0
\$\begingroup\$

I am designing a board to power a WS2815 style LED strip. This strip has three inputs: 12V, GND, Di. This wire -> LED strip will be up to 60ft away from the control board. As such, I want to filter it to prevent common-mode noise on that wire. I will be using 20AWG twisted pair for the 12V/GND pair. I also plan on using a common mode choke. The Di wire will be smaller, say 26AWG.

Due to size constraints, it is not possible to have the Di line in the twisted pair bundle with the 12V/GND, and they will be run next to each other.

My question is, since I am filtering CM on the 12V/GND, do I also need to do anything with the single Di line? My thought is yes, since CM noise could propagate off of my control board, travel on the Di line, and radiate (or noise couples to the Di, and travels back to my control board.)

If I do need to filter it, what options do I have? Other than having a isolated termination on one side of the data line, I can't just have a single CM filtered line of course. I suppose a ferrite or a PI filter with a cutoff above the data transmission speed could also be an option?

\$\endgroup\$
4
  • \$\begingroup\$ Explain to me why common-mode noise is an issue. Common mode means that the difference signal is not affected and what does a WS2815 respond to: the voltage difference. To me, it sounds like you learned about "common mode" and now just assume it is a problem that needs fixing. So I'm here's me saying that common mode is not an issue. But feel free to prove me wrong and show that it is an issue. You might want to concentrate on the voltage drop across your cable, have you done any calculations on that? \$\endgroup\$ Mar 8, 2021 at 21:30
  • \$\begingroup\$ @Bimpelrekkie I suppose I could have been more clear. I'm concerned about the CM more so from a EMC aspect rather than an SI. This control board has a lot of high-speed digital signals (USB/Ethernet/HDMI/etc) on it. I do not want to radiate on this 60ft long cable. You worry about CM not differential in that circumstance. \$\endgroup\$ Mar 8, 2021 at 22:24
  • \$\begingroup\$ Just FYI, disregarding the data concerns, long power runs at low voltage are to be avoided if possible, especially if you have a significant load. You might want to consider using a higher voltage for the 60ft run and adding a buck converter close to the LED strip if you have a significant load. Search "wire volt drop calculator" if you need to figure out what voltage you'll get at the end of a certain size wire with a given amount of current. \$\endgroup\$
    – K H
    Mar 9, 2021 at 2:09
  • \$\begingroup\$ So I am trying to accomplish the same thing that you are with setting up long data runs for the WS2815 on the Di lines - DId you wind up using the higher power and use a Buck converter close to the strip for the data ? I am trying to keep converters away from the strips as those are in wetboxes. \$\endgroup\$
    – LGTZ73
    Aug 30, 2021 at 15:13

2 Answers 2

2
\$\begingroup\$

The 3-wire run forms a current loop. Obviously, the +12 current will return on GND, but what does Di do? Its changes of state will induce return currents on both +12 and GND whenever it toggles.

As long as the 3 wires are run together things should be ok (try to do this if you can.) This minimizes loop area, and thus, EMI. It also helps ensure a consistent impedance for Di to improve its signal integrity.

You could also consider a ferrite bead on the 3 wire run, near the source, to suppress power supply common mode noise. This will probably be more trouble than the Di pin.

You can also consider adding a series terminator on Di, to back-terminate any reflections. This benefits both EMI and signal integrity.

\$\endgroup\$
0
\$\begingroup\$

Start concerning about EMC measurements only if you have access to a certified lab.

A control board like yours with many high speed interfaces is not an easy task to put it in compliance.

If your board is DC powered, let's say 12 V, the certified lab will provide you a clean source at 12 V.

After that, you will have to connect your board, called Device Under Test DUT, to any and all the physical interfaces you declare: USB, HDMI etc.

That's gonna be a bunch of cables all around.

The physical interfaces will have to work during the tests. By that I mean that the technician of the lab will require you to flash and run the latest version of your software.

The lab wants the cables to carry actual signals.

Let's answer your question:

In a Common Mode test the terminal involved are 3: Line, Neutral and Earth.

Earth is connected to a large circular plate under the wood table where your DUT is placed on.

Happens that a small portion of the CONDUCTION current entering the 110 V Line terminal doesn't return in the Neutral terminal.

It returns as a DISPLACEMENT current from the Earth terminal.

Common mode filters reduce the current flowing into the Earth terminal and hopefully put in compliance your DUT.

How to pick the correct CM filter?

Bring in the lab some CM filters, a solder iron, some tin and try.

I did that, and after a day of tests I was able to put in compliance my 55 W, switching AC/DC converter with 93% efficiency. The main MOSFET was Infineon's IPL60R360P6SATMA1 which comes in an SMD package.

I also had to shortend some wires in order to gain 3 or 4 dB.

\$\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.