• State Resolved
  • Locked Locked
  • Replies 9 replies
  • Subscribers 49 subscribers
  • Views 1293 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.

INA201 - Is it OK to make the output saturate ?

Thomas LE BIVIC
Intellectual 500 points
Other Parts Discussed in Thread: INA201, INA200, INA202

Hello all,

I'm thinking of using INA201 to check some power supply status (ON/OFF) based on measured current.

Let's say I have to trig if current is about 10% of max current. To get more that 20mV at that point I have to use a large Rsense value which causes saturation of the amplifier at about 30% of max current. It's not a problem for me but will it cause unexpected behavior / premature aging / over-consumption / any failure ?

Thank you for you help,

Thomas.

P.S. Reducing the gain by choosing INA200 is also an option. For your information I see no difference between INA200 and INA201 in simulation using LTSPICE ;(

  • 0
    Expert 1095 points

    Hi Thomas,

     I guess you are trying to make a comparator. Check this http://e2e.ti.com/blogs_/archives/b/thesignal/archive/2012/03/14/op-amps-used-as-comparators-is-it-okay.aspx

  • 0 in reply to Deepak%20murali
    Intellectual 500 points

    Thanks for you reply, 

    No it's not what I'm trying to do. This chip (http://www.ti.com.cn/cn/lit/ds/symlink/ina202.pdf) includes an amplifier and a comparator. Figure 5 shows saturation situation of the amplifier, I imagine it is not a problem but I'd like to be sure :)

  • 0 in reply to Thomas LE BIVIC
    TI__Genius 14825 points

    Thomas,


    The INA202 (and INA201, INA200) have a max Full Scale Sense Input Voltage of (Vs-0.25)/Gain (page 3 of the data sheet).  So at 18V, 100V/V gain, that's 177.5mV across the shunt.  As I understand it, if this is your power supply, then you hit 177.5mV at 30% input current.  Might I inquire as to what your Vs is, and your Rshunt, just to make sure everything is good?  It sounds like you're doing it just fine though.


    That said, your maximum differential input (the voltage across the shunt resistor on the inputs of the device, page 2 of the data sheet) is between -18V and +18V.  So if 30% is 177.5mV, then I suspect your 100% current across the shunt is 591.7mV, which is well within the -18V to +18V range.  In that regard, the device should be fine as the maximum limits are not stressed.

    Figure 4, page 6 of the data sheet, shows that the Vout will just clip as the differential input increases, as you know, and that goes all the way out to 900mV on the differential, which is well beyond the 177.5mV as well.

    Please let me know if I have any of my numbers egregiously wrong.

    So all in all, I think you're fine, but I will confirm with my colleagues that there should be no long-term reliability problems with the output saturated.

    Thank you for letting us know about the problems with the INA200 and INA201 models.  I will investigate the cause and correct these models promptly!


    Thank you, and I'll reply here again once I have confirmation on the reliability issue.

    -Jason Bridgmon

  • 0 in reply to Jason Bridgmon
    Intellectual 500 points

    Hello Jason, 

    You perfectly understood the situation. I'm pretty sure too, that it's OK to make the output saturate as long as the differential input is within the limits but I prefer to hear confirmation that no edge effect could be induced by this mode.

    Some figures : Vss 5V / Vcm 5V / Rshunt 470mOhm / Max current 500mA / Trig current 100mA / Resistance divider 4k64 and 1k5. The amplifier output starts saturate at about 200mA. 

    Concerning the Spice model, I checked with TINA simulation model of the INA200 and gets the expected x20 gain. I compared with the Pspice model and there is NO electrical difference ! So my guess is that LTspice is misinterpreting something for this Pspice model but I've not been able to find what :/

    Thanks for your support !

  • 0 in reply to Thomas LE BIVIC
    TI__Genius 14825 points

    Thomas,

    I spoke with three engineers here and they all said the same thing--while we don't do lifetest with a saturated input so we can't totally confirm there will be no problem, but so long as the absolute maximums are observed, we expect no problems.

    I've compared the TINA models first and next I'll check out the models in LTspice.  The INA200 definitely gets x20 gain and the INA201 gets the x50 in TINA.


    Jason Bridgmon

  • 0 in reply to Jason Bridgmon
    TI__Genius 14825 points

    Thomas,

    I downloaded the INA200 and INA201 models for spice off TI.com and simulated them in LTspice.  I see the appropriate gain of x20 and x50 (see plots below).  Could you share with me any information regarding the problem you're seeing to help me track this down?

    Thank you,

    Jason Bridgmon

  • 0 in reply to Jason Bridgmon
    Intellectual 500 points

    Hello, on this simulation problem I got a good lead! When run separately: normal behavior (x20 and x50). When run in the same schematic both get x50 gain. I made a package for you with my symbol / schematic and library. Hope it helps :)

    7103.ina20x.zip

  • 0 in reply to Thomas LE BIVIC
    TI__Genius 14825 points

    Thomas,

    I tested your schematic out and see the behavior, then I modified your schematic with my own symbols and models from the web and saw the same thing.  I've passed this on to the model team and we'll see what they say.

    Thank you again, I'll keep you posted.

    - Jason

  • 0 in reply to Jason Bridgmon
    TI__Genius 14825 points

    Thomas,

    Our model team has posited that there are subckts in the macromodel that have the same name.  Some simulators treat everything as a separate entity and we don't see any crossover effects.  However, some other simulators take one or the other subckt for the whole schematic for simulation, which appears to be what is happening here.  We're looking at making our subckts all different names across models to avoid these issues in the future.

    In the mean time, a quick check/repair to this simulation would be to rename the subckts yourself in the model and see if the problems still occur.

    Again, thank you for calling this to our attention!

    - Jason Bridgmon