Table of Contents
Validation of SPICE modeling approach
Properties of Inductor
Detailed simulation with Altair Activate SPICE with experimental verification
Model Based Development for system simulation
Summary
Background
As mention in part I of this blog, i would like to do a modeling and simulation approach for this experiment with inductor.
In part 1 i have mentioned a few variants of modeling (0D model, 1D ODE model, modelica and SPICE modeling). In part 2, we narrow down to SPICE modeling in details.
Validation
this is to validate the simulator itself. to do this i need to have a validation circuit. One of the most common usage of inductor is in dc-dc conversion.
I have chosen to use the LM2596 based buck converter, where an inductor is a vital component in the design.
The Tab table below details my test setup in tab 1, SPICE simulation in Altair Activate in tab 2 and validation simulation LTSPICE in tab 3
| {tabbedtable} Tab Label | Tab Content |
|---|---|
| Test Circuit | LM2596 Buck Converter There is a LM2596 based module readily available on the market. My module has an inductor with 47 μH. Take note that this is not the inductor from the inductor kit, and is the original inductor that comes in the module. subsequently in other experiments after this section, i hack the module and change it to the inductor from the inductor kit. The is capable of driving a 3.0A load with excellent line and load regulation. This device is available in adjustable output version and it is internally compensated to minimize the number of external components to simplify the power supply design. It has terminal block connector for easy to connecting of power wires. Specifications: Input voltage: 4.5-35V Output voltage: 1.5-35V(Adjustable) Output current: Rated current is 2A, maximum 3A(Additional heat sink is required) Conversion Efficiency: Up to 92% (higher the output voltage, the higher the efficiency) Short Circuit Protection: Current limiting, since the recovery Load Regulation: ± 0.5% Voltage Regulation: ± 2.5% i have decided to model this module and hack it for my inductor experiment
my specification is input voltage: 12 V output voltage: 7 V the reason i choose 7 V is so that i can use it to power a motor and solenoid actuator that requires at least 6V, and use another 5V voltage regulator to power up the MCU that is used for controller only the actuator circuit is to be simulated here
I have two version of these module, one of which with 3 digit voltage display.
Figure Typical circuit of LM2596-Adj
note: the pin 5 on/off allows the switching regulator circuit to be shut down using logic signals thus dropping the total input supply current to approximately 80 μA. Pulling this pin below a threshold voltage of approximately 1.3 V turns the regulator on, and pulling this pin above 1.3 V (up to a maximum of 25 V) shuts the regulator down. If this shutdown feature is not required, the ON/OFF pin can be wired to the ground pin or it can be left open. In this module, it it wired to ground hence it is always on
to achieve 7V output, first turn CCW the variable resistor all the way until voltage start dropping. i use a power supply to supply 12V to it. after that i power down and i measure the resistor of the variable resistor. for the module with display, i get around 0.9kΩ, whereas for the module without display i get 1.5kΩ
after this, i try to use SPICE simulation to get to the same value
Experimental MeasurementPicolog and a picoscope 5444D is used to capture the data
which can capture data in table form as well as plot graph
|
| Hyperspice in Altair Activate | Altair Activate Hyperspice solver A simplified SPICE modeling is performed by using schematic editor on spice in Altair. Note: an alternative way is to use SPICE netlist directly as i explained in section SPICE simulation later on
and the sample output is
|
| LTSpice | LTSpice A freeware SPICE III simulator by Linear Technology, now part of ADI. While it is popular and fast, a limitation is it only contains converter manufactured by itself. If one use a different manufacturer, need to use other brand specific simulator. For example, WEBENCH
enlarged view
|
Detailed Analysis with experimental verification
In this experiment we want to see the effect of different inductor to the buck converter using LM2596
Experiment BOM
LM2596 buck converter module with hacked 22μH inductor SBCP-14HY220B
LM2596 buck converter module with 47μH inductor (brand unsure)
test load device #1: DC motor with 0.02A current ratings at 7V
test load device #2 : DC motor with 2.7A current ratings at 7V
DC power supply and banana clips
multimeter
test load #1 and test load 2
Seen in the photo below is the choice of KEMET inductor that i want to replace with the 47μH inductor on the LM2596 module.
however i do face a confused dilemma, when i can't find the exact model in website. as you can see, the inductor looks different. However, after click into the datasheet
http://www.farnell.com/datasheets/2736814.pdf , i find that the reference is correct on element14 site, but the photo of the device is wrong.
