Max’s Cool Beans
By Max the Magnificent
Flashing LEDs and drooling engineers – Part 28
A
s you may recall, in a previous
column (PE, May 2022) I made
mention of a modern incarnation of an analogue computer called
The Analog Thing (THAT) (https://bit.
ly/370WEZH). While mulling over this
fascinating device, I was reminded of
some nuggets of knowledge and tidbits
of trivia that I wanted to share with
someone, and it looks like you aren’t
in any position to get away, so...
When analogue ruled and
digital drooled
Way back in the mists of time, the first
true electronic systems were created
using vacuum tubes (‘valves’ in the
British vernacular and ‘tubes’ in North
American parlance). These little rascals
could act like diodes or switches or amplifiers. Early tube-based systems, like
1920s radios, were analogue in nature.
I have the greatest respect for the people
who used to conceive and realise the
most amazing analogue systems, often
with naught more than a couple of tubes
and a handful of resistors, capacitors
and inductors. As I’ve mentioned on
many an occasion, I’m a digital design
engineer by trade and I find the wibblywobbly nature of analogue systems to be
disquieting. In the case of even the simplest tube-based system, I can peruse and
ponder its schematic diagram for hours
without having a clue as to its purpose in
life or the way in which it was intended
to perform its magic. By comparison, any
of my chums at the Alabama Historical
Radio Society (ALHRS.org) could glance
at the self-same schematic and then discourse for hours on the topic of, ‘flobulating retrograde inductive couplings
titivating the phase-change modulation’
(or words to that effect).
The thing about analogue signal processing (ASP) is that you can do a tremendous amount of stuff with very little
stuff, if you see what I mean. For example, adding two sine waves while filtering out any high-frequency noise is ‘easy
peasy lemon squeezy’ using analogue
techniques. By comparison, implementing the same functionality using digital
signal processing (DSP) requires you to
develop an appropriate algorithm and
40
tends to require an overly enthusiastic
quantity of digital logic gates and registers, thereby leaving you ‘stressed depressed lemon zest,’ as it were.
On the other hand, as we’ve previously
discussed, there are several issues with
analogue systems, not least that – in the
case of computers – you rarely get the
same answer twice (there’s always some
difference, no matter how small). Another aspect of analogue is the way in
which it tends to degrade information.
In the days of video tapes, for example, if
you took a video of grandma’s birthday,
and you made a copy for your aunt, and
she made a copy for her friend, and... it
wasn’t long before the downstream copies
became unwatchable. By comparison,
when performing processing using digital techniques, you always get the same
answer (unless you’ve messed something
up), which I find to be comforting. Even
better, the use of error checking and correction (ECC) codes makes it possible to
make copies of copies of copies, with the
last being identical to the first (hurray!).
But we digress...
The rise of digital
Early digital computers, circa the beginning of the 1940s, were typically
electromechanical (using relays) and/or
electronic (using tubes). Most of these
machines worked in decimal (base-10)
and operated only on integers. Interestingly, a German engineer called Konrad
Zuse had a fully mechanical, binary,
floating-point computer called the Z1 up
and running in 1938, but that’s a story
for another day.
Not surprisingly, machines based on
these antediluvian technologies were
fulsome in size and greedy for power.
As I wrote in Bebop Bytes Back (https://
amzn.to/3MdgfFh), the Harvard Mark
I, which was developed in America between 1939 and 1944, was constructed
out of switches, relays, rotating shafts, and
clutches, and was described as sounding
like ‘a roomful of ladies knitting.’ This
machine contained more than 750,000
components, was 50-feet (15m) long,
8-feet (2.5m) tall, and weighed approximately five tons! Similarly, the Electronic Numerical Integrator and Computer
(ENIAC), which was constructed at the
University of Pennsylvania between 1943
and 1946, was 10-feet (3m) tall, occupied
1,000 square feet (90m2) of floor-space,
weighed in at approximately 30 tons, and
used more than 70,000 resistors, 10,000
capacitors, 6,000 switches and 18,000
vacuum tubes. The final machine required
150kW of power, which was enough to
light a small town at that time.
One of the greatest problems with computers built from vacuum tubes was reliability. For example, 90% of ENIAC’s
down-time was attributed to locating and
replacing burnt-out tubes. Records from
1952 show that approximately 19,000
vacuum tubes had to be replaced in that
year alone, which averages out to about
50 tubes a day!
