This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

LM4050-N-Q1: injected rf-signal at EMI tests vary referenz voltage significantly

Part Number: LM4050-N-Q1
Other Parts Discussed in Thread: LM4040

Hello,

I am using a LM4050QAEM3-2.5/NOPB  2.5V voltage referenz in conjunction with an opamp (single supply 12V) as a current source for 1.1mA. See schematic attached.
This current is consumed by a resistor about 0.5 meters away. The sinal-lne to the resistor is placed in parallel to the 220V AC Power Supply mains lines for my device in the same cable.
I can not change this.

At EMI tests with injecting rf signal of 150kHz-80MHz Amplitude Modulation with tone 1kHz at 80% Modulation at 20V to my Mains Power Supply lines by a coupling device,
I find, that the referenz voltage of my LM4050 will rise and drop significantly from its nominal value of 2.5V (between 0V and supply voltage to the voltage Referenz)
at different rf signal resonant frequencies.

What I have done so far:

- I used  a 2.2nF ceramic capacitor in parallel to the LM4050
- I used an inductor 680uH in the signal-line to the resistor
- I used a 2.2nF ceramic capacitor from current source signal line to ground

Question:
1. Is that behaviour of influencing the voltage referenc value by injected rf signals known?
2. What can i do to avoid this behaviour?
3. Is there an other voltage referenz which does not have such behaviour?

Any ideas?

 

  • Hi Ulf,

    In extremely high frequency testing above 10Mhz, the LM4050 does experience a voltage offset. Some research has been done on this here for LM4040 which has a similar architecture and in IEEE for general shunts. Typically shunts are more susceptible than series voltage references. If EMI Is a large concern, you can make a current source using a series voltage reference that typically do not experience this offset at high frequencies.

    -Marcoo

  • Hi Ulf,

    your signal line is in parallel to the mains cable?? Running in the same cable?? But that's totally idiotic! Spend the signal line an own cable. Use a screened cable with two conductors, one for the signal line and one for the ground return signal. Connect the screen to GND of your circuit. No signal current must flow on the screen. This is the fundamental requirement for a successful performance!

    The next issue is the topology of your current source. If there's a need to add filtering to the current source output, it's no good idea to have the load sitting inside the feedback loop of an operational amplifier. Use a different current generator concept then, for instance the standard high side current source with this well known PMOS-FET at the output.

    Kai