• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 47 subscribers
  • Views 515 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

OPA818: Measurement noise much higher than datasheet specification

Jinhao Lu
Prodigy 10 points
Part Number: OPA818

Tool/software:

Hello TI engineer team,

I designed a charge amplifier with OPA818 but the noise measurement was found about 20 times higher than simulation results. So, I designed the following voltage amplifier as a test circuit dedicated to check the Op-Amp built-in voltage noise level. The input of this 500x voltage amplifier is short to ground in order to measure only the noise. 

A straight forward calculation is Vn = 2.2nV/sqrt(Hz) * 500 = 1.1uV/sqrt(Hz). Accounting for current noise and thermal noise, simulation gives 1.32uV/sqrt(Hz) at 100kHz. However, the measurement is 20.4uV/sqrt(Hz), 15 times higher.

Regarding test equipment, I used Keysight DSOX1202G oscilloscope with a noise floor of 0.4uV/sqrt(Hz). OPA818 is powered by +5V and -5V. The voltage source has a noise floor of <1uV/sqrt(Hz), measured on PCB after bypass capacitors. All measurement used ground spring.

Regarding PCB assembly, proper ESD protection was used. The reflow temperature is well below 250C because of low temperature solder (127C melting temp). I have tried with and without IPA cleaning, the noise measurement results were same. 

Regarding IC supplies, I bought from two different suppliers, Mouser and Digikey. They both gave consistently high noise results.

Please educate me if my design, assembly or test setup would cause the excessive noise. I would appreciate any help and advice. Otherwise, I have to suspect the OPA818 ICs do not meet the datasheet specification. Or perhaps this fabrication batch went wrong. 

P.S. I did not buy and test the OPA818 evaluation board is because my current measurement equipment has a higher noise floor than that 7V/V eval board. 

  • 0
    Prodigy 10 points

    I wanted to provide an update on my previous post. After further reading and adjusting my setup, I realized my initial measurement approach was incorrect. I switched from using an oscilloscope to a spectrum analyzer, and now the results align closely with the simulation (1.6 µV/√Hz vs. 1.3 µV/√Hz).

    I apologize for any confusion caused by my earlier suspicion regarding the datasheet's accuracy. I now understand that my setup was the issue.

    I'm still curious about why an oscilloscope, which has a noise floor an order of magnitude lower, isn't suitable for this particular noise measurement. Any insights on this would be greatly appreciated.

  • +1 in reply to Jinhao Lu
    TI__Mastermind 21577 points

    Hello Jinhao,

    First I want to thank you for your details and dedication to sharing all the important information in your post.  Your post makes it easier to help as you wrote good explanations.

    I also am glad to hear you were able to improve your measurement technique and achieve results close enough to simulation to validate your design.  

    Noise is frequency-dependent and can be better examined in the frequency domain (via a spectrum analyzer) than in the time domain (oscilloscope).  Even though the oscilloscope has a low noise floor, the input stage does not have much filtering as the scope wants to be able to measure signals at different frequencies depending on what you are probing/debugging.  The spectrum analyzer is designed to investigate and show aspects of the frequency domain, including the full power and frequency range of a signal.  

    This app note goes into detail on the spectral aspects of noise and how to measure/identify them.  In general, when measuring noise, gain, and/or phase, the spectrum analyzer will be a better instrument compared to an oscilloscope for those measurements.

    https://www.ti.com/lit/ml/slyy061/slyy061.pdf

    Best,

    Alec