Another consideration was that engineers who were conversant with the
continuous nature of the analogue world
could often handle the concept of asynchronous digital logic, but they found it
difficult to comprehend the nuances of
synchronous (clocked) digital systems.
Things weren’t helped by the fact that
digital techniques themselves were not
well understood at that time.
Life for computer pioneers started to
get easier with the invention of semiconductor diodes and transistors, and even
easier with the invention of integrated
circuits. It also didn’t hurt that we gained
a better understanding of digital logic and
its wily ways. Once we started building
semiconductor-based processors, things
like DSP really started to take off. These
days, we can build chips with hundreds
of millions of transistors (some have billions, a few have trillions), which explains why we now do so much processing using digital techniques.
Change is scary
On my meandering journey through life,
I’ve met engineers who had black belts
in designing with vacuum tubes, but
who couldn’t wrap their brains around
using semiconductor devices like discrete (individually packaged) transistors.
I’ve also come into contact with analogue designers who could happily create
masterpieces using discrete transistors,
resistors and capacitors, but who couldn’t
Practical Electronics | June | 2022
grasp the concept of digital functions like
the primitive logic gates (NOT, AND, OR,
NAND, NOR...) and register elements
(latches, flip-flops...) provided in jellybean integrated circuits (ICs) such as
the Texas Instruments 7400-series chips.
Even today, I know some digital engineers who are as happy as clams creating assembly language programs to run
bare metal on 8-bit microcontrollers,
but who would run for the hills if you
asked them to write a program in C and/
or create an application to run under an
operating system. (Incidentally, the term
‘bare metal’ refers to a computer system
without a base operating system or any
installed applications.)
Reports of analogue’s death
I can’t tell you how many times I’ve
heard the old chestnut that a newspaper
prematurely published an obituary for
Samuel Clemens, who we know by his
pen name of Mark Twain, and that he
responded by saying, ‘The reports of my
death are greatly exaggerated.’ It’s unfortunate that this story is itself something
of an exaggeration, but – as I learned at
my mother’s knee – one should never
let facts get in the way of a good story.
I think it was around the 1980s when
I first started to notice industry pundits
proclaiming the demise of analogue electronics. ‘Digital is destined to rule the
world,’ they cried. Unfortunately, a lot
of young people believed them, leading
to a lack of students wishing to learn the
analogue arts.
Paradoxically, the resulting dearth of
analogue engineers in the 1990s made
these people valuable commodities, capable of demanding high wages. Even
more paradoxically, engineers skilled in
analogue are in ever-increasing demand
today. Part of this is because – as analogue
buffs delight in telling anyone who will
listen – everything is analogue, including digital, which they regard as being a
sub-set of analogue (the cheeky scamps).
Much as it pains me to admit it, this
is true as far as it goes. Although digital
designers like to think of their signals as
instantaneously transitioning between
0 and 1 values with razor-sharp edges,
these signals take time to transition
(think ‘slopes’) in the real world. Also,
the faster we try to convey data through
the copper wires on our circuit boards,
the more their signals succumb to analogue effects, with their envelopes degrading before our very eyes. The bottom
line is that, without the help of analogue
engineers, we digital designers would
be unable to perform our own magic.
Fig.1. The purely analogue artificial neural network AML100 uses near-zero power to
detect events and make inferences and conclusions (Image source: Aspinity).
and machine learning (ML) tasks are realised using digital techniques. By some
strange quirk of fate, however, I’ve recently been seeing an increasing number
of analogue implementations. For example, the guys and gals at Aspinity.com
recently launched the first member of
their AnalogML family, the AML100,
which they describe as: ‘The industry’s
first and only tiny machine learning solution operating completely within the
analog domain’ (Fig.1).
The AML100, which is presented in
a small 7 x 7 mm 48-pin QFN package,
uses near-zero power to ito detect events
and make inferences and conclusions (it
actually consumes <20µA while alwayssensing). With up to three inputs from
sensors like MEMS microphones or accelerometers, the ANN in the AML100
can be trained to recognise things like
the sound of glass breaking or a human
speaking (it doesn’t understand the words,
to be clear, but it can tell the difference
between a person talking and a cat meowing or a dog barking, for example).
