The signal generator is based on the AD9851 DDS chip
The board contains:
- AD9851 chip and passives
- 2 5V LDO’s for the digital and analog DDS power supplies
- footprint and passives for a local 30MHz oscillator, for stand-alone use
- 50MHz low-pass filter copied from the N2PK VNA project:
- several footprints for atenuators
- a 1:2 50ohm power spliter
- 2 gain blocks based on the MSA-0286 MMIC amplifiers
- different connectors for RF / CLOCK / power / DDS data
Schematic: ad9851
The 3 atenuator groups on the board, made by R33,R34,R36 / R35,R38,R41 and R37,R40,R42 are used to control output power.
On the final boards the first group is bypassed and the other 2 are configured for 5dB.
In normal use one of the 2 outputs is terminated into 50ohm.
The boards where designed with 2 outputs to allow for some other experiments.
The initial design plan was for the DDS to output as close to 0dBm as possible, but I did not read the MMIC datasheet carefully enough :
“Output Power at 1 dB Gain Compression: 4.5dBm” => the signal was distorted for a 0dBm output.
I had to tune RV1 “DDS iset” to get an clean sine-wave at an output power of ~ -3dBm
The DDS chip gets quite warm and it will have a small heat-sink stuck on with thermal adhesive or thermal paste and hot-glue.
The board is double-layer FR4 on a 100x50mm template that I chose for all boards in this project.
The top layer is used for signal routing, the bottom one is mostly ground with some short traces:
The final board looks like (with flux residue, etc):