Post Go back to editing

Question Re: AD8343 S-Parameter Data

Hi all,

I've been trying for some time to figure out how to use the AD8343 (first time trying anything RF, actually), and I just wanted to verify a few things. I've been using the S-Parameter file supplied by AD to do impedance matching (using LVDS data output from an FPGA as the LO input and ~300 MHz signal from a function generator as the IF input), and when I simulate in Cadence, after impedance matching the IF and RF ports to 50 Ohms (so that at least S11 and S22 are minimized at my frequencies of interest), I get that the magnitude of S21 is on the order of 0.13 (before impedance matching, it was <<0.1). As far as I understand, S21 = VRFOut/VIFIn, so I'm a little confused as to what I'm doing wrong, if the mixer is speced for a typical conversion gain of 7dB. Is there any way to know what kind of conversion gain to expect at an output frequency of around 300-350MHz if the conversion gain # can't be used? If I'm not mistaken, the LO power also affects the gain, so is it appropriate to expect that the LO power used to test was the "typical" -10dBm?

Is the S-parameter data file actually sufficient to be able to create a matching network, or would it be very much advised to get impedance measurements for myself? Also, I'm having trouble understanding how I should go about verifying that the mixer is common-mode stable at the output, since I'm doing a differential to single ended conversion there.

Are there better recommendations for mixers operating at that frequency requiring differential inputs/outputs with 50Ohm impedances?

Thanks for the help,

Angie

  • Hi Angie,

    Is there any reason you chose AD8343 over its next generations: AD8342 and ADL5801? Both AD8342 and ADL5801are higher performance active mixers, and much easier to use because of their boardband characteristics.

    AD8343 s-parameters are provided for purposes of input/output matching, but not freqeuncy conversion simulation. If you look into the s2p file, you will see all 4 parameters were measured under the same corresponding freqeuncy. This means no freqeuncy translation was involved, or LO was zero. In this case, s12 and s21(aka gain for its magnitude) are not useful data for any mixer simulation, but s11 and s22 have valuable service for impedance matching. Once you find the proper matching for the mixer at desired band, you would need to implement the matching on a prototype board rather than relying on any simulation.

    I'm not sure what's the typical LO drive, but page 22 of AD8343 data sheet does suggest minimum of -12dBm. Your -10dBm guess is in the range, and has plenty of margin for good conversion gain per Figure 65.

    Regarding the stability consideration, page 21 and 22 have the full discussion. I'm not sure if I can add anything else. Once you set up the mixer on board, you should be able to see if there is any oscillation to verify your stabilty.

    Regards,

    Rui

  • Hi Rui,

    Thanks for the response! I was hesitant to use the AD8342 and ADL5801, because, despite being broadband mixers, they seemed to be best used as downconverting mixers. The IF output port for the ADL5801 can go down to LF, but the datasheet specifies the minimum RF input port frequency as 10MHz. I wanted to use the mixer as an upconverting mixer where signals with ~10KHz fundamental frequencies at minimum could modulate ~300MHz carriers. Is it possible with the ADL5801? (Impedance matching would be much less of an issue if I could use it...) It seems like the 8342 works to arbitrarily low frequencies at all of its ports, so I guess it is adequate for ~10KHz. Is there any benefit/drawback to using differential or single-ended mixers (CMRR, etc.)? Also, I've never tried hand-soldering LFCSP packages, so I don't know how feasible it is to use the chips without doing any kind of sketchy soldering job.

  • Hi Angie,

    You were right that ADL5801's RF inputs don't go below 10MHz. AD8342 should still work though I doubt if any work has been done for kHz range input. It is more popular to pick modulators for upconverting baseband signals.

    Compared to single-ended mixers, differential ones usually provide better even order distortion performance and better isolation in many cases.

    Soldering down LFCSP packages may take some practice. If you have solder paste and hot plate (or heat gun), things would get easier.

    Regards,

    Rui

  • Hi Rui,

    Thanks for clarifying. I was wondering if you might also have any s-parameter data for the AD8351 that I want to use at the output of the mixer. The specified 5KOhm || .8pF @ the input and 150Ohm || .8pF at the output were measured at 70 MHz, so I was hoping to get a little more detailed data than just TPC 19/20 (also, from a quick calculation, the above impedances don't seem to make sense with those graphs at 70MHz...)

  • Hello Angie,

    My apologies for the long wait on the S-parameter data for the AD8351.  Attached is the data file.

    Qui

  • 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