3
\$\begingroup\$

I have a Siglent SDS1104X-E, and I'm trying to debug an issue with an LED driver. I was probing the switching node of the driver on one channel and the PWM'd output of a transistor driver that shunts the output to ground to control the brightness. The scope probes ground clips are connected in two different spots, specifically to two different mounting holes on the board that are connected to circuit ground. I've tried reversing them and attaching them to the same spot to see if it would make a difference, and it did not. I've noticed that when I probe only the PWM signal, I see what I expect: a square wave between 0 and 12V.

Signal at transistor driver output,

However, when I am probing both the transistor driver output and the switching node, I see this:

Signal at transistor driver output and switching node.

Changing the trigger channel doesn't make a difference, I still see what appears to be the signal on channel 1 bleeding into channel 2. Is this normal? Is there a way to prevent it happening?

[edit] After further experimentation, I've found that I can greatly decrease the amount of signal bleed by removing the hookup wires that I had used to attach the ground clips to the ground points, and just clipping the ground points directly. I'm still a bit puzzled as to why the length of the ground lead would have this effect.

[edit 2] Here are two additional traces. In the first, I attached the ground clips directly to the mount points on the board, which are attached to a (shared) circuit ground.

Signals with shorter ground leads.

The second trace is from a few minutes later, after I shut off the device and then powered it up again. The ground traces are short as from before, but the signal bleed is back.

Signals with shorter ground leads. Bleed returns.

Here is an image of the probing setup. The grounding points are circled and red, and the two probe points are inside the green circle.

Photo of the probing setup.

Here are some images of the full board layout, and a closeup of the driver portion (of one of the drivers). Just for context, there are three drivers on the board, but I only have one active for this test (the center one, which drives the green LEDs). The other two are disabled via the EN pin (the drivers are LM3409s). I've also added a schematic for the drivers.

Full board layout.

Driver layout.

enter image description here

\$\endgroup\$
11
  • \$\begingroup\$ Maybe the probes are connected in a way they should not be. Impossible to say without more info, like a photo of the setup or probed nodes noted in a circuit diagram. Including where the ground leads are connected, which might be the first suspect here. \$\endgroup\$
    – Justme
    Jul 28, 2022 at 21:55
  • \$\begingroup\$ How are the scope ground clips connected? Are you using one clip or both? To the same point or different ground points? True circuit ground or some other voltage? Pop the details into your question. \$\endgroup\$
    – Transistor
    Jul 28, 2022 at 21:56
  • \$\begingroup\$ Added an explanation of the grounding. I could upload a photo but it's a bit... chaotic. \$\endgroup\$
    – flimsy
    Jul 28, 2022 at 22:08
  • 1
    \$\begingroup\$ "I'm still a bit puzzled as to why the length of the ground lead would have this effect." - how long was it? \$\endgroup\$ Jul 28, 2022 at 22:40
  • 2
    \$\begingroup\$ I appreciate the fact that you put a lot of effort into this question. Definitely an above average question. \$\endgroup\$
    – user57037
    Jul 29, 2022 at 4:02

1 Answer 1

4
\$\begingroup\$

I think this is caused by loop area in the ground lead. It may seem kind of random because it depends on exactly how you move the ground wire on the probe. Essentially, magnetic fields pass through the loop formed by the GND wire and induce a voltage on the wire.

Often, when trying to get the best performance from passive probes, people use spring clips around the probe body as a GND lead.

enter image description here

Photo came from here. Follow the link and read the whole article.

https://www.analog.com/en/technical-articles/short-ground-leads-make-better-scope-photos.html

\$\endgroup\$
3
  • \$\begingroup\$ Ah interesting, so basically the ground loop from the prove is acting like an antenna for EMI from my board? Though in that case why would it only happen when I had both leads connected? Is the ground loop from the second probe somehow causing additional EMI? \$\endgroup\$
    – flimsy
    Jul 29, 2022 at 18:41
  • \$\begingroup\$ A good question. I am not 100% sure. I guess the probe monitoring the clock is kind of a loop-transmitter and the one monitoring the data is a loop receiver. Another thing I should have mentioned, though is that I see this phenomenon all the time and just ignore it because I know it is a measurement artifact. If you probe as shown in the picture you will probably get rid of most of this phenomenon. \$\endgroup\$
    – user57037
    Jul 29, 2022 at 20:34
  • 1
    \$\begingroup\$ Ok, that's great to hear. As long as it's a normal thing I can just work around. Thanks! \$\endgroup\$
    – flimsy
    Jul 31, 2022 at 17:59

This site is temporarily in read-only mode and not accepting new answers.

Not the answer you're looking for? Browse other questions tagged .