This is only a preview of the November 2024 issue of Practical Electronics. You can view 0 of the 80 pages in the full issue. Articles in this series:
Articles in this series:
Articles in this series:
Articles in this series:
Items relevant to "Multi-Channel Volume Control, part one":
Articles in this series:
Articles in this series:
Articles in this series:
Articles in this series:
|
Constructional Project
Project by Tim Blythman
When designing or testing a device that runs
from a coin cell, you need to know how
much current it draws to determine the
cell’s life. That can be difficult given the
low currents often involved. This device
will power such a circuit while showing the
voltage, current and other helpful statistics.
Coin Cell Emulator
W
e have published many designs
powered by coin cells (usually
the CR2032). They must not draw an
excessive current; a high current draw
reduces the cell life and causes its volt
age to sag due to internal resistance.
Coin cells also exhibit a reduced ca
pacity at high discharge rates, com
pounding the effect.
While many circuits can be char
acterised with a standard multimeter,
that doesn’t work well for this type of
circuit. A typical multimeter’s shunt
on the microamp range has quite a
high resistance; values around 100W
are typical.
That is OK for readings in the micro
amp range, but when the current draw
might briefly jump to 5mA or so, the
meter is suddenly dropping half a
volt, which can change the circuit be
haviour substantially. In other words,
the burden voltage starts to dominate
the reading.
One possible solution is the Micro
Current DMM Adaptor from the May
Features & Specifications
2011 issue. That article discusses
burden voltage in detail.
However, this Coin Cell Emulator
does more than just measure current.
It can accumulate the current readings
to calculate a capacity value in mAh.
It can also produce a varying voltage,
so you can test how your circuit be
haves as the cell discharges.
The Emulator can also mimic some
of the non-ideal characteristics of
coin cells, such as internal resist
ance & voltage fall-off as the battery
discharges.
Design
Like the MicroCurrent DMM Adaptor
mentioned earlier, the Coin Cell Em
ulator uses the MAX4238/MAX4239
ultra-low offset, low noise precision
op amp to sense very small currents
without influencing them.
This op amp has a typical input
offset of 0.1µV and an input offset cur
rent of 1pA. These are a few orders of
magnitude lower than we are trying to
» Emulates the properties of a coin cell, including internal resistance and
discharge over time
» Emulates reduced capacity at high currents
» Adjustable voltage
» Current and charge measurement
» Stopwatch/Timer
» Automatically stops when a threshold voltage is reached
» Dummy PCB can be slotted into a coin cell holder
» Voltage setting: in 0.1V steps
» Typical accuracy: 1%
» Current measurement: 0.1μA resolution up to 200mA
» Charge measurement: 1μAh resolution up to 9Ah
» Voltage measurement: 0.01V resolution up to 3.4V
» Time measurement: 1s resolution up to 999 days
52
measure, so they are unlikely to inter
fere with our readings.
This is a necessary feature but in
sufficient to ensure we can measure a
wide range of currents. Our design has
an upper limit of around 200mA but
can measure down to 0.1µA. To do this
across a single range would require an
ADC (analog-to-digital converter) with
21 bits of resolution.
Instead, our design uses two ranges
and a 12-bit ADC that’s built into the
microcontroller. Oversampling (making
multiple measurements and averag
ing them) gives us a few more bits of
resolution, providing the necessary
dynamic range.
Circuit details
Fig.1 is the circuit diagram for the
Emulator. 5V power comes in via
mini-USB connector CON1, with
a 10µF capacitor providing boardlevel supply bypassing. IC1 is an eightbit PIC16F18146 microcontroller that
controls and monitors the Emulator’s
operation.
IC1’s internal DAC (digital-to-analog
converter) can deliver 0-4V from pin
17. Unlike some older PICs, the DAC
on the PIC16F18146 has an internal
buffer and thus has a reasonable drive
strength.
The DAC voltage goes to NPN tran
sistor Q1’s base via a 1kW resistor. Q1
is configured as an emitter follower,
so its emitter ranges from 0V to 3.4V,
about one diode drop below the 4V
maximum from the DAC. Its collector
is connected to the 5V rail.
The emitter-follower relies on the
reasonably constant base-emitter for
ward voltage of around 0.6V. Assum
ing the base voltage is constant, the
Practical Electronics | November | 2024
Coin Cell Emulator
Fig.1: the cunning part of this circuit is the op amp feeding current back
into the output through the 10kW resistor to cancel out the voltage drop
across the 22W resistor. This allows the circuit to work with two current
measuring resistances of vastly differing values, giving it a very wide current
measurement range.
transistor switches on harder if the
voltage at the emitter drops, increasing
the collector-emitter current and rais
ing the voltage at the emitter.
If the emitter voltage rises, the tran
sistor base current decreases, and less
current comes in through the collec
tor. So the circuit maintains the emit
ter voltage at a steady ~0.6V below the
base voltage.
A 1µF capacitor provides some fil
tering and can provide brief bursts of
current to the load. The 1kW emitter
resistor provides a stable load and en
sures that the output voltage decreases
if the base voltage decreases.
