Anda di halaman 1dari 24

MK5PFC

CIRCUIT
ANALYSIS

2008 Slot Tech Magazine


All rights reserved

Slot Tech Feature Article

SETEC MK5PFC
Circuit Analysis

R212

31

14

C59

R125

R163

R159

R195

C93

R188

R143

R189

228

C88

L8

C96

C91
R88

D19

207

C100

R245

12

R244

R87

203

R236

83

C98

F3

R242
163

161

171

R104

82

R95

159
U12

172
160

R98

C105

R89

Slot Tech Magazine

E3

E2
227

67

X5b

62
226

D18

58
Q7

Q6

162
R102

225

169

57

C102

66

224

R217

209

63

C5

C104
56

C106

T1

R218

D42

R91

R211

223

S1

R77

R202

13

C95
D30

C101

59

Page 2

60

15

R1

L4

X6b

C99

R171

R96

17

Q5

C14

11

61

C97

L7

32

R177

R111

R142

R182
C119

R103

C89

R92
R93

R97

C94

R90

18

R197

79

C16

E1

X10

167

92
Q8

C92

19

94
Q3

E4

213

C2

Q2

98

204

C125

R23

165
D33

D6

R84

R243

24

99R130

R193

U9

69

10
4

U1
164

R181

100
168

208

R81

D38

91

R209

R183

R156

114

115

R196

R208

R172

181 146
D7

K1
72

158

U15

30
U13

D5218 138
22

157

97

217
D31

R99

R85

C57

R108

R124

111

113

H1

U7

55

44

47

D1

R126

R162

R235
R240

U4

D39

R94

R200

Q11
198

R192

110

U11

85

R173

R239

180
109

C90

C70

C108

C72

145

D8

199

Q4

148

R115

R131

144

D40

143

R141

R230

U3

178

75

R179

D11

48

D21

C68

R237

R146

R119

R238

R228

R234

R170

R178
R241

196

R231

191

R113

C73

C87

R114

187
136

156

R151

46

R123

R232

C86

C66

R222

R117

D22

D27

R157

96

R216

U5 80
74

R187

R152

95R191201R161

R190

86 Q9
87

R158
C114 90

U10

R229

188

C84

106

R150

R165

Q10

103

X9

D12

197

Q12

R204

C122

Q1

33

D4

R184
206
23

U8

176R160
D28

R176

U14

147

C61

20

R203

102

D26

R86

D3

R233

88 R194

U6

73
C67
76

139
177

D36

C116

195

R118

C82

R220

190

184

R120

D43

C58 205

34

C118

U2

R174

174

140
C79

R223

C78

D35
3

141

142

183

C117

16

C81

C56

R186

C53

211

R205

R219

C76

154

R168

175

212

173

194

193
C74

192

182

R221

35

C55

R185

R135

133

R226

131
R137

R116

130
R139

D41

185

200

R132

129
R140

D37

124
R154

R122
123

R148

128
R149

132

134

R82

R144

R136

26

L5
42

C54

R121
122

127
R133

C52

D16

D17

210

L6

38

R198

R213
121

126
R134

C63

21

29

R127
118

R214
120

125
R147

C120

27

28

R128
117

R199
119

C121

C64

R129
116

R138

25

C75

222

E5

R100

florescent lamps or shut


down the monitor and you
begin to see its complexity.
Top it all off with overcurrent protection and
over-voltage protection,
sprinkle a little temperature
control on top and you have
a complex system that
quickly reminds you that
there are still some very
interesting analog circuits
in this mostly digital world.

he SETEC MK5PFC
power supply is one of
the most complex
power supplies we see in
the gaming industry. It
incorporates the latest
advancements in modern
power supply design including Power Factor Correction
and DC to DC conversion.
Add to that some very gaming-oriented features such
as the ability to control the

X8

X7

SETEC MK5PFC

AC (mains) Input
Filtering and Monitor Control
Switched-mode power supplies are noisy. They generate a lot of electromagnetic
interference (EMI). Like all
modern power supplies,
this one starts off with an
EMI filter on the mains
input (Although we dont
use the term in North
America, in the rest of the
world, household AC current is referred to as the
mains.). This filter system
lives on the appropriately
named Filter Board. It is
comprised of a network of
capacitors and an inductor
(C1-C6, L1) and prevents
EMI from escaping the
power supply and taking a
free ride on the mains
where it might wreck havoc
on the rest of the system. It
doesnt actually change the
AC current in any significant way.
R8 is a varistor, the selfsacrificing surge protector
that absorbs the energy
from things like lightning
strikes and freak power
surges. If you see that it
has blown up, you will
naturally have to replace it.

But dont think badly of it.


Dont think it failed. It likely
was just doing its job. You
might want to look at everything else plugged into the
same power source. Its
usually a one-shot device. If
its blown open, you might
not even realize it because
it did its job and the game
still works perfectly. However, youve lost a layer of
protection and the next hit
will likely be fatal (and
maybe costly).

throw mains switch completely isolates both the hot


and neutral lines when its
in the off position. When
the switch is in the on
position, it applies the
mains to the Main Board of
the power supply through
yet another line filter circuit, this time comprised of
coils L2 and L3, and capacitors C7-C11. It also applies
power to the monitor but
not directly.
Zero-Crossing

From here, the AC current


passes through fuse F2 to
the female mains connector. This is an unswitched
mains outlet. At the same
time, a double pole, single

Before applying AC power to


the monitor, the AC current
passes through a zerocrossing circuit made from
U1 and a TRIAC.

The zero-crossing is the instant in time that the voltage and current are both
zero. By switching only during this time, AC loads such as the monitor can be
safely connected to the mains without excessive inrush current.
SETEC MK5PFC

Slot Tech Magazine

Page 3

A zero-crossing circuit
assures that the only time
that the monitor will actually make the connection to
the mains is when the
instantaneous voltage of the
mains is precisely zero. At
that exact moment in time,
there is no current flowing
through the mains and the
monitor can be connected
safely without causing
excessive inrush current or
freaking out the degaussing
circuit in a CRT monitor.
Essentially, its like the
difference between hot
plugging something and not
hot plugging it. If the
monitors power is connected only when the voltage is zero, it can ramp up
slowly as the sine wave
voltage increases. Its a
much nicer and less destructive way to turn things
on and prevents things like
fuses that seem to blow for
no reason or destruction of
input rectifiers.

zero crossing.
The rest of the monitors
mains control circuit is
straightforward. The hot
side of the AC power passes
from pin 2 of the mains
switch, through fuse F1
and, when gated, through
the TRIAC to the monitor
receptacle. The neutral
connection is between pin 1
of the mains switch and the
monitor receptacle.