Consider a smart speaker like an
Amazon Echo, Apple HomePod, or
Google Home. Such a device typically
boasts a hierarchy of processors. Since
it isn’t doing anything for most of the
time, the main application processor (AP)
spends much of its time asleep, leaving
a smaller, lower-power processor to keep
a watchful eye (well, ear) open in the
hope of hearing the system’s wake-word
spoken. This wake-word is ‘Alexa’ in the
case of the Echo, and I know two people
called Alexa who, much like Queen Victoria, are not amused. Although the wakeword processor consumes less power
than the AP, it’s still digital in nature,
which means it’s probably consuming
more power than we would like it to, if
we were given a choice.
There’s also the fact that, out of all the
sounds going on in my office (air conditioner blowing, keyboard clicking, Geiger
counter beeping, my head banging on the
wall, the sound of a man sobbing...), a
human speaking occurs only a small portion of the time because I’m alone in the
Fig.2. Mantis AIS SoC next
to a 19mm-diameter US
one-cent ‘penny’ piece.
(Image source, AIStorm).
Analogue AI
At the time of this writing, most of the
artificial neural networks (ANNs) used
to implement artificial intelligence (AI)
Practical Electronics | June | 2022
41
Fig.3. Audio Weaver’s intuitive interface (Image source: DSP Concepts).
office and I don’t talk to myself (or, if I
do, I don’t listen to myself). In fact, any
human speech is pretty much confined
to me warbling on a video conference
call or saying, ‘Alexa, set a reminder.’
Now, suppose we were to add an
AML100 into the hierarchy. In this case,
the wake-word processor can itself take
a well-deserved snooze. The analogue
AI in the always-on AML100 listens
to everything that’s going on around
it, only rousing the wake-word processor when it detects someone speaking.
Pretty clever, huh?
Another example of analogue AI is
presented by the chaps and chapesses at AIStorm.ai. These are jolly nice
people, but it has to be admitted that
they are overly fond of saying ‘An AI
storm is coming.’ Consider a traditional
CMOS imaging sensor setup in which the
output from the sensor is passed through
an analogue-to-digital converter (ADC),
which feeds an image signal processor
(ISP), which passes the data to a digital
AI running in a digital signal processor (DSP), microcontroller unit (MCU),
graphics processing unit (GPU) or fieldprogrammable gate array (FPGA), none
of which are lightweights on the powerguzzling front.
By comparison, AIStorm’s Mantis AIin-Sensor (AIS) System-on-Chip (SoC)
transforms the sensor into the input
layer of an analogue AI. Discarding the
42
digitisation stage, the Mantis (Fig.2) uses
the sensor charge to directly couple to
the first of multiple layers of analogue
neurons. The bottom line is that, using a
miniscule amount of energy, the Mantis
can evaluate the current situation and
make an appropriate determination (‘A
human just entered the room’) before a
digital AI would have even laid its metaphorical eyes on the first byte of data
from the sensor.
Weaving audio designs
What? You want more? Well, I’m just
the man for the task. Let’s turn from
analogue to digital. I recently spent a
happy afternoon chatting with the folks
at DSPConcepts.com. Not being an audio
buff myself, I hadn’t really thought about
this prior to our conversation, but they
pointed out that there are frameworks
and design environments to help us work
efficiently and productively in almost
every engineering discipline other than
audio. It’s true! If you wish to create a
website, you take advantage of a modern
web development platform; if you desire
to construct a touch panel, you reach for
an existing graphics library and widget
toolkit; if you ache to implement an AI
system, you leverage an appropriate
framework; and on it goes.
And then there’s audio. I don’t know
if you’ve noticed, but compared to the
way things were just a few short years
ago, awesome audio experiences and
sumptuous soundscapes now feature
prominently in our daily lives. The problem is that developers of audio systems
have been largely obliged to design everything from the ground up, which is
a painstaking and time-consuming way
of going about things to say the least. To
address this problem, the folks at DSP
Concepts have created an awesome product called Audio Weaver (Fig.3).
All I can say is that the things I heard
about Audio Weaver and its intuitive,
drag-and-drop graphical user interface
(GUI) were music to my ears (I’m sorry,
I couldn’t help myself). Suppose you
wish to create something like a smart
speaker that involves an array of microphones and one or more loudspeakers. This is where Audio Weaver leaps
to the fore with a shemozzle of Zeusaphones. You can drag-and-drop functional blocks like noise reduction, echo
cancellation, equalisation, compression,
beamforming... (there are 500+ modules created by leading audio designers) and connect them together as you
wish. Audio Weaver also boasts a runtime core equipped with highly optimised target-specific libraries for use on
a wide range of processors from ARM,
Cadence, CEVA, TI, Qualcomm… (once
again, the list goes onand on).
