Page 1
4707 DEY ROAD LIVERPOOL, NY 13088
PHONE: (315) 701-6751 | FAX: (315) 701-6752
M.S. KENNEDY CORPORATION MSK Web Site: http://www.mskennedy.com/
MSK5059RH
Evaluation Board
User’s Guide
By Bob Abel & Paul Musil, MS Kennedy Corp.; 02/2011
Introduction
The MSK 5059RH is a radiation hardened 500 kHz switching regulator controller
capable of delivering up to 4.5A of current to the load. A fixed 500 kHz switching
frequency allows the use of smaller inductors reducing required board space for a given
design. The 4.5A integrated switch leaves only a few application specific components to
be selected by the designer. The MSK 5059RH simplifies design of high efficiency
radiation hardened switching regulators that use a minimum amount of board space. The
device is packaged in a hermetically sealed 16 pin flatpack and is available with straight
or gull wing leads.
The evaluation board provides a platform from which to evaluate new designs with ample
real estate to make changes and evaluate results. Evaluation early in the design phase
reduces the likelihood of excess ripple, instability, or other issues, from becoming a
problem at the application PCB level.
This application note is intended to be used in conjunction with the MSK5059RH data
sheet and the LT1959 data sheet. Reference those documents for additional application
information and specifications.
Setup
Use the standard turret terminals to connect to your power supply and test equipment.
Connect a power supply across the Vin and GND
output load between the VOUT and GND
terminals. Use separate or Kelvin connections
2
to connect input and output monitoring equipment. When measuring output ripple voltage
with an oscilloscope probe, the wire from the probe to the ground clip will act as an
terminals (see note 1). Connect the
1
Page 1 of 9
Page 2
antenna, picking up excessive noise. For improved results, the test hook should be
removed from the tip of the probe. The tip should be touched against the output turret,
with the bare ground shield pressed against the ground turret. This reduces the noise seen
on the waveform.
Note 1: The MSK5059RH has a typical minimum on time requirement of 300nS
corresponding to a minimum duty cycle of 15% at 500kHz switching frequency. Forcing
the device to operate at less than the minimum on time may result in irregular switching
waveforms and present the appearance of instability. The default configuration for this
evaluation card is 1.8V out and it may present irregular switching waveforms at input
voltages greater than 12V. When configured for an output voltage of 2.5V or greater the
MSK5059RH will function normally with input voltages up to the maximum rating of
15V. If operating the MSK5059RH at less than the minimum on time is required greater
than typical compensation can reduce the irregular switching.
Output Voltage Programming
V
= VFB * (1+R1/R2)
OUT
R1 = R2 * (V
OUT/VFB
-1)
Given: V
= 1.21V Typ.
REF
Factory Configuration: R1 = 1.21K, R2 = 2.49K
V
= 1.21 * (1+1.21/2.49) = 1.8V
OUT
Efficiency
Typical efficiency curves for 1.8V and 3.3V output voltages with 5VIN are shown in
Figure 1.
Vin = 5V
L = 6.8µH
Boost Pin
The Boost pin provides drive voltage greater than V
Using a voltage greater than V
improving overall efficiency. Connect a capacitor between Boost and SW to store a
charge. Connect a diode between V
Page 2 of 9
Figure 1
to the base of the power transistor.
IN
ensures hard saturation of the power switch significantly
IN
and Boost to charge the capacitor during the off
IN
Page 3
time of the power switch. A boost voltage of at least 2.8V is required throughout the on-
time of the switch to guarantee that it remains saturated. The boost components chosen
for the evaluation board are a 0.33µF capacitor (C2), and a 1N914 or 1N4148 diode
(CR2). The anode is connected to the unregulated input voltage. This generates a voltage
across the boost capacitor nearly identical to the input. In applications having output
voltages greater than 2.8V and significantly higher input voltages, the anode may be
connected to the output voltage to further improve efficiency. The default configuration
is with the anode on the input. Remove CR2 and install CR3 to connect the boost diode to
the output voltage. Efficiency is not affected by the capacitor value, but the capacitor
should have an ESR of less than 1Ω to ensure that it can be recharged fully under the
worst-case condition of minimum input voltage. Almost any type of film or ceramic
capacitor will work fine.
