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.

OPA141: The lowest noise amplifier in the 3V3 range?

Part Number: OPA141
Other Parts Discussed in Thread: LMP7711

Hello all.

My application requires to pick up very minute signals from a high output-impedance source, and so the first stage has to be as low-noise as possible. I have been looking for an appropriate amplifier for quite some time now, and the best I could find was OPA141. Unfortunately, it runs on a minimum unipolar supply of 4V5, and my application can provide no more than 3V3. So, I am stuck there, again!

Does TI manufacture any low-noise op-amp with similar, or even better, noise specs and CMMR? I have spent a few days now putting in different parameters into TI's product selector, but could not find anything impressive.

Please note that my application spares me very little space, and since the input impedance is high, the system cannot afford input bias currents of more than 20pA.

  • Below please find the list of op amps meeting all your stated requirements:

  • Thanks, Michael, for the suggestion.

    Saw that. It has much higher current noise and much lower CMRR. I am afraid my application can't afford that.

  • Thank you, Marek, for taking the troubles.

    Checked each one of those. None is as good as OPA141, especially in terms of input referred current noise and CMRR.

  • You are apparently just reading specs. 

    Perhaps you should construct a total noise and output DC error calculation and see if the current noise and CMRR are dominant contributors. likely not. 

  • For the CMMR, yes! I am just reading the specs.

    But, as I mentioned, the output impedance of the source is very high, of the order of 10MΩ. So, input current noise plays a very significant role. Cannot afford a current noise of more than 1fA, the lesser the better.

  • If it is 10Mohm resistive, what is the Johnson noise of that? 

    These low current noises are actually never measured, just estimated by leakage DC currents - to measure them, the R noise dominates. 

  • Yes, the thermal noise for the 10M source resistance dominates.

    Now, here comes another challenge into the picture. The source impedance is not constant!

    It depends on a lot of system parameters and varies between 10K and 10M. And so, I need to consider the worst-case scenarios.

    The source presents an output resistance of about 10K or so for about 60% of its operation time, and I need to ensure that the system noise is minimum under all states of operation.

  • Abbas,

    Without getting into the merits of your concern about the input current noise contribution (thermal noise will dominate the total noise for high input source impedance), I would like to point out that the input current noise is a function of the input bias current and at low frequency Inoise=sq-rt(2*q*IB).  Therefore, the lower IB results in lower input current noise.  Thus, OPA141 input current noise of 0.8fA/rt-Hz shown in the datasheet is a calculated value and the Inoise values in LMP77xx datasheets are in error - all showing very high Inoise of 0.01pA/rt-Hz but it would take 31nA of the input bias current to generate such noise while the actual IB is in 1pA range.

    Below please find the calculated values for some of the parts based on their typical IB.  You may see that as expected that Inoise of OPA141 is the highest because its typical IB is the highest.

    In addition, as you pointed out yourself, the minimum supply voltage for OPA141 is 4.5V so you may NOT power it from 3.3V; 

    Please note that the input common-mode voltage range of OPA141 requires the input 3.5V or more below positve rail so on 3.3V single supply, in a buffer configuration, the input would have to be Vin<-0.2V, which would begin to forward-bias ESD protection diode between the input and the negative rail, thus dramatically increasing IB and Inoise.  On 4.5V single supply, OPA141 linear input voltage range is only from -0.1V to 1V.

  • Marek Lis said:

    I would like to point out that the input current noise is a function of the input bias current and at low frequency Inoise=sq-rt(2*q*IB).  Therefore, the lower IB results in lower input current noise.  Thus, OPA141 input current noise of 0.8fA/rt-Hz shown in the datasheet is a calculated value and the Inoise values in LMP77xx datasheets are in error - all showing very high Inoise of 0.01pA/rt-Hz but it would take 31nA of the input bias current to generate such noise while the actual IB is in 1pA range.

    Below please find the calculated values for some of the parts based on their typical IB.  You may see that as expected that Inoise of OPA141 is the highest because its typical IB is the highest.

    Thanks for bringing that point up. It really helped me, especially in properly directing my searches.

    I am truly grateful.

    Now, the question arises, where did they get those figures from, if those figures are not the actual current noise? Such erroneous data is not expected of TI!

  • So really the next point to make is the input bias current quoted also increases with temperature doubling every 10C normally. So every 20C, the input current noise has also doubled. 

  • Yes.

    But that is not of much concern to us. The temperature is maintained within a fairly decent range. As the temperature increases beyond that range, the output impedance of the source starts decreasing.