If you are interested in learning
more, you might start with the Audio
Practical Electronics | June | 2022
Weaver video from CES 2022 (https://bit.
ly/3JOEcRI). After this, if you wish to explore further, the folks at DSP Concepts
say that Audio Weaver is for everyone,
from hobbyists to seasoned audio professionals. They currently offer a 30-day
free trial for anyone interested in getting
started (https://bit.ly/3M7V5Zf). Furthermore, with every download of Audio
Weaver, they include example audio
signal flows and algorithms (loudspeaker
processing, SPL meters, reverb...) so that
new users don’t have to build their audio
processing signal chains from scratch.
Are you taking notes?
For more years than I care to remember,
I’ve been using the Notepad editor that
comes with Windows for a variety of different tasks, such as taking the contents
of a Word document and stripping out all
of the hidden control codes, for example.
As useful as Notepad is, however, there
are a lot of things it could do better. To
be honest, I hadn’t really given much
thought to this until a few days ago
when my eyes were opened to a world
of possibilities. In an earlier column
(PE, December 2021), I introduced my
chum Guido Bonelli (aka Dr Duino). If
you visit Guido’s website (DrDuino.com)
you’ll see his Dr.Duino Pioneer and Explorer kits, which are awesome when it
comes to developing and debugging Arduino Uno and/or Nano-based projects.
Guido has started hosting a monthly
online video get together for his customers, during which he shares tips and tricks
and new projects. This is followed by a
member of the audience demonstrating
one of his or her Arduino-based projects. For the inaugural meeting, Guido
asked me if I could present one of my
own projects to give them an idea of
what he’s looking for on the participation front. Just for giggles and grins, I
walked them through my 12x12 Pingpong Array project, which we first started discussing here in PE, March 2020.
The thing is, as part of his portion of
the proceedings, Guido mentioned a free
editor called Notepad++ (pronounced
‘Notepad-Plus-Plus’), which you can
download from Notepad-Plus-Plus.org
(a word to the wise, make sure you click
only the ‘Download’ button located
under the image of the cardboard box
with ‘Notepad++’ written on the side –
don’t be distracted with any of the other
‘Download’ or ‘Start Now’ buttons that
are there to tempt the unwary).
All I can say is Notepad++ is the answer
to my text editor dreams, not least that
you can have multiple files open simultaneously, and you can switch back and
forth between them by simply clicking
their tabs (Fig.4). Furthermore, Notepad++ is a context-sensitive, color-coded editor that inherently understands
Practical Electronics | June | 2022
Fig.4. Notepad-Plus-Plus is the answer to my text editor dreams.
languages like BASIC and C/C++. Even
better, it allows you to open Arduino
sketch (.ino) files and library (.h and
.cpp) files.
Although I didn’t know it or have use
for it at the time (this has changed as we
will discuss), Notepad++ also understands things like PIC microcontroller
assembly source code (.asm) files and
machine code (.hex) files.
Before we proceed, one little trick
that can make your life a lot easier is
that – assuming you are using Windows
(there is no Notepad++ for Macs) – you
can simply drag-and-drop a file from
a File Explorer window into the Notepad++ working area to open that file (I
just ran a quick test to discover that we
can do the same thing with the classic
Notepad. I tell you; I learn something
new every day).
The wonderful world of PICs
Are you familiar with PIC (‘pronounced
‘pick’) microcontrollers from Microchip
Technology? These are derived from the
8-bit PIC1650, which was originally developed by General Instruments, who
introduced it to the market in 1976. Initially, PIC stood for ‘Peripheral Interface Controller,’ but it now stands for
‘Programmable Intelligent Computer’…
probably… possibly… the world hasn’t
quite agreed on this, so we stick with
just ‘PIC’.
These days, there is a baffling and
bewildering assortment of 8-bit, 16-bit,
and 32-bit PICs available, each offering a
different collection of features and functions, such as the number of generalpurpose input/output (GPIO) pins, the
number of analogue-to-digital converters
(ADCs), the number of counter/timer/
PWM functions, and... the list goes on.
There are many wonderful things about
PICs, such as the fact they support a wide
range of supply voltages (2.3V to 5.5V in
the case of the 8-bit devices I’m currently
working with) and they are available in
both surface-mount technology (SMT)
packages and lead through-hole (LTH)
dual in-line packages (known as DIPs or
DILs). There are also some not-no-wonderful facets to these devices, especially
the earlier 8-bit ones, whose Byzantine
RAM and register banking architecture
can bring a strong man to tears (I’m dabbing my eyes as I write these words).