For maximum efficiency, switch rise and fall times are made as short as possible. To
prevent radiation and high frequency resonance problems, proper layout of the
components connected to the switch node is essential.
Loop Stability
The compensation for MSK5059RH evaluation board is a 1000pF capacitor in parallel
with a series RC consisting of a 1500pF capacitor and a 10k resistor. This
compensation was selected for use with the default components on this evaluation board.
New values may have to be selected if different components are used. The values for loop
compensation components depend on parameters which are not always well controlled.
These include inductor value (±30% due to production tolerance, load current and ripple
current variations), output capacitance (±20% to ±50% due to production tolerance,
temperature, aging and changes at the load), output capacitor ESR (±200% due to
production tolerance, temperature and aging), and finally, DC input voltage and output
load current. This makes it important to check out the final design to ensure that it is
stable and tolerant of all these variations.
Phase margin and gain margin are measures of stability in closed loop systems. Phase
margin indicates relative stability, the tendency to oscillate during its damped response to
an input change such as a step function. Moreover, the phase margin measures how
much phase variation is needed at the gain crossover frequency to lose stability. Gain
margin is also an indication of relative stability. Gain margin measures how much the
gain of the system can increase before the system becomes unstable. Together, these
two numbers give an estimate of the safety margin for closed-loop stability. The smaller
the stability margins, the more likely the circuit will become unstable.
One method for measuring the stability of a feedback circuit is a network analyzer. Use
an isolation transformer / adapter to isolate the grounded output analyzer from the
feedback network. Remove the jumper across R4 and connect the output of the isolation
transformer across R4 using TP1 and TP2 terminals. Use 1M-ohm or greater probes to
connect the inputs of the analyzer to TP1 and TP2. Use GND3 for the ground reference
for the network analyzer inputs. Inject a swept frequency signal into the feedback loop,
Page 3 of 9
Page 4
and plot the loop’s gain and phase response between 1 kHz and 1 MHz. This provides a
full picture of the situation on both sides of the unity gain frequency (20 kHz in this
case). Figure 2 illustrates typical results for the default configuration. The phase margin
is the phase value at the unity gain frequency, or about 68 Deg. The gain margin is the
gain at the 0° phase frequency, or approximately 32.5dB.
80
200
60
40
20
0
TR1/dB
-20
Gain Margin
-40
-60
-80
3
10
TR1: Mag(Gain) TR2: Phase(Gain)
Phase Margin
4
10
f/Hz
5
10
150
100
50
TR2/°
0
-50
-100
-150
-200
6
10
Figure 2
An alternate method is to step the output load and monitor the response of the system to
the transient. Low pass filtering may be required to remove switching frequency
components of the signal to make the small transients more visible. A well behaved loop
will settle back quickly and smoothly (Figure 3-C) and is termed critically damped,
whereas a loop with poor phase or gain margin will either ring as it settles (Figure 3-B)
under damped, or take too long to achieve the setpoint (Figure 3-A) over damped. The
number of rings indicates the degree of stability, and the frequency of the ringing shows
the approximate unity-gain frequency of the loop. The amplitude of the signal is not
particularly important, as long as the amplitude is not so high that the loop behaves
nonlinearly. This method is easy to implement in labs not equipped with network
analyzers, but it does not indicate gain margin or evidence of conditional stability. In
these situations, a small shift in gain or phase caused by production tolerances or
temperature could cause instability even though the circuit functioned properly in
development.
Figure 3-A Figure 3-B Figure 3-C
Page 4 of 9
Page 5
Figures 4-A and B illustrate typical results for a step load function between 500ma and
2.0A. Notice that the ringing in Figure 4-B is evident for the design intentionally
modified to be less stable versus the critically damped response in Figure 4-A.
