LMX2592: calculating the value for the RFout matching capacitors

Prodigy 40 points

Replies: 3

Views: 39

Part Number: LMX2592

I am looking for a better understanding of the RFOUT for this part and the matching that is suggested.  In the datasheet I have they show 100pF as the DC bi-pass and the AC coupling capacitors.  I see other values being used for this part of circuit as well.  I understand that it can vary with frequency.  We are using the 50 Ohm Output Pull-up technique as described on page 33 of the datasheet.  Is there a table or technique for determining the value of these three capacitors and are all three always the same value (whatever the value is determined to be)?  We are using a 125 MHz OSCIN to create a 1.9 GHz reference. I believe that the inital value of 10pF may be too small for that frequency.  I see 100pF in the evaluation design and other reference designs on your website.  I believer we are getting attenuation or return loss from that value.  Do you have a recommended method for calculating that value?  Let me know.  Thank you

3 Replies

  • Hello Tom,

    The 100 pF value for these capacitors should be okay for the AC-coupling caps and bypass capacitor to GND on the outputs for the 1.9 GHz output. What makes you think there is some attenuation? Please elaborate on what you see. Are you basing this on the output power perhaps?

    Thanks,

    Vibhu

  • In reply to Vibhu Vanjari:

    Vibhu,

     

    Thanks for getting back to me so soon.    The reason I believe I was getting attenuation is this signal feeds a mixer. We were seeing more harmonics then we thought we should see and a  Little less conversion gain then we thought we should see. I don't believe that everything we are seeing from our system as a whole is tied to this one element being the LO  Level. One of my colleagues went through the spur mitigation process as described in the user manual. We believe our set up is optimized and not generating more spurs the necessary. I think if we change a few parameters in different gain stages in front of the mixer (AD5801) we will have a better looking spectrum, free of unnecessary harmonics. 

    I believe optimizing the AC coupling capacitors for the frequency output of the LMX2592 (Local Oscillator) and the bypass capacitor used at the 50  Ohm pull up network, will have an impact on the spectrum.  As mentioned above, better distribution among the variable gain amplifiers will avoid applying too much gain at one stage, and not enough at another.  These little changes will be may be enough to have an impact on the overall performance of the down-conversion process. 

    I did not design the circuit and I'm only responsible for evaluating the performance.  A better understanding of the mandates for each input and output elements of the LMX2592 will help me to make suggestions to modify the design, should I find something that works better than the current set up.

    I understand the function of the AC coupling capacitors for the RF output. I do not completely understand the function of the bypass capacitor from the 50 Ω pull up network. Do you know how that particular component works in this paradigm? Thank you

  • In reply to TOM FRENCH:

    Hi Tom,

    The EVM is design for generic use so we put 100pF as the decoupling capacitor as it can support both low and high frequencies.

    If your application interested only in 1.9GHz, 10pF is a good choice for decoupling purpose. 

    The bypass capacitor at pull-up is also an AC-short. Similar to the DC-block at RF output.

    What kind of harmonic did you see? Harmonic of 1.9GHz?