I'm using a AD9837BCPZ waveform generator to generate a fixed frequency sinewave. We are finding that the output is going full swing +5V when first initialized and put into reset. Once the output is configured and released from reset the output goes to the correct bias and then generates the sinewave. This initial 5V output is causing some issues in our circuit and I would like to figure out how to avoid this.
We control power and configure it via a MCU. The first time after power on we enable the power and then configure the SPI interface. Then, configure the part to generate the waveform and hold in reset for a delay of around 10ms.
After we are done with the function we configure the SPI I/O to ground and then turn off power to the waveform generator. Subsequent cycles work as expected and we don't get the +5V output but rather the output's steady state is the correct bias voltage. The issue seems to only happen to the first cycle after a power on reset.
Attached is a screenshot of an oscilloscope of the problem.
The Yellow waveform is the signal directly at the out pin of the IC.
The Green is at the output of an amplifier stage which is AC coupled after the out pin.
You can see the first cycle where the out pin is 5V before the sinewave is generated. The next two cycles you can see that the out pin is biased correctly before the sizewave.