Figure 4-A (Typical Response) Figure 4-B (Modified Response)
Current Sharing and Synchronization
There are several advantages to using a multiple switcher approach compared to a single
larger switcher. The inductor size is considerably reduced. Three 4A inductors store less
energy (LI2/2) than one 12A coil so are far smaller. In addition, synchronizing three
converters 120° out of phase with each other reduces input and output ripple currents.
This reduces the ripple rating, size and cost of filter capacitors. If the SYNC pin is not
used in the application, tie it to ground. To synchronize switching to an external clock,
apply a logic-level signal to the SYNC pin. The amplitude must be from a logic low to
greater than 2.2V, with a duty cycle between 10% and 90%. The synchronization
frequency must be greater than the free-running oscillator frequency and less than
1 MHz. This means that minimum practical sync frequency is equal to the worst-case
high self-oscillating frequency (560 kHz), not the typical operating frequency of
500 kHz. Caution should be used when synchronizing above 700 kHz because at higher
sync frequencies the amplitude of the internal slope compensation used to prevent
subharmonic switching is reduced. Additional circuitry may be required to prevent
subharmonic oscillation
Shutdown
For normal operation, the SHDN pin can be left floating. SHDN has two output-disable
modes: lockout and shutdown. When the pin is taken below the lockout threshold,
switching is disabled. This is typically used for input undervoltage lockout. Grounding
the SHDN pin places the RH1959 in shutdown mode. This reduces total board supply
current to 20µA.
Input/Output Capacitors
The input capacitors C7A and B are 47µF tantalum capacitors and were chosen due to
their low ESR, and effective low frequency filtering. The input ripple current for a buck
Page 5 of 9
Page 6
converter is high, typically I
/2. Tantalum capacitors become resistive at higher
OUT
frequencies, requiring careful ripple-rating selection to prevent excessive heating.
Measure the capacitor case rise above ambient in the worst case thermal environment of
the application, and if it exceeds 10°C, increase the voltage rating or lower the ESR
rating. Ceramic capacitors’ ESL (effective series inductance) tends to dominate their
ESR, making them less susceptible to ripple-induced heating. Ceramic capacitors filter
high frequencies well, and C1A and B were chosen for that purpose.
The output capacitors C5A,B and C are AVX TAZ series 220uF tantalum capacitors.
AVX TAZ series capacitors were chosen to provide a design starting point using high
reliability MIL-PFR-55365/4 qualified capacitors. Ceramic capacitance is not
recommended as the main output capacitor, since loop stability relies on a resistive
characteristic at higher frequencies to form a zero. At switching frequencies, ripple
voltage is more a function of ESR than of absolute capacitance value. If lower output
ripple voltage is required, reduce the ESR by choosing a different capacitor or placing
more capacitors in parallel. For very low ripple, an additional LC filter in the output may
be a less expensive solution. Re-compensation of the loop may be required if the output
capacitance is altered. The output contains very narrow voltage spikes because of the
parasitic inductance of C5. Ceramic capacitors C6A and B remove these spikes on the
demo board. In application, trace inductance and local bypass capacitors will perform this
function.
Catch Diode CR1 and L1
Use diodes designed for switching applications, with adequate current rating and fast
turn-on times, such as Schottky or ultrafast diodes. The parameters of interest are forward
voltage, maximum reverse voltage, reverse leakage current, reverse recovery, average
operating current, and peak current. Lower forward voltage yields higher circuit
efficiency and lowers power dissipation in the diode. The reverse voltage rating must be
greater than the input voltage. Average diode current is always less than output current,
but under a shorted output condition, diode current can equal the switch current limit. If
the application must withstand this condition, the diode must be rated for maximum
switch current. There are a number of tradeoffs to consider when selecting a coil for your
application. The inductance value determines the peak to peak ripple current under
various operating conditions. A common starting point for the peak to peak current ripple
is 20% of the load current. The equation below shows how to select and inductor value
based on desired ripple current and circuit parameters.
