• State
  • Locked Locked
  • Replies 6 replies
  • Subscribers 48 subscribers
  • Views 1632 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.

Harmonic distrotion problem with OPA847

ali ozgun
Prodigy 30 points
Other Parts Discussed in Thread: THS4031, OPA847

   Dear all,

I'm using OPA847 in IF stage of a downconverter for 10kHz-10MHz band. I attached  the circuit schematic and it is as recommended in datasheet for 20 V/V noninverting gain circuit. In datasheet -74 dBc 2nd harmonic distortion is given for 2 Vpp at 5 MHz. I used exactly the same values and circuit elements but I could obtain only 48 dBc for 2nd harmonic. Measurement results are also attached for both 5 MHz and 10 MHz measurements. I also tried OPA687 and THS4031 opamps on the same circuit but the result did not recover. For 10kHz-10MHZ band I need about 70 dBc dynamic range. how can I recover distortion level.

Thanks.

 

circuit schematic

measurment @10 MHz input Vo=10dBm

measurment @5 MHz input Vo=10dBm

2nd harmonic measurement for 5MHz

  • 0
    TI__Genius 12631 points

    Hello Ali,

      I have a few suggestions to help you improve on the distortion issues you are facing:

    1. The 10MHz frequency- As you may or may not know there is a 10MHz internal reference frequency on many signal sources and analyzers. This internal reference sometimes tends to alter our readings and make them look worse. Because of this we usually offset our frequencies a little bit. I would suggest using , say 4.95MHz with HD2 @ 9.9MHz and retest the part and see if it shows any improvement.

    2. Output termination - To get good impedance matching to reduce distortion we work in 50Ohm environments. Please uninstall R73 and R74 and make R98 a 0Ohm resistor. Your spectrum analyzer is 50Ohm terminated. Also please make the connection from the output to the spectrum analyzer as short as possible and use coax cables if necessary. Note that you will see 4dBm fundamental power at the spectrum analyzer because of the 50Ohm termination.

    3. Spectrum Analyzer settings - In order to maximize dynamic range of the SA, we only look at small spans when making distortion measurements. When measuring the fundamental, I would keep the frequency center at 4.95MHz and span at 5KHz and adjust the Resolution Bandwidth to around 500Hz.. Similarly when measuring HD2, I would try it with frequency center at 9.9MHz and retain the span at 5KHz and the Resolution Bandwidth to 500Hz..

    4. HD2 of the source - Sometimes the signal source has large HD2 as well which could corrupt the results. If you havent already done so, please check the distortion of the source itself (without the DUT) and make sure it isnt too high.

    Please try the above steps and let me know if it helps in improving your distortion measurement.

    Samir 

  • 0 in reply to Samir Cherian
    TI__Genius 15975 points

    Hello Ali,

    Adding to Samir's comments, it seems that the internal circuitry for the spectrum analyzer might be pretty distorted due to the 10dBm output level into the spectrum analyzer. Currently I see the Attenuation to be around 20dB. Please increase the attenuation setting on the spectrum analyzer to around 36dB which should bring down the HD2 distortion products dramatically.

    Best Regards,

    Rohit

  • 0 in reply to Rohit Bhat
    Prodigy 30 points

       Hello,

    Thank you for your quick response.Here are my comments:

    1. The 10MHz frequency- As you may or may not know there is a 10MHz internal reference frequency on many signal sources and analyzers. This internal reference sometimes tends to alter our readings and make them look worse. Because of this we usually offset our frequencies a little bit. I would suggest using , say 4.95MHz with HD2 @ 9.9MHz and retest the part and see if it shows any improvement.

    I have already tried other frequnecies for the measurements. I also measure @4.95 MHz after your suggestion but the results are same with the one I have sent above.

     

     

     

     

    2. Output termination - To get good impedance matching to reduce distortion we work in 50Ohm environments. Please uninstall R73 and R74 and make R98 a 0Ohm resistor. Your spectrum analyzer is 50Ohm terminated. Also please make the connection from the output to the spectrum analyzer as short as possible and use coax cables if necessary. Note that you will see 4dBm fundamental power at the spectrum analyzer because of the 50Ohm termination.

