PHILIPS AN10971 Tl applications with nxp ballast controller Datasheet

AN10971
TL applications with NXP ballast controllers
Rev. 1 — 15 September 2011
Application note
Document information
Info
Content
Keywords
Half-bridge, electronic ballast, tube light
Abstract
This application note describes how to design an application to drive a
fluorescent lamp with a half-bridge circuit combined with an NXP
Semiconductors ballast controller.
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
Revision history
Rev
Date
Description
v.1
20110915
first draft
Contact information
For more information, please visit: http://www.nxp.com
For sales office addresses, please send an email to: [email protected]
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
2 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
1. Introduction
This application note describes the use of an NXP Semiconductors half-bridge driver IC
for High Frequency (HF) Tube Light (TL) single tube, fixed DC bus voltage applications.
2. Scope
This application note is organized as follows:
•
•
•
•
•
•
Section 3 describes the basic operation of a half-bridge ballast for fluorescent tubes.
Section 4 describes how to select MOSFETs for the half-bridge.
Section 5 describes how to select a resonant coil and capacitor.
Section 6 describes how to design magnetic components.
Section 7 describes the feedback control loop used in the dimmable controllers.
Section 8 describes system performance.
3. Lamp characteristics and half-bridge principles
3.1 Introduction
Electronic ballasts must preheat, ignite, control and monitor the condition of the TL
according to the TL specification. In cases of no ignition or lamp end of life, the ballast
must shut down to avoid damage to the following:
• the ballast
• or the overheating of the lamp electrodes
Depending on the lamp type, tube ignition occurs at different voltage levels. For example,
a Compact Fluorescent Lamp (CFL) ignition occurs at around 500 V and up to 1200 V for
a TL. Following ignition a lamp is in its operating state. During operation lamp voltage is
dependent on the shape and content of the tube. For TL lamps, voltages range from 80 V
to 220 V (RMS).
A fluorescent lamp has a so called "negative incremental impedance". The more current
flowing through a fluorescent lamp the lower the voltage (unlike a resistor). Driving a
fluorescent lamp with a Constant Voltage (CV) would result in an unstable system. In
addition, a breakdown of the supply and/or tube within a few milliseconds also occurs.
To stabilize the current through the lamp, a series impedance is needed typically a
resistor. However, in practice a coil is used to reduce losses as shown in Figure 1. In
magnetic 50 Hz to 60 Hz Low Frequency (LF) systems, see Figure 2 coils are large.
However, in HF TL 40 kHz to 80 kHz systems coils are smaller as the impedance of a coil
increases linearly with the frequency.
In steady state, at frequencies higher than 10 kHz a fluorescent lamp can be considered
to be a resistor. At high frequencies Ohms law V = I  R is applicable to fluorescent tubes.
The resistance (R) varies with the power supplied to the lamp. For example, a 36 W T8
lamp with an operating voltage of 100 V can be considered as a 277  resistor. This
condition is true when operated at its nominal power.
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
3 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
+
lamp
ILa
+ VLa
Vm
dne
>0
dt
120
110
dne
<0
dt
100
100
200
VLa
300
+ VLa
VAB
Vi
starting voltage Vi
VR
dne
>0
dt
dne
<0
dt
lamp potential (V)
lamp potential (V)
200
lamp
+ VR -
-
-
VAB
ILa
VLa
400 500
I (mA)
iss
I (A)
(a)
(b)
a. Discharge potential drop versus
current
019aaa652
b. Effect of series resistance in
stabilizing lamp current
(1) dne = delta in the number of free electrons in th plasma
Fig 1.
Lamp characteristics
+ Ib
+
B
-
+
Vi
-
La
Vm
0
Fig 2.
S
Il
-
019aaa654
Magnetic 50/60 Hz TL ballast
3.2 Lamp filaments
Each lamp has two filaments (or electrodes), each consisting of a coated tungsten wire as
shown in Figure 3. The resistance of the tungsten wire is directly related to its
temperature. Figure 4 identifies the relationship between electrode temperature and the
ratio of Resistance hot (Rh) and Resistance cold (Rc) of a typical tungsten wire.
filament
length D
019aaa657
Fig 3.
AN10971
Application note
Lamp filament
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
4 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
019aaa658
12
ratio
Rh/Rc
8
4
0
0
500
1000
1500
2000
T (°K)
Fig 4.
Rh/Rc ratio versus temperature for 100 % wolfram filaments
Lamp electrodes are preheated before the lamp is ignited or lamp life time is considerably
reduced. During preheat, electrode resistance increases typically by a factor of between
four to six. To generate sufficient and an even spread of heat before ignition, preheat time
must not be lower than 0.5 s. Preheat times longer than 1.5 s to 1.7 s are considered
undesirable as heat is lost to the gasses in the lamp tube.
Figure 5 identifies the operating area of fixed preheat current for a T8 36 W burner with
two current levels. Figure 5 also shows that if the current is too high the optimal ignition
temperature is reached before 0.5 s. As a result, either ignition takes place too early or the
filament is overheated before ignition.
019aaa661
705.5 mA
Rhc
6.25
operating
area
4.25
467.0 mA
1.0
0
0
0.5
1.0
1.5
2.0
t (s)
Fig 5.
AN10971
Application note
Fixed preheat current window
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
5 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
3.3 HF TL half-bridge principles
High frequency TL applications typically run at 40 kHz to 60 kHz instead of 50 Hz to 60 Hz
for magnetic ballast applications. In HF TL applications, the impedance of the coil is linear
with its frequency. The size of a coil for an HF TL application is only a fraction of the coil
required for 50 Hz to 60 Hz magnetic ballast applications.
Figure 6 (left) shows a theoretical 50 kHz ballast circuit with a 50 kHz sine wave
generator. Figure 6 (right) shows a real 50 kHz ballast circuit with a DC blocking capacitor
and square wave generator.
A 50 kHz sine wave is not easy to produce, therefore a 50 kHz square wave generator is
used as shown in Figure 6 (right). The higher harmonics of the square wave hardly pass
the coil whereas the lower harmonics, specifically the first harmonic does pass the coil.
The square wave generator in combination with the coil in practice acts as a stable sine
wave source. A square wave of 50 % duty factor and an amplitude between 0 V and
Vbridge has a DC component of Vbridge / 2. This condition causes an unlimited DC current
passing through the lamp. A DC blocking capacitor is used to filter out the DC voltage
component to stop the DC current. The voltage on the DC blocking capacitor is Vbridge / 2.
The resonance frequency of the LC combination is used to ignite the lamp. Following
ignition, the amplitude and frequency of the square wave together with the L and C values
determine the lamp power.
The RMS value of the first harmonic of a square wave is:
2
V bridge   ------- 
 
Therefore, a square wave of 400 V corresponds with a sine wave of approximately
180 V (RMS). The higher harmonics (3,5,7) only contribute < 5 % to lamp current and can
be ignored.
resonance L
2 mH
resonance L
1
2 mH
2
A1 A2
+
-
2
A1 A2
resonance
capacitor
8.2 nF
TL T8
36 W
1
resonance
capacitor
8.2 nF
TL T8
36 W
+
50 kHz
180 V
B1 B2
3
−
4
50 kHz
400 V
B1 B2
3
4
DC blocking
capacitor
68 nF
019aaa662
Fig 6.
AN10971
Application note
019aaa663
Half-bridge circuit overview
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
6 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
Ilamp
50 kHz
S1
Vi, II
+
CF = 1.6
II
Cd
Vi
Lr
CB
lamp
S2
-
Ilamp
60 Hz
PF = 0.4 to 0.6
Cr
10 ms
019aaa664
a. Typical HF TL half-bridge ballast using rectified mains
S1
Vi, Il
+
Il
Cd
Vi
-
60 Hz
PF = 1
PFC
Ilamp
Lr
CB
S2
lamp
Ilamp
Cr
CF = 1.4
019aaa665
b. Typical HF TL half-bridge ballast using a PFC
Fig 7.
