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.

INA213: About gain adjustment and noise countermeasure

Part Number: INA213
Other Parts Discussed in Thread: INA240

Hi,

My customer is considering INA213.
(Customers plan to connect Vref to GND.)
I got three questions from customers.
Could you give me your advice?

1. Can this device adjust the GAIN by connecting 22 KΩ in series with IN + pin and IN - pin respectively?
    Customers want to adjust the gain to 23.8.
    1MΩ/(20KΩ+22KΩ)=23.8

2. INA213 can not install an low pass filter in the background with built-in resistor.
    For that reason, Noise is amplified by gain.
    Please tell me how to counter it.

3. As related to question 2,
    Is there a possibility of oscillation due to power supply noise or input signal noise?

Best Regards,
Yusuke / Japan Disty

  • Dear Tsukui,

    1. The gain of the device is fixed and cannot be adjusted like this. The illustration of the INA213 in the datahseet leaves out internal details. If you need a smaller gain, you can reduce the value of the shunt resistor.

    2. What noise are you referring to? Are you talking about noise from the bus current or noise of the INA213 (e_v = input referred 25nV/rtHz)? If you are concerned e_v getting gained up by the device, this is unavoidable. This value is actually measured as output referred noise and then divided by the gain of the DUT (device under test) so that e_v can apply to all gain variants of INA21x in datasheet. To reduce this noise you have to place a low-pass filter on the output of the INA213 with a cutt-off frequency (fc) less than the bandwidth of the part. Although 25nV/rtHz is a pretty low noise level and for the INA213 results in an overall output noise of:

    Vout_noise_RMS = sqrt(80,000Hz * 1.57)* 25nV/rtHz = 8.86 uV rms,

    where 80kHz is the BW and 1.57 is the brick wall correction factor.

    If you are trying to counter noise from the current source, then you can implement a differential input filter as discussed in the datasheet (see below).

    3. Input noise will not put the device into instability. If the power source is not oscillating significantly and with a magnitude of something (volts), then this should also not throw the amplifier into instability. The most likely source of instability will come from driving too large of a capacitive load or RC filter on the output. You can read about our analysis of closed-loop stability with our current shunt monitors here for a move involved discussion.

    Hope this helps.

    Sincerely,

    Peter Iliya

    Current Sense Amplifier Applications

  • Peter- 

    I'm looking for the gain of INA213 in your description above and I do not see the gain factor for the output noise.  INA213 gain is 50V/V, so: 

    Vout_noise_RMS = 50V/V * sqrt(80,000Hz * 1.57)* 25nV/rtHz = 50* 8.86 uV rms = 443 uV rms

    Do I have this correct?

    I'm doing a similar calculation for INA240, and want to make sure I'm doing it correctly. 

    Thanks

    Jake

  • Jake,

    I actually forgot to add the gain to my initial calculation. Correct calculation for ouput RMS noise for INA213 is:

    Vout_noise_rms = Gain * input-referred noise spectral density * sqrt(BW_-3dB * Kn), where Kn is the brick-wall correction factor.
    Vout_noise_rms = 50V/V * 30nV/rtHz * sqrt( 80 kHz * 1.57) = 531uV RMS

    I chose 30nV/rtHz for the input noise level because that is what we have shown as the input-referred noise for the INA213 in Figure 15 of the datahseet. I chose Kn=1.57 because I am approximating the INA213 gain with a 1st order roll-off (-20dB/dec). This is actually somewhat of a conservative calculation because from looking at Figure 7 (Gain vs. Frequency) the INA213 gain actually looks to be rolling off faster than this, maybe closer to -30dB/decade, which means Kn is probably closer to 1.4.

    If you want to know the 6-sigma, peak-to-peak noise you can expect to see on the output of the INA213, then you multiply the RMS value by 6 to get 3.186mVpp. Meaning 99.99966% of the time, your output peak-to-peak noise will be < 3.186mVpp

    Sincerely,
    Peter Illiya
    Current Sensing Applications