This is only a preview of the September 2025 issue of Practical Electronics. You can view 0 of the 80 pages in the full issue. Articles in this series:
Items relevant to "Compact Hi-Fi Headphone Amplifier, part one":
Articles in this series:
Articles in this series:
Articles in this series:
Articles in this series:
Articles in this series:
|
Constructional Project
Discrete Ideal
This deceptively simple circuit uses just a handful
of transistors, diodes and resistors. But it still
provides a very useful function: active rectification of
the output of a centre-tapped transformer or combining
two DC supplies with low losses. It is much more efficient
than a bridge rectifier or diodes at higher currents, producing
less heat without costing much more.
Project by Phil Prosser & Ian Ashford
T
he Ideal Diode Bridge Rectifiers
project, published in the November 2024 issue, included six different
PCB designs to suit different situations.
It was popular, with many built, but
two aspects of that design bothered
me (and others).
Firstly, it used a pretty expensive
custom IC, with the SMD version being
a bit tricky to solder. Secondly, despite
that expense, it could only handle rectifying the output of a single transformer
secondary. So you couldn’t use it at all
with a centre-tapped secondary, and
two complete boards were required to
derive split rails from a transformer
with separate secondaries, doubling
the cost.
Wouldn’t it be nice to have a direct
drop-in replacement for a bridge rectifier that could handle single, dual or
tapped secondaries? And it’d be great
to use standard parts, so we don’t need
to source that expensive IC.
Reader Ian Ashford sent us a circuit design he uses for dual-rail rectification but didn’t have a PCB design.
When the Editor asked me if I wanted
to turn it into a full-on project, there
was only one possible answer to that!
Ian and I performed further testing,
development and tweaking, finally arriving at this very flexible, robust and
useful circuit.
So, this project is a collaboration
that follows the ideal rectifier theme
but with a different focus from the previous design.
When to use this design
As well as rectifying a transformer's
output(s), this design is also suitable
for combining DC supplies with low
losses, eg, combining the output of a
solar panel and a battery, or a solar
panel and wind generator.
While it costs a little more than a
bridge rectifier to build, it is significantly more efficient at higher currents
and has a much lower voltage loss. So
it’s ideal for high-power devices like
power supplies and audio amplifiers.
Its only real drawbacks are a limited voltage handling capability (up
to ±40V or +80V) and the fact that it’s
larger than a 35A bridge rectifier, so
you’ll need room to fit the PCB.
This project uses high-current, low
RDS(on) Mosfets. To keep the circuit
simple, we have used P-channel Mosfets on the positive rail and N-channel
Mosfets on the negative rail. If your
current demands are only modest, you
could use the ubiquitous IRF9540 (Pchannel) and IRF540 (N-channel) power
Mosfets, which are available from Farnell, among other retailers. They can
handle up to about 5A.
Much more significant currents can
be handled using the devices in the
parts list, which are not all that expensive but are unlikely to be available from your local shop (but kits are
available). All the other parts in the
design are bog-standard, and you will
surely have them in your parts drawer
or at your local shop.
Design process
Between Ian’s initial email with the
circuit he uses in DC and low-power
Figs.1(a) & (b): the two main ways to use the Discrete Ideal
Bridge Rectifier. At the top, a centre-tapped transformer
secondary winding is used to generate split (positive and
negative) rails. Two separate secondaries can also be used if
they are connected in series. The connections at right show
how to use the same board to combine the outputs of two DC
supplies (the solar panel and battery are just examples). OUT+ will be fed by whichever has a higher voltage.
18
Practical Electronics | September | 2025
Discrete Ideal Bridge Rectifier
Bridge Rectifier
» Generates split rails (positive and negative DC supplies) from a single
centre-tapped transformer secondary (or two secondaries wired in
series)
» It can also be used to combine two DC supplies (whichever has a higher voltage
feeds the load)
» Maximum output voltage: ±40V or +80V (transformer applications), +40V (combining
DC supplies)
» Maximum current: 10A RMS without heatsinking, more with heatsinking
» Typical voltage drop: <100mV input-to-output
» Typical dissipation: 1.7W <at> 5A RMS, 6.8W <at> 10A RMS
AC applications and the final design
presented here, we exchanged many
ideas, questions and refinements. Some
requirements we decided on are:
; A low part count was important.
