• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 48 subscribers
  • Views 1377 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.

Op amp with low peak-to-peak noise over 0.05-150Hz (searching for)

Leon Balchin
Prodigy 190 points
Other Parts Discussed in Thread: OPA314, OPA333, TLC2274, OPA378, OPA188, OPA1611, OPA4377, OPA377, OPA320, OPA322, OPA316, OPA4316

I am looking for a cheap op-amp to use as an input buffer for an ECG application.

The difficult feature to search for is low Vpp noise in the range 0.05-150Hz. It is not listed in the standard "compare" parts feature of the TI website. The "Vn" column is not that helpful, since it is at 1kHz. For example, OPA333 has Vn 55nV/√Hz at 1kHz, but only 1.1 μVpp at 0.1-10Hz, while OPA314 has only Vn 14nV/√Hz at 1kHz but has 5 µVpp at 0.1-10Hz.

How can I search for such features, other than checking hundreds of datasheets manually?

How can I calculate the Vpp over the range 0.05-150Hz? The noise graphs don't usually go below 10Hz for some reason.

I certainly need a Vpp of less than 5 μVpp over the range 0.05-150Hz, and am really looking for something more like 1-2 μVpp.

Other highly desirable features include:
Rail-to-rail input and output
Low bias current
Low power
Suitable for 3V supply

TLC2274 has a low Vpp but input is only to 1.5V from positive supply, it drains a lot of current and requires 4.4V supply.

  • 0
    TI__Guru 62905 points

    Hi Bob,

    Unfortunately the search tool doesn't feature a way to sort by noise over the 0.1-10Hz bandwidth.  I'll make a note to our marketing team that this would be a good improvement to the search tool.  

    Regarding recommended op amps, take a look at the OPA378 chopper amplifier which features 400nVpp over the 0.1-10Hz bandwidth.  This device is similar to the OPA333 and does not have 1/f noise due to the chopper architecture.  It features rail-to-rail input/output, bias currents of 150pA, and quiescent currents of 125uA.

    The OPA188 is another option but is better suited for higher-supply voltages.  The OPA188 features 250nVpp over the 0.1-10Hz bandwidth but does not have a rail-to-rail input stage and is only specified from V- to (V+ - 1.5V). 

    Also, while it's not a good fit for this application, a low-noise non-chopper amplifier to look at would be the OPA1611.  It does not feature a rail-to-rail input stage and draws 3.6mA, but features 1.2uVpp noise over a 20Hz - 20kHz bandwidth and roughly 600nVpp in the 0.1-10Hz bandwidth.

    To boost your op amp noise analysis knowledge I would recommend going through the Noise Analysis videos in the TI Precision Labs (TIPL) online lecture series.  The videos will describe how to calculate the noise for any op amp circuit based on the 1/f, broadband noise, resistor configuration and bandwidth of the circuit.  As long as the 1/f corner frequency is included in the noise graph then you will be able to calculate the noise at frequencies lower than what is shown in the graph since the 1/f noise has a slope of "1/f".  The equations to do this are shown in the TIPL presentations. 

  • 0 in reply to Collin Wells
    Prodigy 190 points
    Thanks. I'll have a look at the videos when I get some time.

    I'm concerned the input biases of those op amps may be a bit too high for ECG buffering (OPA378 datasheet says the max is ±670pA), and getting too expensive. The device has to pass a test with a 51kOhm source impedance mismatch (between different buffers) without affecting the CMRR much (which should stay around 100dB) and a 20uVpp sine wave needs to be visible.

    I found OPA4377 which seems like a good choice, apart from that the offset changes when it gets within 1.3V of +Vs. If it went into this range during operation, would the offset jump suddenly?
  • 0 in reply to Leon Balchin
    TI__Guru 63890 points

    Bob432,

    Exceeding the OPA377 input common-mode voltage range will not typically result in a large change in the op amp DC performance like the input voltage offset BUT may greatly effect AC parameters like a significant decrease in the bandwidth and slew rate as well as the increase in the input voltage noise especially when the common-mode voltage is closer than 1V below positive rail (Vcm>Vcc-1V).

    For more detailed information on this subject, please read my blog under following link:  https://e2e.ti.com/blogs_/b/precisionhub/archive/2014/11/18/perils-of-using-op-amps-outside-the-specified-conditions

    Please consider using OPA320 or OPA322 with the rail-to-rail input common-mode voltage range AND 0.1-10Hz input voltage noise of 2.8uVpp.  If quiescent current is a concern, you may also consider using OPA316 with RR input range and 0.1-10Hz input noise of 3uVpp but with lower CMRR figure. 

  • 0 in reply to Marek Lis
    Prodigy 190 points

    Thanks for your suggestions.  I will probably give OPA4316 a go, and the other two are good to know about; the shutdown feature of OPA4322S could come in useful.  However, unfortunately none of them is as good as TLC2774 noise-wise (1.4uVpp + ~1.6Vpp for 150Hz BW).  If I were to boost the power supply slightly to compensate for a limited RRIO, is there an op-amp with lower Iq or cost than TLC2274 that I should consider?

    Regarding OPA377, Fig 22 from the datasheet appears to show a sudden jump in the offset when the input is 0.9V from the positive supply.  Would this not cause a sudden jump/drop of 1mV in the output?  (The ECG signal itself is only +-5mV.)

    Thanks.

  • 0 in reply to Marek Lis
    Prodigy 190 points

    Will the offset jump suddenly when Vcm reaches about 0.9V from positive supply when using OPA4377?  If not, what does the graph mean?  Thanks.

  • 0 in reply to Leon Balchin
    TI__Guru 63890 points

    Bob,

    OPA4377 has a crude second input differential pair (n-channel) that takes over from p-channel diff pair when the input common-mode voltage is closer than 1V below positive rail. This diff pair is not meant to be used for normal operation (since its gm is much lower and thus its bandwidth would be much lower) but it is there in order to prevent a phase inversion. It typically takes several hundred millivolts (300mV-500mV) of change in the input common-mode voltage to transition from p-channel (normal operation) to n-channel (crude operation) and the result is random change in Vos between p-channel pair and n-channel pair as shown on the offset vs Vcm diagram above: Vos may go from low to high, high to low and anywhere in between, or it may hardly change across the entire Vcm range (its behavior is random like the input offsets of two independent op amps). It is important to understand that only p-channel pair Vos is tested to +/-1mV limit since n-channel pair operates only beyond maximum specified Vcm.