ComCell Manual v3.5

Manual
ComCell and ComCell Analyzer for MatLAB

Version 3.5, 16.10.2017
KDEE ComCell v3.5 16.10.2017
Content
1. Introduction .......................................................................... 4
2. General description and first steps .......................................... 4

2.1. Installation ......................................................................................................... 4
2.1.1. Setup of NI-Board .............................................................................................................5
2.1.2. Setup of KDEE_v2 and KDEE_MfPG devices ...............................................................5
2.1.3. Communication-setup of Tektronix TDS7104 oscilloscope ...............................................6
2.1.4. Communication-setup of LeCroy-oscilloscope ...................................................................6
2.1.5. Communication-setup of Yokogawa-oscilloscope ..............................................................7
2.1.6. Setup oscilloscope-presets for switching- and conduction-measurements .........................8
2.1.7. Calibrating an analogue FuG-voltage-source .................................................................. 14
2.2. Starting the application and data-structure .....................................................15
2.3. Creating and managing setups ..........................................................................16
2.4. Equipment setup ...............................................................................................17
2.5. Test setup .........................................................................................................19
2.6. Conduction-loss setup .......................................................................................20
2.7. Switching-loss setup ..........................................................................................23
3. Measurement....................................................................... 26
3.1. Setup necessary values ......................................................................................26
3.2. Connecting the hardware ..................................................................................29
3.3. Running the test and storing the results ..........................................................31
3.4. Analyzing the results ........................................................................................32
3.5. Adding measurement-points to an existing set of data ....................................33
3.6. Auto-zero-measurement for diode-switching .....................................................33
3.7. Dual-device-mode ..............................................................................................34
3.8. Distributed current measurement .....................................................................35
3.9. Combined SwitchingOn and DiodeSwitching with 8ch-Oscilloscopes ...............35
4. ComCell Analyzing tools ...................................................... 36

4.1. Conduction-loss analyzer ..................................................................................36
4.2. Switching-loss analyzer .....................................................................................38
4.3. Calculation of voltage-drop curves and report ..................................................40
4.4. Calculation of switching-losses and report ........................................................40
4.5. Export to Excel-Datasheet ................................................................................41
Page 2 of 61
4.6. Export to KDEE TopBench .............................................................................42
5. KDEE_MfPG Pulse-Generator.............................................. 44
5.1. General description ...........................................................................................44
5.2. Possible output-signals ......................................................................................44
5.3. Communication protocol ...................................................................................44
5.4. Connection of peripheral devices ......................................................................47
5.4.1. Gate-controller ................................................................................................................ 47
5.4.2. Thermo-couples ............................................................................................................... 47
5.4.3. OMRON E5CC ............................................................................................................... 47
5.4.4. FuG voltage-source ......................................................................................................... 48
5.4.5. ThermoStream ................................................................................................................ 48
5.4.6. Interlock .......................................................................................................................... 49
6. Additional content ............................................................... 50

6.1. ComCell configuration file ................................................................................50
6.2. Calculating pulse-widths of double-impulse ......................................................51
6.3. Calculating conduction-losses ...........................................................................53
6.4. Calculating switching-losses and dynamic parameters .....................................53
6.4.1. Switching-loss integration-borders .................................................................................. 54
6.4.2. Recovery-loss integration borders ................................................................................... 55
6.4.3. Dynamic parameters of the switching-process ................................................................ 55
6.4.4. Post-Processing Deskew .................................................................................................. 56
6.5. Available options to start ComCell and the analyzers .....................................57
6.5.1. General starting options ................................................................................................. 57
6.5.2. Starting conduction-loss analyzer ................................................................................... 57
6.5.3. Starting switching-loss-analyzer ...................................................................................... 58
6.6. Voltage- and Temperature control....................................................................59
6.7. Important Code lines of main-program.............................................................60
Page 3 of 61
1. Introduction
ComCell is a MatLAB toolbox for automated measuring of semiconductor-switch-
ing- and conduction-losses. Due to the fact that these kinds of losses will change their
values with temperature, DC-link voltage and gate-resistance, the semiconductor losses
are measured under several conditions.
The ComCell toolbox will help to analyze and verify the losses of a semiconductor
and create an interactive datasheet of the tested device. After setting up the parameters
for the measurement, the program will record data of different points of operation and
will create an ELoss(IC)-curve at the end. Afterwards all measured data can be exported
as an interactive Excel-report-sheet or XML-file for documentation and further calcula-
tion.
The following hardware can be used with the ComCell-software:
• Oscilloscopes
o Tektronix TDS7104 (more models possible)
o Yokogawa (several models)
o LeCroy (most models with GPIB-connection)
• DC-Link-Source
o FuG MCP 350-2000 (automated over NI-board or KDEE_MfPG)
o all other DC-sources (manual mode)
• Pulse-Generator
o NI-board (minimum resolution 1µs)
o KDEE generator (100ns resolution)
o 3rd party generators (minimum code-modifications required)
• Temperature
o Omron E5CC temperature controller with RS485 connection
o 3rd party temperature controller (minimum code-modifications required)
2. General description and first steps

2.1. Installation
The NI-DAQmx-, NI-MAX-software and the NI-libraries have to be installed us-

ing the delivered installation files (or from http://ftp.ni.com/support/softlib/visa/NI-
VISA/16.0/Windows/NIVISA1600full.exe) to communicate with the desired oscilloscopes
Page 4 of 61
over the VISA-connection of MatLAB. Run “NIVISA1600full.exe” and install the NI-
VISA software into a directory on the computer. Important: if you are using a domain-
based network-computer, the installation of NI-DAQmx-software may abort with errors.
If this is the case please log-in as a local user on this computer and run the installation
again.
If you received and want to use the standalone-version of the ComCell-programs

(without MatLAB) you have to install the MatLAB-runtimes first. For this start the
ComCell_FullSetup.exe or ComCell_WebSetup.exe. Afterwards the standalone pro-
grams can be run with a double-click.
2.1.1. Setup of NI-Board
After installing the NI-VISA-toolboxes the path to the file “NIDAQmx.h” has to
be set within the configuration-file “ComCell.ini” under the tag “NIHeaderPath”. If you
used the standard-settings this path should be:
NIHeaderPath:
“C:\Program Files\National Instruments\NI-DAQ\DAQmx ANSI C Dev\include\NIDAQmx.h”
Then run the installed NI-MAX software and check the device-ID of the con-
nected NI-board. It should be “Dev1” if this is the first NI-board on the computer. If it
has a different device-ID, please update this ID within the configuration-file
“ComCell.ini” under the tag “NIBoardID”:
NIBoardID:
Dev1
2.1.2. Setup of KDEE_v2 and KDEE_MfPG devices
The KDEE_v2 pulsegenerator as well as the KDEE_MfPG control-device uses both the
well-known FTDI-chip for the USB-communication. Please download and install the re-
quired driver for your computer from http://www.ftdichip.com and extract the down-
loaded driver-file “CDM vX.YY.ZZ WHQL Certified.zip” to a directory. After extraction
please use the Windows-driver-wizard and direct it to the extracted driver-files. After
installing the driver a new COM-Port will be shown within the Windows-device-manage.
Warning: this COM-Port may change if you use another USB-Port or other devices using
a FTDI-USB-Chip on the same computer!
Page 5 of 61
2.1.3. Communication-setup of Tektronix TDS7104 oscilloscope
The Tektronix-MatLAB-Device-Driver “tektronix_tds7104.mdd” has to be copied

from the folder “%ComCell%\Requirements\” to the instrument-control-toolbox-driver-
path:
“%MatLAB%\toolbox\instrument\instrument\drivers\tektronix_tds7104.mdd”
Within the configuration-file “ComCell.ini” you have to set the communication-

parameters for the scope:
When using a direct-USB-connection, please enter the USB-address of the scope

under the tag “USB address”. The address consists of the USB-Product-ID, -Vendor-ID
and the USB-serial-number of the device. All three values can be found in the VISA-
Explorer or the Windows device-manager:
USB address:
USB0::0xVID::0xPID::SERIAL::0
Using a GPIB-connection the GPIB-address has to be set under the tag “GPIB
address”. The address can be set in the range of 1 … 30:
GPIB address:
GPIB0::GPIBADDRESS::INSTR
Using TCP/IP (Ethernet) the IP-Address of the scope has to be set under the
tag “TCPIP address” in the form IPADDRESS=123.123.123.123:
TCPIP address:
TCPIP::IPADDRESS::INSTR
2.1.4. Communication-setup of LeCroy-oscilloscope
The LeCroy-MatLAB-Device-Driver “lecroy_basic_driver.mdd” has to be copied

from the folder “%ComCell%\Requirements\” to the instrument-control-toolbox-driver-
path:
“%MatLAB%\toolbox\instrument\instrument\drivers\lecroy_basic_driver.mdd”
Within the configuration-file “ComCell.ini” the IP-address and TCP-port to the

LeCroy-oscilloscope has to be set under the tag “TCPIP address” and “TCPIP port”:
TCPIP address:
IPADDRESS
Page 6 of 61
TCPIP port:
IPPORT
When using a direct-USB-connection, please enter the USB-address of the scope

under the tag “USB address”. The address consists of the USB-Product-ID, -Vendor-ID
and the USB-serial-number of the device. All three values can be found in the VISA-
Explorer, the Windows device-manager or the entire configuration-string within the spe-
cific settings-page in the oscilloscope:
USB address:
USB0::0xVID::0xPID::SERIAL::INSTR
Afterwards within the LeCroy-oscilloscope the option “SONSTIGES  [x] Enable

HTTP Screen Capture” has to be enabled so that the ComCell.m is allowed to download
screenshots over HTTP-protocol. Furthermore the option “ERFASSUNG  Offset Setting
Constant In: [x] Div” and “ERFASSUNG  Delay Setting Constant In: [x] Div” has to
be set.
2.1.5. Communication-setup of Yokogawa-oscilloscope
First the Yokogawa-MatLAB-ToolKit-software “DL-MatLAB” has to be installed.