As a result, the last thing you want to
do if you are new to PICs is to try to program them in PIC assembly language. All
of which brings us to the fact that, for
reasons too tortuous to talk about here,
I find myself in the position of having to
maintain and enhance code whose purpose I may not disclose for a company
whose name must remain unspoken …
but I fear I’ve said too much.
Wouldn’t you know it, this code, which
dates back to the dawn of time, was captured in PIC assembly language by someone to whom programming did not come
naturally. When I originally undertook
this task, I was informed that the code
was heavily commented. This is certainly true as far as it goes. Unfortunately,
the code evolved over time, while the
comments... didn’t.
‘Oh dear,’ I said to myself (or words to
that effect). I now have code for multiple
members of the PIC dynasty to maintain.
I’ve programmed in a variety of assembly languages – I’ve even designed my
own – so I downloaded the 279-page
manual for the first device (if there’s one
thing you can say about Microchip, it’s
that they have an awesome archive of
documentation). Earlier, I used the word
‘Byzantine’; I fear this was understating
the complexity of the task.
I must admit that I was starting to feel
a tad overwhelmed. ‘Alas, alack, my end
is in sight,’ I thought to myself, but then
I remembered... Joe.
43
Fig.5. Hex file size comparison: original
(left) vs. new (right).
Are you Positron?
My old friend Joe Farr in the UK has
a black belt when it comes to creating
PIC-based gadgets and gizmos. He can
whip up a printed circuit board (PCB)
design for a PIC-controlled thingamajig
or thingamabob and have it programmed
and ready for action while the rest of us
are still rooting through the dictionary
to check the spelling of ‘thingamabob.’
So, I set up a Zoom call with Joe,
during which I explained my predicament. After looking at the assembly code
in question, Joe suggested that we take
another tack, which was to rewrite everything in BASIC. I’m sure your eyebrows
just started quivering on full alert as did
mine. However, Joe went on to explain
that – when it comes to programming
PICs – he’s a huge fan of the Positron
PIC BASIC Compilers created by Les
Johnson (https://bit.ly/3rtDcw8). A onetime PayPal payment of only £39.99 will
gain you access to the latest and greatest
compilers, along with any and all future
updates and upgrades.
I immediately reached out to Les and
placed my order. It really is a small
world because Les quickly responded,
acknowledging my order and saying that
he had read my book, Bebop to the Boolean Boogie (https://amzn.to/3jIt5PU). It
seems that Les also reads my columns
here in PE, which means he will be reading this one, which means I’d better say
only nice things about him.
One point that puzzled me was the
use of ‘compilers’ (plural). Les explained
that he supplies different compilers for
8-bit and 16-bit PICs, and that both are
included. The accompanying integrated development environment (IDE) selects the appropriate compiler based on
the device specified in the code, all of
which is invisible to the user, which is
just the way I like it.
Following an email exchange, Les also
told me how his dear old dad started telling him about electronics ‘from when I
44
was able to walk and talk.’ As a result,
Les could name the parts in a television
set when he was 3-years old, and he was
making radios and amplifiers and suchlike by the time he was 7-years old.
The story behind the evolution of the
Positron Compilers is captivating. Sadly,
we don’t have the time to go into it here.
Suffice it to say that these bodacious
beauties have been designed from the
ground up to produce the most efficient
PIC assembly code possible. As an example, I just used my spiffy Notepad++
editor to view and compare the .hex file
generated from the original hand-coded
assembly with the .hex file resulting from
Joe’s BASIC program implementing the
same underlying algorithm (I blurred
the results to conceal the captivating
cunningness of this code). As we see,
the new version is around half the size
of the original, and fewer instructions
generally results in higher performance,
all of which leaves my visage adorned
with a gleeful grin (Fig.5).
What’s that bit do?
I have so much to talk about regarding
my adventures with PICs but – unlike
Doctor Who – I’m somewhat limited
in terms of time and space. One consideration is that I need some way to
program the little ragamuffins. I ended
up purchasing a couple of PICkit 3 programmers. I opted for PICkit 3s (https://
amzn.to/3vm0IMR) rather than PICkit
4s (https://amzn.to/3jKKwz4) because
(a) they were much cheaper (£18.79 vs.
