Basic function principles

The measuring functions of the EPP3356-0022 can be described as follows:

General notes

Signal flow diagram

Basic function principles 2:

The EPP3356-0022 processes the data in the following order:

Basic function principles 3:

Measurement principle of delta-sigma (ΔΣ) converter

The measurement principle employed in the EPP3356-0022, with real sampling in MHz range, shifts aliasing effects into a very high frequency range, so that normally no such effects are to be expected in the kHz range.

Averager

In order to make use of the high data rates of the Analog-to-Digital converter (ADC) even with slow cycle times, a mean value filter is connected after the ADC. This determines the sliding mean value of the last 4 measured values. This function can be deactivated for each mode via the CoE object "Mode X enable averager".

Software filter

The EPP3356-0022 is equipped with a digital software filter which, depending on its settings, can adopt the characteristics of a Finite Impulse Response filter (FIR filter), or an Infinite Impulse Response filter (IIR filter). The filter is activated by default as 50 Hz-FIR.

In the respective measuring mode the filter can be activated (0x8000:01, 0x8000:02) and parameterized (0x8000:11, 0x8000:12).

FIR 50/60 Hz

PDO filter

Basic function principles 4:
Notch characteristic/amplitude curve and step response of the FIR filter

IIR-Filter 1 to 8

Basic function principles 5:
Step response and Bode diagramm of the IIR filter

Overview of conversion times

Filter Settings

Value

PDO update time

Filter property

Limit frequency (-3 dB) [Hz] (typ.)

Comment

Rise time 10-90 % [s] (typ.)

Filter deactivated

-

Cycle-synchronous,, min. 100 µs

-

-

-

-

0

FIR 50 Hz

312.5  µs

50 Hz notch filter

22 Hz

Typ. conversion time 312.5 µs

0.013

1

FIR 60 Hz

260.4  µs

60 Hz notch filter

25 Hz

Typ. conversion time 260.4 µs

0.016

2

IIR1

Cycle-synchronous (up to min. 100  µs)

Low-pass

2000 Hz

a0=1/21 = 0.5

0.0003

3

IIR2

Low-pass

500 Hz

a0=1/22 = 0.25

0.0008

4

IIR3

Low-pass

125 Hz

a0=1/24 = 62.5e-3

0.0035

5

IIR4

Low-pass

30 Hz

a0=1/26 = 15.6e-3

0.014

6

IIR5

Low-pass

8 Hz

a0=1/28 = 3.91e-3

0.056

7

IIR6

Low-pass

2 Hz

a0=1/210 = 977e-6

0.225

8

IIR7

Low-pass

0.5 Hz

a0=1/212 = 244e-6

0.9

9

IIR8

Low-pass

0.1 Hz

a0=1/214 = 61.0e-6

3.6

10

Dynamic IIR

The filter changes dynamically between the filters IIR1 to IIR8

11

PDO Filter frequency

1/PDO Value[Hz]*64

Notch filter with adjustable frequency

ca. 0,443 * PDO Value [Hz]

-

-

Basic function principles 6:

Filter and cycle time

If the FIR filters (50 Hz or 60 Hz) are switched on, the process data are updated maximally with the specified conversion time (see table). The IIR filter works cycle-synchronously. Hence, a new measured value is available in each PLC cycle.

At which point the filters can be adjusted is described in the chapter “Object description and parameterization” for example under index 0x8000:12.

Basic function principles 7:

IIR filter

Differential equation: Yn = Xn * a0 + Yn-1 * b1 with a0 + b1 = 1

a0 = (see table), b1 = 1 - a0

Dynamic IIR Filter

The dynamic IIR filter automatically switches through the 8 different IIR filters depending on the weight change. The idea:

Parameterization takes place via the CoE entries 0x8000:13 and 0x8000:14. Evaluation takes place according to 2 parameters:

Example:

The dynamic filter is to be adjusted in such a manner that a maximum slope of 0.5 digits per 100 ms (5 digits per second) is possible without the filter being opened. This results in a "calm" measured value. In the case of a faster change, however, it should be possible to immediately follow the load.

