• Resolved

TRS3122E: High Common Mode Noise Immunity Configuration for TRS3122E

Part Number: TRS3122E

We encounter communication problem when we connected two TRS3122E together.  One TRS3122E is on the uC board and another TRS3122 is 28" away on the communication hub.  They are connected via a RJ45 cable.

The communication hub side is close to metal chassis and the common mode noise is injected into ground of the uC and the metal chassis up to 1000V p-p sine waveform.

I am try to find way to increase the noise immunity and please comment on the following questions:

1.  We supply Vcc with 3.3V now.  Is it better to use 5V supply?

2. For capacitor selection with Vcc = 3.3V, we currently use 100nF for all C1 - C5, is it better to use C1 = 47nF, C2 = 470nF, C3 = 100nF, C4=C5=470nF,

3. Is it better to use 4.7uF Cbypas for Vcc and Vl than 0.1uF for noise immunity?

4. For all test in done at Rload = 3kOhm, and the datasheet shows receiver input resistance in 5kOhm, is that means we can add another Rload in parallel with 5kOhm to 3kOhm to increase the current?  Or

If you can think of any ways to improve the noise immunity of configuring the TRS3122E, we appreciate your comments.

Thank you very much,

Frank W

  • In reply to frank wong:

    Have you captured any signal with common mode noise? Can you check the characteristic of the noise? You mentioned it's 1000V(?) peak to peak sine wave. What's the frequency? And what's your data rate? If the noise is high frequency, you may want to add some RC filter on the bus to smooth it out. If it's more like DC, you can try to increase your signal to noise ratio by increasing the output voltage.
    Back to your questions, 1) in general 5V supply might generate more output voltage; 2) using C1 = 47nF, C2 = 470nF, C3 = 100nF, C4=C5=470nF makes the output voltage have less ripples with 5V supply; I'm not sure it helps in your case; 3) Larger decoupling cap should help stabilize the supply, therefore maybe the signal as well; 4) I'm not sure the reason you added extra resistor load. RS232 standard requires 5kOhm. More load means more current, but it might not help the output voltage.
    Please let me know if you have more questions.
  • In reply to Hao L:

    Thanks, Hao.