The software can be downloaded from the Yokogawa-website:
“http://tmi.yokogawa.com/de/products/oscilloscopes/anwendungssoftware-fr-oszillos-
kope/matlab-tool-kit-for-dl-series-701991/”
Afterwards the path where the software has been installed (mostly “C:\Program
Files\Yokogawa\matlab\dl20”) has to be set as MatLAB-path: within MatLAB click on
“Set Path” click on “Add Folder…” and enter the specific path to the DL-ToolKit.
Within the configuration-file “ComCell.ini” you have to set the communication-

parameters for the scope:
When using a direct-USB-connection, please enter the serial-number of the scope

under the tag “USB address”:
USB address:
SERIALNUMBER
Using a GPIB-connection the GPIB-address has to be set under the tag “GPIB
address” in the range of 1 … 30:
Page 7 of 61
GPIB address:
GPIBADDRESS
Using TCP/IP (Ethernet) the IP-Address of the scope has to be set under the
tag “TCPIP address” in the form 123.123.123.123:
TCPIP address:
IPADDRESS
2.1.6. Setup oscilloscope-presets for switching- and conduction-measurements
For every single measurement-type a preconfigured template/preset has to be

stored within the oscilloscope. Tektronix-, LeCroy- and Yokogawa-scopes (follow chapter
2.1.6.6) have a number of preset-slots for this. Within the configuration-file “ComCell.ini”
these presets have to be configured for switching-on, switching-off, diode-switching, con-
duction and diode-conduction as well as the auto-zero-measurement for diodes.
It is important that the trigger-points and the timebases are set in such way, that
both the (gate-)voltages and currents are stable at the screen-borders, as the integration
borders, du/dt- and di/dt-values are based on the first and last 5% of the single curves.
The timebase depends on the type of switch/diode to be measured. So it is rec-

ommended to test the measurement-steps with the minimum and maximum voltage and
current manually to get the limits of the device under test. Afterwards the presets within
the oscilloscope have to be adjusted to the specific switch or diode.
2.1.6.1. Preset for switching-on
During switching-on the blocking-voltage is falling down, while the current is ris-
ing. The trigger-point has to be set to the falling collector-emitter-voltage (drain-source-
voltage). An A/B-trigger is recommended so that the second impulse of the double-pulse
is triggered safely (see Fig. 2.1):
• Trigger A (the first one) should trigger on rising edge of the blocking-voltage at
a level of 50V to 100V (depending on the maximum blocking-voltage)
• Trigger B (the second one) should trigger on falling edge at the same voltage
range as trigger A
• As there is no trigger-delay or –counter necessary set the delay and counter to 1
so that both trigger-events are used directly
Page 8 of 61
Fig. 2.1: Triggering switching-on-events
2.1.6.2. Preset for switching-off
Triggering switching-off events only use a single-event triggering, as the desired

switching-off current is reached at the end of the first pulse. Here only trigger A is used
at a voltage of about 100V to 150V (depending on the maximum blocking voltage). Fig.
2.2 shows the trigger-event.
Page 9 of 61
Fig. 2.2: Triggering switching-off-events
2.1.6.3. Preset for diode-switching
The reverse-recovery-current can be measured using an A/B-triggering again. Fig.

2.3 shows the corresponding triggering-event to the beginning of the second impulse:
• Trigger A (the first one) should trigger on falling edge of the diode-blocking-
voltage at a level of 100V to 150V (depending on the maximum blocking-voltage)
• Trigger B (the second one) should trigger on rising edge at the same voltage range
as trigger A
• As there is no trigger-delay or –counter necessary set the delay and counter to 1
so that both trigger-events are used directly
Page 10 of 61
Fig. 2.3: Triggering diode-switching-events
2.1.6.4. Preset for conduction
The voltage-drop of a conducting switch can be measured using two channels for
each collector-current and collector-emitter-voltagedrop. Fig. 2.4 shows the trigger-set-
tings for a 450A switch:
• If the measured current is inverted, trigger A is set to rising edge, to trigger at

the end of the current-ramp
• If the measured current is non-inverted, trigger A ist set to falling edge
The timebase of the measurement will be set by ComCell.m to get an optimized resolu-
tion.
Page 11 of 61
Fig. 2.4: Triggering conduction-event
2.1.6.5. Preset for diode-conduction
The voltage-drop of a conducting diode can be measured using two channels for
each forward-current and diode-voltagedrop. Fig. 2.5 shows the trigger-settings for a
450A diode:
• If the measured current is inverted, trigger A is set to rising edge, to trigger at

the end of the current-ramp
• If the measured current is non-inverted, trigger A ist set to falling edge
The timebase of the measurement will be set by ComCell.m to get an optimized resolu-
tion.
Page 12 of 61
Fig. 2.5: Triggering diode-conduction-event
2.1.6.6. Presets on Yokogawa-oscilloscopes
As the Yokogawa-oscilloscopes uses filenames instead of numbered preset-slots,

the settings in ComCell.ini will be translated to filenames to be compatible to the Tek-
Tronix- and LeCroy-scopes. The following table will show the equivalent filenames which
have to be stored on the Yokogawa-oscilloscope to match the settings in ComCell.ini:
Presetnumber Filename on Yoko- Description

gawa
1 SWITCHINGON See chapter 2.1.6.1
2 SWITCHINGOFF See chapter 2.1.6.2
3 DIODESWITCHING See chapter 2.1.6.3
4 CONDUCTION See chapter 2.1.6.4 and 2.1.6.5
5 AUTOZERO See chapter 3.6
6,7,… PRESET6, PRESET7,… Optional storages
Page 13 of 61
2.1.7. Calibrating the FuG-voltage-source
Important: the behavior of the calibration-values within the file “ValueCorrec-

tion.ini” changed with ComCell-version 3.5.7. In older versions the correction-factors
are stored, but in newer versions the final values are stored for better usability!
The FuG-device has to be calibrated to compensate non-linear-effects of the spe-

cific hardware. For this do the following steps for a new FuG-voltage-source:
Calibrate measured values from FuG-device (Input):
• Within the configuration-file “ComCell.ini” set the tag “DCSourceDevice” to

“Calibrate”
• Connect the analog-cable of the FuG-device to the ComCell-hardware-box
• Disconnect both HVDC-cables from the FuG-device
• Start “ComCell.m”
• Unlock and enable the output of the FuG, set to local-mode and set several volt-
ages between 0V and the maximum output-voltage and record the voltage dis-
played in the FuG-display and the measured value within the ComCell-software
into a temporary file.
• Open the file “ValueCorrection.ini” with a text-editor like Notepad++ and attach
the correlating measured values in the following form:
DeviceName device-name
0.0,0.25,0.5,0.75,1.0\n received input-values from 0.0 to 1.0
0.0,250.0,500.0,750.0,1000.0\n Corresponding voltage measured at hardware
x,x,x,x,x,x\n only for output
x,x,x,x,x,x\n only for output
Calibrate set-values to FuG-device (Output):
• Within the configuration-file “ComCell.ini” set the tag “DCSourceDevice” to

“Calibrate”
• Disable the FuG-output (we only need the displayed values) and activate the
remote-control
• Press the button “Set value” to display the desired voltage
• Start “ComCell.m” and switch to the page manual testing. Enter voltage-values
between 0.0V and 1.0V in the area “Temperatures / System-Test” and click “Set
DC-Link to >” in the area “DC-Link Voltage Control”.
Page 14 of 61
• Open the file “ValueCorrection.ini” with a text-editor like Notepad++ and attach
the correlating measured values in the following form:
DeviceName device-name
x,x,x,x,x\n only for input
x,x,x,x,x\n only for input
0.0,250.0,500.0,750.0,1000.0\n Desired output voltage
0.0,0.25,0.5,0.75,1.0\n Output-value that has to be sent to device
• Within the configuration-file “ComCell.ini” set the tag “DCSourceDevice” back
to the desired FuG-device (e.g. “FuG2.0kV” or “FuG6.5kV”)
The amount of sampling-points can be changed, but have to be at least 2 or 3 to allow

creating an extrapolation of the values.
2.1.8. Calibrating ThermoCouples
The received values of the six thermo-couples can be corrected, too. For this the entry
“ThermoCouples” within the file ValueCorrection.ini can be used. Every two lines are in
correlation to each other. The first line is the original value read by the hardware and
the second line is the corrected value:
ThermoCouples
0,10,25,50,75,100\n
0,10,25,50,75,100\n
0,10,25,50,75,100\n
0,10,25,50,75,100\n
0,10,25,50,75,100\n
0,10,25,50,75,100\n
0,10,25,50,75,100\n
0,10,25,50,75,100\n
0,10,25,50,75,100\n
0,10,25,50,75,100\n
0,10,25,50,75,100\n
0,10,25,50,75,100\n
2.2. Starting the application and data-structure
Start the ComCell program by calling “ComCell.m“ out of the MatLAB-command

line. Make sure that the current folder of MatLAB is set to the ComCell-path. The
program will show the welcome screen.
Page 15 of 61
Equipment- Equipment- Equipment-