But this so-called Filter


Board has another function as well and its really
important to realize that
the monitor power is not
only controlled by the zerocrossing circuit but that
the zero-crossing circuit
itself is controlled. It is
controlled by the all-powerful Low Power signal that
comes from the game itself.
The Low Power signal is an
active low signal that

At the heart of the zerocrossing circuit is the zerocrossing detector itself, an


MOC3083. The MOC3083
consist of a normal, infrared light emitting diode
thats optically coupled to a
detector. Sounds like a
normal opto-isolator,
doesnt it? It would be except that on the same little
silicon chip that contains
the photo-detector, there is
a bit of circuitry that detects the zero crossing and
gates the TRIAC (labeled D1
on the schematic but thats
an odd designation for a
TRIAC) with an output from
pin 6 only at the moment of
Page 4

Slot Tech Magazine

SETEC MK5PFC

comes from an open collector output on game itself.


When the signal is low, it
turns on the LED in the
opto-detector-zero-crossingIC-thingy, activating (gating) the TRIAC and turning
on the monitor. The bottom
line is this, if this signal is
not active, your monitor will
not have AC power. Neither
will the florescent lights but
more about that later.

watts) before it reaches the


bridge rectifier, D1 (although I suppose it would
be more precisely correct to
say that the resistors are
actually in the return path,
between the bridge and the
neutral side of the AC line
and that the AC current
passes through Fuse F3,
through the line filter to the
top side of the bridge and
returns through the resistors to neutral).

Inrush Current Limiting


From the filter board, the
mains is connected to the
AC input of the main board
of the power supply. After
passing through Fuse F3,
the AC passes through yet
another line filter (L4 and
associated capacitors). Now
its time to apply the AC to
the input rectifiersor is it?
Not quite yet. One of the
manufacturers specifications for this power supply
(driven by OEM requirements, I assume) is that the
inrush current be less than
25 Amps peak when turned
on at either 120 Vac or 240
Vac. In order to help accomplish this, the MK5PFC
uses a couple of series
resistors and a relay to
achieve a two step, soft
start procedure.
When power is first applied
to the unit, relay contact K1
is open. You can see that
the contact is drawn on the
schematic in its normally
open position. The AC
input must pass through
two, high-wattage, ceramic,
wire wound resistors (R77
and R202, each 9.1 ohms, 7
SETEC MK5PFC

However you want to look


at it, there is a total of 18.2
ohms of resistance in series
with the AC input, limiting
the AC current to less than
.5 Amp before the 7 watt
dissipation rating of the
resistors is exceeded.
The concept is this: Power
is applied to the unit. At
first, the current-limited
power is applied, activating
the power factor correction
circuitry which, as you will
read anon, controls the
charge rate of the primary
electrolytic capacitor, the
main culprit in the generation of both high inrush
current and third harmonics, a pair of nasty phenomena we can do without,
thank you very much.
After a few seconds, when
the primary filter capacitor
(C52) is fully charged and
everything has stabilized,
relay K1 energizes <click>
and the resistors are bypassed by the relay contact.
The mains is now connected directly to the power
supply which then proceeds
to connect power to the
Slot Tech Magazine

loads such as switching on


the monitor AC and the +24
Vdc outputs for the florescent lamps. Please keep in
mind that the game circuitry has ultimate power
through the Low Power
signal and that the loads
will not be energized until
the signal is pulled low.
Also notice that there is a
130 degree Celsius thermal
switch (S1) in series with
the resistors as well. It
opens at 130 degrees Celsius. Thats 266 degrees
Fahrenheit. The thermal
switch is visible in the lower
left corner of the PCB. It is
physically mounted directly
on top of the two inrush
current limiting resistors,
R70 and R202. Its a safety.
If the power supply doesnt
fire up immediately and
energize relay K1, these
resistors will get hot. If the
temperature exceeds 130 C,
S1 opens and the current
flow stops. Keep this in
mind when youre troubleshooting. If these resistors
are hot, dont assume that
something is shorted on the
AC input, drawing too
much current through the

The two inrush current limiting


resistors with the 130 degree
thermal switch mounted on top.
Page 5

resistors. ANYTHING that


prevents the power supply
from energizing relay K1
(and thats just about anything that fails in the power
supply) will result in these
resistors getting hot.
Power Failure Detection
Before we leave the AC area
and enter the DC world,
look back at the top of the
line filter L4 and find diode
D3, a common 1000 volt, 1
amp 1N4007. This is the
start of our power fail
detector. It is connected to
the hot side of the mains
and passes half the AC
cycle through a string of
voltage dividing resistors to
a comparator circuit that
well look at later on. If the
comparator sees a single
missing cycle or even a few
cycles, it will do nothing at
all. However, if the power
really has failed (perhaps a
half-dozen missing cycles) it
generates a PFAIL signal.
This signals the games
computer which then does
all the housekeeping necessary to retain the integrity
of the game prior to the
imminent loss of power. The
power supply itself will
remain operational long
enough to accomplish these
tasks (even to the point of
operating the mechanical
meters). It does this by
virtue of the energy stored
in the primary filter capacitor, extended by instantly
shedding loads such as the
florescent lamps. This is
one smart power supply
and it does it all with analog electronics as youll see
Page 6

later.
To the Bridge and Beyond
We have finally made it to
the bridge rectifier, D1. Of
course the bridge rectifies
the AC input, turning it into
full-wave, pulsating DC.
Following D1, we find a
departure from traditional
power supplies. The primary filter capacitor is not
connected immediately
following the bridge rectifier
as we have seen in all
power supply designs of the
past. Instead, a power factor correction circuit is
inserted between the output
of the bridge rectifier and
the primary filter capacitor,
C52.
Lets follow the positive
output of the bridge rectifier
and see where it leads.
There are two paths here
for the current to flow. One
path passes through diode
D43 and then to C52. But
why do we need the diode?
Its already DC, isnt it?
Sure it is. Its the output of
a bridge rectifier and bridge
rectifiers turn AC into DC.
Is the current being
double-rectified or something? Seems mysterious,
doesnt it?
The answer lies down the
other path so lets go back
to the positive output of the
bridge rectifier and follow it
straight across to coil L5.
This is a large toroidal coil.
From the right side of L5,
we can follow the current
path through diodes D12
and D11 and then to the
Slot Tech Magazine

positive lead of the C52, the


primary filter capacitor.
What is going on here? Why
are there two paths and
why do we have the extra
diodes?
Harmonic Currents and
Active Power Factor Correction
If youre a regular reader of
Slot Tech Magazine, you
know all about harmonics
and switched-mode power
supplies. You know about
the power-sapping third
harmonic and how it robs
your casino of power. If you
need a refresher, the topic
was covered extensively in
the August 2004 issue.
Harmonic currents are a
direct result of the way in
which a switched-mode
power supply (SMPS) draws
current from the system.
The input circuit of an
SMPS is a bridge rectifier
that changes the 120 volt
AC input to DC. A capacitor
smoothes this DC to eliminate voltage ripples and the
resultant DC bus has a
voltage of about 170 volts
when the AC rms input is
120 volts. Although the AC
voltage is a sine wave, the
rectifier draws its current in
spikes. These spikes require
that the AC supply system
provide harmonic currents,
primarily 3rd, 5th and 7th.
These harmonic currents do
not provide power to the
SMPS, but they do take up
distribution system capacity. The principal harmonic
current is the 3rd (180 Hz)
and the amplitude of this
current can be equal to or
SETEC MK5PFC