L = D*(Vin – Vout)/(f
* Ipp)
sw
Given:
D = Duty cycle, approximately Vout/Vin
I
= Peak to peak ripple current, typically 0.2 * Iout DC
pp
fsw = Switching frequency in Hz
L = Inductor value in Henries
Page 6 of 9
Page 7
Current Limitations
Peak current for a buck converter is limited by the maximum switch current rating. This
current rating is 4.5A up to 50% duty cycle (DC), decreasing to 3.7A at 80% duty cycle
for the MSK5059.
Figure 5
This is shown graphically in Figure 5, and can be calculated using the formula below:
IP = 4.5A for DC < 50%
IP = 3.21 + 5.95(DC) – 6.75(DC)2 for 50% < DC < 90%
DC = Duty cycle = V
OUT/VIN
Maximum output current is then reduced by one-half peak-to-peak inductor current.
I
= IP – (V
MAX
)(VIN – V
OUT
)/2(L)(f)(VIN)
OUT
Example: with VOUT = 5V, VIN = 8V; DC = 5/8 = 0.625, L = 3.3µH
IP = 3.21 + 5.95(0.625) – 6.75(0.625)2 = 4.3A
I
= 4.3 – (5)(8-5)/2(3.3µH)(500kHz)(8) = 3.73A (Figure 6)
MAX
Current rating decreases with duty cycle because the RH1959 has internal slope
compensation to prevent current mode subharmonic switching. The RH1959 has
nonlinear slope compensation, which gives better compensation with less reduction in
current limit.
Page 7 of 9
Figure 6
Page 8
MSK5059 Evaluation Board Schematic
Typical Performance
Parameter Conditions Units Typical
Output Voltage Vin = 5.0V, I
Switching Frequency Vin = 5.0V, I
Output Ripple Voltage Vin = 5.0V, I
Line Regulation 4.3V ≤ V
Load Regulation Vin = 5.0V, I
≤ 15V, I
in
OUT
Efficiency Vin = 5.0V, I
= 2.0A V 1.8V (Factory Default)
OUT
= 2.0A kHz 500
OUT
= 2.0A mVp-p 25
OUT
= 2.0A % -0.1
OUT
= 50mA to 2.0A % -0.3
= 2.0A % 75
OUT
Current Limit Vin = 5.0V A 5.5
Gain Margin Vin = 5.0V, I
Phase Margin Vin = 5.0V, I
= 2.0A dB 34
OUT
= 2.0A Deg 65
OUT
PCB Artwork
Top Side Bottom Side
Page 8 of 9
Page 9
Bill Of Materials
0
RefDes Description Manufacturer PartNumber
U1 SwitchingRegulator MSKennedyCo rp. MSK5059RHG
C1A 1210Ceramiccap1.0uF AV X 12103C105KAT
C1B 1210Ceramiccap1.0uF AVX 12103C105K AT
C2 8050CeramicCap.33uF A V X 08053C334KA T
C3 8050CeramicCap1000pF AV X 08053A 102JAT
C4 8050CeramicCap1500pF AV X 08053A 152FAT
C5A 220uFLowESRtantalum AVX TAZH227K010L(CWR29FC227K)
C5B 220uFLowESRtantalum A VX TAZH227K010L(CWR29FC227K)
C5C 220uFLowESRtantalum AVX TAZH227K010L(CWR29F C227K)
C5D N/A
C5E N/A
C5F N/A
C6A 1210Ceramiccap1.0uF AVX 12103C105KAT
C6B 8050Ceramiccap0.1uF AV X 08053C104KA T
C7A 47uFLowESR
C7B 47uFLowESRtantalum AVX TAZH476K020L(CWR29JC476K)
C8 N/A
C9 N/A
R1 Resistor1.21K, 1/8W
R2 Resistor2.49K, 1/8W
R3 Resistor10.0K, 1/8W
R4 Resistor2
R5 N/A
CR1 Fairch ild Fairchild FYD0504SATM
CR2 1N4148or1N914 AN Y 1N4148or1N914
CR3 N/A
L1 6.8uHinductor Coilcraft DO3316P‐682MLB
tantalum AVX TAZH476K020L(CWR29JC476K)
Ω,1/8W
Page 9 of 9
Loading...