Setup I Setup II Setup III
Test- Test- Test-

Setup A Setup B Setup C
Conduction- Switching- Conduction- Switching- Conduction- Switching-

Setup A1 Setup A1 Setup B1 Setup B1 Setup C... Setup C...
Conduction- Switching- Conduction- Switching-

Setup A2 Setup A2 Setup B2 Setup B2
Conduction- Switching- Conduction- Switching-

Setup A... Setup A... Setup B... Setup B...
Fig. 2.6: Data-structure of ComCell v3.x
Fig. 2.6 shows the data-structure of ComCell in version 3.x. Within the equip-
ment-setup all information about the used measurement-devices and the inductor will be
stored. The test-setup will then store the information about the D.U.T. and the safety-
limits. On creating a new test-setup (e.g. for a new D.U.T.) a new set of conduction- and
switching-setups are created. On saving the received data of the oscilloscope, all infor-
mation of equipment- and test-setup will be stored within the conduction- and switching-
data, too. So the analyzers are able to show the used equipment even when these data
are not available anymore.
2.3. Creating and managing setups
As the data-structure shown in the previous chapter 2.2 allows multiple equip-
ment-setups, test-, conduction- and switching-setups the ComCell-software is able to act
as a data-management-software for the measured DUTs as well as the used measurement
hardware. Each of the four setup-screens described in the following chapters 2.4 to 2.7
have the ability to create and navigate through multiple setups. If you are using different
voltage-probes for different types of DUTs than create two or more setups for each hard-
ware-configuration.
Fig. 2.7: setup-control-buttons
Page 16 of 61
With a click on the button “+” and “-“ (Fig. 2.7) you can create and delete
additional setups that can be browsed with the two buttons “<-“ and “->”:
Fig. 2.8: Browsing through single setups
The button “STORE” will store the values of all input-boxes to the current setup.
This is necessary if a value has been changed (input-boxes with red background). Click
the “Copy”-button to mark the current setup for copying. Navigate with the “<-“ or “-
>” to another setup and click the “Paste”-button and all values (except the unique IDs)
of the marked setup will be inserted to the new destination.
For each DUT create an own setup with the specific free-wheeling-diode within
the screen “2: Test setup”. So each DUT will be stored with the corresponding safety-
limits. In Fig. 2.6 it is shown, that each test-setup will create its own groups of conduc-
tion- and switching-setups. So if you creates a test-setup for switch A, you will see the
conduction- and switching-setups of switch A only within the screens “3: conduction-
setup” and “4: switching-setup”. Navigating to the test-setup for switch B will show up
the conduction- and switching-setups of switch B only. Indeed all measured values will
be stored to a common measurement-folder “ComCellResults” but the data-management
is much easier using this setup-management.
2.4. Equipment setup
To set up the measurement equipment please click on the button “1: Equipment
setup” to enter information about the used hardware for documentation. Fig. 2.1 shows
some demo-values within the equipment setup screen:
Page 17 of 61
Fig. 2.1: Equipment setup screen
Most of these values are only for documentation-purpose. Solely the values of the
lime-colored fields for the inductors, the intermediate circuit capacitance and the current-
measurement will be used to calculate the necessary pulse-widths for the double impulses.
So make sure that these input-fields have the correct values used in your hardware. More
information about the pulse-calculation can be found in chapter 6.2.
Page 18 of 61
2.5. Test setup
Fig. 2.2: Test setup screen
When you have finished the configuration of the equipment, you can click on the
button “2: Test setup” to configure the DUT (device under test). Here the names and
maximum ratings for the used switch and diode have to be set.
Fig. 2.9: Configuration of name for switch and diode
Fig. 2.9 shows the input-fields of the name for switch and diode. Please enter the
name without special characters (“,;.:-_#’+*~´`\/()[]{}&%$§”!”) as these values will be
used for creating the folder-structure on the computer.
Fig. 2.10: Defining safety-limits
Page 19 of 61
The safety-limits (Fig. 2.10) will be compared with the values measured by the
oscilloscope:
• “max. DC-Voltage” will be checked during switching-off-loss-measurements.

Please be aware that the device-overvoltage cannot be measured during switching-
on or conduction-loss-measurements. So make sure that the overvoltages will not
exceed the device-limits by checking the maximum current and DC-link-voltage
with manual pulses before you start the automated measurement!
• “max. positive/negative Gate-Voltage” will be tested during switching-loss-meas-
urements only. As the gate-voltage can oscillate on fast switches a lot, you may
increase these values slightly above the safe device-limits
• “max. Current” will be checked after each measurement-step. If channel 3 (7 in
dual device mode) will rise above the maximum value, the automated measure-
ment will abort with an error
The buttons “+”, “-“, “STORE”, “Copy” and “Paste” have the same functions as
described in chapter 2.3. Fig. 2.2 shows the screen of the test setup with some demo
values.
2.6. Conduction-loss setup
Within the conduction-loss setup shown in Fig. 2.3 you can enter a minimum and
maximum value and type of distribution for the gate-voltage and temperature, the DC-
link voltage, as well as the current together with a number of steps (generally the number
of current-steps should be set to 1 for the conduction-loss-measurements). With these
values the program will calculate all single measurement steps. The gate-voltage- and
temperature-distribution can be chosen between “Linear” and “Self-defined”. In linear-
mode the minimum and maximum value will be used together with the number of steps
to create an individual vector. Choosing “Self-defined” a MatLAB-vector like “[5:5:15; -
5:-5:-15]” (gate-voltage, [GATEON; GATEOFF]) or “[25, 75, 150]” (temperature) can be
defined.
Page 20 of 61
Fig. 2.3: Conduction-loss setup screen
• “ Measure negative current” will invert the measured current. This is necessary
if a shunt is used within the inverted current-path
• “ Dual device mode (Ch1..3 and Ch5..7)” will measure two DUTs at the same
time using an oscilloscope with 8 instead of 4 channels (see chapter 3.7)
• “ Record all channels (Ch1..4 (Ch5..8))” will download and save the values of
channel 1 and 2 (and 5 and 6 in dual device mode) even they are not used by the
conduction-loss-measurement. This will take more time but the measured chan-
nels can be used for additional signals
• “ Convert voltage to current on Ch3/7 (Ch 2/6)” will use the set current-sensor-
values of the tab “1: Equipment setup” to auto-scale the oscilloscope-channels 3
and 7 (as well as channels 2 and 6 on distributed current measurements). The
measured shunt- or Rogowski-voltage will be converted to a current using the
resistance of the shunt or the mV/A-ratio of the Rogowski-coil. If you are using
a shunt, set the mV/A-ratio to 0mV/A or if you are using a Rogowski-coil set the
shunt-resistance to 0Ohm. The scaling will be used for both directions: reading
values from the oscilloscope and sending scaling-settings to the scope.
Page 21 of 61
• “ Distributed current measurement (Ch2+3 / 6+7)” will allow to measure to

