Interference with Ws2812b LED strips when multiple control wires are in one cable?

OoPatrick1912oOByOoPatrick1912oO

Hi,

I have a small project where I control 2 LED strips separately with 2 ESP32 C3.

I transmit the control line for each strip using a CAT-7 cable approximately 10 m long. I use a level converter for amplification.

If I route the control lines over 2 individual cables, everything works perfectly.

However, to save space, I want to use only the single CAT cable. However, both strips start flickering and displaying incorrect colors. It seems as if they're interfering with each other in the cable.

I don't know how to solve the problem. CAT-7 cable is designed for frequencies higher than 800 kHz.

I have a logical error and need your help.

Thank you for your answers.

(1 votes)
Loading...

Similar Posts

Change the water filter in an LG side-by-side refrigerator?

ByDaZePaul709

Good day, I bought a new LG refrigerator a few days ago and now it says that I should change my water filter but I have a water tank in the refrigerator. On the internet I can only find something where the refrigerator is connected to the water network and you can believe it or…

Subscribe
Notify of
7 Answers
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Ginpanse
Ginpanse
8 months ago

10m? du weißt schon, dass es sich hier um Kleinspannungen von 3.3V handelt ja? Spannungsabfall? Wundert mich, dass es überhaupt funktioniert hat. So lange Strecken an den Datenleitungen kannst du vergessen. Das kann nicht stabil funktionieren. häng die esp’s direkt in unmittelbare Nähe der strips. die versorgungsleitungen kannst du ja so lang machen wie du willst indem du entweder die zuleitung von 230v zum netzteil verlängerst, oder eine hohe spannung von zb 12v nimmst und dann einen 5v spannungswandler direkt an den esp’s setzt.

0
OoPatrick1912oO
Author
OoPatrick1912oO
8 months ago
Reply to  Ginpanse

Das ist alles ein Versuchsaufbau gewesen mit dem Zeug was ich noch da hatte. Ich wollte schauen, ob mit einem Pegelwandler der LED-Streifen bei 10m Länge Probleme bekommt. (-> Ohne Pegelwandler ja, mit Pegelwandler ging es einwandfrei).

0
Ginpanse
Ginpanse
8 months ago
Reply to  OoPatrick1912oO

Dann ist der pegelwandler das problem und zu instabil. auch induktion wird hier ein faktor sein.

0
Transistor3055
Transistor3055
8 months ago

There seems to be an induction (interference pulses, also called hum loop in audio technology) in the control lines.

So an inductive problem .

Why and how to avoid:

  • By connecting the ground wire multiple times (especially with the shielding), a 10m long coil with one turn is formed. The negative terminal of the control wire and the shielding form this "unwanted" coil. When current flows through a coil, a magnetic field is created. If there is a second coil (= also a coil with one turn, e.g., a second control wire or even the data wire itself), it is like a transformer with one turn secondary and one turn primary. The transformation ratio is 1:1, so theoretically a 3V voltage pulse is induced from the primary side into the secondary side. BUT! Fortunately, we don't have an iron core (like a transformer), but only an air-core coil, which results in significant losses. NOTE: The higher the frequency, the lower these transmission losses, so these interference signals become stronger with increasing frequency.
  • How to avoid this? Simply "cut open" the coil. In this case, connect the shielding only on one side, only on the control unit, not on the LED side (this is called floating shielding). This interrupts the coil, and no current flows through the CAT cable shielding.
  • Did you know? A uniformly twisted pair of wires (like in network cables or unshielded telephone lines) is very resistant to external interference. This is because the interference cancels out due to the constant change in wire orientation, so the total interference is almost zero. Yes, all DSL/VDSL telephone lines have this twist (and shielding, if present, is NOT connected… floating shielding).
  • The current surges on the +/- supply voltage (5V) leads may have the same "induction effects." In this case, it helps to connect a larger , faster capacitor to the LED side of the supply line. Use two capacitors in parallel, approximately 100uF to 470uF electrolytic capacitors (as larger capacitors), and a film capacitor of approximately 100nF (as a fast capacitor). Connecting both capacitors in parallel to the supply voltage on the LED side ensures that the current level stabilizes. This is because the high current pulses from the LEDs are primarily powered by these capacitors. The capacitors are then recharged with a nearly constant charging current. A constant current does NOT cause an alternating field and therefore no induction on this 10m long supply line or control lines.

Good luck!

1
OoPatrick1912oO
Author
OoPatrick1912oO
8 months ago
Reply to  Transistor3055

Thank you for your reply. I hadn't considered the coil issue. ^^
I had supplied the power via a different cable, which is only about 0.5m long, so that shouldn't have any influence.

The inductive problem should also occur if I only supply one LED strip via the cable, but everything works perfectly there.

This is all just a test setup with the stuff I have at the moment to understand and learn for future projects.

0
Transistor3055
Transistor3055
8 months ago
Reply to  OoPatrick1912oO

Unfortunately, I can only give you hints.

I had supplied the power via a different cable, which is only about 0.5m long, so that shouldn't have any influence.

A cable length of 50cm is fine. But even at 50cm, these supply lines can cause minor interference. A capacitor between the positive and negative terminals (directly on the LED chain) will absorb this interference. As I explained, I would try electrolytic capacitors and film capacitors if you have them. Maybe that will help.

The inductive problem should also occur if I only supply one LED strip via the cable, but everything works perfectly there.

You're wrong. Two different signals are needed for interference. A single signal alone can't interfere with itself (only reflections are possible). Mutual interference is much more likely only when a second signal of a similar frequency is added.

It could be that your level converter (3.3V to 5V) is already at its limit, because the description of the level converter (on Aliexpress) explains:

'3. Stable communication baud rate within 28800bps

This means a cutoff frequency of 28.8 kHz. It means that signals with higher frequencies (depending on the timing of your control) may not be transmitted cleanly. The 5V control signal (TTL) then no longer quite reaches 0V for low and 5V for high. Instead, at too high a frequency, the square wave signal becomes a rounded triangle/sine signal with flat edges (see here: https://de.wikipedia.org/wiki/Anstiegs-_und_Abfallzeit ). Whether the technical specifications from Aliexpress are accurate is also questionable.

My conclusion is:

It is the simultaneous occurrence of 3 factors that could cause this misbehavior:

  1. Induction of the pulsed LED currents of the 5V supply line (50cm)
  2. At least two similar signals would be needed to interfere with each other. One signal alone can hardly interfere with itself.
  3. The level converter is running at too high a frequency. Bits could flip (a 1 is not recognized) because the TTL level reaches critical (=indistinct) values.

PS:

Aliexpress also writes (in an incomprehensible way) something about induction/"floating shielding" in its instructions. So, don't double-wire the ground (=GND) to avoid induction (=coil effect). Quote:

1. If the power supply in the high voltage range and the low voltage range have a common ground, the GND in this module only needs to be connected to one ground.

By high voltage, my dear Ali… actually, I mean the 5V voltage. I would call it the 5V side, because in Europe, we understand high voltage to mean something completely different (several thousand volts).

Good luck!

1
OoPatrick1912oO
Author
OoPatrick1912oO
8 months ago
Reply to  OoPatrick1912oO

Zur Leistung: Der LED-Streifen ist jeweils nur 1m lang.

0