgroup\$

5 Answers 5

5
\$\begingroup\$

Try adding a noise source that represents the difference in potential between two points "supposedly" connected to the same earth ground point: -

enter image description here

  • The dotted line represents the actual ground/earth loop that you refer to
  • \$V_N\$ is the potential difference between ground/earth points that are connected
  • They may be connected but, earth is also used by other equipment in an electrical installation
  • Other equipment is outside of your measurement control
  • This can drive a current (can be amps of AC) that flows down your oscilloscope probe wires
  • It screws-up the measurement because there will be a volt-drop down the cable due to its non-zero impedance
\$\endgroup\$
6
  • \$\begingroup\$ Hi, Thank you for your comment. I agree with you. So you think that it would be possible to not have the correct measurement due to large current which could flow into the ground of the probes depending on the environnement of the application. I mean we could see this ground loop as the loop of the squirrel cage into an induction motor. What do you think about this comparaison ? \$\endgroup\$
    – Jess
    Nov 23, 2023 at 10:11
  • 1
    \$\begingroup\$ @Jess the biggest and most annoying culprit in the modern world can be quite high frequency noise from many switch mode power supplies; they use ground to dump their common mode current so that they can pass EMC testing. It gets dumped from live and neutral via EMI suppression circuits and, a lot of spectral content is around 100 kHz. The currents will tend to mostly take the most advantageous path back to solid earth but, a significant fraction will pass via any parallel earth path. In your case that current can flow down the scope probe screen and pass through your "BOARD". \$\endgroup\$
    – Andy aka
    Nov 23, 2023 at 11:16
  • \$\begingroup\$ Also, stuff like motor fault currents can be lethal to measurement circuits because of the high currents involved. They wouldn't normally be present but there can be the odd glitch on some days that disrupt measurement readings that were perfectly normal the day before. I'm not directly answering your comment because I don't exactly understand what you said here: the loop of the squirrel cage into an induction motor <-- the squirrel cage on the rotor is isolated (hence my confusion about what you said). \$\endgroup\$
    – Andy aka
    Nov 23, 2023 at 11:19
  • \$\begingroup\$ Thank you for these precisions. I mean we could think the squirel cage as the ground loop and the 3 phase of the induction motor being the electromagnectic field inducing the current into the squirrel cage. I mean the ground loop of the oscilloscope probes could be like a squirrel cage and the electromagnectic fields of the environnement could be as high as the one produces by the 3 phases of the induction motor. Could it be possible ? \$\endgroup\$
    – Jess
    Nov 23, 2023 at 11:44
  • 1
    \$\begingroup\$ @Jess I think you are thinking too hard about this. It boils down to the need to prevent extraneous external earth currents flowing into and through your target "board" and also disrupting your measurements. A CM choke will not stop those currents getting onto your "board" and returning to earth. Thus they will disrupt your board's operation and possibly damage it. I think you have some kind of scenario in mind that I am not aware of and, if this is so, learn from the answers given and possibly raise a new question on this unknow-to-me scenario. \$\endgroup\$
    – Andy aka
    Nov 23, 2023 at 12:24
5
\$\begingroup\$

Ground loops become more of a problem when there's a current flow. It gets even more serious if the loop has a nonnegligible resistance.

Imagine there are multiple, unnecessary return paths for a power rail on your board, and one of the return paths is also used as a reference (ground pour) by one part of the circuit. If the return paths carry good amount of current then the reference pour will no longer be 0V i.e. it'll fluctuate and that's a problem.

For your case, as long as the BOARD and the scope don't share the same ground before their connection you're fine. Let's assume the SCOPE is powered from mains, so its ground clip will be connected to the EARTH.

  • If the BOARD is powered from non-isolated off-the-line (mains-powered) power supply and uses EARTH or N or L as a reference (e.g. mains-powered buck regulator having bridge rectifier's negative as DC ground) then that can be a problem.
  • If the BOARD is powered from a totally isolated power supply (e.g. battery) you're fine.
  • If the BOARD is powered from isolated off-the-line power supply (e.g. cell phone charger) then you're generally fine.
\$\endgroup\$
3
  • \$\begingroup\$ Hi, Thank you for your comment. Without talking about earth, the BNC of the oscilloscope are all connected together, so if i connect two probes to the same ground. I make a ground loop, whatever the oscilloscope is connected to the earth or not. Also, how can I know if there will be current through the ground loop as it depends on the magnestism into the environnement. The loop is acting as a transformer I mean. \$\endgroup\$
    – Jess
    Nov 23, 2023 at 10:00
  • \$\begingroup\$ @Jess if you worry about the environment (e.g. there's a potential source in the centre of the loop like this) then there's going to be a problem. In this case, use only one probe with its clip, and remove the others'. \$\endgroup\$ Nov 23, 2023 at 10:23
  • \$\begingroup\$ The problem is that I am always worried :D So I just do things that are always "robust" but I do not know if it is still necessary. Thank you for your picture :) \$\endgroup\$
    – Jess
    Nov 23, 2023 at 10:46
4
\$\begingroup\$

So why this kind of ground loop has so little effect ?

Does anyone know when ground loop become relevant into a system ?

You don't have "high" currents in the ground loop of the probes, so it is not meaningful.
Anyway, you should wire only one ground of the probes (on one board) ... and not the two.

\$\endgroup\$
13
  • \$\begingroup\$ Hi, Thank you for your comment. How can you say that there is no current in the ground loop of the probes. If I have an antenna next to the ground loop, I can make flowing a lof of current through the ground loop, as there is no résistance into the loop. We can see it as an Induction motor with the squirrel cage. \$\endgroup\$
    – Jess
    Nov 23, 2023 at 10:05
  • \$\begingroup\$ If you use only one wire ground, there is not a ground loop. Anyway, there should be only one ground on the target board. \$\endgroup\$
    – Antonio51
    Nov 23, 2023 at 10:15
  • \$\begingroup\$ Ok thank you for your comment ! I agree with you :) \$\endgroup\$
    – Jess
    Nov 23, 2023 at 10:18
  • \$\begingroup\$ @Antonio51 If those were e.g. logic level square waves, you would not leave the second probe ground disconnected. You would always connect grounds of all probes you want to do measurements. \$\endgroup\$
    – Justme
    Nov 23, 2023 at 12:08
  • \$\begingroup\$ @Justme At the same "ground" point, yes. Although, there is another "common" point on the scope. \$\endgroup\$
    – Antonio51
    Nov 23, 2023 at 15:31
2
\$\begingroup\$

Ground loops are affected by inductive coupling (mutual inductance) with rapid current variations in the immediate environment. This leads to noise in the ground loops. The same applies to signal-to-ground loops.

enter image description here

\$\endgroup\$
2
\$\begingroup\$

Ground loops are, as you are learning here, greatly overrated as a problem. Indeed, it is often better to "let ground abound": have a dense net of low impedance ground loops forcing everything to a common potential.

Here, the key is that the probe cable keeps the return current close to the signal current, on the cable shield, so the cable faithfully transmits the potential difference from the board to the scope. A good scope will be insensitive to loop voltage/current between the probe returns. Your board apparently has a good ground plane, which makes it insensitive to loop current between the probe returns.

\$\endgroup\$

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.