Signal Processing: Block Average

The Block Average Module improves the fidelity of noisy repetitive signals. Multiple repetitive acquisitions with very small dead-time are accumulated and averaged. Random noise is reduced by the averaging process improving the visibility of the repetitive signal.

Features

• Signal Averaging in Hardware
• Maximum waveform length for one channel between 128 kSamples (M4i) and 1 MSamples (M5i)
• Maximum average waveform throughput of more than 5,000,000 waveforms per second
• All channels are averaged
• Fully compatible with synchronization
• Low dead time between waveforms:
  • 80 ns at 500 MS/s (M4i.44xx series)
  • 64 ns at 5 GS/s (M4i.22xx series)
  • 55 ns at 6.4 GS/s (M5i.33xx series)
• Fully compatible with Spectrum API which allows usage with C/C++, Delphi, Basic, LabVIEW, MATLAB and many more
• Fully integrated into SBench 6 Professional software
• Available for all high-speed digitizers based on M4i technology
• Fast data calculation in hardware reduces the needed transfer bandwidth extremely
• Calculation can run on all channels simultaneously

The complete averaging process is done inside the FPGA of the digitizer generating no CPU load at all. The amount of data is greatly decreased as well as the needed transfer bandwidth is heavily reduced. The signal processing firmware also includes the standard firmware that allows normal digitizer operation with no limitations.

Example

The right side screenshot shows the improvement of a signal that is completely overlayed by a random noise when using different averaging factors.

While the source signal is not even visible in the original single-shot signal a 10 times average of the signal already shows that there is a signal with 5 peaks. Doing a block average of 1000 times improves the signal quality extremely showing us the real shape of the signal including all secondary maximum and minimum peaks.

The example was taken with 500 MS/s (2 ns)  timescale and 14 bit ADC resolution having a signal level of about 2 mV.

TDA Mode (M5i cards only)

Certain applications have a very low-level signal of interest with its rare events sitting barely above a noisy baseline. One such application is time of flight mass spectroscopy (TOFMS).

The process of averaging in itself already reduces random noise with each accumulation step. To further enhance the visibility of the signal of interest, even when synchronous noise is present, the M5i averager includes the Threshold Defined Averaging (TDA) mode. This mode allows samples below a set threshold to either be excluded from the accumulation process, or replaced with a user defined (hence known).
value.

The output of the "level selection" block, shown in the averager's block diagram using AVRGMODE_TDA_HIGH would look the image on the right.

All samples that are above a threshold level are forwarded to the accumulator as is, while all samples below are replaced with the defined value, which is then forwarded.

So, the accumulator would see data segments on its input that are similar to the waveform shown at the bottom of the image. For best results, the threshold level should be set outside the level of noise

Links

• Download M4i Block Averaging Data sheet
• Download M5i.33xx Series Data Sheet