The measured value curve is shown below for a slower (left) and faster (right) change.

Basic function principles 8:
Effect of dynamic IIR filters

Calculating the weight

Each measurement of the analog inputs is followed by the calculation of the resulting weight or the resulting force, which is made up of the ratio of the measuring signal to the reference signal and of several calibrations.

YR = (UDiff / Uref) x Ai

(1.0)

Calculation of the raw value in mV/V

YL = ( (YR – CZB) / (Cn – CZB) ) * Emax

(1.1)

Calculation of the weight

YS = YL * AS

(1.2)

Scaling factor (e.g. factor 1000 for rescaling from kg to g)

YG = YS * (G / 9.80665)

(1.3)

Influence of acceleration of gravity

YAUS = YG x Gain - Tara

(1.4)

Gain and Tare

Legend

Name

Designation

CoE Index

UDiff

Bridge voltage/differential voltage of the sensor element, after averager and filter

 

Uref

Bridge supply voltage/reference signal of the sensor element, after averager and filter

 

Ai

Internal gain, not changeable. This factor accounts fort he unit standardisation from mV to V and the different full-scale deflections of the input channels

 

Cn

Nominal characteristic value of the sensor element (unit mV/V, e.g. nominally 2 mV/V or 2.0234 mV/V according to calibration protocol)

8000:23

CZB

Zero balance of the sensor element (unit mV/V, e.g. -0.0142 according to calibration protocol)

8000:25

Emax

Nominal load of the sensor element

8000:24

AS

Scaling factor (e.g. factor 1000 for rescaling from kg to g)

8000:27

G

Acceleration of gravity in m/s^2 (default: 9.80665 ms/s^2)

8000:26

Gain

 

8000:21

Tare

 

8000:22

Conversion mode

The so–called conversion mode determines the speed and latency of the analog measurement in the EPP3356-0022. The characteristics:

Mode

Meaning

typ. latency

typ. current consumption

0

High precision

Analog conversion at 10.5 kSps (samples per second) Slow conversion and thus high accuracy

7,2 ms

70 %

(see Technical data regarding nominal value)

1

High speed / low latency Analog conversion at 105.5 kSps (samples per second) Fast conversion with low latency

0,72 ms

100 %

(see Technical data regarding nominal value)

Due to the conversion principle of the EPP3356-0022, the analog voltage is only available as a digital value after a defined time. This is shown in figure below.

A step signal 0->1 is applied to the input. The measured value is reached and readable within the defined accuracy after 7.2 ms or 0.72 ms, depending on the conversion mode 0/1. At this time the timestamp is also acquired in Distributed Clocks mode. In real operation a step signal is not normally connected, but rather a higher frequency but constant signal. The EPP3356-0022 then maps the input signal with the corresponding latency for further processing, for which reason faster querying of the sampling unit at shorter intervals than the latency (EPP3356-0022 allows up to 100 µs) makes sense for true-to-detail mapping of the analog input signal.

Basic function principles 9:
Latency of the Analog-to-Digital converter

It is not possible to change the specified latency.

Beyond that the following are individually adjustable in each mode via CoE

Basic function principles 10:
Setting parameters in CoE belonging to the individual modes

Mode change

In particular for dynamic weighing procedures it may make sense to considerably change the measuring characteristic during the weighing procedure. For example, if a bulk material is filled by the sack within 5 seconds, a very open filter should initially be used so that the measured value quickly follows the fill level. During this phase it is of no importance that the measured value is very inaccurate and subject to high fluctuations. If the sack is >90 % full, filling must be slowed down and the loading must be followed with higher accuracy; the filter must be closed. Therefore the two conversion modes can be switched via the process data bit "Sample mode" in the EPP3356-0022 in relation to the processing of the analog values.

The mode change takes about 30 ms, during which time the measured values are invalid and indicate this by the status byte.

Basic function principles 11:
Sample mode switching