current paths at the same time with two separated current measurements. The
spare channel 2 (and 6 in dual device mode) will be used to measure the second
current. Afterwards the analyzers will auto-zero both channels independently and
calculate the summarized value.
• “ Use averaging” will allow an automated re-run of all steps with the given
amount of runs. Afterwards the conduction-loss-analyzer (see chapter 4.1) will
calculate the mean-value of all measured values to minimize the measurement-
error
described in chapter 2.3.
Fig. 2.11: Choosing test-mode
A click on the button “Run test” will start the test-sequence either for the switch
if “[x] Switch” is selected in the box “Test options” or for the diode if “[x] Diode” is
chosen. The button “View results” will open the conduction-loss-analyzer for analyzing
already measured values.
Page 22 of 61
2.7. Switching-loss setup
Fig. 2.4: Switching-loss setup screen
Fig. 2.4 shows the screen of the switching-loss setup. A minimum and maximum
value for the gate-resistance, positive/negative values, number of measurement-steps and
step-distribution for gate-voltage, temperature, the DC-link voltage and the current can
be set. With these values the program will calculate all single measurement steps. Similar
to the conduction-loss setup the distribution of the parameters can be switched between
“Linear”-mode and a “Self-defined”-mode. The gate-resistance- and gate-voltage-vectors
have to be defined in the following style “[RGON; RGOFF]” and “[UGON; UGOFF]”.
The temperature, DC-link-voltage and the current can be define as a standard vector:
“[MIN:STEP:MAX]” or “[VALUE1, VALUE2, MIN:STEP:MAX]”
• “ Measure negative current” will invert the measured current. This is necessary
if a shunt is used within the inverted current-path
• “ Dual device mode (Ch1..3 and Ch5..7)” will measure two DUTs at the same
time using an oscilloscope with 8 instead of 4 channels (see chapter 3.7)
• “ Record all channels Ch1..4 (Ch5..8)” will download and save the value of
channel 4 (and 8 in dual device mode) even it is not used by the switching-loss-
measurement. This will take more time but the measured channel can be used for
an additional signal
Page 23 of 61
• “ Convert voltage to current on Ch3/7 (Ch4/8)” will use the set current-sensor-
values of the tab “1: Equipment setup” to auto-scale the oscilloscope-channels 3
and 7 (as well as channels 4 and 8 on distributed current measurements). The
measured shunt- or Rogowski-voltage will be converted to a current using the
resistance of the shunt or the mV/A-ratio of the Rogowski-coil. If you are using
a shunt, set the mV/A-ratio to 0mV/A or if you are using a Rogowski-coil set the
shunt-resistance to 0Ohm. The scaling will be used for both directions: reading
values from the oscilloscope and sending scaling-settings to the scope.
• “ Distributed current measurement (Ch3+4 / 7+8)” will allow to measure to
current paths at the same time with two separated current measurements. The
spare channel 4 (and 8 in dual device mode) will be used to measure the second
current. Afterwards the analyzers will auto-zero both channels independently and
calculate the summarized value (see chapter 3.8).
• “ SwitchingOn (Ch1..3) + DiodeSwitching (Ch6..7)” will record Eon and Erec at
the same time using an 8ch-oscilloscope (see chapter TODO)
• “ Compensate voltage drop over DC-link” will calculate the expected current-
based voltage drop over the DC-link. This is done by increasing the DC-link by
the expected voltage-drop before the pulse is started. Using this function a weak
DC-capacitance can be used for higher currents
• “ Use averaging” will allow an automated re-run of all steps with the given
amount of runs. Afterwards the switching-loss-analyzer (see chapter 4.2) will cal-
culate the mean-value of all measured values to minimize the measurement-error
Fig. 2.12: Preview of pulse lengths
In the preview-area you can plot the calculated impulse-lengths for all DC-link
voltages and currents with the button “Plot calculated impulse length” to get an idea of
the pulse-signals the program will send to the hardware. Two graphs will be drawn show-
ing the pulse-lengths over steps (left graph) and the voltage-error over steps (right
Page 24 of 61
graph). Disabling the checkbox “ Steps” will switch both graphs from “steps” to “Cur-
rent in A”. Furthermore you can switch the right graph to absolute DC-link-voltage with
the checkbox “ DC-Link-Voltage”.
Fig. 2.13: Manual output section
Within the section “Manual pulse output” you can configure and start a manual
double-impulse for debugging and first setup of the DUT. Enter the desired values for
pulse-duration 1, pause-duration and pulse-duration 2 in microseconds and click the but-
ton “Fire manual pulse”. You may enter comma values in the form “5.1”. Important: on
manual pulses the oscilloscope will not be read by the software automatically so the
limits-section of the test-setup will have no effect!
• “Set FuG to 0V” will turn-off the remote-controlled FuG-voltage source

• “Set FuG to >” will set the remote-controlled FuG-voltage source to a voltage
specified in the field on the right side of this button for manual tests
described in chapter 2.3.
A click on the button “Run test” will start the test-sequence either for the switch-
ing-on or -off if “[x] SwitchingOn” or “[x] SwitchingOff” is selected in the box “Test
options” or for the diode if “[x] Diode” is chosen. Switching-on and –off can be measured
together while the diode-measurement needs a different hardware-configuration. Im-
portant: read the information about the safety-limits in chapter 2.5. The button “View
results” will open the switching-loss-analyzer for analyzing already measured values.
Page 25 of 61
3. Measurement
3.1. Setup necessary values
To measure the switching- or conduction losses, only a few settings have to be

made. These settings are highlighted in lime-colored input-boxes. All other settings are
only for documentation and will be stored in the reports afterwards.
Step 1: Enter Inductance
For switching-losses the program will use the inductance in µH and its resistance
in Ohm from page “1: Equipment setup” to calculate the necessary pulse-length.
Fig. 3.1: Entered values of the used inductances

and the intermediate circuit capacitance
Step 2: Enter names of DUT and safety-limits
From tab “2: Test setup” the program will use the name of the DUT (Switch-
Name combined with Diode-Name, see chapter 3.3) to create the measurement-folders.
Furthermore the “Safety limits” will be used, to check if the values of the DC-sources
and the dynamic device voltages are within the range of the DUT.
Fig. 3.2: Enter names for DUT
Page 26 of 61
Step 3: Configure conduction-loss-measurement
For measuring conduction-losses all lime-colored boxes in “3: Conduction” have

to be set: the temperature range and number of steps (highlighted in red), the DC-link
voltage (highlighted in yellow) as well as the current range together with number of steps
(highlighted in green). The length of the pulse has to be set as well as the number of
averaging runs (highlighted in blue). You can enable or disable this averaging function
by selecting the checkbox “Averaging enabled”.
Fig. 3.3: configure conduction-loss measurement
Step 4: Configuring switching-loss-measurement
For measuring switching-losses all lime-colored boxes in “4: Switching” have to be

set: the gate-drive-settings (highlighted in black), the temperature range and number of
steps (highlighted in red), the DC-link voltage range and number of steps (highlighted
in yellow) as well as the current range together with number of steps (highlighted in
green). The length of the second pulse has to be set as well as the number of averaging
runs (highlighted in blue). You can enable or disable this averaging function by selecting
the checkbox “Averaging enabled”. At the end you can select which type of test you
want to execute (highlighted in orange).
Page 27 of 61
Fig. 3.4: configure switching-loss measurement
With the options called “R_Gate distribution”, “Voltage distribution”, “DC link
distribution” and “Current distribution” the behavior of the gate-resistance-, gate-
drive-voltage, DC-link voltage and current step-distribution can be configured. It can
be chosen between “Linear” and “Self-Defined”. With “Self-Defined” a MatLAB vector
can be set like “[1:1:9, 10:5:25, 30:10:100]”. For the current two additional options
“Non-Linear 1” and “Non-Linear 2” are available for a better current distribution. The
following figures will show the different distribution-methods for the current:
Fig. 3.5: Current distribution using „Linear“
Page 28 of 61
Fig. 3.6: Current distribution using „Non-Linear 1“
Fig. 3.7: Current distribution using "Non-Linear 2"
Fig. 3.8: Current distribution using "Self-Defined" = [1:1:9, 10:5:35, 40:10:100]
3.2. Connecting the hardware
In the ComCell program you can choose between several connection-types of the
oscilloscope in the upper right corner of the program. The options shown within this box
depends on the selected oscilloscope within the configuration-file “ComCell.ini”. The os-
cilloscope can be connected by clicking on “Connect Osci”.
Page 29 of 61
Fig. 3.9: Example for supported connections to an oscilloscope
For measuring the conduction-losses the probes have to be connected as following:
Channel 1: not used Channel 3: Current probe (Shunt/Rogowski)

Channel 2: not used / Current Channel 4: Voltage probe (Low-Voltage)
probe*
For measuring the switching-losses the probes have to be connected as following:
Channel 1: Gate-Emitter-Voltage Channel 3: Current probe (Shunt/Rogowski)

Channel 2: Collector-Emitter-Volt- Channel 4: not used / Current probe*
age
* This current will be measured in “Distributed current measurement”-mode only. See

chapter 2.6 and 2.7 for a description of this option.
Within the oscilloscope several presets are used to recall the trigger- and probe-
setups. There are several preset-slots within the Tektronix-, LeCroy- or Yokogawa-oscil-
loscope that can be used together with the ComCell-software. Open the file “ComCell.ini”
to setup the desired preset-slot for each measurement-type (see Fig. 3.10). The given
preset number stored in ComCell.ini will be used to load the preset-slot out of the oscil-
loscope-memory during the test.
Page 30 of 61
Fig. 3.10: Preset setup within "ComCell.ini"
3.3. Running the test and storing the results
By clicking “Run test” the ComCell program tries to set the gate-resistance, the
gate-voltage, the device-temperature of the DUT, the DC-sources and the oscilloscope.
Some of the functions are not automated yet – here a message will open and give instruc-
tions what to do. Then the software fires the impulses and will read out the waveforms
out of the oscilloscope. Afterwards the software calculates the values of the losses curve
and stores all values to the following folder-structure:
ROOT\ComCellResults\DUTS_DUTD\ID\RGON_RGOFF\UGON_UGOFF\TEMPERATURE\UDC\MEASURE-
MENT\RUNx
• ROOT is the location of the database. You can see this location at the lower left
corner of the program:
Fig. 3.11: location of current database
• ComCellResults is the fixed name of the results folder

• DUTS is the name of the switch and DUTD the name of the diode
• ID is an automatically generated identifier like
“id_68695626_7acd_4dde_8ffa_6dc19642b249”. So you can create several measure-
ment-presets with identical setup for same devices without overwriting data of
other presets. Important: conduction- and switching-loss measurements will have
different IDs.
• RGON and RGOFF are the set gate-resistances for switch-on and switch-off
• UGON and UGOFF are the used gate-voltages for switch-on and switch-off
Page 31 of 61
• TEMPERATURE is one of the several temperature-steps