The 22W resistor acts as a current
measuring shunt, with two of the mi
crocontroller’s ADC pins monitor
ing the voltage across the shunt via
10kW resistors. Each ADC pin also
has a 100nF capacitor to ground to
present a low impedance to the ADC
sampling stage.
The ADC pins are labelled VSHUNT,
upstream of the shunt resistor, and
VOUT, downstream.
Practical Electronics | November | 2024
The downstream side of the 22W re
sistor is the positive side of the emulat
ed coin cell, with circuit ground being
the negative side. This is available at
a pair of 2-pin connectors (CON3 and
CON4) and a couple of large pads on
a circular part of the PCB. This part of
the PCB has a pad on each side and
can be slotted into some 2032-sized
coin cell holders.
Op amp IC2 has its input pins (pins 2
and 3) connected across the 22W shunt,
with its output (pin 6) feeding back into
the low side of the shunt via diode D1
and a 10kW resistor. A third ADC pin of
IC1 (pin 10; labelled ILSENSE) moni
tors the voltage at the diode’s cathode
via another 10kW resistor and 100nF
capacitor arrangement.
A 100nF capacitor bypasses IC2’s
5V supply (pin 7) and ground (pin 4).
Pin 1 (SHDN) is also pulled up to the
5V rail, allowing the op amp to oper
ate normally when powered.
Op amp operation
If a small current flows through the
22W resistor, a voltage appears across
the op amp’s input terminals and its
output rises. Current flows through
the diode and 10kW resistor back to its
inverting input and the downstream
end of the shunt resistor. The diode
ensures the op amp can only source
and not sink current.
Effectively, the op amp overrules
the shunt and supplies current to the
output of the Emulator. Smaller cur
rents can be sensed by measuring the
voltage across the 10kW resistor and
applying Ohm’s Law.
Eventually, the op amp output satu
rates and cannot supply enough cur
rent. It has nearly rail-to-rail operation,
so its maximum output is around 4.9V.
Assuming the Emulator output is at
about 3V, there is around 1.3V across
the 10kW resistor, with the op amp
supplying around 130µA.
The voltage across the 22W resistor
can now develop and is measured by
the ADC channels connected across
it. We can thus measure across a wide
dynamic range since we are effectively
53
Constructional Project
with the output short-circuited, but it
was not damaged.
1/2W resistors are available in this
size, so that’s what we’re
specifying. That allows
the Emulator to handle
a short circuit on its
output indefinitely.
ADC input
impedance
The Coin Cell Emulator is a
compact but handy development
and testing tool. Even if you
don’t design circuits for coin cell
operation, it’s a useful low-voltage
PSU with current monitoring.
using two shunts with vastly different
resistances. Combining the currents is
as simple as adding them.
Using a high-side measuring shunt
also means that the ground circuit
is uninterrupted and can be shared
with any other gear that needs to be
attached (programmers, debugging
gear or other meters) without affect
ing current readings. This is handy,
especially if you are running every
thing from a computer.
The test point labelled RST was
originally included to allow the Emu
lator to control a connected circuit by
pulsing its reset line low. But since the
Emulator can power cycle the circuit,
we did not implement this feature.
Instead, a nominal 1Hz clock signal
is available at this pin. This can be
used to trim IC1’s internal timer for
accurate timekeeping.
Short circuit handling
Let’s examine what happens when a
short circuit is applied to the output of
the Emulator. With the DAC set to its
maximum of 4V, around 140mA flows
through the 22W resistor.
With a typical transistor β (gain) of
around 400, the base current is around
350μA and the 1kW resistor on Q1’s
base drops 0.35V, so the voltage at the
emitter falls from 3.4V to around 3V.
The transistor thus dissipates around
280mW (2V × 140mA), comfortably
within its 500mW rating.
The remaining voltage is across the
22W resistor and it dissipates around
400mW. That’s a bit on the high
side for the typical 1/4W rating of an
M3216/1206-size SMD part. Our proto
type got quite hot around that resistor
54
One design consideration was en
suring that the ADC sampling did not
unduly load the Emulator’s output. A
load of even 1MW to ground would be
measurable, as it would draw 3µA at 3V.
Two ADC channels are fed directly
from low-impedance sources and un
affected by loads; transistor Q1 and
op amp IC2 drive the VSHUNT and
ILSENSE lines, respectively. Effec
tively, they are upstream of their re
spective shunts.
On the other hand, any load applied
to the VOUT line would be indistin
guishable from a load at the Emulator
output. The ADC input used to sense
the VOUT voltage is such a load.
The ADC input consists of a small
capacitor, nominally 28pF, which is
connected to the ADC pin to sample
the voltage. The capacitor is then con
nected to the internal ADC circuitry
(and disconnected from the pin) to
perform the conversion.
The ‘switched capacitor’ model can
be used to calculate an equivalent DC
resistance. A switched capacitor is
simply a capacitor that is switched
between two different connections at
a known frequency. The resistance of
such an arrangement is simply 1/CF,
where C is the capacitance in farads
and F is the frequency in hertz.