; The design had to ‘just work’ without tweaks.
; Reverse current when Mosfets
switch on and off had to be minimal
in all applications.
Many ideas were shared, and challenges were presented in every direction. In the process, the conceptual
circuit grew to something larger and
more complicated than was strictly
necessary. It was at this stage that we
tabled those design goals.
Ian was keen to keep the size of the
board down, so we designed a throughhole version and an alternative that uses
some SMDs to fit in tighter spaces. We
realised that this would never be the
size of a conventional bridge rectifier,
so we just aimed to produce reasonably-
sized boards that would likely fit into
an existing chassis but that aren’t too
fiddly to build either.
The final design is vastly ‘tighter’
than the test board. I often build a prototype board that is purely functional
and worry about improving the layout
later, once I’ve proven it works.
In discussing what changes were
warranted to Ian’s concept, achieving
a design that ‘just worked’ became important. That led to the introduction
of constant current sources as loads
in the design. It makes the operation
largely independent of supply rail voltage and allows constructors to use the
Ideal Bridge with 9-25V AC transformers without any changes.
We also changed the sense circuit
Practical Electronics | September | 2025
to only switch on the Mosfet when
the input is at a programmed voltage above the rectified output. This
blocks reverse currents and allows it
to be safely used for combining DC
supplies, which might be very close
in voltage at times.
The resulting circuit is simple and
works well. We'll get to a couple of
subtleties later in the description,
once we’ve gone over its operating
principle.
Two versions
There are two PCBs for this project:
an SMD version and a through-hole
version. They use the same circuit.
The SMD version is smaller than the
through-hole version, which may be
helpful in some circumstances. It
doesn’t use any tiny parts (the resistors are M3216/1206 and there are
SOT23 transistors), so it isn’t hard
to assemble.
Both versions use the same TO-220
(through-hole) Mosfets. That is because
it makes it easy to add flag heatsinks if
necessary for your application. Highcurrent SMD Mosfets are available, but
they are trickier to heatsink if necessary and will take up more room than
a TO-220 in this application.
Design limitations
This circuit is suitable for rectifying the output of dual or split secondary transformers where the junction of the windings from the ground
point for output capacitors, as shown
in Fig.1(a).
This design will work if you have
a transformer with a single secondary
winding, but the switching could be
noisy. ICs like the LT4320 used in the
November 2024 designs switch the
bottom Mosfets on for a full half-cycle
to ensure clean switching. So, for that
sort of application, we recommend you
build one of the designs we published
then (kits are available at siliconchip.
au/Shop/?article=16043).
Regarding how much current the
board can handle, P-channel Mosfets
typically have a higher RDS(on) figure
than N-channel Mosfets. This means
that the positive-rail Mosfets will be
the limiting factor in how much current can be drawn due to their voltage drop and consequent power dissipation.
We have avoided the complexity
of a gate drive boost circuit there.
Using one would have allowed us to
use four identical N-channel Mosfets,
but we didn’t think that was worth
the extra parts and possibly new failure modes.
Up to about 10A, the Mosfets will
not require heatsinking, although it
wouldn’t hurt to add small flag heatsinks above 5A. Above 10A, you must
add a substantial flag heatsink on each
Mosfet. Decent flag heatsinks should
let it handle at least 15A. Beyond that,
you might need a more serious cooling solution, like forced airflow over
heatsinks.
Circuit details
The circuit is shown in Fig.2. Unlike
the previous Ideal Bridge Rectifier, this
circuit can have its inputs connected
across a single secondary or a pair of
series-connected secondaries to generate split supply rails. In those cases, the
secondary winding’s centre tap does
19
Constructional Project
not connect to this circuit. Instead, it
connects to the output capacitor bank
ground and the load’s ground, as shown
in Fig.1(a).