• UDC is one of the several DC-link-voltage steps
• MEASUREMENT is either “SwitchingOn”, “SwitchingOff”, “DiodeSwitching” or “Conduc-
tion”, “DiodeConduction”
• RUNx contains the measurement-results. “Run1” contains the original measure-
ment. “Run2” and above contains all additional averaging measurements.
Fig. 3.1 Folder structure
For each measurement-step the program will create a single csv file like “step_1
@ 1A, 4µs.csv” and the original oscilloscope-screenshot as png-file within the RUNx
folder. There are two additional files within each “MEASUREMENT”-folder called
“EquipmentSetup.csv” and “TestSetup.csv”. These two files contain detailed information
about the used equipment and test setup you’ve setup in the ComCell program.
3.4. Analyzing the results
After the measurement is finished you have two options for analyzing the data:
• either you can use the csv-files out of the described folder-structure above for a
manual processing (in Excel for example)
• or you can use the “Analyzer”-tools to analyze the data and export a report au-
tomatically. These two programs will be described in chapter 4.
Page 32 of 61
3.5. Adding measurement-points to an existing set of data
If you are missing some points of data (for example an additional temperature, or
another set with different gate-resistors) you can do the following:
• Open ComCell and load the database you want to supplement

• Load your settings within “1: Equipment setup” and “2: Test setup” and make
sure that the name of the DUT is the same as you used before. Otherwise ComCell
will create a new DUT-folder
• Load the necessary settings within “4: Switching” and setup all points as you need
• Click on “Run switching-loss test”
The new results will be inserted into the current folder-structure. If a set of old
measurement results with the same settings (same gate-resistance, same gate-voltage,
same temperature and same dc-link-voltage) is existing the folder will be renamed to
“OLDNAME_backup (DATE TIME)” to prevent an interruption of the automatic test.
You will be informed with an info-message within the MatLAB command-window only!
3.6. Auto-zero-measurement for diode-switching
While the first/last 5% of the measured current of switching-on and –off of an

IGBT are used for auto-zero-compensation the measured current of the diodes is clipped
for a better resolution of the reverse-recovery-current. So an auto-zero-compensation
based on the measured current is not possible after the measurement is done. If the auto-
zero-compensation is necessary, within the configuration-file “ComCell.ini” the a preset-
number under the tag “AutoZero” in the area “[Oscilloscope presets]” has to be set to a
value above -1.
A value above -1 results in an additional measurement before the main diode-

switching-measurement. The oscilloscope-preset has to be configured to trigger on the
AC-grid so that the oscilloscope will trigger immediately after loading the preset. Then
the current will be measured and the mean-value will be used as dc-offset for the main-
measurement of the diode-current.
Page 33 of 61
3.7. Dual-device-mode
Version 3 of the ComCell-program supports the measurement of two devices in

parallel using a oscilloscope with 8-channels. For this you have to activate the option
“Dual device mode” in the tab “3: Conduction” or “4: Switching”. Then the oscilloscope
will record 8 instead of 4 channels. The channel-configuration is analog to the 4-channel-
mode:
For measuring the conduction-losses the probes have to be connected as following:

probe*

probe*
For measuring the switching-losses the probes have to be connected as following:

age

age

The conductionloss-analyzer and switchingloss-analyzer will recognize the dual-

device-mode and show the measured curves of both devices within one graph. All ana-
lyzing-functions will be available for both measured devices. The export of the switching-
energies can be switched between exporting the summarized energies and exporting both
values in single excel-sheets. Conduction-loss-values will be exported in two separated
excel-sheets anyway.
Page 34 of 61
3.8. Distributed current measurement
On some high-current IGBT-packages the current-path is split, so that current-

measuring with a single Rogowski-coil is not possible. If the “Distributed current meas-
urement” is activated in conduction- or switching-loss-measurement, the spare channels
2 and 6 (for conduction) as well as 4 and 8 (switching) (see chapter 3.2 and 3.7) are used
for the second current measurement. The software will consider the halved current for
each path. On demand both Analyzers (conduction and switching) will auto-zero both
currents independently, summarize them and show them as one single device-current. All
calculations are done with the summarized device-current.
3.9. Combined SwitchingOn and DiodeSwitching with 8ch-Oscilloscopes
Using an 8ch-oscilloscope you may record the switching-on- and diode-switching-

losses at the same time. For measuring the losses the probes have to be connected as
following:

age
Channel 5: not used ** Channel 7: Current probe (Shunt/Rogowski)

Channel 6: Diode-Voltage Channel 8: not used / Current probe*

** The information on this channel will be stored within the csv-file but ignored by the
analyzers.
Page 35 of 61
4. ComCell Analyzing tools

The ComCell analyzer-tools are a powerful set of tools to analyze the measured
curves of the ComCell mainprogram. After clicking on “View results” in the main pro-
gram, the “ComCell Analyzer” will open in conduction-loss- or switching-loss-mode.
4.1. Conduction-loss analyzer
The conduction-loss-analyzer will present the measured curves of collector-current

IC(t), voltage-drop Uce(t) and the calculated voltage-drop over current Uce(IC). In the
lower left corner you can select, which measurement you want to view on. You can select
the temperature, the dc-link-voltage and the measurement itself. For each measurement
you can browse through the measurement-step (buttons “Step+” and “Step-“). If the
averaging-function was enabled during the measurement, you can browse through the
averaging-runs, too (buttons “AVG+” and “AVG-”). A click on “View Screenshot” shows
the original oscilloscope screenshot for comparison. On the right you can change settings
for displaying and calculating:
Fig. 4.1: Settings in conduction-loss-analyzer
• “ Y-Autoscale” optimizes the vertical axes

• “ Compensate Voltage Drop over Shunt” can be activated if the voltage over
the shunt is measured together with the DUT. Then the resistance of the field
“Current sensor (Shunt) in the Tab “1: Equipment setup” is used for the compen-
sation
• “ Zero compensate current” will remove a DC-offset within the current. For this
the mean-current of the first 5% of the current-waveform will be used as offset.
• “Number of calculation steps” is the number of steps used for calculating the
Uce(Ic) curve
• “Type of step distribution” allows alternative distribution of the Uce(Ic) curve-
points. As with higher current the curve will flatten the “Nonlinear 2” method is
preferred. In the dropdown-field you can choose, which style of calculation-point
Page 36 of 61
distribution you prefer. The distribution “Nonlinear 2” is the standard-value to

get a maximum resolution at low current-values with less calculation-points.
General information about the conduction loss-calculation can be found in chapter 6.2.
Fig. 4.1 ComCell conduction-loss-analyzer
The conduction-loss analyzer can be started without the ComCell-mainprogram.

Call “ConductionLossAnalyzer.m” in the MatLAB-command and select an ID-folder of
the desired conduction-loss-measurement.
The following settings can be used:
• “ Autoscale Y-axis” shows the individual curves in an optimal scaling (has no

effect on the calculation)
• “ Compensate Voltage Drop over Shunt” takes the resistance-value of the shunt
(1: Equipment Setup, see 2.3) and removes the voltage drop over this shunt from
voltage-measurement in channel 2
• “ Autozero current” and “ Autozero voltage” will use the first 5% of the
current- or voltage-curve to detect a probe-offset and compensate this offset au-
tomatically
• “ Low-pass-filter voltage” will reduce spikes or disturbances on the voltage by
using a low-pass-filter
Page 37 of 61
4.2. Switching-loss analyzer
In five diagrams you can see the following data. Fig. 4.2 shows the GUI of the
Analyzer.
• Gate-Emitter-voltage Uge(t)
• Collector-Emitter-voltage Uce(t)
• Current of the shunt IC(t)
• Values of the calculated math-channel Psw(t)
• and the calculated losses curve Esw(IC)
Like in the conduction-loss analyzer you can select the measurement you want to
view in the lower left corner. You can select the gate-resistance, the temperature, the dc-
link-voltage and the measurement. For each measurement you can browse through each
measurement-step (buttons “Step+” and “Step-“). If the averaging-function was enabled
during the measurement, you can browse through the averaging-runs, too (buttons
“AVG+” and “AVG-”). A click on “View Screenshot” shows the original oscilloscope
screenshot for comparison.
Fig. 4.2 ComCell switching loss-analyzer
The following settings can be made:
Page 38 of 61
Fig. 4.2: Settings in switching-loss-analyzer
• “ Use measured current for…”. If checked, the software will use the measured
current for x-axis of the energy-curve: for switch-on (Eon) the software will cal-
culate a mean-value over the last 10% of the current-waveform. For switching-off
(Eoff) it will search for the peak-current and calculates a mean-value of 5 meas-
urement-points left to 5 measurement-points right of the peak. For diode-switch-
ing (Err) it takes the first 10% of the current as mean-value. If deactivated, the
analyzer will take the calculated reference-current which is necessary if parts of
the current are clipped by the scope (e.g. for reverse-recovery-current).
• “ Comp. Voltage Drop over Shunt” takes the resistance-value of the shunt (1:
Equipment Setup, see 2.3) and removes the voltage drop over this shunt from
voltage-measurement in channel 2
• “ Autozero current” will remove a DC-offset within the current. For this the
mean-current of the first (switching-on) or last (switching-off) 5% of the current-
waveform will be used as offset. For diode-switching the zero-compensation can
be activated within the configuration-file “ComCell.ini”. In chapter 3.6 this fea-
ture will be explained in detail.
• “ Autozero voltage” will remove a DC-offset within the voltage. For this the
mean-voltage of the last (switching-on) or first (switching-off) 5% of the voltage-
waveform will be used as offset-reference.
• “ Zoom dydt by 10x” enlarges the green signals by factor 10 drawn below chan-
nel 2 and channel 3 (only for display – no changes for calculation). They are
showing the du/dt for channel 2 and di/dt for channel 3.
• “ Export summarized energies” will export the summarized calculated energies
of channels 1-4 and channels 5-8 if the ComCell was set to dual-device-mode dur-
ing measurement
Page 39 of 61
• “ Don’t export steps with errorneous voltages” allows neglecting measurement-

steps whose voltage dropped below a 3% tolerance in relation to the desired DC-
Link-voltage. This may be necessary for long turn-on times on low-voltage or
using a too small DC-Link-capacitance.
The switching-loss analyzer can be started without the ComCell-mainprogram.

