Post Go back to editing

Doubt in Direct conversion/Zero-IF operation using ADF4360-9 and AD8348

Hi!

Greetings!

It's been few days I have been struggling with these EVM modules. Please have a glimpse of my query and share your suggestions for 0-IF operation of AD8348.

I have doubt in Direct conversion/ Zero-IF operation using ADF4360-9 as synthesizer and AD8348 as IQ demodulator. I have acquired EVM kits for both the chips. I have given 360MHz to LO and 180MHz to MIX inputs for 0-IF operation.

My setup:

>> For ADF4360-9 I have used EVM kit's default connections with very few changes; RFout frequency: 360MHz, PFD frequency: 1600KHz, REFIN freq=19.2MHz, DIV OUT counter =Divede by /A counter, DIVOUT freq=180MHZ,  RF output power current: 11mA (0dBm)

>> RFOUT=360MHz from ADF4360-9 is fed to single ended LO input of AD8348 using coaxial cable of length 22cm.

>> The other complement RFOUT' is terminated in 50ohm amphenol standard.

>> AD8348 switch SW11 is in enabled position, SW12 is in MIXIN position (Vgin, input VGA: not used)

>> As my baseband operation is DC, two capacitors COI, COQ connected at IOFS, QOFS are removed.

>> Vref is connected to IOFS and QOFS using jumpers LK5I and LK5Q as suggested in datasheet pages: 18-19 under the section: BASEBAND OFFSET CANCELLATION.

>> No filters are used before baseband application: IMXO is directly connected to IAIN using jumpers LK1I and LK3I; similarly QMXO is directly connected to QAIN using jumpers LK1Q and LK3Q.

>> Vref is directly connected to VCMO using jumper LK11

>> Low pass filters are connected in the differential I/Q outputs. All of them have cut off frequency=159Hz.

>> 180MHz DIVOUT frequency from AD8348 is given to single ended MXIP input of ADF4360-9 using say 'x180' length of coaxial cable

Observation:

ADF 4360-9 output frequency generation is confirmed at 180MHz and 360MHz using spectrum analyzer.

For no 180MHz to MIX input, differential DC outputs of AD8348 are Vi=-409mV and Vq=-364mV

Case-1: AD4360-9 RF output current=3.5mA (output power=-9dBm)

For cable length 'x180'=29cm=> differential DC outputs Vi=1840mV and Vq=-1828mV

For cable length 'x180'=24cm=> differential DC outputs Vi=1846mV and Vq=-1831mV

Case-2: AD4360-9 RF output current=11mA (output power=0dBm)

For cable length 'x180'=29cm=> differential DC outputs Vi=-1780mV and Vq=1604mV

For cable length 'x180'=24cm=> differential DC outputs Vi=-1772mV and Vq=1575mV

Question:

As in the 50ohm coax cables the signal will not attenuate and only phase will change due to length difference, we can calculate the phase difference as follows:

Vi=mag*cos (phi), Vq=mag*sin (phi)

case-1:

phi_29cm=arctan(-1828/1840)= -44.81 deg (in fourth quadrant due to sine component is negative)

phi_24cm=arctan(-1831/1846)= -44.76 deg (in fourth quadrant due to sine component is negative)

hence delta_phi ~0, as I cannot see any voltage change in the response!!!

similarly, for case-2 it turns out to be: delta_phi ~0!

Calculation:

effective length difference= 29cm-24cm=5cm

calculated phase change/cm in coaxial Teflon media of the 180MHz signal=3.13degrees/cm

Expected phase change delta_phi= 3.13deg/cm * 5cm = 15.65 degrees

WHY THERE WASN'T ANY VOLTAGE CHANGE OBSERVED IN THE I/Q OUTPUTS DUE TO PHASE DIFFERENCE?

DO I HAVE IMPLEMENTED THE SYSTEM IN CORRECT WAY? PLEASE LET ME KNOW IF I DO NOT MAKE SENSE OR ELSE IF YOU NEED MORE INFORMATION.

Regards,

Sanket

Parents
  • Hi Sanket,

    The only area that may be questionable is the DC offset nulling on the AD8348.  Your application seems to require an accurate DC measurement therefore the DC offset nulling feature is a nuisance here. By installing jumpers LK5I/Q this should disable the offset compensation loop.

    A better way to more accurately test this is to use two signal generators locked in phase to a 10 MHz reference.  Change the phase on one signal generator relative to the other and see the change in output voltage of the AD8348.  Try this for large step sizes.  This is the best suggestion I can offer.  Everything looks like it is configured correctly.

    Regards,

    Sefa

Reply
  • Hi Sanket,

    The only area that may be questionable is the DC offset nulling on the AD8348.  Your application seems to require an accurate DC measurement therefore the DC offset nulling feature is a nuisance here. By installing jumpers LK5I/Q this should disable the offset compensation loop.

    A better way to more accurately test this is to use two signal generators locked in phase to a 10 MHz reference.  Change the phase on one signal generator relative to the other and see the change in output voltage of the AD8348.  Try this for large step sizes.  This is the best suggestion I can offer.  Everything looks like it is configured correctly.

    Regards,

    Sefa

Children
No Data