Hello Analog Devices =). I need a bit of help when noise is concerned.
I have been testing a board made my me for the AD9834 DDS with this 75 MHz reference oscillator. The problem is always the noise. I am capturing the output of the DDS with a TDS5034B 5 GS/s Oscilloscope and i notice lots of noise on top of the signal and also a bit shake (maybe phase noise). I attached a 1 kHz plot and also a video for you to notice "the shake". I have a layout like it is shown for the evaluation board. Lots of decoupling capacitors (1 nF, 10 nF, 0.1 uF ceramics and 10 uF 10 V tantalum). Just one in the reference oscillator (0.1 uF). The schematic is also attached. Could you tell me what is wrong with this schematic? And what about the SPI clock frequency. Does it limit the output frequency of the DDS?
I wanted to have a stable and low noise signal because i want to amplify the signal =/. How can i achieve a more pure signal getting rid of most of the noise?
By the way, i want to change between sinewave, triangle wave and square wave. Since the square wave is output from the SIGN BIT OUT i thought about using the Analog Multiplexer ADG704 and control the output wave. Although after testing this solution i noticed that the input impedance of the ADG704 changes the amplitude range of the output of the DDS (when i put the ADG704 after the DDS). Is the ADG704 a good solution?
After this ADG704 Mux i added a 1st order HP filter to cut the DC component (with a cut off freq equal to 1 Hz) and 3rd Order Low Pass Chebishev II filter for cutting after my maximum frequency (3 MHz) wich is syntonized to 9 MHz (for it to cut the 6 MHz and 9 MHz components abruptly and then also the digital system componentes) This was simulated and supposelly works. However what about when i output the lower frequencies? This will not cut the unwanted components ... will it? This is only for 3 MHz situation isn't it? And i thought this filter would also cut the noise but it doesn't =/.