ADL5513 log amp undershoot

ADL5513 pulse response

I am using this device at 2.4 GHz driven with a balanced feed from a Minicircuits balun, to avoid log amp tails.

However, i see an annoying undershoot at the end of the pulse. The output of the log amp is buffered with an AD8051 and the undershoot appears at all power levels, with or without 22pF CLPF cap. Running off 5Vdc. Test source is an R&S SML03 sig gen, 400ns pulses between -60 and 0 dBm. Undershoot is present at the input of the video buffer amp.

The problem with the undershoot is that it eats into the rising edge of a second pulse that comes after the first.

Any ideas or is it to be expected?

    • You may be running into the ADL5513's own slew rate limitations; video BW will be dominated by the 3pF cap that is internal to the device on the CLPF pin
    • Have you just been looking at the output of the AD8051 buffer? Have you traced the undershoot to the ADL5513 by probing right at the VOUT/VSET (I am assuming you have these shorted) node with a high impedance oscilloscope probe?
    • Is the 400ns the pulse width?  How long until the next pulse starts?
  • No I checked it was not the buffer amp and yes Vout/Vset are shorted. Using a tektronix 200 MHz 10:1 probe. The 400ns is the minimum PW I ever need to see.

    I have found that if I use a 10pF CLPF cap, the undershoot is reduced and the risetime is still good enough for the thinnest pulse to reach peak amplitude, so I can use what I have prototyped for the production circuit. The actual circuit will be on Rogers RO4350B as I am working at 13 GHz in other parts of the circuit. I prototyped on FR4 using a flood groundplane on the component side, but there are no vias to the main groundplane, so perhaps that has something to do with it.

    Looking again very closely, I do notice the undershoot is actually the first cycle of a ripple, which changes amplitude and shape when I ramp up the power, so perhaps there is an RF reflection before the log amp which it is correctly demodulating.

    Will let you know how it goes when I get the real PCBs made up and thanks for your input.

    Here's hoping that your ADF5000 prescaler , Hitite 13GHz VCO and ADF4107 run smoothly, wont know this until the boards arrive.


  • 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