Typical HF TL half-bridge ballast circuits
Figure 7 represents a typical HF TL half-bridge ballast circuit, so called as only two
switches (S1 and S2) are used. Full-bridge designs, use four switches. Coil impedance, Lr
as shown in Figure 7 limits the current in the circuit and can be calculated using
Equation 1:
Z = 2    frequency  L coil
(1)
As the impedance is linear to the frequency, a 1000 times higher frequency requires a
1000 times smaller coil value for the same impedance. As a result, a high frequency
ballast is much smaller and lighter than a 50 Hz to 60 Hz magnetic ballast.
Switches S1 and S2 generate a square wave in the 40 kHz to 100 kHz range. The
average voltage of a square wave is equal to half its amplitude. Capacitor (Cd) is a DC
blocking capacitor, without it a DC current would flow through the coil and lamp. Cr and Lr
together have a resonant frequency where the voltage sweeps up to a point where the
lamp can be ignited. Capacitor Cr is placed between the filaments to preheat the filaments
before ignition.
For more detailed information on how to select suitable resonant tank LC values see
Section 5.
The supply feeding the half-bridge can be either rectified mains or a DC voltage generated
by a pre-conditioner circuit. This type of circuits pulls current from the mains resembling
the resistive load (so called Power Factor) of an incandescent bulb. In most countries
higher than 25 W, a mains voltage-current source relation is mandatory.
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
7 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
Power Factor Correction (PFC) circuits are either active (controller) or passive. The main
passive PFC topologies used include, valley fill and charge pump. This application note
assumes either rectified mains as shown in Figure 7 [A] or an active PFC as shown in
Figure 7 [B].
If the square wave is a fixed frequency, and the voltage on the electrolytic capacitor
fluctuates as in Figure 7 [A] the lamp current also fluctuates. The so-called crest factor of
the lamp current is the peak/RMS value. To maximize the lifetime of the lamp, the crest
factor must remain lower than 1.7.
3.4 Start-up, ignition and operation sequence
V
fstart
(1)
(2)
fB
preheat state
ignition
state
burn state
time (s)
019aaa666
(1) Frequency versus time
(2) Lamp voltage versus time
Fig 8.
Start-up sequence
Figure 8 shows an example of the start-up sequence for the UBA2021 controller which is
similar to the start-up sequence of other NXP Semiconductors controllers.

At the start of the sequence, the controller starts at its maximum frequency (fstart) typically
100 kHz. The frequency then drops quickly to the preheat frequency of around
70 kHz to 80 kHz. The controller then waits for the preheat time to pass (typically
0.8 s to 1.5 s) then drops to a frequency of around 40 kHz (fmin).
The ignition (LC tank resonance) frequency of between 45 kHz to 55 kHz must occur
between the end of preheat period and the minimum frequency of fmin. See Figure 8.
Following ignition, that is, during the burn state the frequency of the half-bridge
determines the power of the lamp. Figure 8 line (2) shows the voltage over the lamp.
3.5 Zero voltage switching of a half-bridge (capacitive mode)
MOSFET switches S1 and S2 need a precisely timed control voltage. The NXP
Semiconductors controller generates this control voltage as shown in Figure 9.
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
8 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
Vi
L1
RRHV
490 kΩ
DS1
DS2
G1
S1
C3
lamp
RHV
CI
13
14
2
CCI
100 nF
S1
R1
8
3
CCP
CP
100 nF
100 nF
L2
mains
supply
FS
CS7
C2
Cboot
C5
UBA2021
12
CCF
CF
100 pF
G2
S2
1
6
10
RREF
RRREF
30 kΩ
DS7
DS3
DS4
VS
C4
CS4
DS6
CS9
5
9 11
7
PGND
SGND
RS
Rshunt
mgs994
Fig 9.
S1 and S2 driven by UBA2021
When the lamp is ignited, the phase shift of the Inductor Resistor Capacitor (LRC) L2, C5
and lamp combination above its resonance frequency is such that the current lags behind.
The voltage current phase shift of an inductor is similar. Figure 10 shows the coil current
amplitude of LRC resonant tank versus frequency while Figure 11 shows the phase
relationship.
Amplitude A
40 kHz
100 kHz
019aaa667
Fig 10. Current amplitude of LRC resonant tank versus frequency
Phase
+90
−90
40 kHz
100 kHz
019aaa668
Fig 11. Current phase of LRC resonant tank versus frequency
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
9 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
The behavior of the load is called inductive mode when the current lags behind on the
voltage. For capacitive mode, the voltage lags behind on the current. These conditions are
similar to the behavior of a capacitor and inductor. Figure 12 and Figure 13 show the
phase relationship of voltage and current flowing through an inductor and a capacitor.
amplitude
V
I
t (s)
019aaa669
+90 °
Fig 12. Current voltage relation of inductor
amplitude
V
I
t (s)
-90 °
019aaa670
Fig 13. Current voltage relation of capacitor
If the current lags behind the voltage (inductive mode), the current charges and
discharges the parasitic capacitance of the half-bridge when both MOSFETs are off. This
state is known as the non-overlap time. The inductor current charges the parasitic
capacitance for the low to high transition. The current flows through the bulk diodes of the
high-side MOSFET until the MOSFET is switched on. For high to low transitions the
opposite is true.
The result is that there are no switching losses in the MOSFETs since the MOSFETs are
opened/closed when there is no voltage over them. This state is known as Zero Voltage
Switching (ZVS), it is the way a half-bridge resonant tank normally operate.
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
10 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
HS gate
LS gate
HS/LS fet
LS diode
Iind
inductive
HS/LS fet
HS diode
half bridge
voltage
normal
inductive mode
019aaa671
Fig 14. Switch timing sequence (inductive mode)
If the resonant tank is run lower than its resonant frequency, then it is in capacitive mode.
In capacitive mode, the MOSFETs remain closed while large voltages exist across them.
This results in large current spikes in the order of 10 A to 15 A. These spikes cause lower
efficiency and can damage the MOSFETs. This condition is commonly known as "hard
switching".
If the bulk diode in one MOSFET is conducting, current while the other MOSFET is
opened the MOSFETs can be damaged within a single stroke. In capacitive mode, there is
always full bus voltage hard switching. However, there can also be an in-between state
with limited hard switching (of a voltage less than the bus voltage). All NXP
Semiconductors controllers react to capacitive mode, albeit in a different way.
HS gate
LS gate
HS/LS fet
LS diode
Iind
capacitive
HS/LS fet
HS diode
half bridge
voltage
capacitive mode
hard switching
019aaa672
Fig 15. Switch timing sequence (capacitive mode)
The UBA2014 reacts to capacitive mode by jumping to its maximum frequency. The
UBA2021 reacts by quickly increasing the frequency until there is no more hard switching.
The UBA2015(A) and UBA2016A react by shutting down as capacitive mode is detected.
3.5.1 None overlap time, fixed versus adaptive
The time period when both MOSFETs are open (conducting) is called the non-overlap
time. The non-overlap time is much larger than necessary to avoid cross conduction,
when for example both MOSFETs are switched on. The time is set such that the resonant
tank coil current has charged/discharged the capacitance on the half-bridge during the
none overlap time. All extra current after this none overlap time, flow through the bulk
diode of the MOSFET that is switched on after the none overlap time ends.
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
11 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
The UBA2014 has an input that senses when the slope of the half-bridge square wave
voltage has ended. The UBA2014 is able to adapt its timing such that it switches on the
MOSFETs after the slope has ended. This condition minimizes the current through the
bulk diodes of the MOSFETs and reduces losses. This principle is called "adaptive none
overlap".
The UBA2015(A), UBA2016A and UBA2021 all have a fixed none overlap time. For a
fixed non-overlap, the charge pump supply capacitor (CS7 in Figure 9) must be fine-tuned
to eliminate any hard switching.
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
12 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
4. Half-bridge MOSFET selection
4.1 Introduction
The half-bridge MOSFETs are both N type switched on/off at precisely the correct time.
The current with which the gates are charged and discharged is also important for
ElectroMagnetic Compatibility (EMC). The main parameters effecting MOSFET selection
for use with half-bridge applications include:
• Breakdown voltage
• Ron resistance - On Resistance
• Low gate charge
The breakdown voltage of a MOSFET is a critical parameter, therefore selection must be
within precise design limits. The Ron resistance value can be selected within limits
depending on the target efficiency of the ballast. The low gate charge keeps the capacitor
on the dV/dt supply small.