even greater than that of


the fundamental current.
We solve this problem with
power factor correction.
Look at the circuit made
from MOSFET Q1 and its
associated driver, U7. It
kind of looks like it is its
own SMPS, doesnt it?
However, the drain of the
MOSFET is connected to
the big toroid coil, L5.
Whats this all about?
This, my friends, is a tricky
little circuit called a boost
power supply. In this case,
its more specifically called
a follower boost. We are
using the coils ability to
store energy, not as a
charge (as we do with a
capacitor) but in the form of
a magnetic field.
Our goal here is to change
the way the monitors filter
capacitor draws current
from the bridge rectifier
and, subsequently, the AC
(mains). Were looking for a
way to boost the pulsating
DC output of the bridge
rectifier so that instead of
charging the filter capacitor
with narrow, harmonicproducing spikes of current, we have a steady flow
of current flowing from the
bridge rectifier into the filter
capacitor.
We accomplish this feat by
pulsing MOSFET Q1. When
Q1 is turned on, current
will flow from the positive
output of the bridge rectifier, through L5 and
through Q1 to the high
voltage return path at the
SETEC MK5PFC

negative side of the bridge


rectifier (not technically
ground in this case as the
entire primary side is isolated, as usual, from the
secondary). The coil is our
load and it builds up a nice
big magnetic field. When Q1
is turned off, the magnetic
field collapses. This rapidly
collapsing magnetic field
slices across the coils of
copper wire and turns the
coil into an electric generator in a process called induction. This newly generated voltage (you can kind
of think of the coil as a
battery for this moment in
time) is now in series with
the output of the bridge
rectifier and, just like two
or more dry cell batteries in
series in a flashlight, the
voltages are added together.
Its called a follower boost
circuit because this newly
generated voltage is added
to the incoming voltage. If
the incoming AC rises, the
boost follows along, rising
as well. We dont care about
regulating the voltage at
this point because were
going to do that next with
the PWM part of the SMPS.
The result is that we are
taking a sine wave in and
producing a constant voltage out and the upshot of
this whole thing is that
instead of charging the filter
capacitor only during the
brief peak period of the AC
sine wave, we can keep a
constant charge on it and
substantially reduce (or
eliminate altogether) the
third-harmonic content of
the system. This is known
Slot Tech Magazine

as active power factor


correction or PFC.
The diodes we were talking
about at the beginning of
this discussion (D11, D12
and D43) are a sort of electronic anti-siphon valve.
They are used to ensure
that the current doesnt
backflow when, for example, the output voltage of
the boost follower circuit is
higher than the output
voltage of the bridge rectifier.
The brains of the outfit is
the UCC38503 combination
PFC/PWM controller IC,
U14. Because modern
power supplies often include active PFC, this IC
simply includes both PFC
and traditional PWM technology in one package. On
the PFC side, the
UCC38503 samples both
the pulsating DC output of
the bridge rectifier and the
voltage at the positive terminal of the filter through a
couple of resistor voltage
divider networks. Youll see
a lot of these in this power
supply, where an extremely
high voltage (as high as
+400 Vdc on the primary) is
passed through a divider
network of five or seven
resistors in order to cut the
voltage down before applying it to the low voltage
input of a voltage comparator or other IC such as this
one. In this case, the voltage sense input is pin 3.
Pin 18 checks to see whats
coming out of the bridge
rectifier. The UCC38503
figures out what to do
Page 7

based on these inputs and


in turn, sends an output
from pin 12 to U7 which in
turn controls Q1, turning it
rapidly on and off, alternately storing and releasing
energy in L5 in order to
maintain a smooth flow of
current into the primary
filter capacitor, C52.

UCC38503 PFC/PWM Controller needs a power source


before it can do anything at
all. So does the MOSFET
driver, U7. Nothings going
to happen unless we get
power to these devices. At
the moment, our only DC
supply is the +400 Vdc at

the positive terminal of the


primary filter capacitor.
To accomplish this task, we
have a power supply within
a power supply here. Its a
remarkable little high voltage, three terminal linear
regulator (U13, a type

Are You High?


Yes. Very high. I am speaking of course, about the
voltage on the primary filter
capacitor. The schematic
diagram pegs it at +400
Vdc. Of course, were eventually going to cut that
down to +24 Vdc. Thats the
output voltage of the unit.
But before we move on to
the secondary side of
things, there is a very important aspect of this design that needs to be
pointed out. U14, the
UCC38503 PFC/PWM Controller

Page 8

Slot Tech Magazine

SETEC MK5PFC

VB408) that takes the +400


volt, unregulated voltage at
the primary and regulates it
to an output voltage that
can be anything between
+1.25 Vdc to just 30 volts
below the input voltage. In
this case, we can create a
nice regulated power supply
of +13.25 volts DC by pegging the reference voltage at
pin 1 with a 12 volt Zener
diode, D38. The output of
the device is 1.25 volt
higher than the reference
input at pin 1. The output
current is limited to just 40
milliamps but its enough to
power the few low voltage
things we need to operate
before the main power
supply output comes on
line, specifically U7 and
U14 as well as an LM339
Quad Comparator (U2) and
the relay, K1 (remember the
inrush current limiting
system and the relay contact that bypasses the resistors? All of that has to
operate BEFORE the power
supply kicks in!).
PWM Controller
We have looked at the PFC
half of the UCC38503
combination PFC/PWM
controller IC. Lets continue
by looking at the other half
of U14. I guess Id have to
say that this is the only
boring part of this power
supply. The PWM controller
part of U14 is totally
normal in every respect. Pin
10 is the Gate 2 output
that controls MOSFET Q5,
the primary switching
transistor that switches the
primary current on and off
SETEC MK5PFC

through the primary


winding of power
transformer T1. Totally
normal.
You can see the typical
regulation feedback
provided by the transistor
half of an opto-isolator (U4)
and you can see that the
LED half of U4 is powered
by the +24 VDC output of
the supply with a
frequency-compensated
TL431 (U9) providing a nice
reference voltage on the
cathode of the device.
We light up the LED in the
opto-isolator with voltage
from the secondary output
of the power supply. The
higher the voltage, the
brighter the LED shines.
We read the brightness of
the LED with the
phototransistor in the optoisolator, which is connected
to the PWM control circuitry
in U14 on the primary side
of the transformer. In this
way, the secondary can
talk to the primary
without actually touching
it.
But in order to maintain
tight voltage regulation, we
need to go just one step
further. We need to control
the brightness of the LED
under a variety of changing
load conditions of both high
and low frequencies. There
is another element in the
chain of regulation and in
this case its U9, a type
TL431. Get to know the
TL431 because there are
four of them in this power
supply.
Slot Tech Magazine

