Transcript
Summary
Study and Improvement of a Light Dimmer Study setting
INSA of Strasbourg in Electrical Engineering with a major in System Student’s name and first name
JACQUIER Thibaud Date of Internship
09 February – 09 August 2015 Company
Hager Controls SAS 33 rue Saint Nicolas 67700 Saverne - FRANCE Abstract My internship (end of studies’ project) was about the study and improvement of a light dimmer. It was a study on the internal structure of the light dimmer sold by HAGER. It was designed on Matlab/Simulink software and its several toolboxes (Simscape, SimElectronics). My goal was to create a reliable and complete model of a phase cutting light dimmer that consists of a MOSFETbased power stage. The aim of that modeling was to optimize the functioning of the light dimmer and also to improve its sturdiness (reliability, heating reaction …) and to speed up the development (by avoiding dichotomy). During my internship I have studied the functioning and the various blocks involved in a light dimmer. For example, I have modeled the power stage and its command and also the voltage and the current synchronization. The results of this modeling are reliable and were compared to Pspice modeling or directly to an existing product. Tutors names: Dr. Guy STURTZER (INSA tutor) ................................. Teaching Assistant in Electrical Engineering, INSA Strasbourg Dr. Eric STURTZER (company tutor) ................................................ Electronic Development Engineer, HAGER Saverne
Keywords – Light Dimmer; Phase Cut; Modeling; Matlab/Simulink; Voltage Synchronization; I. INTRODUCTION Since a few years there is a development of load especially with the arrival of LEDs on the market. That is the reason why the light dimmer has to anticipate these changes. At the beginning light dimmer structure was based on TRIAC but this technology does not allow dimming all kind of load. That is why HAGER chooses a MOSFET structure. With this more “universal” architecture, the control of all kinds of loads (halogen, CFL, LED, etc.) is possible. The aim of this study is to build a relevant and complete model of the company light dimmer. This model should be able to simulate thermal constraint on electronics component (heterogeneous simulation). The final target is to optimize and reduce development time by simulating the comportment of the light dimmer. This resume will first present the global principle of a light dimmer: the phase cut. Then the dimming brick is detailed with the presentation of the light dimmer synoptic. As the light dimmer is introduced, it is now possible to present the Matlab/Simuling modeling results with the detail of the Voltage Synchronization function. The conclusion will summarize the results and open the study to perspectives. II. WORKING PRINCIPLE: PHASE CUTTINGS The principle used in the different HAGER light dimmer is the phase cutting1. The load is supplied during a fraction of time of mains. It exists also others principle to dim a luminous load. For example the dimming principle used wide dimming range electronic ballast is the frequency variation (cf [3]). The TRIAC even it is becoming an obsolete component in the light dimming they are able to dim other kind of load different than incandescent load. For example the two wires electronic ballast presented in [4] [5] introduce the complexity to dim new generation of load like LED or CFL by using TRIAC structure. It exist two cuttings mode according to the nature of the load connected to the dimmer output. These two modes are the “trailing edge” mode and the “leading edge” mode. These two modes are illustrated in Figure 1. A. Trailing Edge The Trailing Edge mode consists to supply the load from the beginning of the mains half period to a certain fraction of it. We use this conduction mode for capacitive loads. By definition these loads are sensitive to brutal voltage variation that is why we have to shut the controlled switch (MOS) when the mains voltage pass by zero. B. Leading Edge The Leading Edge mode consists to supply the load from a certain angle to the end of the mains half period. We use this conduction mode for inductive loads. By definition these loads are sensitive to brutal current variation that is why we have to shut the controlled switch (MOS) when the mains current pass by zero.
