Texas Equipment ADC14X250EVM, User Manual, Page: 10 / 19

Optimizing Evaluation Results
www.ti.com
3 Optimizing Evaluation Results
This section assists with optimizing the performance during evaluation of the product.
3.1 Clocking Optimization
The sampling clock provided to the ADC must have very low noise to achieve optimal results. By default,
the ADC14X250EVM is configured to require a device clock from an external signal source via the CLK
SMA connector to allow for the best possible measured performance.
Filtering of the signal generator is required to remove broadband noise and non-harmonic spurs. Narrow-
band filters are recommended to remove as much noise as possible. If a signal generator output is used
directly without filtering, significant degradation in SNR will result.
Measure the power of the clock signal at the EVM clock input (50- Ω input). Optimize the clock amplitude
so that it is as large as possible without exceeding the maximum voltage swing specified in the datasheet
(
SLASE49 ). This maximizes the clock edge rate giving better SNR. The insertion loss of the filter in the
clock signal path can significantly impact the voltage swing at the ADC clock input.
3.2 Coherent Input Source
A Rectangular window function can be applied to the captured data when the sample rate and the input
frequency are set precisely to capture an integer number of cycles of the input frequency (referrer to as a
coherent frequency). The clock and analog inputs must be frequency locked (such as through 10-MHz
references) in order to achieve coherency. If this cannot be achieved, then a window function like
Blackman must be applied to the data for fast-Fourier transform (FFT) analysis in the HSDC Pro Software.
3.3 HSDC Pro Settings
The HSDC Pro options highlighted in Table 3 can help improve the performance measurements.
Table 3. HSDC Pro Options for Optimal Analysis Results
HSDC Pro Feature Description
Analysis Window Selects the number of samples to include in the selected test analysis. Collect more data to improve
(samples) frequency resolution of FFT analysis.
Data Windowing Select the desired windowing function applied to the data for FFT analysis. Select ‘Blackman’ when
Function sampling a non-coherent input signal or ‘Rectangle’ when sampling a coherent input signal.
Test Options → Select bins to be removed from the spectrum and back-filled with the average noise level. One may also
Notch Frequency Bins customize the harmonics and spurs that are considered in SNR and THD calculations and select the
method for calculating spur power.
Test Options
→
Enable markers to narrow the Single-Tone FFT test analysis to a specific bandwidth.
Bandwidth Integration
Markers
Data Capture Options → Configure the number of contiguous samples per capture (capture depth). May also enable Continuous
Capture Options Capture and FFT Averaging.
10 ADC14X250EVM Evaluation Module SLAU625 – November 2015
Submit Documentation Feedback
Copyright © 2015, Texas Instruments Incorporated