So that it can produce split rails, it
contains two similar sections stacked
on top of each other. They would be
identical except that they have opposite polarities to handle current flowing
in opposite directions. The upper section uses two P-channel Mosfets, four
PNP bipolar junction transistors (BJTs)
and two NPN BJTs. The lower section
has two N-channel Mosfets, four NPN
BJTs and two PNP BJTs.
Each of the four sections senses the
input AC voltage at one terminal. When
it is about 34mV greater in magnitude
than the output voltage (higher than
the positive rail or lower than the negative rail), the corresponding Mosfet is
switched on by driving its gate with
an appropriate voltage.
We only want the Mosfet on when
the input exceeds the output by a small
margin to ensure that the Mosfet is off
when these voltages are equal and that
there is no chance the Mosfet is on as
the input voltage magnitude drops
below the output.
If that were to occur, current would
reverse and flow from the capacitor
bank through the transformer, creating current spikes and a great deal of
electrical noise, plus possibly overheating the Mosfets.
DC+
Q1 SUP70101EL
AC1
D
S
68 W
AC2
D
S
68W
G
G
A
A
C
B
B
K
A
A
K
A
D4
1N4148
(WS)
K
K
ZD2
Q 10
Q9
12V
B
C
x
5
6
B
C
x
5
6
E
E
A
C
D5
1N 4148
(WS)
ZD1 12V
K
A
D2
1N 4148
(WS) K
K
100k W
Q6
Q5
E BCx56 BCx56 E
D1
1N4148
(WS)
D3
1N4148
(WS)
Q2 SUP70101EL
47kW
C
D6
1N 4148
(WS)
TG1
B
B
K
100k W
A
C
TG2
22k W
22k W
2 2 kW
22kW
CURRENT SINK
C
Q7
MPSA42/
MMBTA42 E
SUP70101EL,
IRFB4410ZPBF
AC IN1
CON3
G
ZD1– ZD4
B
C
D
D
B
E
330 W
S
Q8
MPSA42/
MMBTA42
(BZX84C12)
MMBTA42, MMBTA92,
BC846, BC856
1N4148WS
K
C
K
A
B
A
E
DC OUT +
DC–
AC IN2
AC2
AC2
CON1
DC+
CON4
330 W
COMPONENTS IN
THIS AREA ARE NOT
REQUIRED FOR
COMBINING DC
SUPPLIES
E
B
Q15
MPSA92/
MMBTA92
E
B
C
A
D9
1N4148
(WS)
BC546, BC556
E
K
C
E
B
C
CON2
E
C
B
Q19 E
Q 17
BCx46 BCx46 100kW
A
D11 A
1N4148
(WS) K
K
2 2 kW
22kW
C
B
DC OUT –
MPSA42, MPSA92
B
A
22k W
K
A
ZD4 12V
K
ZD1– ZD4,
D1-D12
CURRENT SOURCE
22k W
D7
1N4148
(WS)
C
Q 16
MPSA92/
MMBTA92
C
K
D8
1N 4148
(WS) A
47kW
68 W
E
D10
1N 4148
(WS)
B
Q 20
Q 18
BCx46 BCx46
A
A
D12
1N4148
(WS) K
K
K
E
ZD3
12V
100k W
A
68W
G
G
C
B
AC2
AC1
D
S
Q3 IRFB4410ZPBF
S
D
Q4 IRFB4410ZPBF
DC–
SC DISCRETE IDEAL BRIDGE RECTIFIER
Ó2024
Fig.2: the Ideal Bridge Rectifier circuit comprises two identical sections at the top to deliver current to the DC OUT+
terminal, with two more sections below to handle current flow through the DC OUT− terminal. The lower sections are
‘mirror images’ of the upper sections, with components of opposite polarity (NPN transistors instead of PNP etc). The
circuit is the same for the TH and SMD versions; the alternative devices are direct equivalents except for their packages.