£142.99 in the UK, and a similar difference in the USA where I currently hang
my hat), (b) they handle all the devices
I currently need to deal with, and (c) Joe
told me this was a good way to go. The
reason I purchased two units is that Joe
says it’s easy to brick these little rascals, in which case you can use one to
flash (re-program) the other (the term
‘brick’ or ‘bricking’ refers to rendering
an electronic device unusable, often as
the result of a failed software or firmware update).
Something else I need is a development board to program and test my
PICs. This board has to be tailored to
verify the cunning algorithms we’ve
loaded into the devices. This includes
the board having its own integrated test
pattern generator, which we can implement in the form of another PIC (I tell
you; my eyes have been opened wide to
the wonderful world of PICs).
Joe and I bounced a few ideas back and
forth, then we called it a day. You can
only imagine my surprise and delight
when we Zoomed again the following
morning to discover that Joe had gone
ahead and designed the board. I will tell
you more about this little beauty (the
board, not Joe) in a future column. The
reason I mention it here is that, as Joe
was walking me through his design, I
asked, ‘what’s that bit do?’ To which Joe
responded, ‘I included my usual power
supply circuit.’
When I enquired further, Joe explained
that he rarely powers his creations using
a USB interface because you can’t count
on it providing you with a solid, spoton 5V supply. Instead, he implements
his own little circuit that can accept a
7 to 25V input, AC or DC (the DC can
be either polarity), and it will generate
rock-solid 5V and 3.3V DC values to
power his electronics.
The fact that the DC can be any polarity struck a chord with me. Most of the
little power jacks with which I come
into contact have a +ve inner and 0V
outer. Having said this, I once picked
up a secondhand pair of tasty computer speakers, which I connected to one
of my old power blocks. Unfortunately,
they expected a 0V inner and a +ve outer,
which I realised around the same time
the smoke started to appear.
While we were chatting about this, it
struck me that this could be a useful circuit for a lot of people, so I asked Joe if
he could whip up a standalone version
for me to share here, and he promptly
did so.
Feel the power!
The circuit diagram for Joe’s handy-dandy power supply unit (PSU) is shown in
Fig.6, while the PCB layout is presented
in Fig.7. Although your knee-jerk reaction may be that there’s nothing revolutionary here, experienced hands will
recognise that the design and physical
realisation reflect years of experience.
We start with a bridge rectifier (BR1),
which provides full-wave rectification
from a two-wire AC input. If the input is
DC, this also allows the power to be connected with the +ve on the inner or the
outer of the jack. The bridge rectifier is
followed by a smoothing capacitor (C1).
The smoothed output from the bridge
rectifier is fed to the input of the first regulator (IC1), and also as a raw value to
the output connector (SK2). The output
from the first regulator is fed as input to
the second regulator (IC2), and also to
the output connector (SK2).
In my case, I need 5V and 3.3V outputs, so that’s what’s reflected in the
schematic. However, by swapping out
the regulators, you can generate a variety of voltage combinations, such as
12V and 5V, for example.
Parts list – Joe Farr PSU
SK1 This can be any PCB barrel-type
connector with the correct footprint.
Alternatively, you can skip the connector and just connect directly to
pads ‘a’ and ‘b’ on the PCB.
Practical Electronics | June | 2022
Fig.6. Schematic diagram for Joe Farr’s handy-dandy PSU.
SK2 A 5-pin Molex-type connector is
ideal as it allows for the board to
be easily disconnected from the
rest of the project. However, you
could use any 5-pin connector
with a standard 0.1-inch / 2.54mm
pitch, or you could solder header
pins, or even solder wires directly
to the PCB.
BR1 Any bridge rectifier with a working
voltage of at least 50V and rated for
a minimum of 1A (always slightly
more than the combined maximum
load of the board) can be used.
C1 This capacitor must have a voltage
rating higher than the maximum
expected board input voltage (see
Note 1 below).
C2,3 Pretty much any 100nF capacitors
with a working voltage greater than
35V can be used.
C4 If IC2 is an LD1117V33 regulator,
a 10µF/16V capacitor is ideal here.
However, if a 78xx regulator is used
for IC2, then this capacitor should
be changed for another 100nF capacitor, identical to C2 and C3.
R1 The resistor value used depends on
what output voltage will be used to
drive LED1 (see Note 2 below). This
resistor should be ¼-watt rated.
LED1 Any 5mm or 3mm LED with a forward voltage of around 2V can be
used here.
