Beckhoff EPP3504-0023 Short Manual page 102

Commissioning
• In the PLC/ C++, on the time axis, convert to the desired sampling rate, e.g., by linear interpolation
based on the timestamp for each input value (sample). Since the EPP3504‑0023 units provide time-
equidistant samples based on distributed clocks, this is easily possible.
For example, a sinusoidal signal decimated with 50/44.1 = 1/0.882 can be represented as follows:
◦ Green: corresponds to original analog signal (input), approx. 432 Hz
◦ Blue (O): corresponds to sampling of the EPP3504‑0023 with f
interval of 100 µs
◦ Red (X): corresponds to signal converted by PLC to 8820 sps (factor 0.882) and thus a time
interval of approx. 113.37.. µs
• Note: The term "decimation" is applied both to the calculation in the box module (see chapter
"Decimation") and to the conversion in the PLC program. The following refers to the conversion in the
PLC.
• Since the time interval of the desired sampling after decimation in the PLC is usually no longer an
integer (finite) number, value/time pairs are used for representation in the PLC/Scope, i.e., an X time
value is assigned to each Y value. Such value/time pairs can easily be displayed with TwinCAT
ScopeView in XY mode. See also infosys.beckhoff.com:
TwinCAT3 → TExxxx | TC3 Engineering → TE13xx | TC3 ScopeView → Configuration → XY‑Graph
• The conversion also has consequences for further processing in PLC/C/ADS:
◦ A PLC/EtherCAT/TwinCAT system tends to be set up such that a constant number of samples is
processed per cycle. Usually this is now no longer the case: a different number of samples has to
be processed from cycle to cycle (specified by the program variable nResultNoOfSamples).
◦ While a time stamp per signal value has so far remained relatively insignificant, the method of
conversion of the decimation process used here, however, means that the respective timestamp
per signal value must be taken into account in an elementary manner.
• The non-constant number of samples is not visible in the TwinCAT XY Scope because some values
are sporadically drawn twice, and this should be taken into account; it may be advisable to use an
intermediate buffer for further processing.
• For orientation of the currently valid number of samples per task cycle, the program provides the
variable nResultNoOfSamples. It indicates which values in the array variable contain valid values in a
task cycle (indicates the field number - 1).
The following sample program, which also contains the XY representation in the TwinCAT Scope, serves
as a guide. Due to the above-mentioned problem relating to the non-constant number of valid samples, the
program returns the array pair aVarDecResult_TS and aVarDecResult for the Scope with the same number
of elements as for the input value aSamples_1 (value = nOVS). If there are fewer values in a task run, the
last value is simply entered repeatedly (similar to "sample & hold"). The ScopeView was configured as
follows for the recording:
Property
ScopeNodeProperties
ChartXYNodeProperties
XYChannelNodeProperties
102
ViewDetauilLevel
Record time
Default Display Width
Max Data Points
Marks
Mark Size
Mark Color
Version: 1.2
of 10,000 sps; a sampling
max
Value
ExtendedXYOnly
00:00:00:05
0,00:00:00,050:000
200000
On
5
(other than line color)
EPP3504-0023