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.

LF353: The differential voltage sampling error of LF353

Part Number: LF353

As shown below, I used the LF253 and the OPA2354AIDDA to build a voltage sample circuit, which is used on a grid-connected inverter for the line-to-line voltage.  Unfortunately, the result is unsatisfactory. The BDin0 port is connected to  my DSP, and the AD's output is shown as following picture.  It seems that the three line-to-line voltage is well,  but the Ud Uq after the pll is bad. 

From the oscilloscope waveforms (as showm below), the power supply of LF353 has a serious common-mode noise, and the output has the same noise. I want to know if the noise will input the DSP and disturb the AD sample result, and if I need change the LF353 to another amp.

  • Hello, 

    I am not sure what you mean by "Ud and Up after pll"... is this somewhere not in the schematic shown? If so, can you point to where in the schematic?

    However, from the oscilloscope picture it would seem that the noise from the 12V power supply (pink) is linked to the DSP input (blue). In the layout, is the power supply and DSP input ever nearby each other (not isolated)? I recommend lowering the noise in the power supply and in turn it will lower the noise in your DSP input. 

    All the best,
    Carolina

  • Oops! C459 should not be connected to signal ground but to the output of OPAmp. Or by other words, in parallel to R615 :-)

    Kai

  • Thanks for your replay,

    The Ud and Uq are calculated in the  DSP by the PLL(phase-lockedloop) program based on the three phase line-to-line voltage. In the layout, the power supply is away from the input of DSP. In my opinion, the noise of power supply is from my inverter because it become worse when I increase my DC voltage which cause a higher dv/dt. 

  • Thans for your reply,

    I will try your advice as soon as possible.Let's wait and see.

  • Thanks for your advice, and it has solved my problem. In my design before this  sample circuit works well, so I have some confusion. Can you explain why I can't place the C459 at the original position, and make it palleled with R615 is efficiency? Furthermore, if my input signal is single-ended, how should I change the sample circuit? 

  • Hi,

    the cap mounted to signal ground destabilizes the OPAmp by ruining its phase margin.

    Best see this TI's "stability" teaching video series:

    There's no need to change anything, if your signal is single ended. Or do I misunderstand you? :-)

    Kai

  • Thanks for your share,

    I think that you understand me. In my design, I also used a current sensor of ACS 723 shown in following picture to measure the current. It outputs 0V~5V voltage signal from VIOU port, and I think it can't use the difference sample circuit we disscussed above. Thus, I change the sample circuit as following picture. This circuit can work, but the result is not that good. Do you have some advice for me? Thank you again.

  • Hi,

    the LF353 works as a differential amplifier and removes common mode noise, concretely spoken the noise between the signal ground of ADC723 and the signal ground of LF353. So, with respect to the single ended output, I see no need to change anything.

    By the way, the corner frequency of your low pass filtering is 1 / (2 x Pi x (10k // 6k) x 2n) = 21kHz and not 7.5kHz :-)

    So you might want to improve the low pass filtering?

    Kai

  • Thanks for your reply,

    Does that mean the difference input sample signal is a better choice for the noise?

    I really need a low pass filter because the current is 50Hz, and the swithing frequency of my inverter is 40kHz which is the worst noise. The 21kHz may be really high, and why the 7.4kHz is chosen? Can I use a lower corner frequency such as 4kHz?

    Thank you very much. 

  • I would do it this way:

    To achieve highest common mode rejection (suppressing of noise between the both signal grounds) the passive components of differential amplifier should be tightly matched. Highest balance results in a highest common mode rejection. Take metal film resistors (thin film !!!) and NP0 caps.

    user_lf353.TSC

    Kai