20
Practical Electronics | September | 2025
Discrete Ideal Bridge Rectifier
As the four separate sections all
work the same way, let’s concentrate
on the one shown in the upper-left
corner of Fig.2. The voltage sensing
circuit comprises diodes D1 and D2
plus PNP transistors Q5 and Q6. Q5
acts as a diode, since its base and collector are joined.
Ignoring the 68W resistor for now,
with a constant current flowing
through these transistors, both will
conduct if the AC input voltage at
CON3 is the same as the DC output
voltage at CON1. If Q6 is on, Mosfet
Q1's gate voltage is high, and it is
off. As the input voltage increases,
Q6 switches off, so the gate of Mosfet
Q1 is pulled low by its 22kW collector resistor – see Scope 1.
The 68W resistor is important as it
alters how the comparator works. The
total current through the two 22kW
collector resistors is determined by a
constant current sink comprising NPN
transistors Q7 and Q8. On the positive
cycle for the AC1 input, about 0.5mA
is drawn through each of these resistors (as well as the matching pair for
Q9 & Q10).
This 0.5mA flows through the transistor and diode pairs Q5/D1 and Q6/
D2, which drop the voltage by about
1.2V, but on the AC input path, it also
flows through the 68W resistor, dropping 34mV or so in the process.
This extra voltage drop means we
draw more current from the base of Q6
than Q5 until the AC input is 34mV
above the output voltage. Mosfet Q1
remains switched off until that condition is met. Once the input exceeds the
output by 34mV, Q6 starts switching
off and the Mosfet switches on. This
charges the output capacitors until they
get to 34mV below the input.
Essentially, the circuit contains a
negative feedback loop, where Q5 and
Q6 try to maintain a 34mV difference
across the Mosfet by controlling its
gate voltage. Without the 68W resistor,
they would try to maintain 0V across
the Mosfet, and due to various tolerances in the circuit, the Mosfet might
be held on all the time, which is not
what we want!
As a result, at lower load currents,
we are not simply switching the Mosfet
hard on and off; instead, it is operating
in linear mode with a low voltage drop
across it due to the negative feedback.
Part of that voltage drop is a result of
the RDS(on) of the Mosfet, while part is
from the gate voltage being moderated,
Practical Electronics | September | 2025
Scope 1: an
oscilloscope grab of
the Ideal Bridge in
operation, showing
rectification of the
voltage at the AC1
terminal. The pink
trace is the output
voltage at 5A, cyan
is the AC input
voltage, and yellow
is the gate drive for
Mosfet Q1, which
peaks at about -8V.
Scope 2: the
Discrete Ideal
Bridge starting
into two 35,000μF
capacitor banks.
This is a pretty
brutal thing to do to
any bridge. Usually,
you would use a
soft-start circuit
to keep the initial
current surge under
control. Still, the
Bridge survived it!
which we can see in the oscilloscope
screen grabs.
As the load current increases, we
see the sense circuit driving the Mosfet
harder, ie, its Vgs increasing until it is
12V, at which point the gate protection
zener diode (ZD1 in this case) conducts
to prevent the Mosfet gate from being
driven beyond its ratings.
If you look at the scope images (especially Scope 4), you will see that when
drawing high currents, the circuit transitions from the linear feedback operation to driving the Mosfet fully on with
12V. This occurs because the voltage
drop across the Mosfet exceeds 34mV
due to its minimum RDS(on).
As a result of the way we are driving the Mosfet, there is little value in
utilising ultra-low RDS(on) Mosfets in a
dual-rail bridge. 10mW or so is fine. We
felt this was the sweet spot at which
the voltage drop across the Mosfets
is defined by the feedback loop up to
about 5-6A. Because of how Mosfets
are made, P-channel Mosfets tend to
have a higher RDS(on).
The constant current sink based
around Q7 & Q8 is a standard two-
transistor current source/sink configuration. We could have tied this to
the output ground and reduced the
dissipation in transistor Q7, but we
chose to tie it to the negative output
rail for the positive rail comparators
and positive rail for the negative
comparators.