4.1.1 MOSFET voltage rating
Voltage overshoots in a half-bridge resonant tank always exist due to parasitics.
Therefore, select a voltage rating of the MOSFETs that is 50 V higher than the bus voltage
of the half-bridge.
4.1.2 MOSFET RON
The RMS current through the MOSFETs and the desired efficiency of the ballast
determines the Ron of the MOSFET. Ron resistance depends on its operating temperature,
for TL a typical value is 90 C and 105 C for CFL. The RMS current through the
MOSFETs is the vector sum of the burner and resonant capacitor current.
As the lamp current is known that is, controlled or determined by a fixed frequency the
resonant capacitor current is equal to Equation 2.
2    f run  V lamp
(2)
As the lamp and capacitor current have a 90 phase shift, the coil current is equal to
Equation 3.
I coil =
 I lamp 2 + I capacitor 2 
(3)
Losses in each of the MOSFETs do occur, the losses can be determined using
Equation 4. Typical temperature values inside a TL ballast ranges from 80 C to 100 C
(100C to 120 C for CFL). Therefore, select the Ron of the MOSFETs so that no
overheating occurs. In addition, always consider the MOSFETs packaging during
selection. Generally, lower Ron values improve ballast performance at increased cost of
the ballast. Always maintain losses in the MOSFETs to less than 1.5 % of the lamp power.
½I
2
coil
 R on
(4)
Example:
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
13 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
Ballast for a TLD 36 W lamp running at 40 kHz, where:
•
•
•
•
LC = 1.9 mH, 8.2 nF
Ilamp = 0.32 A (RMS)
Icapacitor = 0.20 A (RMS)
Icoil = 0.38 A (RMS)
A pair of MOSFETs each with a Ron of 3  when at operating temperature have a power
loss of approximately 0.5 W (0.25 W per MOSFET).
4.2 Boot strap diode and capacitor
As both MOSFETs are N type, a voltage of approximately 12 V higher than bridge voltage
(Vbridge) is required to drive the upper MOSFET. In addition, an external bootstrap
capacitor is required to store the 12 V buffer charge. Figure 16 shows an example of an
NXP Semiconductors half-bridge ballast controller IC with integrated bootstrap diodes that
provide such a voltage. External capacitor C5 is used to store the buffer charge. Typically,
a 100 nF, 50 V capacitor is sufficient for all similar applications.
F1
1A
9 FVDD
BOOTSTRAP
VDD 7
SUPPLY
DRIVER
CONTROL
10 GH
HIGH SIDE
DRIVER
LOW SIDE
DRIVER
C5
100 nF
11 SH
6 GL
019aaa673
Fig 16. Boot strap
4.3 MOSFET gate drive RLC circuit
The gate drive circuit consists of a Resistor Inductor Capacitor (RLC) series circuit as
shown in Figure 17. Resistance (R) is derived from the built-in resistance of the controller
IC which is normally sufficiently high enough in most designs. However, an additional
external resistor can be added if necessary. See Section 4.3.1.
Inductance (L) consists of parasitic PCB inductance coupled with the internal inductance
inside the MOSFET. The capacitance (C) is the gate capacitance added with parasitic
PCB.
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
14 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
IC
C1
L
R
C2
019aab528
Fig 17. MOSFET gate drive circuit
For an RLC series circuit, the mathematical condition where no oscillation/undershoot in
step responses exists is as Equation 5. See also Figure 18 line 4.
R 2  C 2 – 4LC > = 0
(5)
019aaa674
600
I(t)
(mA)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
400
200
0
−200
0
4
8
12
16
t (s)
(1) The critical damping plot is the bold red curve. The plots are normalized for L = 1 and C = 1
Fig 18. Under/overdamped responses of an RLC circuit
4.3.1 Extra resistance
As the L and C are known (given the layout), a minimum value of the series resistance (R)
can be calculated. Where extra resistance is needed the extra series resistor (R) must be
low enough for the gate drive not to open itself due to Miller capacitance. Hard switching is
unavoidable at start-up. As an example, the upper MOSFET is switched on it must not
lead to opening of the lower MOSFET due to the Miller capacitance. In such conditions,
cross conduction occurs.
NXP Semiconductors controllers have internal resistors such that in almost all cases no
additional external resistors are needed.
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
15 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
Miller capacitance is known from the MOSFET specification, the inductor is the result of
layout parasitic inductance of the MOSFET. Spice simulation and circuit measurements
are the most effective method to achieve optimal component values. If it is not possible to
find a resistor value that suits both conditions, then a resistor diode combination can be
used. See Figure 19.
6
8
12
13
5
GL
PCS
ACM
33 Ω
4.7 Ω
LVS
GND
Lpar
T2
Lpar
Lpar
RS
1Ω
019aaa675
Fig 19. Extra resistance, lower MOSFET of half-bridge for example
The most important PCB layout parameters for a half-bridge are the loops of the MOSFET
gate drives and the loop through the electrolytic bridge capacitor. The electrolytic bridge
capacitor must be close to the half-bridge MOSFETs rather than close to the PFC
MOSFET diode. Where this layout arrangement is not possible, add a 100 nF capacitor
over the bus PFC voltage close to the half-bridge MOSFETs.
For more layout advice, refer to the SMPS layout and EMC guidelines.
5. Resonant tank
5.1 Introduction
The controller starts at its maximum frequency then quickly drops to the preheat
frequency. The controller then waits for the preheat time to pass and then drops to the
ignition frequency. Following ignition, the frequency of the half-bridge controls the power
to the lamp. For example, dimmable controllers like the UBA2014, UBA2015(A) and
UBA2016A all use a feedback control loop. The UBA2021 operates at a fixed frequency.
See Section 3.4.
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
16 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
5.2 Frequency versus power
019aaa676
105
Vlamp
(V)
trace1
104
103
(3b)
(3a)
(4)
102
trace2
(2)
(5)
(1)
(6)
10
103
104
105
106
f (Hz)
(1) Red line trace 1
(2) Blue line trace 2
Fig 20. Resonant tank frequency versus power
Key points for Figure 20
•
•
•
•
•
•
•
•
1 to 5 = Frequency sweep range
1 = Start
2 = Preheat (yes/no)
3a = Ignition warm
3b = Ignition cold
4 = Burn transition
5 = Burn
6 = Dimming
The resonant coil is the main dominating component of lamp power and the resonant
capacitor an influential factor. The coil, resonant capacitor and lamp form the resonant
tank. The coil and capacitor determine the resonant frequency.
In series resonant topology, as in Figure 1 the first selected component is the resonant
capacitor. The resonant capacitor determines the current through the filaments during run
time. In addition, it also determines the maximum preheat current given a maximum lamp
voltage during preheat.
The more filament current that is needed the higher the resonant capacitor value must be.
As an example, a value of 1.5 nF is practical for a small CFL and up to 10 nF for a TLD/T8
tube. A larger capacitor value also means that enough energy remains in the resonant
tank for a single successful ignition attempt. For further details, see Section 5.3.
The next selected component is the resonant tank inductor (L) as shown in Figure 21. The
lamp power depends on the frequency according to Equation 6.
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
17 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
L
Vhb =
DC block
2
· Vbus
Vlamp
C
019aaa677
(1) See Equation 6
Fig 21. Resonant tank with a burner driven by a square wave voltage
V lamp 2
V bus 2 2
V hb 2
V lamp
P out = ---------------  ------------- -------  –  1 –  2 LC  2 = -------------  -------------  –  1 –  2 LC  2



L
V lamp 
L
V lamp 
(6)
Where: w = 2    f
The same equation can be mathematically converted to determine frequency versus lamp
power as shown in Equation 7.