You can think of the TL431


as a sort of programmable
Zener diode. It is a shunt
regulator that can be
programmed to be any
voltage from a minimum of
2.5 volts to a maximum of
37 volts. Inside the device,
an internal 2.5 volt
reference is compared to
the voltage that is applied
at the reference pin input.
This reference voltage is
derived by a resistor voltage
divider (R96, R97, and
R100). The TL431 provides
the gain that is needed at
low frequencies so that the
LED in the opto-isolator will
produce enough of a change
in brightness in order to
signal the primary side and
compensate for the low
frequency changes in the
load.
But this gain is not needed
at high frequencies. The
gain of the opto-isolator
itself (the CTR or current
transfer ratio) works just
fine, without any assistance
from the TL431s gain,
thank you very much. This
leads to a sort of
engineering dilemma where
high frequency changes in
load can produce larger
voltage swings than low
frequency loads, making
tight regulation impossible.
Compensation
In this case compensation
is frequency
compensation which is the
way we can control the
frequency response of
various circuitry. By using
a combination of resistors
Page 9

and capacitors primarily,


we can integrate the various
load frequencies and tell
the TL431 how to behave at
certain frequencies. In this
power supply, the
compensation network is
made from C119, R103 and
C89, a 1 uf, bipolar
capacitor. The
compensation network
allows the TL431 to
maximize its contribution at
very low frequencies and to
remove its influence at
higher frequencies. The
connection of C89 and
R103 between the cathode
and the reference terminal
of the TL431 allows
maximum loop gain at DC
for the best voltage
regulation.
What else can you say
about this totally normal
SMPS design? The output of
power transformer T1 is
rectified by D30 and filtered
by C88 (2200 mF 35 V).
After passing through a
choke (L8), C106 (also 2200
mf 35V) provides additional
filtering. At the same time,
of course, C88 and C106
are reservoir capacitors
that, along with the energy
stored in C52 (the primary
filter capacitor) will be all
the energy that the power
supply has in case of a
power failure.
Cut the Juice, Bruce!
Do you remember the Low
Power signal from last
month? Remember that the
Low Power signal must be
dragged to ground in order
to energize the monitor AC
Page 10

power. Well, the Low Power


signal is doing something
else at the same time,
something sort of unrelated
(electronically speaking) to
monitor control but
something that a slot
machine needs to take care
of and that is surviving a
blackout. In the case of an
immediate and unpredicted
loss of AC power, a slot
machine has some serious
business to attend to before
the energy stored in the
power supplys electrolytic
filter capacitors is fully
dissipated. Mostly, the CPU
simply has to store a small
amount of data (things like
customer credits and
current game condition)
and perform an orderly
shutdown but in some
cases, the slot machine
might even want to run long
enough to continue to
increment the hard coin
meters (the
electromechanical units
themselves) until the
correct count is obtained.
The MK5PFC makes this
possible by quickly
shedding some of the +24
VDC load and it does it with
a remarkable little device
called a UCC3913 Negative
Voltage Hot Swap Manager,
also known as a circuit
breaker! Both of these
terms are familiar to us, of
course. We know all about
hot swapping (and the
damage it can cause in a
slot machine) and a circuit
breaker is, well, a circuit
breaker.

Slot Tech Magazine

When the Low Power signal


(it comes from the slot
machine, remember?) goes
high, the LED in U1 on the
filter board turns off. We
covered that in part one.
But what we didnt cover is
where the LED in U1 gets
its power source at the
anode. This is very clever.
Lets look at the entire
control circuit.
U13 is the UCC3913. Lets
follow the green path. This
is the +24 VDC power bus.
Its the output from the
secondary winding of T1,
rectified by D30 and filtered
by C88. Its the actual
output of the power supply,
its Raison dtre. The +24
VDC bus is connected to
the florescent lamps at
connectors X7 and X8, pin
4. However (and here comes
the interesting part) we
need a return path to
complete the circuit for the
florescent lamps. The
return path is through pin
2. Follow the checkered
green path, remembering as
you do that this is the
return path and were
headed for ground
somewhere, the shorter the
path (least resistance) the
better. In this case, the
shortest path (the only
path) is through MOSFET
Q7 and its source resistor
R236 (just 15 milliohms
used, naturally, for overcurrent sensing) to ground.
Naturally, Q7 has to be
turned on for this to
happen and you can see
where all this is headed.
U13 is the thingy that
SETEC MK5PFC

controls the gate voltage of


Q7. Obviously, U13 has to
have power in order for this
to happen. That power
source (Vdd-pin3) comes
from the +24 VDC bus
through transistor, Q6. Its
a PNP transistor with its
emitter connected to the
+24 VDC bus and its base
(which must be dragged
down in voltage in order to
energize Q6 because its
PNP) connected toTA
DA!the Low Power signal!
Thats an interesting
connection between the
circuit that controls the
monitor AC and the circuit
that controls the florescent
lamps. At the same time
that the Low Power signal is
lighting the LED in U1
(thus controlling the
monitor) it is pulling down
the voltage at the base of
Q6, shooting the power
source to U13 which
energizes Q7, completing
the return path to ground
and lighting the florescent
lamps.
In the case of a power
failure, the Low Power
signal goes high. This
instantly shuts off the
TRIAC providing AC power
to the monitor (which at
first glance seems sort of
silly since there actually is
no AC at the moment but
its part of the whole load
control circuit) and, at the
same time, turns off the
florescent lamps. By
shedding some of the load
from the +24 VDC power
bus, the energy stored in
the power supply will be
sufficient to take care of
SETEC MK5PFC

business before everything


decays to zero.
Correction
Actually, its more of an
addendum but I admit to
an error in our discussion
of U13, the type VB408. I
had mentioned that the
output current is limited to
just 40 milliamps but
mentioned that it was
enough to power the few
low voltage things we need
to operate before the main
power supply output comes
on line. It is not. Not
without some assistance.
What I had failed to
mention is a pretty darned
clever little part of this
power supply within a
power supply.
If you recall from our
discussion, coil L5 is an
energy storage device used
in the PFC circuit. We
looked at L5 as part of the
PFC circuit but there is yet
another winding on L5,
wound around the same
toroidal ferrite core. Its
connected to pins 5 and 6
and can be found on the
schematic diagram just to
the right and below the
voltage regulator, U13.
Once the PFC circuit kicks
in (and Q1 is operating)
there is a ton of energy in
L5. The winding between
pins 5 and 6 simply picks
up some of this energy.
Sounds like a transformer,
doesnt it? For all intents
and purposes, it is. Its just
a really cool and efficient,
high-frequency, toroidal
transformer that keeps all
Slot Tech Magazine

of its precious energy


tightly held within its
donut-shaped core until we
tap into it when we need to.
And we need it now. Once
the PFC circuit has kicked
in, we take this low voltage
output from L5 and rectify
it with D17. From there, the
current passes through R82
and onto the power bus.
Diode D19 prevents the
current from flowing
backward when the output
voltage of this little supply
within a supply is greater
than the output voltage of
the regulator. Also, we cant
allow the output voltage of
U13 to exceed the input
voltage as can occur when
the main AC power is
removed. The primary filter
capacitor can (will)
discharge faster than the
secondary filter capacitors
because it uses the last bit
of its energy charging them!
In general, voltage
regulators dont do well
when the output voltage
exceeds the input voltage.
By their Grounds Ye Shall
Know Them
As with all switched-mode
power supplies, it is very
important to realize the
isolation between the
primary circuits and the
secondary circuits. In this
power supply, there is quite
a bit of low-voltage circuitry
connected to the primary
return (0v) which is totally
and completely isolated
from the secondary ground!
This hot return (meaning
that although it represents
Page 11