This is because it gives maximum
gate drive to the Mosfets for low-voltage
operation, especially during startup
when massive currents are often drawn
for charging capacitor banks. This
This version of the
Ideal Bridge Rectifier
uses all through-hole
components. Note the
pairs of transistors joined
face-to-face so they track
thermally.
21
Constructional Project
Scope 3: a close-up of
the rectified output.
Again, pink is the
output, cyan is the
AC input, and yellow
is the gate drive. This
neatly shows the
Mosfet switching as
the AC input voltage
slightly exceeds the
DC output voltage.
Scope 4: the negative
rail behaviour, which
is a ‘mirror image’
of the positive rail.
Driving the rectified
12V AC into a 1W
load is clearly giving
the transformer a
workout, as seen by
the flattened top and
bottom of the cyan
waveform. Under
these conditions, it
would be advisable to
mount a flag heatsink
to each Mosfet as
they individually
dissipate about 1.5W.
reduces dissipation in the Mosfets
during these high-stress phases of operation. It also has the benefit of the
PCB not needing a GND connection.
If you look at Scope 2, which shows
the startup behaviour, the Mosfet has
over 10V of gate drive in the first cycle
of operation. One benefit of using a
constant current source/sink is that the
circuit’s behaviour is mostly independent of the operating voltage, as long as
it’s above the minimum threshold required to bias on the Mosfets.
The 22kW resistors in the circuit
allow one current source/sink to drive
the sense amplifiers for both input rails.
The actual value of these resistors is
not that important, although we don’t
Application Max current
Low-Current Full Bridge
2-3A no heatsink
want a large voltage drop across them
so that we can use the Ideal Bridge at
modest AC input voltages.
For the PCB layout we need to consider the thermal characteristics of D1,
D2, Q5 and Q6 (the sense amplifier).
Silicon diodes have a -2.1mV/°C thermal coefficient for their forward voltage drop, so for every 1°C increase in
temperature of a diode junction, its
forward voltage falls by 2.1mV.
This means that if one diode is hotter
than the other, we will get an error in
the switching voltage. A similar effect
is seen with the base-emitter voltages of Q5 and Q6. For this reason, we
have placed the diodes right next to
one another, and placed the transisMax voltage
N-channel
P-channel
Source/comments
40V
IRF540
IRF9540
Farnell 8648298
IRFB4410ZPBF
SUP70101EL-GE3
IRF135B203
IXTP76P10T
±40V
High-Current Full Bridge 10A no heatsink
DC Combining 5A no heatsink
DC Combining 10A no heatsink
22
tors so they can be glued together.
This will ensure our switching margins are stable even as the board heats
and cools during use.
The ‘sense’ transistors (Q5 & Q6, Q9
& Q10 etc) only ever have 12V across
their collector-emitter junctions, so
we have specified standard BC546-9
or 556-9 devices (or their SMD equivalents, BC8xx).
However, the current source/sink
transistors will have the full dual rail
voltage across them, which could be
up to ±40V or 80V total. Therefore,
we have specified MPSA42/92 transistors for these (or the SMD equivalents, MMBTAx2).
These standard high-voltage, lowpower devices are available from all
the larger online suppliers. If you
have ±25V or lower voltage rails,
you could use BC546/556/846/856
transistors there instead. It is important to consider that the BC546/
BC556 have the opposite pinout to
the MPSA42/92 transitors, so you
would need to install them backwards if you do this.
Luckily, for the SMD transistors, the
BC846/856 series SMD pinouts are the
same as the MMBTA42/92 pinouts, so
they are a direct swap for applications
below ±25V.
Note that the 47kW resistor values
were chosen to allow operation from
low voltages to about ±40V at the
output. At the upper limit, the 47kW
resistors will dissipate 130mW each.
While that is well within the ratings of
a 1/4W resistor, we have specified 1/2W
resistors just to be safe.
