• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 48 subscribers
  • Views 252 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

INA186-Q1: Output Oscilation

Alexander Skulsky
Prodigy 40 points
Part Number: INA186-Q1
Other Parts Discussed in Thread: INA186, INA190, INA213, INA296B, INA185, INA280, INA281, INA181

Hi

I have encountered an issue with INA186A2-Q1 I am using in one of the designs.

 

The schematic is attached bellow. I use a DCDC from 5V to 3.28V

The Max Current is 4.48A and typical is 2.56A.

I have placed a voltage divider on the output of the current sense amplifier to protect the ADC input from high voltage (max 0.6V)

DCDC switching frequency is 2.3MHz

The output of the DCDC measured on C329 @4.48A is relatively clean and has 7mV ripple with 20MHz BW of the Scope:

You can see the 2.3MHz on the spectrum view and additional 146KHz

On the other hand the output of the INA186 is very noise and has a ripple as high as 32mV measured on R272

I have measured the input on the INA186 with a differential (orange) probe and it can clearly be seen that there is no noise at 91KHz or110KHz but the output is strongly oscillating at this frequency

One of the test I made was adding a LPF (suggested in section 8.1.3 of the datasheet) on the INA input with series 100Ohm resistors and 47nF capacitor.

This reduce the output oscillation to 7mV but the 110Khz and 98KHz are still there but lowered by ~30dB

Additional finding on the INA186 output ringing.

The ringing frequency changes based on the supply Voltage

 

Supply Voltage

Frequency

1.75V

30.5KHz

1.8V

9.3KHz and 18.5KHz

1.85V

8.5KHz and17.8KHz

1.9V

21KHz

2V

40KHz

 

Attached oscilloscope print screens with spectrum view

Supply Change.zip

  • 0
    TI__Mastermind 28385 points

    Hello,

    I can tell you right away that the spike in noise at around 100kHz is due to the internal switching frequency of the device, which is typically 100kHz, but could have significant variation (+/- 20%) as it is not a tightly controlled parameter.

    As for the other spikes, they seem to be shifting/getting aliased into higher frequency spectrum as Vs increases past 1.8V. Now why it goes in opposite direction at 1.75V I do not know, but there are many non-linearities involved with the front end of this device. It may be that the device noise is lower with VS closer to minimum 1.7V because this is closer to optimal gate voltages; however this is not known nor easily determinable.

    The frequency spikes (<100kHz) are probably coming from the shunt current or are getting picked up from nearby noisy voltage sources. 

    The only way to reduce noise is to:

    1. Limit shunt noise with the differential input LPF. Maximize fc_diff by setting fc_diff = INA190_BW and keeping Rin < 1kOhm. Cdiff should be at least 1nF.

    2. Put LPF on output. Also set fc to INA190 BW so no signal BW is lost. Choose RC values so the impedance seen by INA190 output does not capacitively cross through it's output impedance to avoid output oscillation. Here is an example of what this LPF impedance should look like where the Riso dominates resistance before Zload crosses Zout:

    3. Consider shielding OUT trace with ground pours or putting in an inner layer in between ground/analog ground pours.

    4. Also make sure that the decoupling capacitance at VS and GND pins it place close as possible to VS pin. Consider adding another decoupling capacitor at 1uF or greater.

    Sincerely,

    Peter

  • 0 in reply to Peter Iliya
    Prodigy 40 points

    Are there alternative Current sense amplifier that don't  suffer from such instability? 

  • 0 in reply to Alexander Skulsky
    TI__Mastermind 28385 points

    Hey Alexander,

    By instability are you referring to the spike in noise density or the output impedance?

    I think all of our analog output current sense amplifiers have similar looking closed-loop output impedance curves, where the Zout looks like an inductor (going up). Most chopping amplifier will exhibit similar behavior while linear amplifiers have a resistance dominated Zout.

    If you are referring to the spike in noise density. the INA296B, INA281, and INA280 are similar performance devices that do not have this lone spike in noise density. The INA181, INA185, INA199A2, and INA213 do have a spike, but it is shifted out closer to 500 kHz.

    Sincerely,

    Peter