• State TI Thinks Resolved
  • Replies 3 replies
  • Answers 2 answers
  • Subscribers 47 subscribers
  • Views 56 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

LMP7721MAEVALMF_NOPB: LMP7721ma nonideal output signal behavior

Cristian Velazquez
Prodigy 10 points

Part Number: LMP7721MAEVALMF_NOPB

Tool/software:

Hello, I was wondering if you can shed some insight into two issues we have been observing with our OpAmp circuit topology. We have a Transimpedance amplifier circuit design (schematic drawn below), that essentially converts a picoamp current from ionized electrons from a VOC gas, into a voltage that we read with an ADC. The gas is ionized via a UV lamp and the ionized electrons induce a current when they travel through the electric field. One plate of the electrode is at -200V and the other plate is connected to the inverting input of the OpAmp.  

1. The first issue we have been seeing is that the output signal of the OpAamp, when responding to gas, has some oscillation behavior. Initially almost all the sensors' output signals when responding to gas seem ideal (picture attached below of ideal response), but then after some prolonged time powered on, most sensors begin to show this oscillation effect. See photo below of oscillation behavior. I am wondering if this could be related to some noise issues on our GUARD net that could be causing an elevated output and nonideal signal response. Technically our Baseline or output of the OpAmp is supposed to be close to ~100mV but when we ignite the lamp, the output Baseline significantly increase to somewhere between 200mV and 350mV.

2. The second issue we have observed is that sometimes some sensors will show an "inverse" response to gas. So when the output of the OpAmp is supposed to swing high, the output will instead drop, but once it reaches a random value above the baseline ref 100mV, we see the signal swing back high. Please see attached photo below of this behavior, PINK signal is showing this problem.

I am wondering if perhaps both of these problems we are observing are related? Possibly noise issues. I was wondering if you had any recommendations as to what to look out for, or any methods you'd suggest for testing and verifying some of these behaviors and how to improve our design. I have written some notes about the circuit and can provide more information if you have any questions.

Notes:

a. This oscillation behavior tends to happen to sensors that have been powered on over time and dont really show this behavior initially. Also, when the sensor was typically reaching say 1.5V when responding to a certain gas concentration, the oscillation behavior will cause the output signal to NOT be able to reach that same level of output voltage and will be lower. 

b. BaseLine is supposed to be at 100mV, but output is always around ~300mV when lamp is ignited. Tends to be roughly closer to the ~100mV when lamp is NOT ignited.

c. Have observed when C15 is broken, missing, or not soldered properly the output signal suffers and doesnt reach the proper voltage levels its supposed to. This is one of the reasons why i believe it could be related to noise issues. Do I need better filtering?

Schematic: link and photo attached

/resized-image/__size/640x240/__key/communityserver-discussions-components-files/14/LMP7721_5F00_Transimpedance.jpg

IDEAL response

Inverted pink response signal

Oscillation behavior

  • 0
    TI__Guru 57986 points

    Hello Christian,

    Remove C13, the capacitor on the output. Never place a capacitor directly on the output of an op-amp - it causes instability. Place it on the ADC side of R50 (may have to adjust the time constant).

    The LMP7721 is a CMOS input op-amp with femtoamp levels of current. The 500M current balancing resistor R9 is not needed - go direct into the input (short out R9).

    Capacitance across the inputs, or on the negative feedback node, can also cause instability. Try removing C17.

    Also make sure your power supplies and reference voltage are stable and clean. At these current levels, even a few mV of supply/reference fluctuation can cause errors (60 Hz hum - a common issue).

    Make sure the diodes are in the dark if they have a glass body and I hope the whole assembly is inside a guarded enclosure.

  • 0 in reply to Paul Grohe
    Prodigy 10 points

    Hello Paul,

    Thank you for your response and insight! I have created some samples that we are currently testing to see if we are able to achieve more ideal response behavior. 

    Do you know where I can find information on transimpedance amplifiers and how to design and understand them better? 


    I will keep you posted on the results of our testing and let you know in a few weeks if we have further questions.

  • 0 in reply to Cristian Velazquez
    TI__Guru 65700 points

    Cristian,

    Your TIA circuit may have DC bias problem caused by large diode leakage current from D8 and D9 that will prevent it from working properly - see below.

    Aside from this, the circuit is NOT properly compensated resulting in sustained oscillation - see below.

    Running AC stability analysis confirms unstable operation with circuit’s phase margin of negative 7 degrees whereas the minimum recommended is +45 degrees.

    Moving C13 to the right of R50 and lowering the feedback capacitor (C14) from 4.7nF down to 400pF increases the phase margin to 49 degrees. 

    Also, as Paul suggested, there is no need for R9 balancing resistor and thus also no need for C19 (both removed - see below)

    Running transient analysis confirms TIA stable operation with the small-signal overshoot of about 25% - see below.

    Below I have attached for your convenience Tina-TI circuit schematics for our own simulations.

    LMP7721 AC stability.TSC

    LMP7721 Transient Stability.TSC