Tektronix OM4006D, Руководство пользователя

Страница: 111 / 205

Core processing software guide
Block processing
The OM4000 core processing software is able to cope with very large record sizes,
for example 250M samples, by breaking down the record into smaller pieces, or
blocks, and processing them in sequence. All functions are designed so that the
output is near-identical if block processing is used compared to performing the
processing in a single block. Each function collects information about the signal
and any relevant internal variables at the end of the time period of one block, and
then effectively tags that information onto the start of the next block. The result
is a seamless transition from one block to the next.
Block processing is an advanced feature. It applies only if the record size is larger
than the block size set in the OUI, and typical computing hardware can cope with
block sizes larger than 100 k samples. A user wishing to customize the core
processing software does not have to be concerned about block processing unless
large record sizes have to be processed.
The complete steps of interaction between the OUI and CoreProcessing, taking
block processing into account, are listed below. Italics are used to highlight
a change compared to the simpler description in
Interaction with OUI . (See
page 79.)
1. OUI launches MATLAB engine
2. OUI writes variables to MATLAB workspace corresponding to settings in the
OUI Analysis Parameters window
3. OUI fetches one block of a record from oscilloscope, and writes to MATLAB
workspace
4. OUI executes EngineCommandInit
5. Loop over blocks until scope record is exhausted
a. OUI executes commands listed in MATLAB Engine Commands window
b. OUI executes EngineCommandBlock
6. End of loop over blocks
7. OUI executes EngineCommandPost
8. OUI retrieves output variables from MATLAB workspace
9. OUI displays output variables as eye diagrams, constellation diagrams,
numerical values, etc.
OM4000D Series Coherent Lightwave Signal Analyzer 87