Page 12

Slot Tech Magazine

SETEC MK5PFC

SETEC MK5PFC

Slot Tech Magazine

Page 13

a common return path (0v)


for all of our primary
circuitry, it is hot in respect
to Earth ground) is marked
on the schematic diagram
with a black and red
checkerboard pattern.
When you are making
voltage measurements in
the primary, you must have
your meter ground
connected to the proper
point. If you are using an
oscilloscope, you must have
the power supply plugged
into an isolation
transformer before
connecting your scope
ground to this point or you
will vaporize portions of
both the power supply and
the oscilloscope ground.
The cold return is the
secondary ground. Its
marked on the schematic
with a blue and black
checkerboard. This is
totally normal and
connected to all of the
grounds throughout the slot
machine, including the
Earth ground and all DC
grounds everywhere.
As you look at the two
return paths (the primary,
with its red checkerboard
and the secondary in blue)
the schematic diagram sort
of resolves itself and you
can more readily visualize
that the circuits in the
lower left hand corner of the
schematic (U1, U3 and
associated components
theyre fault detection
circuits that well get to
shortly) actually belong on
the right side of the
schematic (to the right of
Page 14

the secondary winding of


power transformer T1) if
you wanted to follow hard
and strict rules of drawing
inputs (the primary circuits)
on the left and outputs (the
secondary(s)) on the right.
Once you realize that, its
much easier to visualize
how the circuits actually
operate and the schematic
diagram is way less
intimidating. Of course,
there is no way to draw the
schematic that way in any
sort of acceptable aspect
ratio. It would be way too
wide. Honestly, its a
miracle the engineers were
able to fit it all on one page
and it is, in fact, a very
well-drawn schematic
diagram. Ive just made it
even better by adding the
colored busses so we can
identify, at a glance, the
overall structure of the
unit.
Fault Detection and
Protection
We want to keep an eye on
a few different functions.
We certainly want to look at
the output voltage to make
sure it doesnt go too high.
If something fails and the
24 VDC output climbs as
high as 28 volts, we want to
do something about it
(pronto) before anything
becomes damaged.
Likewise, if the output
voltage drops below 22.5
VDC, wed like to know that
as well. Its not likely to
cause any damage but if
the voltage is dropping, we
will want to start shedding
loads and at the same time,
Slot Tech Magazine

inform the CPU so it can


start an orderly shutdown.
We also measure the
internal temperature as well
(youll see why in a minute).
All of these are analog
measurements but we
really dont need to
measure these actual
values so much as we
simply need to know when
we have crossed a preset
threshold. As long as the
output voltage is between
22.5 and 28, we really dont
care what the actual voltage
is.
This comparison between
two values (our sampled
value and the preset)
perfectly describes the
function of a comparator
and thats what U3 is. Its
an LM339 quad
comparator. The quad
part means that there are
actually four identical
comparators in a single
package so what appears at
first to be four individual
devices is really just one
component.
In a single-ended
configuration like this one,
its operation couldnt be
simpler. There are two
inputs and one output. The
two inputs are labeled +
and -. The + input is also
known as the noninverting input while the
input is also called the
inverting input. The
LM339 compares the two
voltages at the inputs. If the
input is a higher voltage
than the + input, the
SETEC MK5PFC

output pin goes to zero


volts (the voltage at pin 1).
However, rather than
thinking of it as putting
out zero volts its much
better to think of the output
as what it really is, a
connection to ground. Its
an open collector output,
Q8 on the schematic of the
comparator. When the
input is higher in voltage
than the + input, the
comparator is activated
and the output is connected
to ground through the
transistor. Anything
connected to the output pin
will become grounded. It is
a current sink. It is NOT a
current source.
On the other hand, if the +
input is a higher voltage
than the input, the output
pin essentially becomes
disconnected from
everything (Q8 is turned off)
and will be just swinging in
the breeze with its open
collector. Of course, the
voltage on the pin will
swing up to the voltage
determined by what ever
pull up resistor or resistor
voltage divider network we
have on the output. We
dont really even care what
this voltage is all the time.
We can often just think of it
in digital terms as being
high or even just not
grounded and leave it at
that.
With this concept firmly in
place, lets start on the left
side with U15, the LM35C
Temperature Sensor. This
power supply has an
SETEC MK5PFC

Inside the LM339 Comparator


Slot Tech Magazine

Page 15

internal temperature
sensor that is part of the
over-current protection
system. Its is a really
great place to see the
comparator in action. In a
nutshell, if too much
current is drawn from the
power supply for an
extended period of time
(causing the power supply
to overheat) or if the
machine is operating in an
environment that exceeds
the maximum temperature
rating (causing the power
supply to overheat) we
want to turn off the power
supply. U15 has just three
leads: a power supply
input, a ground and an
output. Want to guess how
it works? If you said the
output voltage changes
with temperature youre
right. It operates in a range
of -55 to +150 C. The higher
the temperature, the higher
the output voltage (it rises
at 10 millivolts per degree
Centigrade). You can see
the output is connected to
pin 10, the input of the
LM339, U3A.
For the other input to the
comparator, we generate a
+2.5 VDC reference voltage
using a TL431. This
reference gives us a precise
and unchanging voltage to
which we can compare
other voltages. In this case,
we set the trip level of the
circuit with a voltage
divider made from R124
and R231.
As long as things are cool,
the input voltage is lower
than the + input and the
Page 16

comparator output sits at


about 3 volts (R142 and
R111 form a voltage divider
that does this). However, if
it becomes too hot inside
the power supply, the
voltage at the input will
exceed the reference voltage
at the + input and the
output at pin 13 will go to
ground.
This low signal is felt at the
input of U3B at pin 6. The
+ input of U3B is tied
directly to the 2.5 VDC
reference voltage so as the
input drops from 3 volts to
zero, the output voltage of
U3B (pin 1) swings high.
This signal is connected to
the + input of U3C at pin 9.
As the voltage rises from 0
(which is what it will be if
the temperature is OK) and
passes the +2.5 VDC
Slot Tech Magazine

reference (connected to the


input of U3C at pin 8) it
will trigger the output of
U3C at pin 14 to go high as
well. As you can see, the
output of U3C is connected
to the gate of MOSFET Q10
so this high signal will turn
on the device.
Please remember that what
is REALLY happening here
is that the gate of Q10 is
normally HELD DOWN TO
ZERO VOLTS by the
ACTIVE output of U3C. It is
only when the temperature
RISES that the output of
U3C goes high (meaning its
internal open collector
transistor is now turned off)
and the gate is allowed to
be pulled high by resistor
R229. This negative logic
is carried out throughout
the design of this power
supply. The active devices
SETEC MK5PFC

