AnsweredAssumed Answered

ADF4355 does not lock and strange waveform in N divider output.

Question asked by DarioM on Apr 16, 2018

Hello, my name is Dario and i am member of DTA ( engineering team.

We are developing an Up-converter block, and we want to use the ADF4355 PLL as Local Oscillator.

I was made a custom PCB for the IC, but i can't make the ADF4355 locks.


I am using a 10MHz reference input, and i want to get a 4.9 GHz output. I placed in my layout the following loop filter:

The registers that i am programming are:

I get in the output the following frequency:

3332MHz (i have an amplifier in the output).


If i put the N dividir in the MUXOUT, i get:

Channel 1: 6.80008 MHz, it is "correct", because i have a N divider of 490, and 6.80008 x 490 = 3332 MHz. I don't know if this waveform is correct, it is the first time i "play" with this PLLs. I expected to see a sinewave or squarewave of 6.8MHz, but maybe this strange waveform is ok...


The Channel 2 show the Voltage in the VTune pin, it is 4.6V approximately (the multimeter show 4.6V, but in the oscilloscope i see a inestable value around 4.6V).


If i put the R divider in MUXOUT:

I can see exactly 10 MHz, but with these "spikes", i don't know if it is ok.


I tried the followings things:


- If i changes the phase detector polarity to NEGATIVE, i can see the Vtune voltage changes to 0.6V, and the output frequency of VCO decreases to 3210 Mhz approximately (and the "phase noise" increases). I look in the loop filter for short or open circuit, but it seems ok. 

- I went up and down the CP current to 0.3 and 2.7, but only changes the Vtune a few mV.

-If i configures other output frequency in the registers, i see always a different frequency in the output in the spectrum analyzer, but never the correct i programmed. And the N divider always give me the correct division in the oscilloscope.

-I open the loop filter and put a variable voltage in the Vtune pin, and i can see the output frequency to change slightly, but always around the 3300 MHz. I suspect in that band i only can change 15Mhz/V (very low variation). In this case, i think the wrong VCO band is select, but in the registers i put the autocalibration bit on, so the band selection should be automatic.

- The VCCs voltage seems ok (3.3v, 3.3vRF, and 5V_VCO)

- I have some strange communications problems. I am programming the adf4355 with a Cypress PSoC5 by SPI. When i disconnect the cables, i can see some variation in the waveforms. Also, sometimes, if i disconnect the SPI connection cables, the adf4355 turn off. Also, when i want to reprogramming the ADF, i have to power off the ADF, power off the PSoC, and power ON both, otherwise the changes are not taked by the ADF. It is very strange. But in the other moments, it is seems are programming ok (the MUXOUT always responds ok when i set the N or R divider, so i don't inquire a lot in this communications issues for now, but i mention it here just in case)



So, the questions are:


1- Are the Vtune voltage correct? How can i know that? The datasheets says 15 MHz/V in the VCO sensitivity, but i don't know what means. If a need 4900 MHz, 15 MHz/V give me a very high voltage. I can't find a plot of Frequency vs. Vtune anyway.


2- Are the waveform of N and R divider ok? It seems very strange for me.


3- If i connect the oscilloscope in the Charge Pump output (CP pin, the input of loop filter), what should I see? And in the Vtune pin? A constant voltage?


4- Are other type of measurements i can do?


I will appreciate so much your help.

Thanks in advance.