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.

LMP2012QML-SP: LMP2012QML-SP

Part Number: LMP2012QML-SP
Other Parts Discussed in Thread: ADS1278, LMP2012, LMP2011, TINA-TI, THS4511, THS4513, THS4304, THS4521, LMP2012QML, ACF2101, IVC102

Hi,

What is output impedance and maximum capacitive load driving capability of the op amp LMP2012QML-SP?

Thanks

Dinakar

  • Hi Dinakar,

    Question:

    What is output impedance and maximum capacitive load driving capability of the op amp LMP2012QML-SP?

    Are you able to share the schematic and application with us? Op Amp does not like to drive capacitive load. If you do, it  has to be properly compensated, otherwise, it may oscillate in a circuit. 

    I need to see what types of load you are driving in an application. I am unable to answer the questions without a schematic information. 

    Best,

    Raymond

  • Hi Raymond,

    Please see my schematic.

    Thanks,

    My Schematic.TSC

    Dinakar

  • Hi Dinakar,

    Your circuit is not working. Please describe what you are trying to do in the following schematic capture, so that I can fix it. 

    It looks like that you want to inject a step constant pulse current into Transimpedance Amplifier, followed by LPF at output and passed onto the next stage. Once this part is working, I will try to figure out the next stage. 

    Best,

    Raymond

     

  • Hi Raymond,

    You are right. It is a TIA followed by a LPF. The output of LPF is converted to a differential signal. Then the signal is applied to ADC ADS1278 after passing it through a charge bucket filter. The current generator is for simulating photodiode signal at 10 Hz rate. The ADC sampling rate is 12.8 kHz and data rate is 100 Hz.

    I defined the current generator as unit step type. This can be changed to periodic linear signal for transient analysis.

    I am not finding open-loop ouput resistance value and capacitive load driving capability of the op-amp in its datasheet.  Those values are required to finalize the charge bucket filter design by ensuring stability of the op-amp and settling time of the signal for up to 16-bit resolution.

    Actually, I want to use LMP2012 op-amp instead of LMP2011. But its macro model is not available with TINA.

    Please provide the required values so that I can do the op-amp stability analysis.

    Thanks,

    Dinakar

  • Hi Dinakar,

    the TINA-TI simulation gives this result:

    dinakar_lmp2011.TSC

    Kai

  • Hi Kai,

    Thanks for the simulation. I understood that the Zol at 10 Hz is 10 Kohm. Also simulation is showing that Zol at 10 Hz is 50 ohm with V+ = 5 V and V- = 0V. Is it power supply dependent? I am new to this particular area. Please clarify.

    Thanks

    Dinakar

  • Hi Dinakar

    I am not still sure what you wanted to do. So I modified your circuit and try to figure out what you were trying to do. 

    1. For TIA, your current step is driving downward, since Vout = -Isensor*R18 (you had a single rail, and you want the output voltage to be positive. 

    2. The LPF has the BW around 1/(2*pi*160k*100nf)=10 Hz, which is near DC response. 

    3. I compensated your TIA with BW set approx. 786kHz. 

    Enclosed is the modified of your circuit. Please simulate it and see if this is what you want. 

    /cfs-file/__key/communityserver-discussions-components-files/14/LMP2011-05052020.TSC

    It is very nice for Kai to provide LMP2011's output impedance simulation for us. Since you are new in this, I am going to briefly explain what the expert did. 

    1T = 1 TeraHenry, only used for simulation purposes. It provide DC bias and DC negative feedback to the op amp, and it blocks all AC frequency, because impedance of inductor is sL or 2pifL. 

    1T = 1 TeraFarad, only used for simulation purposes (1TH or 1TF does not exist in the real world, at least not on earth). The capacitor's impedance is 1/sC. It blocks DC signal, and allow AC signals to pass. 

    When you want to measure output impedance of a node, you can inject a pulse current or sink a pulse current from that node. A momentary voltage changes are measured vs. the changes in a programmed pulse current, see Kai's captured image and simulation. The rate of change in dVout/di is the  op Amp's output impedance by definition. 

    If you have any additional questions, please let us know. 