IC1 Any 78xx regulator can be used
(see Note 1).
IC2 The original design is specified as
using an LD1117V33 LDO (Low
Dropout) regulator. Make sure that
the input voltage to the regulator
(coming direct from the bridge
rectifier or from IC1), is within an
acceptable range for this regulator
(see Note 1).
To capture this design, Joe used the DipTrace schematic and PCB design software
package (diptrace.com). Joe has kindly
made all of his design files available.
You can download a compressed ZIP
(file CB-June22-01.zip) containing the
schematic and foil files from the June
Practical Electronics | June | 2022
Fig.7. PCB layout for Joe Farr’s handy-dandy PSU.
2022 page of the PE website (https://bit.
ly/3oouhbl). Also, file CB-June22-02.
zip contains all the Gerber and other
files required to fabricate the board. If
you wish, you can download a free version of DipTrace that will allow you to
open, modify, and print the schematic
and layout files and also generate updated Gerber files (https://bit.ly/3uQfZ9B).
Joe has designed this board to be singlesided, so those with their own facilities
can make the PCB. Alternatively, you
can ask a PCB house to take the Gerber
files and do all the hard work for you.
The voltage regulators don’t really
need to be bolted to the PCB unless
you are attaching heatsinks, which can
be homemade if required. On the other
hand, bolting the regulators to the PCB
does add to the mechanical stability, so
it’s usually a good idea. However, be
very careful to not allow the mounting
bolts or heatsinks to come in contact
with anything electrically connected
to the circuit.
Note that some regulators – the
LD1117V33 being a prime example –
have the metal mounting tab connected
to the device’s output pin. By comparison, 78xx regulators have the mounting
tab connected to the center (ground) pin.
PSU design notes
Joe provided the following design notes
to help you get the most out of his PSU:
Note 1: The regulator used for IC1 can be
any 78xx-style regulator. Do not attempt
to use a 79xx series regulator anywhere
on this board as the pin-out is different.
The regulator works by taking the input
voltage and lowering it to match its specified output voltage. The difference between the input and output voltage is
dissipated as heat and the regulator can
get extremely (‘burn the skin off your
finger’) hot. To minimise the heat being
dissipated, try to match the input voltage so that it’s around 3V higher than
the output of the regulator. When running the board from DC, allow for the
bridge rectifier to drop your input voltage by around 1V. You should make sure
that the new voltage is sufficient for the
input of the regulator you have chosen.
When running the board from AC, you
need to be careful since the rectified and
smoothed DC produced by the bridge
rectifier in combination with the capacitor C1 can exceed the maximum voltage
rating of the capacitor, especially when
there is little or no loading on the PSU
output. Always check the datasheet for
the specific voltage regulator(s) you are
using to make sure you comply with
recommended voltage input ranges. If
the input voltage to either regulator is
too high, that regulator may run hot,
with the temperature increasing as you
increase the current draw. If the input
voltage is too low, the regulator will be
unstable and not supply a regulated voltage correctly.
Note 2: Depending on the input voltage
applied to the LED, the R1 resistor value
needs to be adjusted to give a reasonable
brightness. Aiming for around 10mA,
which gives a nice brightness without stressing the LED, some suggested
45
Your best bet since MAPLIN
Chock-a-Block with Stock
Visit: www.cricklewoodelectronics.com
O r p h o n e o u r f r i e n d l y kn
o w l e d g e a b l e st a f f o n 020 8452 0161
Components • Audio • Video • Connectors • Cables
Arduino • Test Equipment etc, etc
chair in the family room using spoken commands (thank you,
Alexa) to control the lights, turn on the television, and search
for programs of interest. I think they would have seen this as
a sign that people were turning into ‘weeds and wets’ as the
celebrated English philosopher Nigel Molesworth (https://
bit.ly/3hEJnIj) might have said regarding his fellow pupils
at the legendary English prep school, named ‘St Custard’s’.
