Post Go back to editing

What is optimum load impedance of ADL5562 ?


We are using ADL5562 in our design as driver for AD9467.

I have downloaded S-parameter of ADL5562 and I have simulate in ADS using input termination as 40.2 ohm on each input.

As per datasheet of ADL5562 gain is 15 dB when load RL = 200 ohm but I can able to get only 9 dB. But when I use RL=6 ohm I can get gain as 15 dB.

What RL should I consider for designing anti aliasing filter for ADC ?

RL will be input impedance for anti aliasing filter.

Regards,

  • Hi,

    I moved your question to RF & Microwave community. Someone there will be able to help you.

    Thanks,

    Phil

  • Hi,

          Adding results and schematic of simulation for better understanding.

    Regards,

  • Hi, Dhavlap,

        The ADL5562 has a differential input impedance of 133 ohm and output impedance of 12 ohm.  For the input, matching to a 133 ohm source will give you max. gain.  For the frequency you use for testing, VSWR shouldn't be an issue unless you have very long cables.  Please clarify which gain your results are showing: the spec is for a 15.5 dB max. DIFFERENTIAL VOLTAGE gain with RL=open.  It seems you are plotting in ADS the power gain from one single-ended 50 ohm port to another single-ended port, which will be lower.

        While the output impedance is 12 ohm, it is not advisable to connect it to a load of 12 ohms, as the characterization was measured with a load of 200 ohm.  Distortion will increase significantly due to the much larger output currents required for any given voltage.

        If possible, a 200 ohm load should be used together with an asymmetrical filter (for 12-ohm in, 200 ohm out).  ADS has a filter design package that would help in the design of such filters per your frequency requirements.

    Benjamin

  • Hi Benjamin,

               Thank you for your reply. My operating range is 50 to 90 MHz. Yes, I am plotting power gain with single-ended 66.66 ohm port(to match input).  Regarding gain of device query is clarified.

               Now, I have to design AAF( Anti-aliasing filter ) which has 12 ohm input impedance (since ADL5562 is driver for AD9467) and 530 ohm output impedance (this is input impedance of AD9467) to avoid passband ripple.

                But as you mentioned "Distortion will increase significantly due to the much larger output currents required for any given voltage." Could you please clarify on that in terms of what should be load of ADL5562 ?

               As I understand from your reply load for ADL5562 should be 200 ohm then my AAF input impedance is 200 ohm and output impedance is 530 ohm. Is it correct ?

    Regards,

  • Hi Benjamin,

                        Thank you. I understand filter design as you explained. But my query is,

            1: As filter is designed with input impedance Rs=12 ohm, which is load impedance of ADL5562. Will it cause any distortion ?

            2: Inductor of value 220 nH @ 70 MHz (center frequency of operating band) with Q=50 will be available in 1206 package size only or higher size only. It will also has ESR of roughly 3 ohm, which will further contribute to voltage loss. To avoid this,Is it wise to choose capacitor as leading component of 3rd order pi filter ?

    Regards,

  • Hi, Dhavalp,

         There are more than one solution to this problem, but let's start with the simplest:

    1. Rs=12 ohm, Rload=530 ohm // 3.5pF (both differential), so we design a half filter for Rs=6 & RL=265.

    The above shows half of the differential filter (mirror along the x-axis to obtain the other half) of Chebychev response,  ripple=0.2 dB, 95 MHz BW, and about 36 dB of rejection at the image frequency of 155 MHz.  Unfortunately, the last cap of 4.16 pF each half means that this design can only "absorb" up to 4.16/2=2.08 pF of differential ADC parasitic capacitance.

        We can reduce the load resistance to increase the amount of parasitic cap that can be absorbed.  E.g., adding a 321 ohm resistor in parallel with the ADC, its differential input impedance becomes 200 ohm.  So now we can design a similar filter for 12 ohm in and 200 ohm out, which produces:

    The last cap is 11.25 pF, which allows for an extra 4.25 pF of real cap, after absorbing the 3.5 pF ADC parasitic cap.  So the

    final component values (each half) are approximately L=0.22uH, C=27pF, L=0.22uH, C=3.9pF (after allowing for ADC and a little bit more parasitics), and the response looks like:

        Note that an even order filter with a leading inductor is chosen, so as not to present a capacitive load to the amplifier (to prevent peaking or even oscillations).  If faster roll-off is desired, then a higher order filter will be needed.  At frequencies lower than the cutoff frequency, the filter disappears, and the voltage loss is that of the resistor divider 12 ohm into 200 ohm, or about 0.5 dB.  The in-band distortion performance will be the same as those on the datasheet, while out-of-band harmonics will be attenuated by the filter.

    Benjamin

  • This question has been assumed as answered either offline via email or with a multi-part answer. This question has now been closed out. If you have an inquiry related to this topic please post a new question in the applicable product forum.

    Thank you,
    EZ Admin