If you will only use this bridge at
the higher end of its voltage range, you
could increase those resistor values
slightly to, say, 68kW. That will reduce
their dissipation to a maximum of
±30V
As above
12-24V
Not required
12-24V
Not required
Table 1 – examples of suitable Mosfets
SUP90P06
Mouser, DigiKey &
Silicon Chip kit
IXTP96P085T
IRF9540
Farnell 8648620
100mV/A drop
SUP90P06-09L-E3
Mouser & DigiKey
7.4mV/A drop
SUP70101EL-GE3
Mouser & DigiKey
11.4mV/A drop
IRF4905
Mouser & DigiKey
Practical Electronics | September | 2025
Discrete Ideal Bridge Rectifier
94mW, so 1/4W resistors should be fine.
You could also lower their values for
low-voltage applications, although that
shouldn’t be necessary.
Startup behaviour
Scope 2 shows the circuit starting
up when AC power is first applied. On
that first cycle, the AC input blue trace
goes negative. This charges the negative capacitor to about 5V, although we
don’t have a plot of the negative rail
here – we know that the negative and
positive rails will be about the same.
The Mosfet body diode conducts on
this cycle in the absence of voltage at
the Mosfet gate (due to the low initial
voltage). Once there are a few volts on
the output rails, the constant current
source/sink and BJT-based voltage sense
circuits kick in. By the time we are into
the first positive excursion of the AC1
input in cyan, we can see the gate drive
pulling the gate low (in yellow), having
already charged the large capacitor bank
enough in the first cycle.
Indeed, the gate voltage on that Pchannel Mosfet goes below 0V, being
pulled toward the negative rail, and
we see a full 12V on that P-channel
Mosfet gate in the first real cycle of
operation. This shows the benefit of
connecting the current source/sink to
the opposite rail rather than ground.
I love the simplicity of circuits like
this, which squeeze more out of a
handful of components than seems
reasonable. I also like going back to
basics and using BJTs in the current
sink and sense amplifier.
PCB layout
We touched on some PCB layout
considerations earlier. There are a few
aspects of the PCB design that are very
important:
Parts List – Discrete Ideal Bridge Rectifier
4 6.3mm pitch PCB-mount vertical spade connectors (CON1-CON4)
2 SUP70101EL 100V 120A P-channel Mosfets, TO-220 (Q1, Q2)
2 IRFB4410ZPBF 100V 97A N-channel Mosfets, TO-220 (Q3, Q4)
Resistors (1% ¼W axial – TH version | 1% ¼W M3216/1206 – SMD version)
8 22kW
2 330W
4 68W
4 100kW 2 47kW 0.5/0.6W (5% OK)
Through-hole version
1 double-sided PCB coded 18108241, 87.5 × 45.5mm
4 BC556/7/8/9 100mA PNP transistors, TO-92 (Q5-Q6, Q9-Q10)
2 MPSA42 300V 500mA NPN transistors, TO-92 (Q7, Q8)
2 MPSA92 300V 500mA PNP transistors, TO-92 (Q15, Q16)
4 BC546/7/8/9 100mA NPN transistors, TO-92 (Q17-Q20)
4 12V 0.4W zener diodes, DO-35 (ZD1-ZD4)
12 1N4148 75V 200mA diodes, DO-35 (D1-D12)
SMD version
1 double-sided PCB coded 18108242, 54.5 × 54.5mm
4 BC856/7/8/9 100mA PNP transistors, SOT-23 (Q5-Q6, Q9-Q10)
2 MMBTA42 300V 500mA NPN transistors, SOT-23 (Q7, Q8)
2 MMBTA92 300V 500mA PNP transistors, SOT-23 (Q15, Q16)
4 BC846/7/8/9 100mA NPN transistors, SOT-23 (Q17-Q20)
4 12V ¼W zener diodes, SOT-23 (ZD1-ZD4) [BZX84C12]
12 1N4148WS 75V 150mA diodes, SOD-323 (D1-D12)
For combining DC supplies, halve the numbers of all components except the
PCB and spade connectors.
– TH version kit (SC6987, ~£15)
– SMD version kit (SC6988, ~£14)
● The layout of the current sense
amplifier with its two transistors, two
1N4148 diodes and 68W resistor is kept
very tight as it must accurately sense
small voltages with relatively low bias
currents.
