DAC Linearity
DAC Linearity measures the gain or gain error of the DAC. In the plot below, you can see the gain of the DAC. For a given DAC output level (in dBFS), we can read the corresponding amplitude in dBV.
In the plot below, we see the gain error. The gain error looks at the gain for each step, and then averages that across all the points and plots the deviation from the average gain at each point. For the range of -100 to -5 dBFS, a 5 dB step was used (blue trace). For the range of -5 to 0 dBFS, a 0.2 dBFS step was used (red trace).
A modern DAC can challenge the dynamic range of the QA40x audio analyzer. Prior to running your test, you need to ensure the full scale input will capture the sweep you have specified. For example, in the first plot above, we can see the maximum output of the DAC is just shy of 12 dBV. And so, we could pick +12 dBV as the full scale input, or we could pick +18 dBV for a bit more margin. If we went with a +24 dBV or higher full scale input, we run the risk of swamping out the DAC low-level performance with attenuator noise in the analyzer.
If you aren't sure where to be begin, you can make several runs to help guide your decision. For example, we know the +12 dBV is the minimum input range. So, let's run the test 3 times with different input range options (12, 18 and 24 dBV), and plot them all together.
The options we'll use are shown below. We'll start at -100 dBFS, and increment by 5 dB to 0 dBFS. We'll use a 1 kHz test tone, and plot the results as gain error. For the first run, we'll use +12 dBV input range on the QA40x, and +18 and +24.
The result of all three combined are shown below. As expected, we can see the +24 dBV full scale input range clouded the true picture at lower levels.
