• State TI Thinks Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 1 answer
  • Subscribers 40 subscribers
  • Views 1788 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.

TINA/Spice/LMP2012QML-SP: Input common mode range when powered by 3.3V and Pspice model.

Michael Ford
Prodigy 160 points
Part Number: LMP2012QML-SP
Other Parts Discussed in Thread: TINA-TI, LMP2012, LMP2011,

Tool/software: TINA-TI or Spice Models

I have recently resimulated a circuit at 3.3V that previously worked at 5V and found it to not simulate properly.

On investigation the issue appears to be that when powered by 3.3V having the LMP2012 amplifiers biased at 3.3V/2 =1.65V does not work (using pspice model LMP2011).

That datasheet indicates specs at VCM=1.35 (half supply) when powered by 2.7V but then list a mush narrower range of −0.3 ≤ VCM ≤ 0.9V when specing CMRR? What is working range?

Is the Pspice model LMP2011 accurate for common mode input voltage range for LMP2012QML-SP at Vsupply=3.3V?  Will LMP2012 function normally at 3.3V with VCM=1.65V? 

I do have some working engineering boards but need to decide if I should be shifting the bias voltage downward from mid scale if this bias level is unreliable.

  • 0 in reply to Michael Ford
    TI__Genius 9775 points
    We're looking into this.

    BTW, since you say that you have engineering board already, do you see the problem at the board level as well or do you only see it on pspice model thus far?
  • 0 in reply to Herman Theodorus
    Prodigy 160 points
    The engineering/Qual boards all worked fine at 3.3V. I only became concerned when I tried to update some simulation based analysis with supply lowered to 3.3V and simulated LMP2011 models failed to operate properly. I have some production board already in process (and already conformal coated) and I need to decide if I need to rework to make component changes to make the common mode voltage lower for reliability or if this is just a simulation model issue.
  • 0 in reply to Michael Ford
    TI__Genius 9775 points
    I suspect strongly this is just simulation issue because even though the datasheet does not specify explicitly CMIR, but it does say through the table that it works on VDD/2. Thus the model should work as well at VDD/2.

    I am checking with our team on this.
  • 0 in reply to Herman Theodorus
    Prodigy 160 points

    Is there any recommendation here?  Will I get normal circuit operation at CMIR VDD/2 when powered by 3.3V (end of life, with radiation, over temp) and ignore the Pspice model or do I need to update my designs to use a common mode input at voltage closer to negative rail.  Currently I am assuming the Pspice model does not reflect the part.

    Thanks,

    Mike

  • 0 in reply to Michael Ford
    TI__Mastermind 27715 points

    Hi Mike,

    It is my recommendation that for the highest reliability in your design across all variables, you should not exceed a CMIR of Vs - 1.8V. While the LMP201x datasheet does not state this explicitly, it implies it in several places.

    • The CMRR spec for Vs = 2.7V is defined for Vcm up to 0.9V, or Vs - 1.8V.
    • The CMRR spec for Vs = 5V is defined for Vcm up to 3.2V, or Vs - 1.8V. 
    • In Figure 4, Vos vs. Vcm for Vs = 5V, Vos starts to increase dramatically when Vcm = 2.2V (Vs - 2.8V) at 125°C and when Vcm = 2.75V (Vs - 2.25V) at 85°C. This trend suggests that the threshold for Vcm at 25°C is just off the plot past 3.5V, or Vs - 1.5V. Vos increasing dramatically is a telltale sign of the input stage becoming nonlinear.


    • In Figure 6, Ib vs. Vcm for Vs=5V, Ib starts to shif downward at Vcm = 3V (Vs - 2V) and then moves downward dramatically at Vcm = 3.5V (Vs - 1.5V). The dotted-line plot past Vcm = 3.5V indicates more unpredictable behavior in this region. This is also a strong indicator of the input stage becoming nonlinear.

    Adjusting the common-mode voltage of the SPICE model down to 1.45V, or Vs - 1.85V, gives the expected results. In fact, I was able to bias the Vcm up to 1.58V, or Vs - 1.72V, before the model behaved nonlinearly. However, I still recommend staying below Vcm - 1.8V for reliable operation.

    LMP2011CMRtest2_iw.TSC

    Depending on your application's maximum temperature, I would recommend pushing Vcm down even further based on the information from the Vos vs. Vcm plot over temperature.

    Best regards,

    Ian Williams
    Applications Engineer/SPICE Model Developer
    Precision Amplifiers