● The sense transistor pairs, like
Q5 and Q6, are face-to-face, so you
can super glue these together to keep
them as tightly thermally coupled as
possible (or add a smear of thermal
paste between them). On the SMD
version, these parts are tight against
one another.
The SMD version of the
Discrete Ideal Bridge
Rectifier is 54.5 ×
54.5mm, while the
through-hole only is
a bit larger at 45.5
× 87.5mm (not
shown to scale).
Practical Electronics | September | 2025
Both kits from Silicon Chip include the
PCB and everything that mounts on it.
● The pairs of 1N4148 diodes (D1
& D2) are right next to one another,
so they stay at similar temperatures.
● The path from the AC inputs
through the Mosfets and to the DC
outputs is kept as short as possible
and uses large copper fills to maximise the current carrying capacity
of the PCB. PCBs do not have a fixed
‘current rating’, but we must ensure
that the voltage drop and heating in
the tracks is reasonable at any current
likely to be drawn. At the AC1 input,
which has the thinnest connection to
the Mosfet, we have parallel copper on
the top and bottom layers of the PCB.
23
Constructional Project
Fig.3(a) & (b): the full-populated through-hole version of the PCB (left) and the reduced version for combining DC supplies
only (right). The full version can also be used to combine DC supplies. Watch the diode and Mosfet orientations, and
remember that Q7/Q8 and Q15/Q16 need to be reversed if you are using BC546/BC556 transistors instead for lower voltage
applications, compared to what’s shown here.
Mosfet selection
We have included 100V low-RDS(on)
Mosfets in the parts list. They only
cost a few dollars each and work well.
If selecting alternative Mosfets, look
for a voltage rating well above the
rail voltage you want; we feel that 80100V is about right. Select an RDS(on)
of 10mW or less.
The P-channel Mosfet will usually have a higher RDS(on); there is little
point in selecting N-channel Mosfets
with a significantly lower on-resistance
than the P-channel devices you will
be using.
For lower currents, you can get away
with less expensive Mosfets. Even
though the savings in dissipation won’t
be as great, the reduction in voltage loss
can still make this design very beneficial in lower-current designs.
For example, we used IRF540/
IRF9540 Mosfets from Altronics in
some tests, and it was fine up to about
3A, still giving a much lower voltage
drop than a conventional bridge. Table
1 includes some advice on Mosfet selection.
Construction
The through-hole version is built on
a double-sided PCB coded 18108241
that measures 54.5 × 87.5mm, while
the SMD version is coded 18108242
and is a bit smaller at 54.5 × 54.5mm.
For the former, refer to the Fig.3(a) PCB
overlay diagram, while Fig.4(a) is the
overlay for the SMD version.
The smallest SMD parts are the
SOT-23 transistors and SOD-323 diodes.
These are large enough that they are
not too challenging if you have a desk
magnifier and a reasonably good soldering iron.
If you are using it to combine solar
panels or DC power sources, you can
leave off all the negative rail parts,
shown in a dashed box in Fig.1. These
Figs.4(a) & (b): the SMD versions of the PCB, with the full version on the left
and the DC combining version only on the right. If substituting BC846/BC856
transistors for the MMBTA types, you don’t need to change how they are fitted
to the board. Only diodes D1-D12 and the Mosfets could be easily installed
backwards, so ensure they aren’t.
24
versions are shown in the alternative
overlay diagrams, Figs.3(b) & 4(b).
Start by fitting all the resistors.
Follow with the diodes, making sure
you orientate them correctly, with the
cathode stripes facing as in the relevant
PCB overlay diagram. We found that
for the SOD-323 SMD diodes we got,
it was tough to tell which end was the
cathode. If unsure, use a magnifier or
a DMM set on diode test mode.
Next, solder the signal transistors
in place. As mentioned earlier, if you
are using this at low voltages only, you
can use all BC546/556/846/856 transistors throughout. If you do this, remember that the through-hole devices
for Q7, Q8, Q15 & Q16 must be rotated
by 180°, as the MPSA42/92 types have
a different pinout.