V
Vsupply Vload t
(a)
V
t
(b) Figure 1: Mains and Load Voltage versus Time in Trailing Edge (a) and Leading Edge (b)
III. DIMMING BRICK The HAGER light dimmers are proposed in the company catalog since several years and are fully operational (mature in its life cycle). The well-known problematic of flickering2 is already mastered. The light dimmer structure is mainly based on the knowledge and the experience of its creator. The dichotomy principle is applied during the selection of component value. In this section will be present the functional synoptic implemented in HAGER light dimmer. After that, each function on this synoptic will be briefly explained. We called that synoptic the “dimming brick”. As we can see in the Figure 2 there is a PIC microcontroller that controls the MOSFET pair in output. This microcontroller is also here to regulate the light dimmer and make the date processing from different sensors. Now, each function of the synoptic will be presented: A. The Power stage is made of two MOSFET. These Nchannel MOSFET have a Drain Gate voltage of 600V with an RdsON resistance of 140 mΩ. These MOSFET will conduct each an alternation of the mains sinusoid in synchronism with the command applied on their common gate. B. The Product protection materialized by the two varistors between Phase (P) and Neutral (N) and between Phase (P) and output load (ɛ) are here to protect the circuit against mains overvoltage but also against a wrong wiring of the user. The one between the Phase (P) and the output is here to prevent overvoltage on the product output. C. The Power supply is based on a Flyback architecture which is going to feed in DC voltage the whole embedded electronic in the light dimmer. There is two different level of voltage 12V and 3,3V. D. The Voltage Synchronization is an electronic circuit which enable an accurate detection of the voltage zero crossing. Like this, we will be able to cut capacitive load accurately at zeros voltage.
1
The phase cutting is a little bit more detail in [1] – [2] but with a TRIAC structure.
Vsupply Vload
2
As presented by C.-S. Wang in [6].
SUPPLY
P FUSE T
CAPTURE
I SYNCHRO
CAPTURE
MOS DRIVER
ON/OFF SLOPE
S/C
I/O
OVERVOLTAGE
COMP IRQ
OVERLOAD
I/O
OVERHEAT
RELAY DRIVER
E
« Standard »
XOR
PB
+ -
2
IRQ ou COMP. IRQ
U SYNCHRO
CURRENT MEASUREMENT VOLTAGE MEASUREMENT
KNX
PFO
POWER
ɛ
PB
MICRO
VCC VDD
N
+ -
ADC I/O ADC I/O ADC I/O
I/O I/O
CONNECTION POINT / DATA PROCESSING
XOR
Rx UART Tx UART
Figure 2:Light dimmer synoptic
E. The Current Synchronization is an electronic circuit which enable an accurate detection of the current zero crossing. Like this, we will be able to cut inductive load accurately at zeros current. F. The MOS command enables a voltage adaptation between microcontroller output and the power stage. It exist a lot of literature concerning command aspect like in [7] who is mainly about the command necessary to control a light dimmer (The difference in this publication is they are using TRIAC). This electronic circuit has also a slope parameter to adjust it when the MOS commute. Nevertheless this slope setting has a direct impact on CEM results especially on radiated emission. That is the reason why we add a capacitance in parallel of the output load (ɛ) and the Phase (P). That way we can reduce this CEM impact. G. The Short circuit is a bloc implemented to protect the light dimmer in case of short circuit on the output load (ɛ). (Typically if the filament brakes in a light bulb). H. Thanks to the Overvoltage circuit we can detect if the network voltage goes repeatedly above a threshold. If this is the case then we detect an overvoltage. That one is generally due to temporary phenomenon (e.g. bad conduction). I. The Overload function embedded in HAGER product to control the output power. If this power is above a critical threshold then we place the product in a fault mode. J. The Overheat is an electronic function based on a resistor evolving in function of the temperature. Like the overload if we are above a critical threshold we put the product into a fault mode. K. Thanks to the Current Measurement circuit we can continuously measure the current flowing into the load. This measure is taken near the MOSFET thanks to two shunt resistor.
L. The Voltage measurement is taken by pulling down the load voltage into a range measurable by the microcontroller. Like this we can continuously control the voltage in the load. M. The Relay driver is implemented in the connected dimmer (KNX bus). It allows evacuating leakage current flowing into the load when there is no order given by the KNX bus. This cutting by block will allow us to model of an easiest way (block by block). My model will be also more understandable by others people which do not use the dimming brick. The next section is going to introduce this modeling and how I designed my model. IV. LIGHT DIMMER MODEL Later in my project once the various blocks were models in Simulink I had to think about a way to architecture my model. For doing that, I started from the previous synoptic (Figure 3) to build my model. The Figure 3 presents the global structure which I chose to develop for my model. We will distinguish three levels of description: The level N is a black box vision of the product, that is to say like we ignore what the light dimmer is made of. This vision corresponds to the dotted line on the Figure 3. The level N-1 is made for someone who has some knowledge on the light dimmer internal structure. It is the functional level. The level N-2 has been developed for a specialist of light dimming that is to say for someone who got solid basis on the topics. This is the expert level.
Level N Level N-1 Level N-2
Figure 3: Model architecture
V. MODELLING OF THE VOLTAGE SYNCHRONIZATION The Voltage Synchronization (also named voltage zero crossing detection) is used to synchronized the command compared to the mains zero, and filter price scale signals. This block of voltage zero crossing detection is very important in our light dimmer because it allows for example to avoid overcurrent when you command a capacitive load.