Specification of SBCP-14HY220B
Photo below shows the replacement of the original 47μH with KEMET SBCP-14HY220B 22μH. Picoscope is used to measure the output voltage
This is the output as captured by Picolog
enlarging the spike, one can see that it reach 7.44V
Background of experiment
According to LM2596's datasheet, choice of inductor is given in the table below
Based on what i have in the inductor kit https://www.element14.com/community/docs/DOC-93857/l/experimenting-with-inductors-about-the-competition?ICID=experiment-…
i have chosen to hack the module with a 22μH inductor. This means, at my chosen operating voltage of 7V, it is able to drive current in the range less than 1.5A. Anything more
it is not gonna work. In other work, it will work for the test load device #1 but fail for test load device #2
Source: page 19 of https://www.ti.com/lit/ds/symlink/lm2596.pdf
Experimental Verification
| Setup | Load #1 | Load #2 |
|---|---|---|
LM2596 buck converter module with hacked 22μH inductor from KEMET SBCP-14HY220B |
Here the photo shows that it is capable of running the load #1 as its current is low 0.01A measured |
the video shows that it is incapable of running the load #2 at 2.4A its measured voltage drop from 7V to 5V
this is due to the inductor is not able to handle the current requirement due to its capacity |
LM2596 buck converter module with original 47μH | ||
| summary | As expected, both buck converter can power up the small test load device as it is only consuming 20mA | only the buck converter with 47μH can power up the dc load that needs 2.4A for the hacked buck converter with 22μH, the saturation current is much lower at less than 1.5A, hence the circuit cannot continue to source the current necessary to drive the load. As a result, voltage is also dropped. |
SPICE simulation
the scenario above can be simulated with SPICE in activate by editing the SPICE netlist
In this model below, there is a superblock which contains the SPICE netlist
we can then edit the netlist to have the parameter we want, for example, here is netlist representing our operating voltage on 12V input, 7V output, with 47μH inductor
BUCK_BASIC.CIR - BASIC BUCK CONVERTER
*
* SWITCH DRIVER
VCTRL 10 0 PULSE(0V 5V 0 0.01US 0.01US 5US 20US)
R10 10 0 1MEG
*
* INPUT VOLTAGE
VIN 1 0 DC 12
*
* CONVERTER
SW1 1 2 10 0 SW
D1 0 2 DSCH
L1 2 3 47UH
C1 3 0 25UF
*
* LOAD
RL 3 0 5
*
*
.MODEL SW VSWITCH(VON=5V VOFF=0V RON=0.01 ROFF=1MEG)
.MODEL DSCH D( IS=0.0002 RS=0.05 CJO=5e-10 )
*
* ANALYSIS
.TRAN 1US 800US
*.TRAN 0.1US 840US 800US 0.1US
*
* VIEW RESULTS
.PLOT TRAN V(2) V(3)
.PROBE
.END
Output Ripple Filter
Another application of inductor is as filter. In fact, for the same LM2596, we can apply output ripple filter.
This can be referred to https://www.ti.com/lit/ds/symlink/lm2596.pdf , page
Properties of Inductor
I attempt to learn the inductor from basics and there are two great youtube videos provided by KEMET
In-depth with inductors with KEMET - Part 1
In-depth with inductors with KEMET - Part 2
from the manufacturer perspective, i learn that KEMET is small size due to usage of metal composite instead of ferrite material. However, there are trade offs and is summarized as below.
rated current
this is defined by the inductor self heating
saturation current
this is defined by how much the current drop due to inductor core material type
source: In-depth with inductors with KEMET - Part 1
in short, we need to take note of the inductance drop at rated current, which is mentioned in datasheet. 20-30% drop in inductance is common. This explains the reason why replacing an inductor with
similar inductance but different manufacturer to a circuit like switch mode power supply (SMPS) might has different behaviour owing to this difference of saturation current.
it is important to know the saturation point as we may not have sufficient energy stored when it is running at higher current for application like SMPS, resulting in drop in accuracy due to ripple
current. for SMPS the ripple current should be as low as possible.
when inductance drop, it results in higher ripple current
this heating effect cause dc couple loss, ac couple and core loss to the inductor. Based on my renewed and basic understanding, among the major reason why inductance decrease
with temperature is due to increase of skin effect (current tends to travel at outer surface of conductor, the higher the temperature the more it does) decreasing current flow.
AS per the datasheet page 18, exceeding an inductor's maximum current rating may cause the inductor to overheat because of the copper wire losses, or the core may saturate. If the inductor begins to saturate, the inductance decreases rapidly and the inductor begins to look mainly resistive (the DC resistance of the winding). This can cause the switch current to rise very rapidly and force the switch into a cycle-by-cycle current limit, thus reducing the DC output load current. This can also result in overheating of the inductor or the LM2596. Different inductor types have different saturation characteristics, so consider this when selecting an inductor.
In In-depth with inductors with KEMET - Part 2, there is an experiment to measure the saturation current.
my curiosity arises as in whether i can simulate such behaviour --> after some studying, i not yet successful in doing simulation of this
Model Based Development for system design
Model Based Development Introduction
In model based development (MBD), computer model is developed to simulate the system behaviour prior to producing the prototype for testing. This methodology allows designer to have insight into
the behaviour of the system and optimize its operation before prototyping. This allows the overal saving of time and cost.
below shows a possible scenario of simulating a door access system with Altair Activate at system level, illustrated using the V-model of MBD. In MBD capable software such as the Altair CAE range of
software that i am using, it is capable of simulating component and subsystem in different physics domain, or what commonly known as multiphysics. In the diagram below, we have power electronics and
the mechanical latch which is modeled in 3D with finite element method (FEM).
the power electronics portion can be simulated by SPICE or modelica model representing the power electronics can be modelled and interface to the 3D FEM model or its representative dynamic 1D plant model in Altair Activate.
The input voltage and output measurement can be simulated and its design parameter optimized.
System Simulation
In this blog, we perform system simulation of our buck converter powering up the test load device
In hyper spice
In modelica
simulation output with 22μH inductor, showing voltage output and speed for dc motor
Summary
I have enjoyed going through this experiment. It has enriched my knowledge of inductor.
what's shortcomings is that i find that i don't know how to do the more advance SPICE simulation but can only do simple ones
but with this new knowledge, i am eager to do more simulation driven design going forward
References
https://sciencedemonstrations.fas.harvard.edu/presentations/audio-filters
KEMET engineering center https://ec.kemet.com/
https://en.wikipedia.org/wiki/Model-based_design
part 1 of this blog Inductor Modeling & Simulation - Experimenting with Inductors - Blog 1 of 2
