0
\$\begingroup\$

While trying to filter the ripple of my switching PSU, I've come to a phenomenon that is probably well known to professionals, but seems absurd to me. Admittedly, my scope is not of the highest quality, but it can sample at 100MHz and gives usually coherent results.
Furthermore, it is only connected to a floating portable computer working on batteries; so, no parasitic current can flow through the ground or through another connection.

The phenomenon is the following: I took an ATX power supply like the one in your desktop computer, and connected the TWO probes of my scope to the ground wire (the black one) of the PS, AT ITS EXTREMITY A (that is, point A is connected to nothing, except the two probes). In my scope (set in AC mode), I could observe a small ripple.

As this seems absurd, I first thought this is an artifact, so I connected the first pin of a toroidal inductor (5mH) to A, the ground probe of the scope to A, and the other probe to the other pin of the inductor (connected to nothing except the probe); surprise ! my scope now show a ripple of up to 40mV, ranging from 10KHz to 20MHz (centered at 4MHz). I was trying to explain this to myself by invoking wave propagation, when the next experience mystified me completely: I connected the ground probe in the other side of the inductor, and the other probe to point A. Then an incredible ripple of up to 0.5V was shown by my scope !

These experiences contradict everything I know on currents, and in particular, why is there an asymmetry regarding the connection of the probes with respect of the two sides of the inductor (the ripple is AC no?)

I would greatly appreciate any insight.

EDITS:

  1. If I shut down the PSU, while maintaining everything connected as is, the scope shows nothing.

  2. If I don't shut down the PSU, but I disconnect the two probes from the ground wire, while maintaining them connected one to the other, the scope shows nothing as well.

  3. Here is what I believe to be a beginning of explanation: the electrical wave is exploring the ground wire, and most of it is reflected at the first pin of the inductor. So, between the PSU and the inductor appears a standing wave. At the other pin, few of the AC components of the wave has passed the inductor, hence the ddp between the two pins of the inductor. If there is no inductor, a similar effect appears (but much less strong) because the two probes have a somewhat different inductance. As convincing as this may seem, this does not explain the asymmetry of the strength of the ripple with respect to the pins of the inductor.

\$\endgroup\$
6
  • 2
    \$\begingroup\$ So you are connecting a makeshift antenna to your scope and wonder that it receives some crap that is in the air? \$\endgroup\$
    – PlasmaHH
    May 13 '16 at 14:12
  • \$\begingroup\$ Here the two probes are connected to the same point of the wire. As far as I know, in antennas, there is a ddp between the ground and one extremity of the antenna. Can you elaborate? Oh, of course, I forgot to write that the phenomenon ceases as soon as I interrupt the PSU. \$\endgroup\$
    – MikeTeX
    May 13 '16 at 14:41
  • \$\begingroup\$ a danish design price? there is crap in the air and the scope probe picks it up, simple as that. You might want to look into construction schematics of a scope probe, they are far more than just a wire from the tip to the scope. \$\endgroup\$
    – PlasmaHH
    May 13 '16 at 15:02
  • \$\begingroup\$ PlasmaHH, if so, this should be observed even when I disconnect the two probes from the ground wire (or the wire from the PSU) while maintaining everything connected as is. But this is not the case (see the edit in my question). Furthermore, even if it were the case, I don't see how this can explain the asymmetry with respect to the inductor. \$\endgroup\$
    – MikeTeX
    May 14 '16 at 19:27
  • \$\begingroup\$ make a small coil of a few turns (you could even use the ground lead) and connect the scopes ground to the probe tip, bring the coil close to any inductors/transformers in the SMPS and you will see ripple, using a coild of ~10 turns around my finger, I can spot where the core CPU regulator in my laptop is. \$\endgroup\$
    – Sam
    May 15 '16 at 0:26
1
\$\begingroup\$

The main concern of the third item in your edit, namely the dependence on the polarity, is explained by the asymmetry of the input impedance of you battery powered scope. While at DC, both the ground connection and the signal connection have practically infinite resistance to ground and can be treated equally, this is no longer the case at RF.

Even if your scope is not grounded, the surroundings of your scope are at a certain potential, even if it is just air. The scope has a capacitance to the environment which is notably different from zero. You might have learnedvin physics education not only how to calculate the capacity of a capacitor made out of two concentric spheres, but also, likely even earlier, how to calculate the capacitance of a single sphere. Now, years later, you observe that this is not just strange theory, but actually a real thing that matters in practice: the scope ground is likely connected to a big shield in the case and has considerable capacitance to ground at high frequencies, shorting them away. The signal input on the other hand is designed to be extremely high impedance even at higher frequencies, and thus much better suited to pick up stray RF. This makes your floating single-ended instrument different from a real differential instrument, in which the impedance of the positive and negative input are matched.

\$\endgroup\$
1
  • \$\begingroup\$ argghh! of course! well catched! \$\endgroup\$
    – MikeTeX
    Aug 9 '16 at 16:35
2
\$\begingroup\$

Your two probe wires connected together at the "far end" (farthest away from the oscilloscope) form a loop antenna. They are picking up magnetic radiation from the transformer in the SMPS (Switch Mode Power Supply).

You could conduct a couple more experiments with the probes inter-connected but NOT connected to the SMPS. And another experiment where you simply power down the SMPS and see what your scope shows without changing the connection.

\$\endgroup\$
8
  • \$\begingroup\$ Richard, I did that, and the scope is showing nothing in both cases: when I power down the SMPS, and when I disconnect the probes from the ground wire of the SMPS while maintaining the two probes connected one to the other. I've edited the question to add this information. \$\endgroup\$
    – MikeTeX
    May 14 '16 at 19:20
  • \$\begingroup\$ Making a measurement like this WITHOUT the scope ground connected to the DUI (Device Under Test) ground yields UNDEFINED results by definition. You have NO common point of reference between your measurement device and your DUT. There are dozens of experiments to conduct in order to discover where your artifact is coming from. The ones I suggested were just the first two. This is only the beginning of your journey. Although, buy definition, your test methodology is flawed. \$\endgroup\$ May 14 '16 at 22:05
  • \$\begingroup\$ Richard, first thank you for answering me. This is not that I like unproductive discussions. But the scope ground was connected to the DUI in my experiences. This is only to check if the probes of the scope form a loop antenna that I've disconnected the ground of the scope from the DUI. And it appears that this is not the cause of the phenomenon. Obviously, there is some capacitive or inductive coupling acting here, but I would like to understand exactly what happens, and why is there an asymmetry of the measurements with respect to the coil. \$\endgroup\$
    – MikeTeX
    May 15 '16 at 5:02
  • \$\begingroup\$ More practically, I would like to know how to get rid of this phenomenon, because I don't want a ddp appearing between the pin of devices connected to the ground of the PSU. \$\endgroup\$
    – MikeTeX
    May 15 '16 at 5:07
  • \$\begingroup\$ The primary way of getting rid of this phenomenon is to properly connect the ground of your test equipment (scope, meter, whatever) to the ground of the DUT. Until you do that, you are chasing phantoms. And with no practical benefit. \$\endgroup\$ May 15 '16 at 5:20
0
\$\begingroup\$

This man seems to have the solution:

the shorted scope probe problem

it's just a wire isn't it?

switching power supplies...

common mode effect

and many other more or less related articles

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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