U Synchro
Figure 4: Voltage Synchronization model
As we can see on the Figure 4 phase and neutral are logically placed in input of this bloc because we want to detect the changing of alternation of the mains. R23 and R24 are two resistors of high value (called drop down resistor) around MΩ to have a voltage drop important enough. That way the voltage level is measurable by the electronic. Z1 diode is a Zener (Simulink gives the same graphical symbol for all diode but the comportment of this one is the right one that is to say a Zener comportment). This Zener is here in order to limit C8 charge at its Zener voltage. On the positive alternation of the mains, D1 diode has for role to rectify the power supply made of C8, D3 and R23 in serial with R24. On the negative alternation this diode is here in order to force C8 capacitor to discharge in the collector of T9 transistor. The C8 capacitor is called a “tank” capacitor; this one is going to charge on the positive alternation and slowly discharge in T9 transistor during the negative one. R27 is
simply a pull up resistor which is going to adjust the high level threshold of the signal in output of the optocoupler. R26 has two functions in the circuit. The first one is to set the low level threshold voltage of the signal. The second one is to make a low pass filter of the first order with C10 capacitor. On the Figure 5 we can check the model is fully functional because when you change from a positive alternation to a negative there is a falling edge on the output voltage synchronization. To be totally sure my model was working I made a Pspice Simulation and a measure. To take this measurement I have used a MSOX3014A oscilloscope, a differential probe and a classic probe to measure the voltage synchronization signal (in green on Figure 5 on the oscilloscope screenshot). As expected the Pspice model and the product measurement give me the same signal shape and the same behavior.
(a) Simulink Model 400V
VI. CONCLUSION My objectives were to build a model for our light dimmer and to take in consideration several physical domains (like thermal for example) which impact our product. I have completed the major part of my objectives by realizing a model who took in account thermal impact. I had also build a model but I did not have the time to fully end this model. So for the future someone can take my work and continue to build the model. This way he will be complete. VII. REFERENCE [1]
J. B. Smith, J. Speakes and M. H. Rashid, "An Overview of the Modern Light Dimmer: Design, Operation, and Application," in Power Symposium, 2005. Proceedings of the 37th Annual North American, 23-25 Oct. 2005.
[2]
J. C. W. Lam, J. C. Y. Hui and P. K. Jain, "A Dimmable High Power Factor Single-Switch Electronic Ballast for Compact Fluorescent Lamps With Incandescent Phase-Cut Dimmers," IEEE Transactions on Industrial Electronics, vol. 59, no. 4, pp. 1879 - 1888, Avril 2012.
[3]
S. Y. R. Hui, L. M. Lee, H. S.-H. Chung and Y. K. Ho, "An Electronic Ballast with Wide Dimming Range, High PF, and Low EMI," IEEE Transactions on Power Electronics, vol. 16, no. 4, pp. 465 - 472, Juillet 2001.
[4]
S. T. S. Lee, H. S.-H. Chung and S. Y. R. Hui, "TRIAC Dimmable Ballast With Power Equalization," IEEE Transactions on Power Electronics, vol. 20, no. 6, pp. 1441 - 1449, Nov 2005.
[5]
A. Tjokrorahardjo, "Simple Triac Dimmable Compact Fluorescent Lamp Ballast and Light Emitting Diode Driver," in Applied Power Electronics Conference and Exposition (APEC), Palm Springs, CA, USA, Février 2010.
[6]
C.-S. Wang, "Flicker-Insensitive Light Dimmer for Incandescent Lamps," IEEE Transactions on Industrial Electronics, vol. 55, no. 2, pp. 767 - 772, Février 2008.
[7]
H.-M. Jung, J.-H. Kim, B.-k. Lee and D.-W. Yoo, "A new PWM dimmer using two active switches for AC LED lamp," in International Power Electronics Conference (IPEC), Sapporo, JAP, 21-24 Juin 2010.
200V
0V -200V -400VV(V1:+) 3.0V 2.0V 1.0V SEL>>
0V0s V(C4:2) 10ms 20ms 30ms 40ms 50ms 60ms 70ms 80ms 90ms100ms Time
(b) Spice Model
(c) Measure Result Figure 5: Voltage synchronization model: Simulink Model (a) Spice Model Leading (b) Product Measurement(c)
