• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Subscribers 50 subscribers
  • Views 913 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.

INA196: Output saturation issue

VIII
Prodigy 220 points
Part Number: INA196
Other Parts Discussed in Thread: TINA-TI, INA193, INA198, INA194, INA195, INA197

Hello,

I'm using INA196 to measure current flowing through the load as per attached picture. I have wide range of currents, so I use INA196 to measure very low currents and then I rely on measuring woltage drop over resistor bank I have in my circuit. I hoped that when my input voltage is more than INA196 is able to output (0.25V) the output will saturate at about 5V. Unfortunately, when input voltage to INA196 is in range between 2.6 and 8.2V the output goes to 0V. Above 8.2V it saturates at about 5V. What may be the reason of output voltage going to 0V in input voltage region 2.6-8.2V ?

Thanks,

Jacek.

  • 0
    TI__Mastermind 26710 points
    Hello Jacek,

    Thanks for posting your question on E2E.

    Are you seeing this problem in simulation or with lab/bench data?

    Just so I’m certain, you see the INA196 rail to ground when the input differential voltage is in between 2.6V and 8.2V? This is definitely not an intended behavior. While we do not characterize the difference for such large input differential sense voltages this does seem unusual. We only state that device will survive such large input voltages.

    Let me talk with my colleagues and test this myself to see what may be happening.

    Sincerely,
    Peter Iliya
  • 0 in reply to Peter Iliya
    Prodigy 220 points

    Hi Peter,

    I have verified this issue with TINA-TI and it just saturates to about Vcc and stays there. What I can see that when it goes from saturation to GND there is short period of  triangle shaped noise near to GND and then it goes to ground. I have tried this on two different PCBs and it's consistent.

    Regards,

    Jacek Meyerhold.

  • 0 in reply to VIII
    Guru 140200 points
    Hi Jacek,

    can you show a scope plot?

    Can this "triangle shaped noise" be seen in static operation or only when the output quickly settles to 0V?

    Kai
  • 0 in reply to kai klaas69
    Prodigy 220 points

    Hi Kai,

    This "triangle shaped noise" actually looks like a bit malformed sinewave. When you vary voltage there is some boundary condition, between saturation nad ground and that is when this noise happens.

    It looks like picture below.

    Regards,

    Jacek.

  • 0 in reply to VIII
    TI__Mastermind 26710 points
    Hello Jacek,

    Thanks for the waveform. Could I just get a little more explanation of what's happening? In your scope shot, the VOUT is centered around 400mV and this is about as low as what the INA196 can swing to ground on a worst-case basis. So the INA196's output pin (OUT) is already saturated in the waveform above.

    What was the input differential voltage (Vdiff or Vsense) into INA196 in your waveform? Keep in mind Vdiff = IN+ - IN-.

    Another interesting thing is that the output frequency is 1.67MHz, which beyond the bandwidth of the device. Is Vdiff a sinusoid waveform as well? Is the common-mode voltage at IN+, IN- pins a stable 12V then entire time? Or does this voltage rail droop as the current increases?

    I am still looking into this, but some more information would be useful.

    Sincerely,
    Peter Iliya
    Current Sensing Applications
  • 0 in reply to Peter Iliya
    Prodigy 220 points

    Hi Peter,

    I have took a sequence of photos. Scope shows INA196 output, voltmeter shows voltage between IN+/IN-.

  • 0 in reply to VIII
    Guru 140200 points
    Hi Jacek,

    very intersting and very nice photos! I didn't know that the INA196 behaves that way. Does a fresh INA196 show the same behaviour?

    Well, that datasheet of INA196 doesn't tell anything about what the output makes when the input voltage range is overdriven in such an extreme way. See figure 3 of datasheet. So, at least, the datasheet isn't lying...

    Why not clamping the input voltage with diodes? If the useful input voltage range is 1V and the strange things begin to happen at voltages above 2.5V, you should be able to clamp the inputs by the help of some diodes. You could put two or three 1N4002 in series for the clamp. If the measuring current is very small you could even take two or three small signal silicon diodes like the 1N4148 of BAV99 in series. These would show less leakage currents and would result in less measuring errors.

    Kai
  • 0 in reply to VIII
    TI__Mastermind 26710 points
    Hey Jacek,

    What you are seeing is actually the device acting normally. When you overdrive the input sense voltage of the INA196, you are overdriving the internal BJT's. At some point there will be a phase crossover from the BJT and this will cause the amplifier output to drive to GND. Eventually you can increase the input voltage and then you will leave the crossover region and the INA196 output will drive back to the postive voltage rail. I have confirmed this behavior with lab data and design guidance.

    We apologize for any inconvenience this has had. This behaviour will only show up for the devices with the same internal BJT topology: INA193, INA194, INA195, INA196, INA197, INA198.

    You could use diode clamps at the inputs, but you want to make sure the leakage currents are minimal at the full-scale input voltage (250mV).

    Hope this help, please post back if you need more information.

    Sincerely,
    Peter Iliya
    Current Sensing Applications