are almost never sourcing


current (Q6 being the
notable exception as it
sources the Vdd for U12).
The current source instead
comes from some sort of
power bus, through a
resistor or some resistors as
a voltage divider. We shunt
this current to ground (or 0
volts or whatever you want
to call it here) when the
comparators input
voltage exceeds the voltage
at the + input.
So, after all of this, here is
where we stand: If the
temperature is normal, Q10
will be off. If it gets too hot,
Q10 will be turned on.
What is Q10 doing and why
is it so important in this
chain of events? The drain
of Q10 is connected to the
cathode of the LED in optoisolator U5. When Q10
turns on, the LED lights up.
Directly above the LED half
of U5 is the phototransistor
half of U5, connected (as
you can see by the red
checkered return path) to
the primary side of the
power supply. Now were
getting somewhere because
we have not only detected
the high temperature
condition but we have a
way to tell the primary side
of things (where all the
action is!) that we have a
problem and that it might
be a really good idea to shut
down before things get any
hotter.
So turn your attention now
to the phototransistor half
of U5 and the + input of U2
SETEC MK5PFC

to which it is connected. As
you can see, if the
phototransistor in U5 is
turned on by the light from
the high temp LED half of
U5, its going to drag the +
input down to zero volts
and, since the input will
now be higher in voltage
than the + input, the
output of U2D at pin 13 will
go low. This will drag down
the gate of MOSFET Q4,
turning it off and when that
happens, finally, at long
last, we arrive at the final
goal of this circuit.
When Q4 is turned off, the
relay, K1, drops out. Do you
remember K1 from part one
of this discussion? Relay
contact K1 shorts out the
18.2 Ohms of series
resistance on the AC input.
If the temperature rises too
high, the relay drops out
and the power supply
enters an operating mode
that places resistors in
series with the AC input
once again.
Its also time to shut down
the power supply. This is
easily accomplished because U14, the UCC38503
PWM Controller has an
Enable input. Pin 4 has to
be high (typically, its
around 8 volts) for the
device to operate. All we
have to do is to drag pin 4
low and the GT2 output at
pin 10 turns off, turning off
the entire primary circuit.
Of course, the +24 VDC
output goes down as well.
As covered previously, when
an over-temperature condiSlot Tech Magazine

tion occurs, U2, pin 13 goes


low. The cathode of diode
D28 is connected to this
pin. The anode of D28 is
connected to the Enable pin
of U14. When U2, pin 13
goes low, it drags down the
Enable pin and shuts down
the power supply.
OVP
In this same area, we find
the over-voltage protection
circuit as well. Like the
Over-Temperature control,
its U3. Its the one remaining section of U3, input
pins 4 and 5 with the output at pin 2. Like an overtemperature condition, an
over-voltage condition in a
power supply needs to be
dealt with swiftly and forcefully. Over-temp might
destroy a power supply but
a severe over-voltage condition has the potential of
creating a lot of damage all
over the system so we need
to take care of it right away.
Pin 4, the inverting input of
the comparator, is always
looking at the +24 VDC
output of the power supply.
Its the green colored buss
on the schematic diagram
(if you dont have last
months schematic handy,
you can download a copy
from tinyurl.com/mk5pfc).
It does it through a voltage
divider made from resistors
R113, R150 and R216. I
really like the way these
engineers think at SITEC
because the total resistance
of all three of these resistors is very close (less than
5% away) to 24K ohms.
Page 17

That makes it easy to see


that you have one volt per
ohm, making the voltage
(when everything is normal)
at pin 4, 2.2 volts. Compare
that to the 2.5 volt reference and you can see that
when all is well, the output
of U3, pin 2 will be high
and we will not be turning
on the LED in U5. As you
recall from the Over-Temperature Sensor discussion,
as soon as we energize the
LED in U5, the power supply is going to shut down
and you can easily see now
how thats going to happen.
When the power supply
output rises above +28
VDC, the voltage at pin 4
rises above 2.5 volts, making the - input higher in
voltage than the reference
at the + input and forcing
the output at pin 2 to go to
ground. Of course, this
completes the ground connection for the LED in U5,
turning it on and subsequently (and immediately)
turning off the power supply as explained earlier in
the Over-Temperature Sensor discussion.
Recap: Bad things such as
over-temperature or over
voltage will turn on U5.
Lets go back to relay K1
and recall that when bad
things happen to the power
supply and energize U5, K1
drops out and no longer
shorts out the 18 ohms
worth of resistance that is
in series with the mains at
the AC input. Lets see how
the power supply is going to
react to this series resisPage 18

tance.
Naturally, putting 18 ohms
in series is going to have a
noticeable effect on things.
Specifically, the unregulated primary voltage is
going to drop. This is where
having the colored busses
really comes in handy because if you follow the sort
of pink colored primary DC
buss all the way to the left,
youll see a voltage divider
made of seven resistors,
R199, R214, R213, R121,
R122, R154 and R156. Our
divided voltage is applied to
pin 6, the inverting input of
one of the four devices
inside integrated circuit U2,
an LM339.
You can see from the fact
that these guys are all .5%
resistors that this is something important. This is
where we keep an eye on
the primary DC voltage.
This is another comparator
circuit. Our reference voltage is at pin 7, connected to
the 2.5 volt reference buss
through precisely 1000
ohms (again, .5% precision
resistors are used).
It is worthy to note that the
12 volt power supply for the
LM339 U2 and the reference voltage generated by
U8, the TL431 comes from
U13, the three-terminal
regulator we discussed in
part 1. Because the 12 volt
output of this regulator is
so low in comparison to the
+360 VDC or so that is its
input (when the mains
input is 240 VAC) the output of the regulator will
Slot Tech Magazine

remain perfectly stable,


regardless of the fact that
we may have dropped the
primary DC voltage somewhat with the inclusion of
the 18 ohms of series resistance or even in the case of
a serious problem with the
mains input. This +12 volt
supply will remain up and
perfect for quite some time,
even as the world collapses
around it as the input only
needs to be a few volts
higher than the output in
order to function perfectly.
Normally (when everything
is operating properly) the
voltage at pin 6 is higher
than pin 7, activating the
output and dragging pin 1
down to ground. Notice that
it is the hot ground on the
primary side since thats
what were looking at here.
This turns on the LED half
of opto-isolator U6.
When we throw the 18
ohms of resistance in series
with AC input, the unregulated primary voltage drops.
The voltage at U2, pin 6
drops below the reference
voltage and the output at
pin 1 will swing high, leaving the LED without a return path, its cathode
swinging in the electronic
breeze without a path to
ground. Of course, this
turns off the LED.
Resistor R160 straps the
output voltage back into the
non-inverting input, assuring that the output at pin 1
remains latched off and
that the circuit doesnt
reassert itself. The specific
SETEC MK5PFC

circumstances of this
reassertion do not warrant
discussion in this article.
So to recap, if the primary
voltage is good, U6 LED is
on. If the primary voltage
begins to drop (whether
caused by the 18 ohms of
series resistance OR by an
actual drop in primary
voltage caused by a significant drop in the mains
voltage) U6 LED is off.
If were driving the LED half
of an opto-isolator we must
be getting ready to talk to
the outside world. We are,
but in a surprise move, all
of this primary voltage
detection that we just performed has absolutely nothing to do with the actual
power supply. That is to
say, it has nothing at all to
do with the generation of
the +24 VDC output of the
power supply. It is a detection circuit that talks to the
slot machine itself. It examines the condition of the
primary DC voltage (from
which it infers the condition
of the mains-when we add
the series resistance, we are
actually tricking the
power supply into thinking
that the mains voltage is
dropping) and signals the
CPU in the slot machine
through the status of the
Power Fail signal. You can
see it on the schematic
diagram at pin 9 of connector X9. Follow it back and
youll see its connected to
the drain of Q2 so obviously, when Q2 turns on,
the PFAIL signal will go low.
SETEC MK5PFC