    Best,

    Raymond

  • Hi Kai,

    So Zol is 10 kohm at 10Hz. Then how to do stability analysis for the above circuit? My frequency of interest is 5 to 10 Hz.

    Thanks,

    Dinakar

  • Hi Raymond,

    The TIA outputs positive voltage only. My input current range is from 0 to 25uA. For the input current of 0A, theTIA output is 2.8V and for the input current of 25uA, the TIA output voltage = 0.3V. I have used this configuration for connecting common cathode type segmented photodiodes with a single supply op-amp in zero bias mode.

    Also, my frequency of interest is only 10Hz. The TIA cut-off frequency (160 Hz) is chosen one decade above the LPF cut-off frequency (10 Hz) to decrease noise gain of the TIA.

    The second stage is a driver (single ended to differential) for the ADC ADS1278.

    Please comment about its stability while driving  the  ADC. The pole and zero formed with the ADC charge bucket filter and the op-amp open loop resistance (10 kohm at 10 Hz as per Kai's simulation) are more than one decade apart.

    Thanks,

    Dinakar

  • Hi Dinakar,

    yes I know, everything seems to be confusing, if you are new to all this  :-)

    I would recommend to see the very nice TI's stability training videos:

    See the whole series!

    By the way, instability usually occurs at the high frequencies. So, the open loop output impedance at very low frequencies might not be of much interest when determining stability.

    Kai

  • Hi Kei,

    Ok thanks. I will see the videos.

    Dinakar

  • Hi Dinakar,

    In your initial circuit, you had 1A step current injected into TIA op amp, which it will saturate the op amp and wrong injection direction. Let us assume a small amplitude step function in current is injected into V-,  how does 160Hz BW Op Amp is going to capture a transient event? 

    Enclosed is an Op Amp stability analysis training video, and I suggest that you take a look at it. There are 7 video clips in the stability series. 

    https://training.ti.com/ti-precision-labs-op-amps-stability-introduction?context=1139747-1139745-14685-1138805-13848

    You have a single ended signal, if you want to convert the single ended to differential input, please see  Figure 90 in ADS1278, which is Delta-Sigma Wide Bandwidth, up to 144ksps ADC. You have 10Hz signal BW after LPF, why you want to use the high speed and high BW ADC at your acquisition circuit. 

    Best,

    Raymond

  • Hi Raymond,

    Correct. The TIA saturates for 1A current input. But, my maximum input current is 25 uA and it ramps up from 0 to 25 uA at 10 Hz rate. That is why I set my LPF cut-off frequency also as 10 Hz for better noise filtering (though 3dB less gain).

    My concern is for the ADC driver and charge bucket filter. The fully differential driver THS4521 is a good suggestion. But I need a space qualified driver. The space qualified ADC drivers THS4511, THS4513 and THS4304 are not suitable for low frequency applications due to high 1/f noise. So, I decided to use single op amp method for single-ended to differential conversion. The LMP2011 op amp is available in space grade version as LMP2012-SP.

    Stability analysis for a single stage output type op amp is discussed thoroughly in the videos. But i think, the LMP2011 is a two stage output type. Its open loop output resistance (Zol) varies with frequency. Its Zol at 10Hz is 10 Kohm (as per Kai's simulation). My charge bucket filter values are R = 100 ohm and C = 1 nF. The filter settling time is good enough for 16-bit resolution (12*RC = 1.2 usec which is less than half of the ADC sampling rate of 128 kHz) . The zero frequency formed by the filter is 1/2*Pi*RC = 1.59 MHz and the pole frequency is 1/2*Pi*(R+Zol)C = 15.9 KHz. Difference between the zero and the pole frequencies is two decades apart. But according to the video 1335, the difference should be less than 1.5 decades maximum. So should I increase the R value to meet this requirement? Is the Zol value I have taken correct? Increase in R value causes settling problems.

    Please correct me if I am wrong in the calculations.

    Thanks,

    Dinakar

  • Hi Raymond,

    I have used out-of-the loop  compensation technique for my calculations. Advantage of this technique is no settling issues as with the double feedback or in-the-loop compensation technique.

