What input resistance for AD7960?

I would like to know what would be the maximum recommended resistances to place between an FDA and AD7960 inputs (R3 and R6 on the figure below). I would like to reduce the capacitor value to reduce power consumption of the LTC6363 and increase the resistances accordingly to keep cutoff frequency. I am thinking of 200R or 300R instead of 100R. The input signal is a amplitude modulated sine wave, 125kHz.

Thank you in advance for your inputs.

  • 0
    •  Analog Employees 
    on Oct 21, 2017 3:38 AM

    Hi,

       We'll check on this and get back to you.

    Regards,

    Jonathan

  • Hello, Jonathan, thank you.

    As a complement to the info above, I do not need absolute accuracy in terms of A/D conversion. I may live with small errors due to input bias current or bias due to AD7960 S/H capacitance. My main concern is the higher resistances degrade SNR too much or add too much error or some other thing I may be missing.

    Regards.

  • 0
    •  Analog Employees 
    on Oct 24, 2017 7:59 AM

    Hi EVC,

    True increasing the R value would hardly degrade  the SNR and THD performance of the ADC. I'd recommend to use lower R or the recommended RC in the UG.

    regards,

    LLoben

  • Hello, LLoben,

    Thank you for your input.

    Can you provide more details? Even an estimate with rationale, based on the AD design, would be useful, even if not tested in production or characterized. 

    THD is not as critical as long as it is stable (not noisy). Noise is more critical though I still have some margin. Neither the data sheet nor the UG provides information on acceptable resistance range, they just show 33R resistors in series with each input. The closest info I could find is the input leakage current (60nA, typical) that would not add too much error for my application with 200R or even 300R. On the other hand, the graph on figure 24 is confusing, actually disturbin, as it would add a huge error even with the 33R of the design reference, assuming I interpreted the meaning of that graph right and the unit (mA) in the vertical axis is right  (is it?).

    Working with lower value resistors as indicated on the DS and UG will require a significant design change because the FDA is supplied with -5V and output voltage may be negative under some circumstances. There is a Maxim FDA that has provisions to limit the output voltage but I'd rather keep the LT part in the design (I assume AD would too  ).

    Again, thank you for your support.

    Elder.

  • 0
    •  Analog Employees 
    on Oct 26, 2017 9:42 AM

    Hello Elder,

    If its okay I'd like to know  what's your end application? We have verified the recommended RC(33Ω, 56pF) filter that offers that optimum performance for both muxed and non-muxed use cases. Increasing R to 200 to 300 will certainly affect the THD performance, not so much SNR. However, I'd strongly recommend to verify the SNR/THD performance on the evaluation board for a given RC combination.

    regards,

    LLoben