Well back up and look at


the entire circuit in just a
moment but I want to
pause and reinforce this
concept to those of you who
have been wondering just
how it is possible that you
can have a power supply
that works perfectly on the
workbench, producing a
perfect +24 volt output
under a massive test load
but fails to operate in a slot
machine. This circuit is
totally separate from the
+24 VDC power supply. The
circuit lives in this box and
on this PCB because this is
the only place where the
slot machine actually physically touches the mains
and, as you have seen and
as you will continue to see
in this article, we need to
watch the mains so that in
case of power failure, we
can tell the slot machine to
store data and perform an
orderly shutdown. This is
one of the ways we do it.
This circuit really has nothing at all to do with generating the +24 VDC output.
It just lives in the same box,
looking at the primary DC
voltage and telling the CPU
what is happening through
the Power Fail output
signal.
Lets go back to opto-isolator U6. This time, lets look
at the transistor half. We
are now on the isolated,
secondary side of course,
where everything in the slot
machine is all grounded
together.
This is an interesting design
in that you see the transisSlot Tech Magazine

tor in the source path


rather than simply acting
as a ground switch. If the
transistor in U6 is on, the
+24 VDC output buss is
connected through the
transistor to pin 6, the
inverting input of U1. A 6.8
volt Zener diode pegs the
voltage at no higher than
6.8 volts. I have an engineering issue with this
circuit as there is no resistive current limiting here
other than semiconductor
junctions. I am going to run
this by the engineers at
SETEC for comment.
This circuit compares the
voltage at pin 6 to the reference voltage at pin 7, the
non-inverting input. As long
as the power is good, pin 6
will be higher than pin 7
and the comparator will be
active, its output voltage
dragged to ground at pin 1.
When the power fails, U6
turns off and the voltage at
the non-inverting input (the
reference voltage) will exceed the voltage at pin 6.
The comparator turns off
and pin 1 is pulled up to
the +24 volt buss by R187.
Of course, you can see that
what we are really doing
here is controlling the gate
of MOSFET Q9, which is
connected directly to the
comparator output at pin 1.
When everything is normal,
pin 1 is low and Q9 is off.
Upon power failure, Q9
turns on.
Now, its on to the next
stage. Its also U1, this time
pins 4 and 5 are the inputs
Page 19

and pin 2 is the output.


Our reference voltage is
connected to the + input
while the voltage at the -
input is determined by Q9.
If Q9 is off (normal operation) the voltage at the
inverting input will be
higher than the reference
and the comparator output
at pin 2 will be low. Upon
power failure (when Q9 is
turned on) the voltage at
the inverting input drops
and the comparator turns
off, allowing the voltage at
pin 2 to rise, pulled up by
R179. This, of course, gates
MOSFET Q2 and when Q2
turns on, it generates the
Power Fail signal, more
correctly referred to as
Power Fail Not as it is an
active low output. Youll
notice that the signal is
labeled PF with a bar over
the top. The bar on top
indicates an active low
signal. Oddly enough, the
same signal is labeled
PFAIL at the connector,
without the bar over the
top. I think thats just an
oversight.
Thats it. That is the whole
point of that circuit. You
can see that it has nothing
to do with the creation of
the +24 VDC output nor
does it have to ability to
perform any sort of shutdown of the power supply
on its own. It can only talk
to the CPU. Its a sort of
tattletale circuit, a Rat
Fink, a Stool Pigeon, a
squealer (I gotta stop
watching old prison movies).
Page 20

Output Monitor
While were in this area,
lets take a look at another
circuit that is used to monitor something, this time the
+24 VDC output itself. We
have already seen how the
power supply will respond
(immediately) to an overvoltage condition by shutting down the power supply
before damage can occur.
On the other hand, in the
case of a loss of the +24
VDC output, we dont have
to worry about damage and
we dont have to shut down
the power supply. Its already going down! However,
if the +24 VDC output isnt
up and running (or, more
specifically, if it is in the
process of failing during
otherwise normal operation)
it would be nice to let the
CPU know about it so it can
do some housekeeping.
The output monitor uses
the remaining two sections
of U1, the LM339 Quad
Comparator. At first glance,
it may look like this is some
sort of circuit without an
input. You can see that the
non-inverting input at pin 9
is connected to the reference buss but whats up
with the - input at pin 8?
It looks like its just hooked
up to the +24 VDC buss
through some more of those
.5% precision resistors,
R173 and R208 (both 1.8 k
in series for a total of 3600
ohms) and R181 and R209,
both 2k ohms in parallel for
a total of 1000 ohms. The
use of precision resistors
tells us were monitoring
Slot Tech Magazine

something important here


and in this case, were
looking at the +24 VDC
buss itself, the same buss
that is powering U1. Thats
our input!
Of course, first semester
electronics students learn
all about the voltage divider
and we have sure seen a lot
of them in this power supply. This is another classic
example of why we learn
about this stuff in school.
The problem is that in
school, we arent shown
WHY or WHERE we use
voltage dividers. This is a
beautiful (and easy to understand) example of voltage division using resistors.
When the output voltage is
+24 VDC, the voltage at pin
8 will be 5.2 volts. This is
higher than the reference
voltage at pin 9 and so the
comparator is active; the
output at pin 14 is zero
volts. You can see that this
output is connected to the
gate of MOSFET Q8. As
long as the output voltage
of the power supply is
greater than 22.5 VDC, Q8
remains off. However, if the
output voltage drops below
22.5 VDC, the reference
voltage at the + input will
be the higher than the
voltage at the - input and
the comparator turns off.
Resistor R189 pulls up the
gate and Q8 turns on.
Recap: +24 VDC output
normal, Q8 off. Output less
than +22.5 VDC, Q8 on.
When Q8 turns on, it completes the ground connecSETEC MK5PFC

tion for resistor R195. What


this is really doing is dragging down pin 10 (as part of
a voltage divider-again!), the
inverting input of the following comparator stage.
Its the normal deal as weve
seen again and again in this
design: pin 11 (the + input)
is tied to a reference voltage
and the condition of the
input determines the output. When the voltage at
pin 10 drops below the
reference voltage at the +
input, the output at pin 13
swings high and gates the
following MOSFET, in this
case Q3.
This creates our Output
Fail signal. Like the Power
Fail signal, its an active
low signal that speaks
directly to the CPU through
a dedicated connection at
X9, pin 2. If the output
voltage drops below 22.5
VDC, this output goes low.
The CPU then decides what
to do next. Again, this signal actually has nothing to
do with the generation of
the +24 VDC directly nor
does it have any sort of
protection or shut-down
facilities of its own. If there
is any shutting-down to do,
these two circuits (Power
Fail and Output Fail) dont
do it directly. They can only
tell the CPU what is happening.
Remember the Mains!
Of all the things that can go
bad while a machine is in
operation, loss of AC power
is by far the most common
occurrence. From the
SETEC MK5PFC

machines point of view,


this happens a lot. It happened a bunch of times
before the machine ever
made it to the slot floor!
Every time the power switch
is turned off or the machine
is unplugged or disconnected from the mains
(accidentally or otherwise)
the machine (obviously)
loses power but as we have
seen, the entire system
doesnt actually lose power
immediately. Because of the
energy stored in the electrolytic capacitors, we have a
little bit of time to put
things in order before we
lose power completely.