    Thanks

    Dinakar

  • Hi Dinakar,

    The TIA saturates for 1A current input. But, my maximum input current is 25 uA and it ramps up from 0 to 25 uA at 10 Hz rate. That is why I set my LPF cut-off frequency also as 10 Hz for better noise filtering (though 3dB less gain).

    I am not sure that I understand what your application is. I need to understand what you are trying to measure. Are you trying to integrate the current? or are you trying capture the entire current ramp up event from 0-25uA in 100msec? Again, the current description is different from your initial circuitry, which is a transient current injection to TIA, I recalled. 

    Out of the loop compensation means open loop; and nothing is special about it. It means that the op amp will be stable (no oscillation) no matter what you connect it at the output, but you will not be able to control the output, because the circuit has no feedback with respect to the input. 

    Here is something you need to be aware. 

    1. If this is for aerospace application, your operating temperature range is required to be -55C to 125 or 155C. 

    2. If you are concerned about 1/f noise, then zero drift (chopper amplifier) has to be place in the analog front end, where injected current is taken place.  

    3. When you closed a feedback loop, the output impedance is lowered by Zol/(1+T), which is totally different from open loop Zol in op amp. 

    4. Can you send me Tina simulation about your Op Amp circuit, so that I can help you to  check the stability of the op amp circuit? It is really difficult to read your texts and not be able to visualize the bode plot. 

    Best,

    Raymond

  • Hi Raymond,

    Sorry, I was not available. So could not reply to you. 

    Yes, I want to integrate current for 100 msec. Current will ramp up from 0 to 25 uA in 50 sec. There was a mistake in the transient current injection in the circuit.

    I will use LMP2012QML for my application. It has operating temperature range from -55 to +125C. LMP2011 is a similar device. I am using LMP2011 for simulations only since I could not find SPICE model in TINA for LMP2012QML.

    I have simulated my single-ended to differential signal conversion circuit with different Riso values. I have used indirect phase margin measurement technique for finding stability by measuring signal overshoot percentage. In case 1, phase margin is < 45 degrees. In case 2, it is ~46 degrees. In case 3, it is ~58 degrees. So can I conclude that case 3 is OK? In all the three cases, signal is settled for 16-bit resolution (38 uV) with in ADC acquisition time (3.8 usec).

    Here I have not included transimpedance circuit.

    Please comment.

    Thanks

    Dinakar

    My Schematic - Simulations.TSC

  • Hi Dinakar,

    Yes, I want to integrate current for 100 msec. Current will ramp up from 0 to 25 uA in 50 sec. There was a mistake in the transient current injection in the circuit.

    Are you sure?

    Please, Dinakar, what exactly is your signal?

    Kai

  • Hi Kai,

    The signal I want to measure looks like this.

    Thanks

    Dinakar

  • Hi Dinakar,

    You will need an integrator for the I vs. time. I have another question. Are you going to integrate 29 msec in a single captured event? or are you going to  capture the details of current rise as a function of time?

    Best,

    Raymond 

  • Hi Raymond,

    I want to capture the current rise details as a function of time.

    Thanks for your patience.

    Dinakar

  • Hi Dinakar,

    I think that you are still able to use your initial TIA circuit, and your output of TIA will be 100kOhm*input_current vs. time (up to 29msec). The TIA's output does not need to go through 10Hz LPF, and you can feed the input voltage directly to your ADC (You may convert single ended to differential input voltage for your ADC) .

    Once your Vout vs. time profile is captured in your ADC, you can handle whatever your application is required from the data. If you are concerned about noise, you may place a LPF before the ADC. I would not recommend to limit your BW right after TIA, since you are interested the current changes over time. 

    Enclosed is an application note for TIA. Also TI  has IVC102 and/or ACF2101 integrated TIA . I don't think that you will need these, but you can use these as a reference when you are designing your TIA front end. 

    http://www.ti.com/lit/an/sboa268a/sboa268a.pdf?&ts=1589903106735

    Best,

    Raymond

  • Hi Raymond

    Thank you very much for the information. I will read the application note.

    Regards,

    Dinakar