P%  2
1
1
- +
f % = ------ ------- – 2  ------------
2 LC
CV 2 %
P%  2
1
------- – 2  ------------
LC
CV 2 %
4V I 
 ---------–
1
2
 V % 
– ------------------------L2C2
(7)
Where:
•
•
•
•
•
•
L = Output stage inductor (H)
C = Output stage capacitor (F)
P% = Lamp power at % dimming level (W)
V% = Lamp voltage amplitude at % dimming level (V)
f% = Frequency corresponding to lamp power at % dimming level (Hz)
Vi = Bus voltage (Vbus)
The DC blocking value is not accounted for in Equation 7, as a low blocking value has no
significant influence over the results. More complex formulas that include the DC blocking
capacitor/value (see Figure 21) into account can be found in Ref. 1
Figure 22 shows the power versus frequency graph using a series resonant PFC. In the
example, L = 1.9 mH, C = 8.2 nF and Vhb = 400 V T8 36 W. Dimming at 10 % is reached
before fmax and operation power is reached a few kHz higher than fmin.
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
18 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
019aaa678
60
Plamp
(W)
400
llamp
(mA)
350
300
(2)
(3)
40
250
(1)
200
150
20
100
50
0
40
50
0
70
60
f (Hz)
(1) Plamp
(2) Ilamp
(3) Pfil  1.5
Fig 22. Power versus frequency
Table 1.
Suggested resonant tank values for different lamp types
Lamp
Resonant
coil Lres
(mH)
Resonant
capacitor
Cres (nF)
fph (kHz) Rsense () Vbus (V) Iph (mA)
(RMS)
TLD 36 W
1.9 mH
8.2 nF
70 kHz
1
400 V
600 mA
Vlamp during
preheat (V)
tph (s)
230 V
1.7 s
TLD 58 W
1.4 mH
10 nF
60 kHz
0.8 
400 V
700 mA
170 V
1.7 s
TL5 HE 14 W
3.9 mH
5.6 nF
54 kHz
3.3 
400 V
225 mA
180 V
1.5 s
TL5 HE 21 W
3.7 mH
5.6 nF
54 kHz
3.3 
400 V
225 mA
180 V
1.5 s
TL5 HE 21 W
4.0 mH
3.9 nF
58 kHz
3.3 
400 V
225 mA
230 V
1.5 s
TL5 HE 35 W
4.0 mH
3.9 nF
58 kHz
3.3 
400 V
225 mA
230 V
1.5 s
TL5 HO 39 W
2.0 mH
10 nF
52 kHz
1
400 V
560 mA
240 V
1.5 s
TL5 HO 49 W
2.6 mH
6.8 nF
52 kHz
2.2 
400 V
370 mA
240 V
1.5 s
TL5 HO 54 ‘W
1.5 mH
10 nF
55 kHz
0.75 
400 V
800 mA
330 V
1.5 s
5.3 Ignition and coil saturation
The resonant tank coil and capacitor value determine the preheat, ignition and operation
conditions. A coils saturation current and losses are important parameters besides its
value in Henries. The critical point for saturation is during lamp ignition. The reason is, the
voltage over the lamp is at its peak, as is the current through the coil/resonant capacitor.
The required ignition voltage together with the chosen resonant capacitor and ignition
frequency determine the level of current the coil must handle before saturation occurs.
Coil saturation causes high currents that damage the ballast. The UBA2015(A) and
UBA2016A have direct coil saturation detection that increase the frequency and starts a
fault timer in case of saturation. The UBA2021 increases its frequency and the UBA2014
jumps to its maximum frequency due to the capacitive mode that is a direct consequence
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
19 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
of coil saturation. A ballast based on UBA2014 and UBA2021 can be extended with a
small OverPower Protection (OPP) circuit. to bring them to standby in case of coil
saturation.
The ignition frequency is the frequency where the resonant tank reaches the ignition
voltage. Ignition frequency can be calculated using Equation 8.
1
F ign = --------------------------------------------------------------------------------------------------------- L res  C res  C DCblock  
2 ---------------------------------------------------------------------------------------------V bridge  2 


 C res + C DCblock   1 + -----------------------
   V ign  
(8)
Where:
• Vbridge is the PFC output voltage
• Lres and Cres is the resonance coil/capacitor
• CDCblock is the DC blocking capacitor
The current through the coil as the resonant tank reaches the ignition voltage can be
calculated using Equation 9:
I coil  max  = V ign  2    F ign  C res
(9)
The saturation value of Lres must be higher than the coil current during ignition. In
addition, in cases of a restart/power dip the coil is still warm which can reduce its
saturation current depending on the magnetic material used.
Under test, run a test with an aged burner as aged burners tend to have a higher ignition
voltage than new ones. In addition, a cold burn has an increased ignition voltage. Finally,
remember that the incoming photons accelerate tube ignition. Therefore the best stress
test is an aged burner, shielded from any environment light straight from the freezer (leave
24 hours in between tests). The presence of a grounded metal plate close to the tube
lowers the ignition voltage. Therefore, testing must take place under near to final design
conditions.
No ignition protection
To protect the ballast TL controllers, the UBA2014, UBA2015(A) and UBA2016A all have
lamp over voltage protection circuits. Lamp over voltage protection circuits limit the
frequency ramp down during the ignition stage if the lamp voltage exceeds a level set by
the external components. Figure 23 shows an example circuit using the UBA2014. The
circuit prevents a high voltage over the lamp avoiding damage to the resonant capacitor,
coil or other components. Following a time-out of about 100 ms the controller switches to
standby.
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
20 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
C24
C8
330 pF
LAMP
LVS
VOLTAGE 13
SENSOR
100 nF
R20
220 kΩ
C19
56 nF
R5
10 kΩ
D4
BYG20E
C17
6.8 nF
C22
8.2 nF
lamp
C23
100 nF
C20
68 nF
R18
180 kΩ
019aab527
Fig 23. Lamp over voltage protection circuit
5.4 Preheat and run time filament current
The amount of preheat current versus time can be found in the burner specification as
shown in Table 2.
Table 2.
Preheat current versus time values
Lamp type
PL-C 4-pin
10 W
13 W
18 W
26 W
RMS Iph Preheat
current (mA)
tph
Rsub ()
(0.5 s)
(1.0 s)
(1.5 s)
(2.0 s)
min (RMS)
295
230
205
190
30 
max (RMS)
390
305
270
250
30 
min (RMS)
300
240
215
200
30 
max (RMS)
395
315
280
265
30 
min (RMS)
375
300
270
255
18 
max (RMS)
490
395
355
335
18 
min (RMS)
560
450
405
380
9
max (RMS)
735
590
535
500
9
The substitution resistor is 3.1 × cold filament resistance value during lamp operation. In
addition, the burner specification gives maximum voltage during preheat without the risk
of so called glow or early ignition. Glow is a discharge current without ignition which
damages the filaments. When no burner specification exists, then as a guideline the
filament resistance at the end of preheat must be 4.75 times the cold (room temperature)
resistance.
The NXP Semiconductors controllers that have preheat, the UBA2021, UBA2014,
UBA2015(A) and UBA2016A all have an input for a current sense resistor. Current sense
resistors regulate the frequency so that the desired preheat current is obtained. The level
of the lamp voltage under these conditions depends on the resonant tank LC components.
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
21 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
During lamp operation, the current through the filament must be within limits as specified
in the burner data sheet.
There are two kinds of specifications:
• The lead wire current versus discharge current.
• Sum of Squares (SoS) specification
5.4.1 Lead wire current versus discharge current
Figure 24 is an example of a specification of lead wire current versus discharge current.
All currents are RMS values.
019aaa679
280
ILH
(mA)
240
200
160
TARGET setting
120
80
40
0
0
40
80
120
160
200
240
280
ID (mA)
Fig 24. Example ILH versus ID burner specification
ILH
ID
ILL
019aaa680
(1) ILH = lead-high (total) current
(2) ID = lamp (discharge) current
(3) ILL = lead-low (heating) current
Fig 25. Definition of lamp currents
5.4.2 SoS specification
The specification can also be in the form of the so-called SoS of the ILH and ILL currents.
SoS is the more recent and more accurate way of specifying filament currents. The
equation for calculating the SoS value is shown in Equation 10.
