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?

Parents
  • 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.

    Julian.

Reply
  • 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.

    Julian.

Children
No Data