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.

OPA690: Harmonic distortion measurement

Part Number: OPA690

Hi all,

I need your help to understand how to reach the THD performances reported in OPA690 datasheet. Below the application circuit:

sys_attachment.do?sys_id=49af0abf9740a954d7ccb4d3f153afd0

 The input signal is 1Vpp, sinewave, 1kHz; the spectrum analyser input impedance is 50ohm. I get this bad FFT:

sys_attachment.do?sys_id=3b009a339780a954d7ccb4d3f153af40

So basically the 2nd and 3rd harmonic are just 84 and 77 dB far from fundamental, while according to OPA690 datasheet I should get at least 100dB, at this amplitude, frequency and load. 

My question is: where is may circuit wrong? I suspect non idealities of 4 resistors, which are standard thick film 0603. Could you provide me guidelines on how to improve the circuit? Which kind of resistors you used to test and report into the datasheet?
 
Thanks
Kind regards
Federico
  • Hi Frederico,

    your feedback resistors are way too low. Increase them to 402R as mentioned in the datasheet. Also see section 8.3.6 of datasheet.

    I would use the circuit of figure 36.

    Kai

  • Hello Federico,

    Kai's point is important: when you are investigating datasheet parameters in the lab, it is usually easier to find success in the investigation by using the datasheet circuits and recommended values.  

    Let us know if you find any additional questions or want to share updated results.  I am curious if the small proposed edits will resolve the discrepancy outright.

    Best,

    Alec

  • Hi guys, thanks for quick reply; I have checked changing to 402R and also 1Kohm, but the results are basically the same. To be fair, the differential amplifier config is not reported in the datasheet, but I do not see why this shouldn't work as expected. In literature there are discussion on which resistors offer the best linearity, especially in audio band, hence my original question "what kind of resistors were used".   

    I'll try also to add a cap between supply pins as suggested in par 8.3.6.

    Federico

  • Morning Federico, 

    your diff. amp configuration suggests might be having its own issues. Is this coming from DAC outputs by any chance? If so, there is a subtle loading mismatch to the DAC that causes mismatch output voltage swings on the DAC pins. Essentially, I am saying the source needs to be very low distortion to begin with. If it is a DAC, we can improve that output swing match with some different resistors. 

    I think you are referring to this plot, note this is only 2Vpp at the op amp output pins, what is your Vpp in this test at the output pins, 

  • Hi Michael, with a high performance differential acquisition card, I have qualified the DAC differential output itself, and when connected to our diff.amp and it is fine (meaning >110dB). The output pin Vpp is 2Vpp, same as in the plot. Lowering the input signal to one tenth, the THD performance remains the same.

    Federico

  • Ciao Federico,

    I would first check the distortion with the circuit shown in figure 36 and then go to the differential amplifier.

    For the resistors in the signal chain I always take thin film resistors, not thick film. Susumu has nice thin film resistors in 0603 or 0805:

    https://www.mouser.de/ProductDetail/Susumu/RG1608P-4020-B-T5?qs=juJhw1yPxAFIb5h9eQrp3g%3D%3D

    https://www.mouser.de/ProductDetail/Susumu/RG2012P-4020-B-T5?qs=1srY7g%252BY2jgkY%2FNOUJV4YQ%3D%3D

    Panasonic also has very good thin film resistors.

    By the way, the layout also has a deep influence on the performance of circuit. A solid ground plane is highly recommended. Can you show your layout?

    And the power supply decoupling can make issues as well. I always use Pi-filters containig ferrite beads in the power supply lines of HF-OPAmp when high perfromance is crucial.

    Kai

  • Hi Federico,

    can you describe your whole circuit / setup being used for the distortion measurement?

    Kai

  • So with the OPA690 circuit connected, try measuring the voltage swings with a scope at the two DAC outputs pin. If they are different amplitudes that might be producing an HD2 term right at the inputs to the circuit. This came up recently here, I have the spreadsheet working now fairly well, not perfect, but if I had more DAC info, could easily apply it. 

    https://e2e.ti.com/support/amplifiers-group/amplifiers/f/amplifiers-forum/1183395/opa690-output-voltage/4457145#4457145

  • Hello Federico,

    There are quite a few factors which could contribute to the issues you are encountering.  Kai & Michael have touched on some important ones, such as PCB layout.

    I would like to add another: are there any filters being used in your setup, at any point along the signal chain?  There could be some amount of attenuation present which is not already accounted for.  Additionally, you may encounter issues overloading the spectrum analyzer by passing in the fundamental frequency as such high gain (to the spectrum analyzer), but then having the subsequent frequency points be much lower in magnitude.  Please check that your instrument is operating within its safe regions.

    I think if we keep swapping information, we will be able to help you out on this topic Slight smile.

    Best,

    Alec

  • Hello Kai, I will try thin film resistor soon. The board has a solid ground, several kinds of supply bypass caps but no special filtering on supply. Consider I work below 10kHz. 

    Thanks

    Federico

  • Hi Kai, full setup is: a PXI-4463 generates the diff signal, directly connected to the OPA690 board. Its output is going to high performance Spectrum analyser through SMA conn and coax cable. Supply is coming from a dedicated low noise bench power supply.

    Thanks

    Federico

  • Hi Alec, I do not use any filtering along the signal chain; about spectrum analyzer it has max +30dBm, so with +20dBm I should be well in spec, but I'll try if with lower power levels it improves.

    Thanks

    Federico

  • Normally, I recall we targeted about -30dBm for the fundamental at the mixer. This means you dial in attenuation internally to the spectrum analyzer. Soon, you run into a noise floor for the spurs, where then you zoom way in on the spur and drop the res. bandwidth as low as possible - doing that, it really helps to lock the spectrum analyzer to source to find the spur. 

    Or, you could add a notch filter at the fundamental after the DUT. 

  • Hi Federico,

    Consider I work below 10kHz.

    Not your low signal frequency of 10kHz is demanding a proper supply voltage filtering but the 500MHz bandwidth of OPA690.

    A simple decoupling cap may work, if you have only one OPA690 hanging on the supply voltage line. But when your circuit contains more than one HF-OPAmp I would recommend to provide Pi-filters. I always use Pi-filters in the supply voltage lines of HF-OPAmps when I work with them. This is such a fundamental advantage over simple decoupling caps...

    Kai

  • Hello Federico,

    Please let us know if you make edits to your setup or circuit so we can track your progress & follow your actions.  I will be in the lab the rest of Monday January 9th and may not respond right away.

    Best,

    Alec

  • Hi Alec, this week I'll have very few available time for this topic. Likely I will try low distortion resistor, measure the voltage amplitude at both P and N inputs to see if this is balanced or not, and I will try with lower input power.  

    Thanks

    Federico

  • Hello Federico,

    Okay; that is alright Slight smile.  I will not forget about you next week.

    Best,

    Alec