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
22 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
019aaa910
0.4
SoS
(A2)
maximum settin
0.3
g
2 × 12
LLmax.
target setting
0.2
minim
um s
etting
0.1
0
0
0.1
0.2
0.3
Ilamp (A)
Fig 26. Graphical representation of the SoS lines of TL5 HO 24 W and 39 W
2
2
(10)
SoS = I LH + I LL
6. Magnetic component design
6.1 Introduction
6.2 Inductor design parameters
An important factor of good ballast design is the quality of the main inductor. To achieve a
high efficiency solution the inductance value, saturation current, proximity, core and ohmic
losses, parasitic capacitance and stray magnetic fields are all important. Not
understanding the functionality and implementing non-optimized components results in
either, inferior performance or an impractical design. The following provides detailed
guidelines for performance and design optimization.
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
23 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
014aaa940
400
3F3
PV
(kW/m3)
f
(kHz)
^
B
(mT)
200
100
400
50
25
100
200
100
300
200
100
0
0
40
80
120
T (°C)
Fig 27. 3F3 specific power loss for several temperature/flux density combinations
For core material, each manufacturer has another code. For applications between 40 kHz
and 100 kHz, 3F3 (Ferroxcube), N87 (EPCOS) or TP4 (TDG), are recommended. Select
the material that has the lowest loss at working temperature. See Table 3.
Table 3.
Ferrite core comparative geometry considerations
Aspect
Pot
core/RM
core
Double
slab core
E core
Ec, ETD
cores
PQ core
EP core
Toroid
Core costs
high
high
low
medium
high
medium
low
Bobbin costs
low
low
low
medium
high
high
none
Winding costs
low
low
low
low
low
low
high
good
excellent
excellent
good
good
fair
Winding flexibility good
Assembly
simple
simple
simple
medium
simple
simple
none
Mounting
flexibility
good
good
good
fair
fair
good
poor
Heat dissipation
poor
good
excellent
good
good
poor
good
Shielding
excellent
good
poor
poor
fair
excellent good
6.3 Core type selection
Core geometry depends on several factors, for example, cost, flexibility, shielding and
utilization factors. A core can have an inner core that results in a round or square winding
shape. Stray inductance can vary with core shape. The maximum stored energy in the
inductor together with the required air gap determines the core size.
A core with a large air gap can store more energy than a core with a small air gap. In
practice, for resonant converters, an optimum design is reached when the core losses and
the winding losses (proximity and skin losses) are balanced. Therefore it is necessary to
compromise between high storable energy levels, low leakage inductance and small
tolerances on the inductance. The maximum energy stored in the inductor can be
calculated using Equation 11:
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
24 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
2
1
E = ---  L  I p
2
(11)
Example:
• L = 357 H
• Ip = 1.48 A
• E = 3.9  104 J
The E series cores are the most commonly used for TL applications. CFL applications
sometimes use different shapes for a mechanical fit. Table 4 and Figure 28 lists and
shows some of the core types.
Table 4.
Core selector
Core type
Typical size T5 HE 35 W
E
E20
EE
EE30
EER
EER25.5
EI
EI25
EPC
EPC27
ETD
ETD19
(1)
(2)
(4)
(3)
(5)
019aab529
(1) EI core
(2) EE core
(3) EER core
(4) ETD core
(5) EPC core
Fig 28. Core types
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
25 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
6.4 Calculate windings
Al is often specified in the data sheet of the core material. It relates to the inductive value
of a single turn on the selected core. Using this figure and knowing the inductance the
number of inductor windings can be calculated using Equation 12:
NL =
L
----AI
(12)
Example:
• AI= 630 nH
• L= 357 H
• NL = 24
A practical value for NL can be obtained by rounding the calculated value to its nearest
integer. As a check, the magnetic material determines the maximum magnetic B-field. The
peak value of B-field reached during operation has a substantial impact on core losses.
Generally, the max B-field in the magnetic material must remain lower than the specified
saturation field Bsat of the material. The B-field can be calculated using Equation 13:
N L3  I p
B max = U e  -------------------Ie
(13)
Example: refer to Equation 14:
•
•
•
•
NL3 = 24
Ip = 1.48
ue = 342
le = 35.6
1.48
B max = 342  24  ---------- = 338mT
35.6
(14)
6.5 Auxiliary winding count
CAUTION
Check that good isolation exists between the primary and secondary windings. As
during ignition, a voltage difference of up to 1.5 KV can occur.
The auxiliary winding can be used for the following three purposes:
• To heat the filaments
• To detect lamp failure
• Generate the required voltage to power the controller
Equation 15 and Equation 16 apply:
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
26 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
L sec
V sec = -------------------------------L prim  V prim
(15)
2
L sec
N sec
----------- = --------------2
L prim
N prim
(16)
6.6 Select wire diameters
Wire diameter selection is a trade-off between available winding area, ohmic, proximity
and skin losses. The use of wire gauge sizes smaller than 1 mm in diameter at operating
frequencies lower than 100 kHz results in negligible skin losses. Wire gauge sizes larger
than 1 mm in diameter, it is recommended that litz wire or multiple strands are used.
Skin depth can be calculated using Equation 17.
 =
2
------------------------------------------------2    f eff  u r  u o
(17)
Where:
• uo = 4    10
•  = resistivity = 1.7  108 (copper)
• Ur (copper) = 1
Example:
• At 50 kHz sinusoidal current, using copper, the skin depth is 0.3 mm.
For the coil in a resonant tank ballast only take the first harmonic into account, as the
current is close to a sine wave. For a coil (transformer) with secondary windings (inductive
mode filament heating) a square voltage wave form is present on the windings. The
secondary current (unlike the one from the main coil) therefore has high frequency
components. For the wire type of these secondary windings, consider only the 3rd and 5th
harmonic.
The peak current value is a dominant parameter when calculating the ohmic losses.
Estimate the value by calculating the wire resistance and calculating the average power
dissipation in the wire. Generally, the current density must be between 300 circular mills
and 500 circular mills per ampere.
Table 5 shows wire sizes relative to current.
Table 5.
Diameter
(mm)
AN10971
Application note
Wire selection table
Nearest AWG Area (mm2)
Area (circular DC res. (/M) Typical
mills)
current level
(A)
0.1
38
0.008
15
2.195
0.04
0.2
32
0.031
62
0.549
0.15
0.25
30
0.049
97
0.351
0.24
0.315
28
0.078
154
0.221
0.38
0.355
27
0.099
195
0.174
0.49
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
27 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
Table 5.
Wire selection table …continued
Diameter
(mm)
Nearest AWG Area (mm2)
Area (circular DC res. (/M) Typical
mills)
current level
(A)
0.4
26
248
0.126
0.137
0.62
0.56
23
0.246
486
0.070
1.22
0.71
21
0.396
781
0.044
1.95
16  0.2
-
0.503
992
0.034
2.48
37  0.2
-
1.162
2294
0.015
5.73
61  0.2
-
1.916
3782
0.009
9.45
During lamp operation, the coil current is at its maximum around the 10 % to 15 %
dimming level of the lamp current. This condition applies to a typical ballast operating at
60 kHz to 70 kHz.
6.7 Half-bridge transformer design for deep dimming
A low capacitive coupling from the primary to the secondary windings is used to reduce
ElectroMagnetic Interference (EMI). In addition, stray current that interfere with the lamp
current measurement are reduced.
As shown in Figure 32 the coil can also be a transformer with two small secondary
windings. Typical values of these secondary windings are 0.5 H for none dimmable
(large capacitor in series) and 25 H for dimmable applications (small capacitor in series).
See Figure 32 for an example schematic.
It is important to maintain low parasitic capacitance for deep dimmable applications as
follows:
• Cstray primary to secondary < 25 pF
• Cstray secondary to secondary < 15 pF
The breakdown voltage across the primary, primary to secondary and secondary to
secondary windings must be higher than the maximum voltage the ignition protection
circuit allows.
6.7.1 Suggested transformer construction
Figure 29 shows an example of suggested transformer construction:
• Winding 6-5 is secondary winding connected to the hot side of the lamp
• Winding 7-8 is secondary winding connected to the cold side of the lamp
• Winding 3-2 is the primary winding; pin 3 connects to switching half-bridge MOSFETs
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
28 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
7
8
5
6
2
3
(a)
5
6
7
8
2
3
(b)
5
6
8
7
gnd
2
3
(c)
(1)
019aaa682
(1) Indicates mechanical start of the winding and electrical phase
Fig 29. Three different half-bridge transformer winding examples
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
29 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
7. Dimmable controller feedback control loops
7.1 Introduction
Dimming ballasts need a feedback control loop to maintain the lamp current stable at the
desired value. None dimmable ballast in principle can run at a fixed frequency. However,
lamp current variations due to lamp temperature and aging can be avoided by using a
controlled system. Figure 30 show an example of a feedback control loop.