     The resistors R73, R74 and R98 compose 3dB pi attenuator so its input and output impedances are also 50Ohm. I'm using them for matching improvement. But the measurements above were taken without this atteuators. In addition, 10 cm-length coax cable is used between output and SA. By the way, what do you mean with '4dBm fundamental power at the spectrum analyzer'?  I took the measurements @10 dBm output power.

     3. Spectrum Analyzer settings - In order to maximize dynamic range of the SA, we only look at small spans when making distortion measurements. When measuring the fundamental, I would keep the frequency center at 4.95MHz and span at 5KHz and adjust the Resolution Bandwidth to around 500Hz.. Similarly when measuring HD2, I would try it with frequency center at 9.9MHz and retain the span at 5KHz and the Resolution Bandwidth to 500Hz..

    I have already measured fundamental and harmonic signals with narrow span. I put the figure of the wide span measurements for visualization purpose. But exactly the same results were obtained for both narrow and wide spans. In order to maximize dynamic range of the SA, I decreased 'max mixer level' of the SA to -40 dBm as indicated on datasheet.

     

    4. HD2 of the source - Sometimes the signal source has large HD2 as well which could corrupt the results. If you havent already done so, please check the distortion of the source itself (without the DUT) and make sure it isnt too high.

     

    HD2 of the source had been measured before the opamp measurements. @5 MHz and @10 MHz fundamental signals, source has 67dBc distortion level which is quite far away from my measurement results. To be sure, I used 10 MHz LP filter for 10 MHz fundamental signal measurement. With this filter, harmonic level became 95 dBc and I used this signal as an input for the opamp circuit. For this measurements, distortion level of the spectrum analyzer was reduced to about 90-100 dBc by decreasing max. mixer level option as indicated in SA's datasheet.

    Adding to Samir's comments, it seems that the internal circuitry for the spectrum analyzer might be pretty distorted due to the 10dBm output level into the spectrum analyzer. Currently I see the Attenuation to be around 20dB. Please increase the attenuation setting on the spectrum analyzer to around 36dB which should bring down the HD2 distortion products dramatically

     I increased the attenuation level from 20dB to 70dB and HD2 level decreased only 2 dB. I think this reduction is due to the measurement tolerances.

     

     

     Since I tried 3 different opamps as I said before, I thougth that the problem is on my layout or circutry. So, I ordered a demo board from e-store. I hope it will arrive next week. I will let you know the measurement results with this demo board as soon as possible.

  • 0 in reply to ali ozgun
    Prodigy 30 points

    Here are my measurement results with new settings:

    9.9MHz fundamental signal

     

    19.8MHz HD2 signal

     

    4.95MHz fundamental signal

     

    9.9MHz HD2 signal

  • 0 in reply to ali ozgun
    TI__Genius 12631 points

    Hello Ali,

       Does your Spec. Analyzer have 50Ohm termination or are you using a high impedance probe to drive the SA? The measurements in the datasheet are for 2Vpp at the output of the amplifier not at the SA input. Hence, if you have a 50Ohm series resistor followed by a 50Ohm Load, when you measure at the SA input you will see an attenuation of 2, i.e 1Vpp. Currently if you are seeing 2Vpp at the SA input, then the amplifier is putting out 4Vpp. Hopefully this will give you about 6dB of improvement.

    It looks like your setup is good. As you mentioned the layout could be a concern. Please let me know how the measurements go once you receive the TI EVM.

    Samir

  • 0 in reply to Samir Cherian
    TI__Mastermind 22825 points

    Hello,

    To add to everybody's great inputs (I apologize if some of these have already been noted or if they do not necessarily apply in this case):

    1. I have found good distortion measurement information in this article:

    4336.EDN Article Improve Layout Distortion.pdf

    2. The sensitive input device pins should be isolated / moved-far-from the supply pins and their decoupling caps (both supply side and ground side connections).

    3. Tie the supply decoupling cap ground side connections to a dedicated return path back to the power supply

    4. Use the above article Figure 2 to maintain 50Ohm back termination:

    Regards,

    Hooman