Call “SwitchingLossAnalyzer.m” in the MatLAB-command and select an ID-folder of the
desired switching-loss-measurement.
4.3. Calculation of voltage-drop curves and report
To calculate the Uce(IC) curve of the conduction-measurement, click on the follow-

ing buttons:
• “Calculate curve of current average step”

This will calculate the Uce(IC) curve of the current averaging step. The curve will
be plotted in the right graph and will be stored as “Results.csv” within the RUNx
folders within the measurement folder.
• “Calculate averaged curve of all averaged steps”
This will calculate an averaged Uce(IC) curve of all averaging steps. The curve will
be plotted in the right graph and will be stored as “Results_Averaged.csv” within
the measurement folder. On averaging the values of two or more measurement-
runs with the same settings and same measurement-devices are taken, a single
curve for each run calculated and afterwards a single averaged voltage-drop-
curved is calculated using the mean()-function for each current-step.
• “Analyze all recorded data and create report” will be described in chapter 4.5
4.4. Calculation of switching-losses and report
To calculate the losses, the program uses some integration borders, which are
shown in the math-channel graph. The ComCell program has already analyzed these
curves and set the integration borders to default values. You can change these pre-set
limits simply by moving the scrollbars under the math-channel plot. To apply the changes
click on “Set borders”. “Autoset borders” will reset your changes to the default values
for the current step. If you like to autoset the borders of all steps, click on “Autoset
Page 40 of 61
broders of all measurements”. When the integration limits are OK for all steps and all
average-runs, you can click on the following buttons:
• “Calculate losses curve of current average step”

This will calculate the losses curve of the current averaging step. The curve will
be plotted in the lower right graph and will be stored as “Results.csv” within the
RUNx folders within the measurement folder.
• “Calculate averaged losses curve of all averaged steps”
This will calculate an averaged losses curve of all averaging steps. The curve will
be plotted in the lower right graph and will be stored as “Results_Averaged.csv”
within the measurement folder. On averaging the values of two or more measure-
ment-runs with the same settings and same measurement-devices are taken, a
single energy-curve for each run calculated and afterwards a single averaged
curved is calculated using the mean()-function for each current-step.
4.5. Export to Excel-Datasheet
With the menu “Export”  “Export to Excel-Datasheet” the data-analyzation

and report-generation can be started. The program will iterate through all tempera-
tures, all DC-link-voltages, all gate-resistances/-voltages and all measurements and fi-
nally calculates the averaged Uce(IC) and Esw(IC) curves. Afterwards it will be saved as
an interactive Excel-Sheet.
If there is an existing Excel-Report made by the conductionloss-analyzer, the file

can be “overwritten” to add the loss-values of the current analyzer – the values of the
conductionloss-analyzer are kept and vice versa.
Page 41 of 61
Fig. 4.3: Excel-Report showing values exported by switching-loss-analyzer
The Excel-Sheet contains all necessary hardware-information entered in the

ComCell-dialogs and contains additional calculated values. Next to the Uce(IC) and the
Esw(IC) values the report contains rising- and falling-times for voltage and current, as well
as the dudtmax and didtmax values. For switch-off the voltage-overshoot will be calculated
as well.
By double-clicking on a single value of the collector-current the ComCell ana-

lyzer will open to allow a detailed look at the measured curves. A double-click on an
energy- or a voltage-drop-value will open the screenshot taken by the oscilloscope for a
fast look at the original data.
4.6. Export to KDEE TopBench
The menu “Export”  “Export to KDEE TopBench” allows the calculation and
export of all measured steps into an XML-based *.cle-file. This file can be opened in the
KDEE TopBench topology-calculation-program to compare and benchmark the meas-
ured semiconductor with other devices.
Page 42 of 61
Fig. 4.4: Exported ComCell-measurement in KDEE TopBench
Page 43 of 61
5. KDEE_MfPG Pulse-Generator
5.1. General description
The pulse-generator has to be connected using a standard 230V power-cord con-

nected to a 230V power-plug with ground fault circuit interrupter and over-current-
protection. The power-supply inside of the device will not supply external devices con-
nected to the pulse-generator. The device is ready for operation when both green LEDs,
marked with “operating voltages” light up.
Attention: The device must not be opened. Inside, there are components that work
with mains voltage. There is a risk of electric shock when opening the appliance and
therefore a life hazard! If the seal is broken, the user waives any claims for damages if
the use of the device causes injury or damage.
5.2. Possible output-signals
The KDEE_MfPG Pulse-Generator offers several operation-modes for signal-out-

put. While using the MatLAB ComCell program, the device will be used in standard
double-pulse configuration, that will output only the following signal:
5.3. Communication protocol
Communication is realized via a conventional USB 2.0 Type B cable, which is

connected directly to the controlling computer. An additional USB isolator is not re-
quired. The device is ready for communication when the green LED labeled “USB” lights
up and the mains-ac is connected and powered-up.
Note: The USB interface of the controller is electrically isolated from the rest of
the board. The USB interface is supplied via the connected computer. If the USB cable
is not connected, the shield port of the USB port is not grounded. The grounding must
be carried out via the controlling PC.
Page 44 of 61
For communication with the device a MatLAB-library can be used which offers
several commands for interacting with the hardware. Here only the commands for gen-
erating the pulses will be explained:
Preparing the pulse-settings with setTime(TimeID, Time in Nanoseconds)
PulseID Description
1 OnTime1
2 OffTime / Pause-Time
3 OnTime2
4 Deadtime
5 PulseID:
0 Single Pulse
22 Double Pulse
44 Synchronous Rectifier pulse
66 SoftSwitching pulse
The pulse can be fired using the command pulseCMD(CmdID), where the
“CmdID” can be chosen from the following options:
CmdID Description
0 Standard pulse (Single, Double, Synchr., SoftSwitch.)
3 PWM with n pulses
4 Turn output on
5 Turn output off
If using the “PWM with n pulses”-mode, the following settings can be done using
setTime(TimeID):
TimeID Description
1 Period time in Nanoseconds
2 When PulseID 0, 22 or 44  desired On-Time in Nanoseconds
When PulseID 66  Phase-Shift in percent
3 Number of Pulses
4 Desired deadtime in Nanoseconds
5 Mode-Selection:
0 Standard-PWM on output 1 and 2
22 Complementary PWM
44 Push-Pull PWM
66 Phase-shifted PWM with fixed 50% duty-cycle
Page 45 of 61
Overview of possible output-signals:
Pulse-Name Picture
Single-Pulse
Double-Pulse
Synchronous Rectifier pulse
SoftSwitching pulse
PWM with n pulses
Standard PWM
Complementary PWM
Push-Pull PWM
Phase-shifted PWM with fixed

50% duty-cycle
Page 46 of 61
5.4. Connection of peripheral devices
5.4.1. Gate-controller
The signals are output via two BNC jacks. The primary and secondary outputs
are marked on the chassis. The housing contact of the BNC sockets is grounded. A
galvanic separation, which may be necessary, must be carried out externally. A successful
pulse output is indicated by the short flashing of the LED marked “Pulse”.
5.4.2. Thermo-couples
The connection of up to 6 thermo-couples is done via connectors on top of the

device. The controller supports only J-type-thermo-elements. The polarity is printed on-
top of the case and has to be followed.
5.4.3. OMRON E5CC
The temperature controller “Omron E5CC-CX3A5M-004” is connected via the

marked connection. The pin assignment is as shown in the figure below. The temperature
controller must be configured for communication via ModBus with a baud rate of 38.400.
Only one controller can be connected at the same time. The data lines must be termi-
nated as close as possible to the temperature controller with a 120Ω / 0.5W resistor.
Please also follow the connection instructions of the specified temperature-controller
Page 47 of 61
5.4.4. FuG voltage-source
The digitally controllable voltage source is connected via a standard RS-232 cable
whose slot is marked on the housing. Only one device can be connected at the same time.
The pin-assignment of the Sub-D-connector is as follows:
Pin Assignment Pin Assignment

1 B 6 n/c
2 A 7 n/c
3 Y 8 n/c
4 Z 9 n/c
5 GND Case GND
A and B are the differential input while Y and Z are the differential output of the FuG
device. So A/B have to be connected to the Y/Z pins of the the FuG while Y/Z have to
be connected to the A/B of the FuG to allow the communication.
The FuG device has the following pin-assignment:

1 Z (Tx-) 6 n/c
2 Y (Tx+) 7 120 Ohm Resistor
3 A (Rx+) 8 120 Ohm Resistor
4 B (Rx-) 9 n/c
5 GND
5.4.5. ThermoStream
The ThermoStream-device will be connected by a standard RS-232-cable as shown

on the case of the pulse-generator. Only one device can be connected at the same time.
The pin-assignment of the Sub-D-connector is as follows:

1 n/c 6 n/c
2 RxD 7 n/c
3 TxD 8 n/c
4 n/c 9 n/c
5 GND Case GND
Page 48 of 61
The ThermoStream device has the following pin-assignment:

1 n/c 6 n/c
2 RxD 7 n/c
3 TxD 8 n/c
4 n/c 9 n/c
5 GND
A “null-modem”-cable has to be used to connect the Pins as follows:
Sub-D-KDEE_MfPG Sub-D-ThermoStream
2 3
3 2
5 5
Additionally, the following settings have to be applied within the ThermoStream:
• Baudrate to 38400  DIP1 to ON and DIP2 to OFF

• Address-Switch to “STANDARD”  DIP4 to OFF
• Enable digital communication  DIP5 to ON
5.4.6. Interlock
Interlock switches are connected via the “Interlock” connector. The two slots are con-
nected in series. To operate connected voltage sources and heating elements in an auto-
mated manner, the connected switches must make an electrical contact between their
connection terminals. If only one safety switch is used, the second slot must be short-
circuited with a plug-in bridge.
As soon as one of the two interlock switches disconnects the contact, the voltage source
and the heaters are switched off if they are connected via their control lines and are
configured for remote control properly. The user is responsible for checking the safety
shutdown before commissioning.
The discharge of existing capacitances and storages is not caused by the controller and
must be carried out via a separate circuit.
Page 49 of 61
6. Additional content
6.1. ComCell configuration file
The hardware address-setup is done within the “ComCell.ini” configuration-file in

the program folder using a text editor like notepad++. Within this “ComCell.ini” you
can define several options concerning the oscilloscope (brand, used MatLAB driver, hard-
ware-addresses), as well as the used presets within the oscilloscope for the specific meas-
urements. These preset-numbers are representing the oscilloscope-preset slots within the
device (Tektronix/LeCroy) itself. As on Yokogawa-Oscilloscopes no preset-slots are avail-
able, setup-files will be loaded with the following file-names:
1 “SWITCHINGON” 4 “CONDUCTION”
2 “SWITCHINGOFF” 5 “AUTOZERO”
3 “DIODESWITCHING” x “PRESETx”
Fig. 6.1: ComCell Configuration-file
Page 50 of 61
6.2. Calculating pulse-widths of double-impulse
The pulse-widths of the double-impulse are mainly dependent on the used induct-
ance and the voltage of the intermediate circuit. The voltage-drop over the inductor and
the shunt will influence the accessible current with increasing current. Likewise the de-
creasing voltage over the capacitors of the intermediate circuit during the discharge will
influence the length of the pulse. So ComCell will compensates these parameters and
increases the length of the pulses dynamically.
The compensation of the decreasing voltage over the capacitors is done by calcu-
lating both the (inductor-limited) current drawn out of the capacitors and the remaining
charge. Fig. 6.2 shows the equivalent circuit for the voltage drop over the intermediate
circuit capacitance, while Fig. 6.3 is showing the voltage- and current-curves on dis-
charging the capacitors.
Fig. 6.2: Equivalent circuit for voltage drop over intermediate circuit
Fig. 6.3: Decreasing voltage and inductor current
ComCell is calculating the drawn current, the remaining charge and afterwards
the remaining voltage for the capacitors for several thousand steps. Then the original
Page 51 of 61
pulse-width and desired current is compared to this capacitor-model and a compensated

pulse-width is calculated:
𝑡𝑠𝑡𝑒𝑝 = 100𝑛𝑠
𝑛 = 100000
𝑈𝑅 = (𝑅𝐿 + 𝑅𝑆ℎ𝑢𝑛𝑡 ) ∗ 𝐼𝐿,𝑛−1
−𝑡𝑠𝑡𝑒𝑝
( 𝐿 )
𝑈𝐶 − 𝑈𝑅 𝑅𝐿 +𝑅𝑆ℎ𝑢𝑛𝑡
𝐼𝐿,𝑛 = 𝐼𝐿,𝑛−1 + ( )∗ 1−𝑒
𝑅𝐿 + 𝑅𝑆ℎ𝑢𝑛𝑡
( )
𝐼𝐿,𝑛 − 𝐼𝐿,𝑛−1
𝑄𝐶 = (𝑈𝐶,𝑛−1 ∗ 𝐶𝑍𝐾 ) − ( + 𝐼𝐿,𝑛−1 ) ∗ 𝑡𝑠𝑡𝑒𝑝
2
𝑄𝐶
𝑈𝐶,𝑛 =
𝐶𝑍𝐾
𝑡𝑜𝑛,𝑛 = 𝑡𝑠𝑡𝑒𝑝 ∗ 𝑛
Fig. 6.4: Example for voltage-drop estimation at ton=1.2ms
Fig. 6.4 shows the original calculated pulse-width of 0.897ms. Due to the voltage-
drop the desired current of 200A could not be reached. Instead a current of 174A is
switched-off by the device under test which is too low. The green line shows the corrected
time of 1.07ms to reach the 200A by the device. This value is calculated by searching
the calculated current-vector for the desired current and using the corresponding value
of the time-vector. If the desired current is out of the capability of the intermediate
circuit a time-value of 0µs is returned.
Page 52 of 61
6.3. Calculating conduction-losses
To reduce the measuring-time to a minimum only one single measurement will

be used to calculate the complete UCE(IC)-curve. For this a discrete multiplication of
voltage-curve and current-curve is implemented (see Fig. 6.5). In the setup-box you
can choose the number of discrete multiplication-steps as well as the distribution.
Fig. 6.5: Discrete multiplication of voltage-drop and current
6.4. Calculating switching-losses and dynamic parameters
The switching-loss analyzer is able to calculate the energy during a commutation-

process as well as dynamic properties of the semiconductor itself (dudtmax, didtmax, Voltage
overshoot, etc.).
Fig. 6.6: Calculated dynamic properties of the DUT (here: switch-off)
Page 53 of 61
6.4.1. Switching-loss integration-borders
You can use multiple types of integration-borders. In standard-configuration the

software uses the “enhanced” rules. These are a combination of several limit-definitions
correlating the gate-emitter-voltage, the current and the collector-emitter-voltage to find
the integration-borders.
Fig. 6.7: Used borders for Switching-loss-energy
Fig. 6.8: IEC60747-9 rules for integration-borders for switching-energy1
1
Source: Semikron Application-Note AN1403, 19.08.2014, Dr. Ulrich Nicolai
Page 54 of 61
6.4.2. Recovery-loss integration borders
Again, you can use two types of integration-borders. The standard type uses the
IEC60747-9 rule and uses the zero-crossing of the current as well as the 2%-limit of the
peak-current to find the borders. The other (simplified) type uses the peak of the
power-curve again. The left border will be set at 0,5% of the peak-power. The right in-
tegration border is set at the first zero crossing to the right of the peak-power (see Fig.
6.10).
Fig. 6.9: IEC60747-9 rule for integration-borders for reverse-recovery-energy2
6.4.3. Dynamic parameters of the switching-process
All dynamic parameters are calculated by using the definition of dynamic param-
eters shown in Fig. 6.10. Turn-on- and turn-off-delays are calculated using the 90%-
/10%-limits of the gate-emitter-voltage- and the collector-current-peak. Fall- and rise-
times of voltage and current are calculated using the 90%-/10%-limits of the peak-values,
too. Fig. 6.10 shows the definition for both an IGBT and a diode.
2
Source: Semikron Application-Note AN1403, 19.08.2014, Dr. Ulrich Nicolai
Page 55 of 61
90%
UGE
10%
t
90% tf,i
IC 90% IF
10% 10% trr
t 90%
td,off tf,i
td,on tr,i 10% 10%
90% Uovershoot UR t
UCE 90% Irr 10%
10% 10%
t 90%
tf,u tr,u
tr,u
Eoff
Eon Erec
Psw Prec
3% 2% 3% 2% 0,5% 0%
t t
t1 t2 t3 t4 t1 t2
(a) (b)
Fig. 6.10: Used definition of switching-losses (simplified 2%/3% type) and

dynamic parameters for IGBT (a) and Diode (b)
6.4.4. Post-Processing Deskew
If your individual signals need to be deskewed by some nanoseconds due to prob-

lems during recording you can apply a post-processing deskew-time for each channel. On
the lower left side of each graph you will find two input-fields, where you may enter a
time in nanoseconds. Use a negative value to shift the curve to the left and a positive
value to shift the curve to the right side. Finally you have to click “Edit” in the menu
and choose the option “Save postprocessing deskew time” to apply the changes. From
then on this deskew time will no only be applied on this current step but for the whole
measurement as the same measurement hardware has been used.
Fig. 6.11: Applying post-processing deskew time
Page 56 of 61
6.5. Available options to start ComCell and the analyzers
6.5.1. General starting options
The ComCell toolbox can be started within MatLAB using the m-files, or without
MatLAB by using the precompiled 32-Bit Windows-executables:
• ComCell.exe
• ConductionLossAnalyzer.exe
• SwitchingLossAnalyzer.exe
ComCell.exe does not require parameters and can be started directly by double-
clicking on ComCell.exe. The program can be used the same way as the MatLAB-version.
6.5.2. Starting conduction-loss analyzer
When starting ConductionLossAnalyzer.exe without any parameters, a folder-di-

alog will be displayed to open an ID-folder containing some measurement-data to display.
To open a specific measurement in the analyzer, use the batch-files StartConductionLos-
sAnalyzer.bat with the following parameters:
StartConductionLossAnalyzer.bat DUTS_DUTD ID UGON TEMPERATURE UDC MEASUREMENT RUN
• DUTS_DUTD = Switch and Diode