Pi
(1)
+
f
f/p
p
LAMP
v
I
V/I
019aaa683
(1) Desired lamp power
Fig 30. Feedback control loop
A desired lamp power is derived from the customer, including:
•
•
•
•
•
0 V to 10 V interface
Dali
Potentiometer
Measurement of environment light
Triac position for CFL ballasts
The UBA2014, UBA2015(A) and UBA2016A all have an input pin with a voltage range of
0 V to 2.5 V to set the desired lamp current. The topology and component values for a
dimmable system can be different than for a non-dimmable.
7.1.1 Resonant tank values for a dimmable system
The frequency to lamp power/current takes place via the resonant tank as described in
Section 5. It is important that the coil is designed with large losses at the deepest dimming
level (highest frequency). A typical dimming curve is from 40 kHz at full power to 75 kHz
at deepest dimming level.
7.1.2 Inductive mode heating
There are two commonly used filament heating topologies for fluorescent ballast
controllers including:
• Classic series resonant, see Figure 31
• Inductive mode, see Figure 32
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
30 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
Lres
1
2
A1
A2
RT1
PTC
Cres1
lamp
TLD36W
B1
B2
3
Cres2
4
CDCblocking
019aaa052
Fig 31. Series resonant heating
T1
25 μH
1.9 mH
33 nF
1
2
A1 A2
U11
TL T8
36 W
8.2 nF
T2
+
-
V1
V (DC)
B1 B2
3
25 μH
over power
protection
1.5 nF
4
33 nF
68 nF
CS-lamp
current sense
0.5 Ω
1.5 nF
10 Ω
019aaa684
Fig 32. Inductive mode heating
The advantage of inductive mode heating is the lamp current can be measured separate
from the current through the resonant capacitor. This advantage allows deeper dimming
than with classic series resonant. The current through the sense resistor in the source of
the lower MOSFET is the coil current which is the lamp current plus the capacitor current.
The deeper the dimming level the higher the resonant capacitor current becomes
compared to the lamp current. For classic series resonant heating, at 15 % of nominal
lamp current the feedback control loop regulates on the resonant capacitor current only.
As a result it leaves the lamp switch-off. Therefore, dimming is limited to around 15 % see
Figure 31.
The inductive mode topology allows dimming down to between 1 % to 2 %. To reach
these deep dimming values, it is important to maintain low parasitic capacitance from
primary to secondary on the coil and on the PCB.
Table 6 contains a list of values for different tubes for inductive mode heating topology.
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
31 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
Table 6.
Suggested resonant tank values for different lamp types
Lamp
Resonant
coil
Resonant
capacitor
Secondary
inductance
Secondary
capacitance
Vbus
Preheat
energy
Preheat time
TLD 18 W
2.14 mH
4.7 nF
11 H
39 nF
400 V
4J
1.2 s
TLD 36 W
1.9 mH
8.2 nF
14 H
33 nF
400 V
4.6 J
1.2 s
TLD 58 W
1.38 mH
8.2 nF
14 H
27 nF
400 V
5J
1.2 s
TLD 70 W
1.3 mH
8.2 nF
17 H
27 nF
400 V
6.5 J
1.2 s
TL5 HO 24 W
2 mH
4.7 nF
10 H
33 nF
400 V
3.2 J
1.2 s
TL5 HO 39 W
1.8 mH
4.7 nF
10 H
27 nF
400 V
3J
1.2 s
TL5 HO 54 W
1.3 mH
4.7 nF
8 H
33 nF
400 V
3.8 J
1.2 s
TL5 HO 49 W
2.7 mH
4.7 nF
20 H
15 nF
450 V
2.8 J
1.2 s
TL5 HO 80 W
1.1 mH
8.2 nF
15 H
22 nF
400 V
5.8 J
1.2 s
TL5 HE 14 W
3.65 mH
4.7 nF
18 H
22 nF
400 V
2.2 J
1.2 s
TL5 HE 21 W
3.56 mH
4.7 nF
18 H
22 nF
400 V
2.2 J
1.2 s
TL5 HE 28 W
3.83 mH
4.7 nF
16 H
22 nF
450 V
2.2 J
1.2 s
TL5 HE 35 W
3.88 mH
4.7 nF
16 H
22 nF
480 V
2.2 J
1.2 s
7.1.3 Loop control speed
Figure 30 shows that the controller is a part of a feedback control loop (for all dimmable
systems, and also for high end non-dimmable).
For deep dimming a fast feedback control loop is needed due to the physics of the lamp. If
the lamp, does not receive power longer than 100 s to 300 s (ionization time constant)
the discharge stops. Reignition is now needed.
NXP Semiconductors controllers only have on pin to set the feedback control loop speed.
For systems with deep dimming two time constants are needed. One for the ignition
sweep and average current setting and one to maintain discharge at low currents levels.
CT
CSW
CF
IREF
VREF
C16
220 nF
R11
1 kΩ
C15
220 nF
C14
100 pF
R9
33 kΩ
1
2
3
4
14
C19
10 nF
C12
10 nF
019aaa685
Fig 33. Fast feedback control loop
Figure 33 shows an example circuit using the UBA2014 with pin CSW configured as a fast
control loop. Resistor R11 and C12 provides a small (a few kHz) and fast control around a
base frequency determined by the voltage on capacitor C15. In Proportional Integral
Differential (PID) controller terms, it is considered an extra pole.
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
32 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
8. System performance
8.1 Introduction
Ballast losses are derived from:
• Losses in the PFC stage (typical efficiency for a designed PFC is around 96 % to
98 %).
• Losses in the switching MOSFETs (1 % to 2 %) see Section 4
• Losses in the resonant coil (1 % to 5 %) see Section 5
• Losses in the resonant capacitor (negligible if a high performance capacitor is chosen,
however between 1 % to 2 % for low-cost capacitors)
• Miscellaneous losses in controller, diodes, sense resistor, and so on, 1 % in total
In total for a ballast optimized for performance the efficiency is around 93 %. A typical
ballast optimized for Bill Of Materials (BOM) cost has an efficiency of approximately 85 %.
The losses in the filaments are typically not considered to be ballast losses as heating the
filaments is an essential part of the lamp operation.
8.2 Inductor losses
The inductor has several loss mechanisms. Calculation of these losses is complex and
not clearly defined how these losses contribute to the overall total inductor losses.
Equation 18 and Equation 19 simply illustrates a number of loss mechanisms within the
inductor.
8.2.1 Ohmic losses
The combination of wire length and thickness causes ohmic losses. The calculation of the
resistance and losses can be calculated with Equation 18 and Equation 19.
1
R DC =   --A
(18)
1 T 2
P DC = ---   I  R DC
T o
1 2
dt = ---  I p  R DC
2
(19)
Example, refer to Equation 20 and Equation 21.
•
•
•
•
AN10971
Application note
Wire length 1 m at diameter 0.56 mm
 (copper) = 1.72  108
A =   R2 = 2.46  107
Ip = 1.48 A
1
- = 70m
R DC =  1.72  10 – 8   -------------------------2.46  10 –7
(20)
1
P DC = ---  1.48 2  70 = 76.5mW
3
(21)
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
33 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
8.2.2 Proximity losses
For proximity losses, the full calculations are outside the scope of this application note.
However, it is important to understand that proximity losses are closely related to the skin
depth and number of windings as shown in Figure 34.
014aaa943
100
5
100
RAC
RDC
40
20
10
4
3
2
10
n = 1 layer
1
0.1
1
10
h /δ
Fig 34. Proximity loss graph
Avoid too many layers of wires with radius that is close to, or below skin-depth. Normally,
the proximity losses are calculated as a factor of the DC wire resistance as shown in the
Equation 22.
R AC = n  R DC
(22)
By maintaining low resistive losses, proximity losses remain minimal.
