Programming Order to Change Attenuator

Question we are testing new IQ board with ADRF6755 IQ mixer. System works almostproperly, but we cannot change setting of RF attenuator even we write correctvalue into appropriate register CR30 and then also read the same required valuefrom the register CR30. We don't detect any change at RF level. Generally wesee about 8 dB lower level then we calculated.Any idea how to solve it? Answer I have set this up in the lab to confirm that no special programming order isneeded to change the attenuator. Below are my steps: 1. Using the customer evaluation software, I initialised the part to defaultusing “write default settings” and lock at 330MHz LO with 10MHz Reference and10MHz PFD and enable the Modulator Power with CR30=0 to give 0dB attenuationa. The LO feedthrough I see at the RFOUT is -45.5dBm2. Click “Read all register settings” which saves register information to theindicators on the right side of the GUI3. Power cycle the board to reset the part4. Write the saved register settings from the indicators on the right side ofthe GUI by clicking “write these register settings”a. The software for this performs the following actions:1. writes values for CR0 – CR30 in that order2. Writes to CR0 and CR27 to trigger PLL calibrationb. Once this is complete the output shows the LO feedthrough at 330MHz at-45.5dBm5. Write to CR30 only to change the attenuator:a. CR30=5, I see LO feedthrough=-50.5dBm (5dB change)b. CR30=30, I see LO feedthrough=-76dBm (30.5dB change) If you have a customer evaluation board you can try to program it using yoursoftware (by attaching wires from your prototype board to the evaluationboard). If the problem happens on the evaluation board then we know thesoftware is the problem.You could also try the opposite, program your prototype board using thecustomer evaluation software by attaching wires from the eval board to theprototype board. The evaluation board has resistors you can remove on the SPIlines if you want to communicate only to the prototype board.

Spurs moving around CW

Question We detected moving andaccidental spurs around CW (measured either at RFOUT with CW generatedfrom IQ or at LOMON). Position of spurs depends on loop filter BWoutside loop filter are marginal. We completely disconnected all switching power supply sources at ourboard and we work only with LDO now. we also disconnected IQ signals andtesting PN at LOMON only to be sure that we are checking spurs at LO. Have you any idea how to resolve it? Answer I have seen this one before. Let me know if the solution below fixes the issue.If not, there are other more complicated things we can try to change. · These transient spurs occur for high modulus PLLs. We have also seen thesespurs on PLLs from our competitors.· They occur when the device operates with a very small Fractional value andfor fraction=0 (as in this case)· Customers have not reported that this causes any negative effect on theirsystem.· In the case where Fraction=0, it should be possible to eliminate the spurs bysetting CR19=0x96. This will disable the “Third-Order FractionalInterpolator” (see image below for position of this block)o If you intend to operate with fractional word not equal to zero, CR19 shouldbe set back to =0x80 to re-enable this block.