compromise and thats


what the MK5PFC does. It
looks at the AC input and
after around a half-dozen
missing cycles decides that
that there is a real and
persistent loss of AC power
and it does something
about it.
And what it does is pretty
darned interesting and now,
since you already know all
about how all the other
stuff in the power supply
operates, we can see how
the AC input detection
works and what it does to
the power supply.

What we need then is a

Lets start with the detection circuit. We want to


know when we have lost AC
power so we use a single
diode (D3, a common
1N4007) to look at the
positive half cycle of the
waveform. Well get 50 or 60
pulses a second here. The
voltage of this pulse is
divided down to a more
reasonable level and applied to the inverting input
of U2 at pin 8. The trick
here is that we want to be
able to miss one or two or
even a few pulses without
triggering the CPU to begin
a machine shutdown. This
is accomplished by hanging
an electrolytic capacitor
(C67-22uf) from pin 8 to
ground. The larger the
value of the capacitor, the
longer the time constant
and so this value was chosen to allow the system to
remain operational with the
loss of a few cycles but
trigger the Power Fail
signal when there is a real

Slot Tech Magazine

Page 21

Our goal here is to detect


when we have lost the AC
input power right away,
before the +24 VDC output
of the power supply fails.
However, we dont want to
jump the gun and start
shutting things down too
soon. We dont want the
loss of a single cycle of AC
(or even two or three cycles)
to trigger a shutdown. A
momentary loss of AC
shouldnt affect the machine if the +24 VDC output
remains perfect. On the
other hand, you dont want
to delay in starting the
shutdown procedure if the
problem is a genuine loss of
AC power. By the time the
output of the power supply
really is affected, it may be
too late to do anything
about all the data in the
CPU if we havent completed housekeeping before
the power supply craps out.

loss of AC power as the


voltage on C67 (and pin 8)
will quickly bleed off
through R212 and the
input structure of U2 when
the AC goes missing. As the
voltage on pin 8 drops
below the reference on pin
9, the output at pin 14 goes
high, gating MOSFET Q11.
At this point, a couple of
things are going to happen
more or less simultaneously. One is that the
unit will generate the
Power Fail signal that tells
the CPU to store data and
shut down properly. The
other is that the power
supply itself will shut down,
killing the +24 VDC output.
Lets look at generating the
Power Fail signal first.
This is easy because weve
already seen how a drop in
the primary DC voltage will
trigger the Power Fail
signal. We just have to see
how the two circuits are
connected and the way
theyre connected is by
diode D26. The anode is
connected to the - input
at U2, pin 6. This is the pin
that is also watching the
primary DC voltage. The
cathode of D26 is connected
to the drain of MOSFET
Q11. When Q11 is energized (due to a loss of a
half-dozen cycles of AC
power) the cathode of D26
is essentially grounded,
dragging the inverting input
at U2 pin 6 down in voltage.
The result is precisely the
same as it is when the
primary DC voltage drops,
eventually resulting in the
Page 22

active low Power Fail


signal being generated.
At the same time, when AC
power is lost and Q11 has
energized, it will drag down
the voltage at U2 pin 4, the
inverting input. This will
turn off the comparator,
allowing Q12 to turn on (it
gets its gate voltage from
R232 and R233).
Once Q12 is turned on,
were really going to kill
some things. Firstly, take a
look at diode D27. The
cathode is connected to the
drain of Q12 and anode is
connected to our old friend,
the Enable input of U14,
the UCC38503 PWM controller IC. When Q12 is
turned on, were going to
kill the +24 VDC power
supply immediately. We are
going to turn off all primary
activity and thats a good
thing because we dont
want to waste any of our
precious stored energy
messing around. By
grounding the Enable input
through Q12, we turn off all
PWM activity and now rely
on the charge stored in the
secondary electrolytic filter
capacitors.
Secondly, you can see
through the connection of
diode D39 between the
exact same drain of Q12
and the gate of Q4 that
when Q12 turns on due to a
loss of AC power, it will also
drop out relay K1. You can
go right back to the very
beginning of this discussion
if you need to refresh yourself on all that that entails.
Slot Tech Magazine

The Stuff That Fails


So, thats how the unit
works. Its hard not to call
it a complex system full of
circuits that create, control,
regulate, signal and shut
down but broken up into its
individual functions (and
with familiarity and an
annotated schematic diagram) it becomes manageable. This is not a cheap,
disposable power supply
but they can fail from years
of 24/7 operation inside a
hot slot machine. Somebody has to repair these
things and now that you
know all about the unit,
that someone can be you.
On the other hand, many
technicians (me included)
are just as happy to have a
list of the stuff that fails as
we are to have a detailed
and exact knowledge of how
the system operates. As a
technician, my job is to fix
stuff as quickly and accurately as possible. My job is
not to prove to my co-workers how smart I am. If
someone can tell me what
to replace, I am happy to
have the advice and if I
have it, I am happy to share
it.
To that end, I asked the
folks at Aristocrat and
SETEC to name the top
dozen faults for MK5PFC.
You can now totally forget
everything you just learned
about this remarkable
power supply.
- STM
SETEC MK5PFC

1944 Falmouth Dr. El Cajon, CA 92020 tel.619.593.6131 slot-techs.com

MK5PFC High Failure Items


Low output voltage -> C89 leaky
Unit unstable -> C52 failed
No power up -> C66 leaky
T1 noisy
No power up -> Q1 failed
No power up -> U7 failed
No power up -> R127 open circuit
R77, R202 wire wound
U13 Voltage Regulator
U14 Surface Mount IC
D1 Bridge Rectifier
D11, D12, D43, D50 Diodes
Pin pushed back in X3
S1
Monitor fuse holder damaged in transit, no monitor O/P

SETEC MK5PFC

Slot Tech Magazine

Page 23

On behalf of Table Mountain Casino I just wanted to express our thanks


to you and your team. I couldnt have asked for anything better.
Brian Rankin - Slot Technical Manager

Camel Rock Casino

On-Site Slot Tech Training


Customized Classes Available

Page 24

Slot Tech Magazine

SETEC MK5PFC