8.2.3 Core losses
The magnetization curve and frequencies determine the core losses in the magnetic
material. At each converter cycle, the magnetic flux density excites the magnetic field in
the core material. This result, produces a curve that is highly non-linear with the saturation
level and hysteresis. The surface area enclosed by the variation in B-field strength at a
certain frequency determines the losses. A bigger core, a higher B-field and a higher
frequency increase these losses. The core material data sheet shows the loss per unit of
volume at given frequencies. See Figure 35 and Figure 36.
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
34 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
B
μ
H
019aaa862
Fig 35. BH curve magnetic material
014aaa942
1
Power loss
density
100 kHz
(W/cm3)
1 MHz
500 kHz
200 kHz
50 kHz
0.1
20 kHz
0.01
0.01
0.3
0.1
Tesla
(Bmax)
Fig 36. Material loss graph
A simple empirical formula that calculates core loss is called a Steinmetz equation as
shown in Equation 23.
P h = K h  f   B max   V core
(23)
Kh and  depend on the core material. Equation 23 can be improved by including the
harmonics of a square waveform as shown in Equation 24.
P nse = K h   2f    B max     1 –  +  1 –   1 –    V core
(24)
Where:
• d = duty cycle
• Bmax = the peak flux density
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
35 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
• f = the frequency of the fundamental
• Vcore = the volume of the core
We can see that a higher frequency, a higher flux density, a smaller duty-factor and a
bigger volume all increase core losses. A bigger core does not always reduce core losses;
if the B-field is already low. The increase in volume counter acts the lower losses due to
reduce flux density.
Example:
•
•
•
•
•
•
•
d = 50 %
Kh = 0.05
f = 80 (kHz)
 = 1.84
Bmax = 0.1 T
=3
Vcore = 2.4 cm3
8.3 Sense resistor losses
For the sense resistor losses, the coil current can be calculated in the same way as the
MOSFET losses. The lower MOSFET is switched on half the time. Sense resistor losses
can be calculated using Equation 25.
½  I coil 2  R sense
AN10971
Application note
(25)
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
36 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
9. Abbreviations
Table 7.
Abbreviations
Acronym
Description
BOM
Bill of Materials
CFL
Compact Fluorescent Lamp
EMC
ElectroMagnetic Compatibility
MOSFET
Metal-Oxide Semiconductor Field-Effect Transistor
HF
High Frequency
LF
Low Frequency
LRC
Inductor Resistor Capacitor
PFC
Power Factor Correction
PID
Proportional Integral and Differential
RLC
Resistor Inductor Capacitor
RMS
Root Mean Square
RON
On Resistance
SoS
Sum of Squares
SMPS
Switched Mode Power Supply
TL
Tube Light
ZVS
Zero Voltage Switching
10. References
[1]
AN10971
Application note
Lamp Model IEEE Transactions on Power Electronics Volume 20, No 5, Sept
2005. — An Improved Design Procedure for LCC Resonant Filter of Dimmable
Electronic Ballasts for Fluorescent Lamps.
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
37 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
11. Legal information
11.1
Definitions
Draft — The document is a draft version only. The content is still under
internal review and subject to formal approval, which may result in
modifications or additions. NXP Semiconductors does not give any
representations or warranties as to the accuracy or completeness of
information included herein and shall have no liability for the consequences of
use of such information.
11.2
NXP Semiconductors does not accept any liability related to any default,
damage, costs or problem which is based on any weakness or default in the
customer’s applications or products, or the application or use by customer’s
third party customer(s). Customer is responsible for doing all necessary
testing for the customer’s applications and products using NXP
Semiconductors products in order to avoid a default of the applications and
the products or of the application or use by customer’s third party
customer(s). NXP does not accept any liability in this respect.
Export control — This document as well as the item(s) described herein
may be subject to export control regulations. Export might require a prior
authorization from national authorities.
Disclaimers
Limited warranty and liability — Information in this document is believed to
be accurate and reliable. However, NXP Semiconductors does not give any
representations or warranties, expressed or implied, as to the accuracy or
completeness of such information and shall have no liability for the
consequences of use of such information.
In no event shall NXP Semiconductors be liable for any indirect, incidental,
punitive, special or consequential damages (including - without limitation - lost
profits, lost savings, business interruption, costs related to the removal or
replacement of any products or rework charges) whether or not such
damages are based on tort (including negligence), warranty, breach of
contract or any other legal theory.
Notwithstanding any damages that customer might incur for any reason
whatsoever, NXP Semiconductors’ aggregate and cumulative liability towards
customer for the products described herein shall be limited in accordance
with the Terms and conditions of commercial sale of NXP Semiconductors.
Right to make changes — NXP Semiconductors reserves the right to make
changes to information published in this document, including without
limitation specifications and product descriptions, at any time and without
notice. This document supersedes and replaces all information supplied prior
to the publication hereof.
Suitability for use — NXP Semiconductors products are not designed,
authorized or warranted to be suitable for use in life support, life-critical or
safety-critical systems or equipment, nor in applications where failure or
malfunction of an NXP Semiconductors product can reasonably be expected
to result in personal injury, death or severe property or environmental
damage. NXP Semiconductors accepts no liability for inclusion and/or use of
NXP Semiconductors products in such equipment or applications and
therefore such inclusion and/or use is at the customer’s own risk.
Applications — Applications that are described herein for any of these
products are for illustrative purposes only. NXP Semiconductors makes no
representation or warranty that such applications will be suitable for the
specified use without further testing or modification.
Customers are responsible for the design and operation of their applications
and products using NXP Semiconductors products, and NXP Semiconductors
accepts no liability for any assistance with applications or customer product
design. It is customer’s sole responsibility to determine whether the NXP
Semiconductors product is suitable and fit for the customer’s applications and
products planned, as well as for the planned application and use of
customer’s third party customer(s). Customers should provide appropriate
design and operating safeguards to minimize the risks associated with their
applications and products.
Evaluation products — This product is provided on an “as is” and “with all
faults” basis for evaluation purposes only. NXP Semiconductors, its affiliates
and their suppliers expressly disclaim all warranties, whether express, implied
or statutory, including but not limited to the implied warranties of
non-infringement, merchantability and fitness for a particular purpose. The
entire risk as to the quality, or arising out of the use or performance, of this
product remains with customer.
In no event shall NXP Semiconductors, its affiliates or their suppliers be liable
to customer for any special, indirect, consequential, punitive or incidental
damages (including without limitation damages for loss of business, business
interruption, loss of use, loss of data or information, and the like) arising out
the use of or inability to use the product, whether or not based on tort
(including negligence), strict liability, breach of contract, breach of warranty or
any other theory, even if advised of the possibility of such damages.
Notwithstanding any damages that customer might incur for any reason
whatsoever (including without limitation, all damages referenced above and
all direct or general damages), the entire liability of NXP Semiconductors, its
affiliates and their suppliers and customer’s exclusive remedy for all of the
foregoing shall be limited to actual damages incurred by customer based on
reasonable reliance up to the greater of the amount actually paid by customer
for the product or five dollars (US$5.00). The foregoing limitations, exclusions
and disclaimers shall apply to the maximum extent permitted by applicable
law, even if any remedy fails of its essential purpose.
Non-automotive qualified products — Unless this data sheet expressly
states that this specific NXP Semiconductors product is automotive qualified,
the product is not suitable for automotive use. It is neither qualified nor tested
in accordance with automotive testing or application requirements. NXP
Semiconductors accepts no liability for inclusion and/or use of
non-automotive qualified products in automotive equipment or applications.
In the event that customer uses the product for design-in and use in
automotive applications to automotive specifications and standards, customer
(a) shall use the product without NXP Semiconductors’ warranty of the
product for such automotive applications, use and specifications, and (b)
whenever customer uses the product for automotive applications beyond
NXP Semiconductors’ specifications such use shall be solely at customer’s
own risk, and (c) customer fully indemnifies NXP Semiconductors for any
liability, damages or failed product claims resulting from customer design and
use of the product for automotive applications beyond NXP Semiconductors’
standard warranty and NXP Semiconductors’ product specifications.