The reason I mention this here is the sentiment, ‘What has
been will be again, what has been done will be done again;
there is nothing new under the sun,’ as the Preacher writes
in Ecclesiastes 1:9. In my previous column (PE, April 2022),
for example, I noted that the pushbutton and toggle switches
of the type with which we are familiar originally appeared on
the scene circa 1880 and 1916, respectively. Well, I recently
ran across an interesting article – At the Interface: The Case
of the Electric Push Button, 1880-1923 – by media studies
scholar Rachel Plotnick (https://bit.ly/3I3Nc4y). In Rachel’s
article we discover that, in 1916, the educational reformer,
social activist, and best-selling American author Dorothy Canfield Fisher warned that ‘There is a great danger of coming to
rely so entirely on the electric button and its slaves that the
wheels of initiative will be broken, or at least become rusty
from long disuse.’ All I know is this this tidbit of trivia is
going to stick in my mind for years to come because I now
think of it whenever I make use of a pushbutton (goodness
only knows what Dorothy would have thought about Alexa).
Don’t lose your head
Visit our Shop, Call or Buy online at:
www.cricklewoodelectronics.com
020 8452 0161
Visit our shop at:
40-42 Cricklewood Broadway
London NW2 3ET
resistor values for different voltages are: 3.3V = 150Ω, 5V =
330Ω, 12V = 1kΩ, 15V = 1.2kΩ. There are two options for
IC2. The original design uses an LD1117V33 3.3V regulator.
These regulators should use the IC2a pads as they have a different pin-out compared with the more common 78xx series
devices. If you wish to use a 78xx for the second regulator,
then the pads labelled IC2b should be used.
Voltage options: Regulator IC1 is always fed directly from
the output of the bridge rectifier and capacitor C1. However,
depending on requirements, there are two possibilities for driving regulator IC2. If you wish, it can be fed directly from the
bridge rectifier (fit jumper C to E). Alternatively, it can be fed
from the output of IC1 (fit jumper D to E). This latter option is
useful if the voltage output of IC1 is higher than the minimum
input voltage required for IC2. This will reduce heat, but it
assumes that regulator IC1 has sufficient capacity to provide
power to your circuit as well as power to regulator IC2.
Problems with pushbuttons
We didn’t have remote control for our television when I was
a young whippersnapper growing up in England. In addition to the fact that there were only two TV channels, called
BBC1 and ITV (at least, this was the case until 1964 when
BBC2 arrived on the scene), people would have considered
it to be the height of laziness to control things from the comfort of the sofa rather than getting up and ambling across the
room. Similarly, we didn’t have things like a dishwasher
(unless you were to count my dad) or a
garage door opener (unless you were to
count me), because these were tasks the
fates had ordained we undertake by the
sweat of our brows. I don’t know what
my grandparents would have thought to
see me today, ensconced in my command
46
I know, I know... I haven’t even mentioned the animatronic
robot head created by my friend Steve Manley with some
(hopefully) helpful suggestions from yours truly.
I also remember that I promised to talk about stepper and servo
motors and topics of that ilk, and I shall, but not today because
I am being buffeted by a maelstrom of happenings and it’s all
I can do you keep my nose above the metaphorical waterline.
What? 1000001? Me?
‘Eeek Alors,’ as my dear friend Shears used to say, because
this is the month – May 2022 – that I reach the ripe old age of
65 (or 1000001 in binary). Shears was my best friend at high
school, and we shared many life experiences, like seeing Pink
Floyd play Dark Side of the Moon at the Knebworth open air
concert in 1975. When I went to university, Shears set off on
an extended tour around Europe, largely living off his wits.
Sometime after I completed my degree and commenced my
first job in Manchester, he came to visit for a weekend and
stayed for two years. Oh, the times we had.
Why ‘Shears’? Well, his real name was Mark Burkinshaw,
so his friends alliteratively renamed him ‘Billy Burkinshaw’,
which (following the release of the Sgt. Pepper’s Lonely Hearts
Club Band album by the Beatles in 1967) led to ‘Billy Shears’,
which was subsequently truncated to ‘Shears’ (my dad used to
call him ‘Scissors’).
Shears and I would both have reached 65 within a few weeks
of each other. Unfortunately, the silly sausage found himself
in close vicinity to the Chernobyl nuclear power station in
Ukraine when it melted down in 1986 (what are the odds?).
Happily, he was able to dance at my wedding when I married my wife (Gina the Gorgeous). Sadly, he passed away a
couple of years later from an unusual type of aggressive brain
tumor. I miss him a lot. I will raise a glass or three to both of
us on my birthday.
Cool bean Max Maxfield (Hawaiian shirt, on the right) is emperor
of all he surveys at CliveMaxfield.com – the go-to site for the
latest and greatest in technological geekdom.
Comments or questions? Email Max at: max<at>CliveMaxfield.com
Practical Electronics | June | 2022