Mount the 12V zener diodes next.
The SMD SOT-23 parts are small and
in the same packages as the bipolar
transistors, so make sure you don’t
mix them up. Place them with tweezers and tack one leg, allowing you
to adjust it (if necessary) by reheating
the initial joint before soldering the
remaining leads.
Fit the power Mosfets next. Watch
the layout here, as they face in alternate directions on the board to optimise the track layout. Also, don’t get
the two different types mixed up. Tack
one leg of each and fiddle them so they
are neatly aligned and the same height,
then solder the remaining leads.
Finally, mount the 6mm connectors.
You could solder wires directly to the
board, but we reckon using crimp spade
lugs is much neater.
Testing
We suggest testing the board in two
Practical Electronics | September | 2025
Discrete Ideal Bridge Rectifier
JTAG Connector Plugs Directly into PCB!!
No Header!
No Brainer!
Our patented range of Plug-of-Nails™ spring-pin cables plug directly
into a tiny footprint of pads and locating holes in your PCB, eliminating
the need for a mating header. Save Cost & Space on Every PCB!!
Solutions for: PIC . dsPIC . ARM . MSP430 . Atmel . Generic JTAG . Altera
Xilinx . BDM . C2000 . SPY-BI-WIRE . SPI / IIC . Altium Mini-HDMI . & More
www.PlugOfNails.com
Tag-Connector footprints as small as 0.02 sq. inch (0.13 sq cm)
halves. The following steps test the
two positive sections.
1. Connect the Bridge outputs to
an electrolytic capacitor of at least
470μF. Make sure you get the polarity correct.
2. Connect the negative of a 12-24V
power supply to the negative of your
capacitor and the positive to either of
the AC inputs. If you can set a current limit, set it to a few hundred
milliamps.
3. Switch on the supply and check
that the capacitor charges up to the
input voltage.
4. Put a 100W 1W resistor (or similar) across the capacitor and check
that the voltage across it does not
droop significantly (no more than
100mV).
This verifies that the appropriate
Mosfets are on; otherwise, the voltage would drop by 600mV or more. It
also confirms there are no catastrophic
shorts, or you would get smoke.
Now test the other AC input using
the same method.
If you run into trouble in either
case, go through the following checklist below:
1. Is your power supply going into
Practical Electronics | September | 2025
current limiting? Use a multimeter to
check for the expected voltage at the
AC input.
2. Are your Mosfets the right way
around?
3. Check that the diodes are all orientated correctly. If any are wrong, the
Rectifier will not work.
4. Check your soldering and look for
solder bridges.
5. Check that the current sink and
source work by measuring the voltage
between the base and emitter pins of Q8
and Q16. The reading should be close
to 0.6V in both cases. Also check for
a ~600mV Vbe on Q7 and Q15. If the
readings are low, check that the associated 47kW resistors are OK.
6. Check the voltage across the zener
diodes. Are they the right way around?
If the capacitor bank is charged up
and there is no load resistor, the voltage across them should be low, while
you should get a reading of several
volts with the 100W resistor across
the capacitor.
7. If the behaviour is correct for one
AC input of the Bridge but not the other,
check the circuitry around the misbehaving input and compare voltages to
the other half.
8. If both inputs don’t work, you
have a systematic problem since they
are essentially independent.
Having tested it with one polarity,
switch off the supply and connect
its positive output to DC OUT+ on
the Bridge and the negative of your
power supply to one of the input terminals. You should see the capacitor
charge up to the input voltage again.
Proceed with testing in this configuration as above.
Using it
Once installed, it will pretty well
look after itself. Refer to Figs.1(a) &
(b) to see how the connections should
be made. If you expect to draw continuous high currents from the power
supply, you will probably want to put
some flag heatsinks on the Mosfets.
Aside from that, you should find that
it just works.
Remember that you may need a
mains soft-starting system if you have
a really substantial capacitor bank and
low-impedance transformer like in a
big audio amplifier.
We published a circuit design for
limiting inrush currents in the April
2013 issue, the “SoftStarter”.
PE
25
|