11.3
Trademarks
Notice: All referenced brands, product names, service names and trademarks
are the property of their respective owners.
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
38 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
12. Tables
Table 1.
Table 2.
Table 3.
Table 4.
Table 5.
Table 6.
Table 7.
Suggested resonant tank values for different 
lamp types . . . . . . . . . . . . . . . . . . . . . . . . . . . .19
Preheat current versus time values . . . . . . . . .21
Ferrite core comparative geometry
considerations . . . . . . . . . . . . . . . . . . . . . . . .24
Core selector . . . . . . . . . . . . . . . . . . . . . . . . . .25
Wire selection table . . . . . . . . . . . . . . . . . . . . .27
Suggested resonant tank values for different 
lamp types . . . . . . . . . . . . . . . . . . . . . . . . . . . .32
Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . .37
continued >>
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
39 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
13. Figures
Fig 1.
Fig 2.
Fig 3.
Fig 4.
Fig 5.
Fig 6.
Fig 7.
Fig 8.
Fig 9.
Fig 10.
Fig 11.
Fig 12.
Fig 13.
Fig 14.
Fig 15.
Fig 16.
Fig 17.
Fig 18.
Fig 19.
Fig 20.
Fig 21.
Fig 22.
Fig 23.
Fig 24.
Fig 25.
Fig 26.
Fig 27.
Fig 28.
Fig 29.
Fig 30.
Fig 31.
Fig 32.
Fig 33.
Fig 34.
Fig 35.
Fig 36.
Lamp characteristics . . . . . . . . . . . . . . . . . . . . . . .4
Magnetic 50/60 Hz TL ballast . . . . . . . . . . . . . . . .4
Lamp filament . . . . . . . . . . . . . . . . . . . . . . . . . . . .4
Rh/Rc ratio versus temperature for 100 % 
wolfram filaments . . . . . . . . . . . . . . . . . . . . . . . . . .5
Fixed preheat current window . . . . . . . . . . . . . . . .5
Half-bridge circuit overview . . . . . . . . . . . . . . . . . .6
Typical HF TL half-bridge ballast circuits . . . . . . . .7
Start-up sequence . . . . . . . . . . . . . . . . . . . . . . . . .8
S1 and S2 driven by UBA2021 . . . . . . . . . . . . . . .9
Current amplitude of LRC resonant tank versus
frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9
Current phase of LRC resonant tank versus
frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9
Current voltage relation of inductor . . . . . . . . . . .10
Current voltage relation of capacitor . . . . . . . . . .10
Switch timing sequence (inductive mode) . . . . . . 11
Switch timing sequence (capacitive mode) . . . . . 11
Boot strap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14
MOSFET gate drive circuit. . . . . . . . . . . . . . . . . .15
Under/overdamped responses of an RLC circuit .15
Extra resistance, lower MOSFET of half-bridge 
for example . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16
Resonant tank frequency versus power. . . . . . . .17
Resonant tank with a burner driven by a square
wave voltage . . . . . . . . . . . . . . . . . . . . . . . . . . . .18
Power versus frequency . . . . . . . . . . . . . . . . . . .19
Lamp over voltage protection circuit . . . . . . . . . .21
Example ILH versus ID burner specification . . . . .22
Definition of lamp currents . . . . . . . . . . . . . . . . . .22
Graphical representation of the SoS lines of 
TL5 HO 24 W and 39 W . . . . . . . . . . . . . . . . . . .23
3F3 specific power loss for several 
temperature/flux density combinations . . . . . . . .24
Core types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25
Three different half-bridge transformer 
winding examples . . . . . . . . . . . . . . . . . . . . . . . .29
Feedback control loop . . . . . . . . . . . . . . . . . . . . .30
Series resonant heating . . . . . . . . . . . . . . . . . . . .31
Inductive mode heating . . . . . . . . . . . . . . . . . . . .31
Fast feedback control loop. . . . . . . . . . . . . . . . . .32
Proximity loss graph . . . . . . . . . . . . . . . . . . . . . .34
BH curve magnetic material. . . . . . . . . . . . . . . . .35
Material loss graph . . . . . . . . . . . . . . . . . . . . . . .35
continued >>
AN10971
Application note
All information provided in this document is subject to legal disclaimers.
Rev. 1 — 15 September 2011
© NXP B.V. 2011. All rights reserved.
40 of 41
AN10971
NXP Semiconductors
TL applications with NXP ballast controllers
14. Contents
1
2
3
3.1
3.2
3.3
3.4
3.5
3.5.1
4
4.1
4.1.1
4.1.2
4.2
4.3
4.3.1
5
5.1
5.2
5.3
9
5.4
5.4.1
5.4.2
6
6.1
6.2
6.3
6.4
6.5
6.6
6.7
6.7.1
7
7.1
7.1.1
7.1.2
7.1.3
8
8.1
8.2
8.2.1
8.2.2
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Lamp characteristics and half-bridge 
principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Lamp filaments . . . . . . . . . . . . . . . . . . . . . . . . . 4
HF TL half-bridge principles . . . . . . . . . . . . . . . 6
Start-up, ignition and operation sequence . . . . 8
Zero voltage switching of a half-bridge 
(capacitive mode) . . . . . . . . . . . . . . . . . . . . . . . 8
None overlap time, fixed versus adaptive. . . . 11
Half-bridge MOSFET selection . . . . . . . . . . . . 13
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . 13
MOSFET voltage rating . . . . . . . . . . . . . . . . . 13
MOSFET RON . . . . . . . . . . . . . . . . . . . . . . . . 13
Boot strap diode and capacitor . . . . . . . . . . . . 14
MOSFET gate drive RLC circuit . . . . . . . . . . . 14
Extra resistance . . . . . . . . . . . . . . . . . . . . . . . 15
Resonant tank . . . . . . . . . . . . . . . . . . . . . . . . . 16
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Frequency versus power . . . . . . . . . . . . . . . . 17
Ignition and coil saturation . . . . . . . . . . . . . . . 19
No ignition protection . . . . . . . . . . . . . . . . . . . 20
Preheat and run time filament current . . . . . . 21
Lead wire current versus discharge current . . 22
SoS specification . . . . . . . . . . . . . . . . . . . . . . 22
Magnetic component design . . . . . . . . . . . . . 23
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Inductor design parameters . . . . . . . . . . . . . . 23
Core type selection . . . . . . . . . . . . . . . . . . . . . 24
Calculate windings . . . . . . . . . . . . . . . . . . . . . 26
Auxiliary winding count . . . . . . . . . . . . . . . . . . 26
Select wire diameters . . . . . . . . . . . . . . . . . . . 27
Half-bridge transformer design for deep 
dimming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Suggested transformer construction. . . . . . . . 28
Dimmable controller feedback control loops 30
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Resonant tank values for a dimmable system 30
Inductive mode heating . . . . . . . . . . . . . . . . . 30
Loop control speed . . . . . . . . . . . . . . . . . . . . . 32
System performance . . . . . . . . . . . . . . . . . . . . 33
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . 33
Inductor losses . . . . . . . . . . . . . . . . . . . . . . . . 33
Ohmic losses . . . . . . . . . . . . . . . . . . . . . . . . . 33
Proximity losses . . . . . . . . . . . . . . . . . . . . . . . 34
8.2.3
8.3
9
10
11
11.1
11.2
11.3
12
13
14
Core losses . . . . . . . . . . . . . . . . . . . . . . . . . .
Sense resistor losses. . . . . . . . . . . . . . . . . . .
Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . .
References. . . . . . . . . . . . . . . . . . . . . . . . . . . .
Legal information . . . . . . . . . . . . . . . . . . . . . .
Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . .
Disclaimers . . . . . . . . . . . . . . . . . . . . . . . . . .
Trademarks . . . . . . . . . . . . . . . . . . . . . . . . . .
Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Contents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
34
36
37
37
38
38
38
38
39
40
41
Please be aware that important notices concerning this document and the product(s)
described herein, have been included in section ‘Legal information’.
© NXP B.V. 2011.
All rights reserved.
For more information, please visit: http://www.nxp.com
For sales office addresses, please send an email to: [email protected]
Date of release: 15 September 2011
Document identifier: AN10971
Similar pages