• ID = ID of measurement
• UGON = gate-voltage
• TEMPERATURE = Temperature of measurement
• UDC = DC-link voltage of measurement
• MEASUREMENT = Type of measurement (SwitchingOn, SwitchingOff, DiodeSwitch-
ing or Conduction)
• RUN = StepNumber
Example using the batch-file:
StartConductionLossAnalyzer.bat IKW40N120H3_IDT16S60C
id_a35f8be4_4b8b_4c71_bfeb_bd7a776d3fef 15 25 130 Conduction 1
Page 57 of 61
This opens the conduction-losses of the IKW40N120H3 at Ug=15V, TJ=25°C and

UDC=130V
Example using the 32-Bit program directly:
(Take care of the used quote signs!)
ConductionLossAnalyzer.m “’C:\ComCell\’ ‘IKW40N120H3_IDT16S60C’

‘id_a35f8be4_4b8b_4c71_bfeb_bd7a776d3fef’ ‘15V’ ‘25°C’ ‘130V’ ‘Conduction’ ‘1’”
Example using the m-file within MatLAB:
(Take care of the used quote signs and brackets!)
ConductionLossAnalyzer([’C:\ComCell\’ ‘IKW40N120H3_IDT16S60C’
‘id_a35f8be4_4b8b_4c71_bfeb_bd7a776d3fef’ ‘15V’ ‘25°C’ ‘130V’ ‘Conduction’ ‘1’])
6.5.3. Starting switching-loss-analyzer
When starting SwitchingLossAnalyzer.exe without any parameters, a folder-dialog

will be displayed to open an ID-folder containing some measurement-data to display. To
open a specific measurement in the analyzer, use the batch-files StartSwitchingLossAna-
lyzer.bat with the following parameters.
StartSwitchingLossAnalyzer.bat DUTS_DUTD ID RGON_RGOFF UGON_UGOFF TEMPERATURE UDC

MEASUREMENT RUN
• DUTS_DUTD = Switch and Diode

• ID = ID of measurement
• RGON_RGOFF = gate-resistors for switching-on and –off
• UGON_UGOFF = gate-voltages for switching-on and -off
• TEMPERATURE = Temperature of measurement
• UDC = DC-link voltage of measurement
Page 58 of 61
• MEASUREMENT = Type of measurement (SwitchingOn, SwitchingOff, DiodeSwitch-

ing or Conduction)
• RUN = StepNumber
Example using the batch-file:
StartSwitchingLossAnalyzer.bat IKW40N120H3_IDT16S60C
id_51eb42b4_8d57_498c_a95a_26395fe26d07 4_4 15_-10 75 300 SwitchingOn 5
This opens the switching-on-losses of the IKW40N120H3 at Ug=+15V/-10V, TJ=75°C

and UDC=300V
Example using the 32-Bit program directly:
(Take care of the used quote signs!)
SwitchingLossAnalyzer.exe “’C:\ComCell\’ ‘IKW40N120H3_IDT16S60C’

‘id_a35f8be4_4b8b_4c71_bfeb_bd7a776d3fef’ ‘4_4Ohm’ ‘15_-10V’ ‘75°C’ ‘300V’
‘SwitchingOn’ ‘5’”
Example using the m-file within MatLAB:
(Take care of the used quote signs and brackets!)
SwitchingLossAnalyzer([’C:\ComCell\’ ‘IKW40N120H3_IDT16S60C’
‘id_a35f8be4_4b8b_4c71_bfeb_bd7a776d3fef’ ‘4_4Ohm’ ‘15_-10V’ ‘75°C’ ‘300V’
‘SwitchingOn’ ‘5’])
6.6. Voltage- and Temperature control
Within the program ComCell.m the DC-Link-voltage and device-temperature can

be controlled automatically. For this a two-level-controller with timer has been imple-
mented. The general function is as shown in the following figure:
Page 59 of 61
Fig. 6.1: Two-Level-Controller for voltage- and temperature-control
6.7. Important Code lines of main-program
Name Line Description

ComCell_OpeningFcn 5836 Called on startup
Contains FuG Correction tables
Figure1_CloseRequest- 6101 Called on exit
Fcn
RunTest 918 Contains main-measurement-loop
RunTest: 1667 Here the scaling-measurements are done. Os-
CurrentScalingMeasure- cilloscope scalings can be changed here
ment
RunTest: 1933 Here the scope-channels are scaled to the op-
Oscilloscope Scalings timal scale. The timebase is set for conduc-
tion-loss, too.
RunTest: 1828, Here the double-impulse is started
FireDoublePulse 2454
RunTest: 1414, Here the DC-Link voltage is tested for +/-
TestDCLinkVoltage 1563 5% tolerance. The sequence pauses until the
voltage is with this tolerance. Afterwards the
voltage is tested to be within a +/-15% tol-
erance.
Page 60 of 61
KDEE
Kompetenzzentrum für dezentrale elektrische Energieversorgungstechnik
Ansprechpartner:
Dr.-Ing. Christian Nöding | 0561/804-6166 | [email protected]
Universität Kassel / KDEE-EVS

Prof. Dr.-Ing. habil. Peter Zacharias
Wilhelmshöher Allee 71, 34121 Kassel Digital signiert von Christian
Noeding
http://www.kdee.uni-kassel.de
AN: C=DE, S=Hessen, L=Kassel,
O=Universitaet Kassel, OU=EECS,
OU=Universitaet Kassel / FB16 /
KDEE-EVS, CN=Christian Noeding
© KDEE 2011-2017 Grund: Ich bin der Verfasser dieses
Dokuments
Ort: Kassel
Datum: 16-10-2017 13:19:58
Page 61 of 61

ComCell Manual v3.5

Uploaded by

Copyright:

Available Formats

ComCell Manual v3.5

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ComCell Manual v3.5

Uploaded by

Copyright:

Available Formats

Manual

ComCell and ComCell Analyzer for MatLAB

2. General description and first steps .......................................... 4

4. ComCell Analyzing tools ...................................................... 36

4.6. Export to KDEE TopBench .............................................................................42

6. Additional content ............................................................... 50

The following hardware can be used with the ComCell-software:

2. General description and first steps

The NI-DAQmx-, NI-MAX-software and the NI-libraries have to be installed us-

If you received and want to use the standalone-version of the ComCell-programs

2.1.1. Setup of NI-Board

2.1.2. Setup of KDEE_v2 and KDEE_MfPG devices

2.1.3. Communication-setup of Tektronix TDS7104 oscilloscope

The Tektronix-MatLAB-Device-Driver “tektronix_tds7104.mdd” has to be copied

Within the configuration-file “ComCell.ini” you have to set the communication-

When using a direct-USB-connection, please enter the USB-address of the scope

2.1.4. Communication-setup of LeCroy-oscilloscope

The LeCroy-MatLAB-Device-Driver “lecroy_basic_driver.mdd” has to be copied

Within the configuration-file “ComCell.ini” the IP-address and TCP-port to the

When using a direct-USB-connection, please enter the USB-address of the scope

Afterwards within the LeCroy-oscilloscope the option “SONSTIGES  [x] Enable

2.1.5. Communication-setup of Yokogawa-oscilloscope

First the Yokogawa-MatLAB-ToolKit-software “DL-MatLAB” has to be installed.

Within the configuration-file “ComCell.ini” you have to set the communication-

When using a direct-USB-connection, please enter the serial-number of the scope

2.1.6. Setup oscilloscope-presets for switching- and conduction-measurements

For every single measurement-type a preconfigured template/preset has to be

The timebase depends on the type of switch/diode to be measured. So it is rec-

2.1.6.1. Preset for switching-on

Fig. 2.1: Triggering switching-on-events

2.1.6.2. Preset for switching-off

Triggering switching-off events only use a single-event triggering, as the desired

Fig. 2.2: Triggering switching-off-events

2.1.6.3. Preset for diode-switching

The reverse-recovery-current can be measured using an A/B-triggering again. Fig.

Fig. 2.3: Triggering diode-switching-events

2.1.6.4. Preset for conduction

• If the measured current is inverted, trigger A is set to rising edge, to trigger at

Fig. 2.4: Triggering conduction-event

2.1.6.5. Preset for diode-conduction

• If the measured current is inverted, trigger A is set to rising edge, to trigger at

Fig. 2.5: Triggering diode-conduction-event

2.1.6.6. Presets on Yokogawa-oscilloscopes

As the Yokogawa-oscilloscopes uses filenames instead of numbered preset-slots,

Presetnumber Filename on Yoko- Description

2.1.7. Calibrating the FuG-voltage-source

Important: the behavior of the calibration-values within the file “ValueCorrec-

The FuG-device has to be calibrated to compensate non-linear-effects of the spe-

Calibrate measured values from FuG-device (Input):

• Within the configuration-file “ComCell.ini” set the tag “DCSourceDevice” to

Calibrate set-values to FuG-device (Output):

• Within the configuration-file “ComCell.ini” set the tag “DCSourceDevice” to

The amount of sampling-points can be changed, but have to be at least 2 or 3 to allow

2.1.8. Calibrating ThermoCouples

2.2. Starting the application and data-structure

Start the ComCell program by calling “ComCell.m“ out of the MatLAB-command

Equipment- Equipment- Equipment-

Test- Test- Test-

Conduction- Switching- Conduction- Switching- Conduction- Switching-

Conduction- Switching- Conduction- Switching-