With our 100Hz sampling, this comes
out to around 350MW, which is more
than high enough. Higher sampling
rates would reduce this apparent re
sistance.
Another point to consider is that the
ADC capacitor is not discharged be
tween samples, so the load presented
by the switched capacitor is not equiv
alent to a load to ground, but rather
as a resistance between the different
sampling points. That raises its effec
tive resistance.
The PIC16F18146 has an ADCC
(analog to digital converter with com
putation) module. We previously used
some of its advanced features in the
Digital Boost Regulator that was pub
lished in the December 2023 issue.
The differential ADC inputs make
it much easier and more accurate to
measure the difference between two
voltages, as we are doing here. The
sampling time is also programmable, so
we have extended it slightly to ensure
the sampling capacitor can fully settle
at the input voltage.
There is also a DIA (device informa
tion area) that holds information such
as the measured value of the chip’s in
ternal voltage references. This means
we can measure voltages against this
reference without a separate calibra
tion step.
The DAC mentioned earlier is an 8-bit
type with a 4.096V (nominal) voltage
reference. It can deliver up to around
4V in 16mV steps and can produce a
voltage with 0.1V precision.
The output voltage at VSHUNT and
VOUT is thus limited by design to
around 3.4V. This works well with cir
cuits using 3.3V microcontrollers that
typically have a 3.6V upper limit. The
MAX4238 op amp specifies a common
mode voltage up to around 3.6V (with
a 5V supply) and the op amp inputs
stay within that range.
Microcontroller and interface
IC1’s pins 2, 3 and 5 connect to
switches S1, S2 and S3, respectively,
with their other sides grounded. The
micro applies an internal weak pullup
current to each, so it can detect button
presses as level changes on those pins.
An I2C OLED module is connected
to IC1’s pins 12 and 13 for the SDA
and SCL signals. The OLED is pow
ered from 5V; it has an onboard 3.3V
regulator with I2C pullups, allowing
it to interface with a microcontroller
running from 3.3V or 5V.
IC1 has a local 100nF bypass capaci
tor between its pin 1 supply and pin
20 ground. Pin 4 (MCLR) is pulled up
to 5V by a 10kW resistor, allowing the
microcontroller to run.
These pins and pins 18 and 19 (PGC
and PGD) are taken to CON2 for in-
circuit serial programming (ICSP) of
the microcontroller.
Coin cell behaviour model
As the saying goes, all models are
incorrect, but some are still useful!
There are several characteristics of
coin cells that we are explicitly mod
elling. We’re not claiming that the
model is comprehensive, but it mimics
the behaviour of a real coin cell well
enough to be useful.
Practical Electronics | November | 2024
Coin Cell Emulator
Our model is based mainly on a
CR2032 cell, as that is what we have
used the most. We fitted graphs pro
vided by several CR2032 manufactur
ers to curves described by simple equa
tions, adjustable by a single parameter.
There is a lot of variation between
manufacturers and even between cells
from the same manufacturer under
different conditions. The default be
haviour of the Emulator is similar to
a typical coin cell.
Firstly, coin cells have internal re
sistance. For CR2032 cells, the value
is around 20W, but it can change with
load and state of charge.
Other 20mm diameter cells, such as
the CR2016 (half as thick as a CR2032
at 1.6mm), appear to have a similar in
ternal resistance. So the Emulator will
also be suitable for thinner cells of the
same diameter but might not be as ac
curate for those with a smaller diameter.
A simple way to model the internal
resistance is with a fixed resistor, and
we chose the 22W part that we have
already explained. One advantage of
using a fixed resistor is that this resis
tor can also be used as a current meas
uring shunt.
The actual circuit appears to have an
internal resistance of around 24.5W, as
the 1kW base resistor carries a current
in proportion to the load current di
vided by the β (gain). So it adds around
2.5W (1000W ÷ 400) of resistance for a
β of around 400.
The next factor is that, like most bat
teries, the terminal voltage drops as the
cell discharges until it is flat. For coin
cells, the voltage drops a little at the
start, then is quite steady for most of
We have used a
socket header to attach
the OLED module in
our prototype, but
the Emulator will be
much more robust if
you solder the display
directly to the main
PCB.
the cell life. Once it starts to fall after
that, it does so quite dramatically.
While we looked at using a curve
to model this, curves that fit all three
stages were complex, and we found
that they weren’t helpful for observing
circuit behaviour as the cell goes flat.
Instead, we have implemented a
simple model that maintains a flat
voltage and then linearly changes the
output voltage as the cell’s state of
charge (SoC) nears its endpoint. For
example, with this set to 10%, the
voltage is flat from 100% to 10% SoC,
then drops to half by 5% SoC. Finally,
the voltage is ramped to zero when the
Emulator determines the cell is flat.
This feature can be turned off (set
to 0%) to disable this behaviour. Fig.2
shows the graph of the data sheet be
haviour compared with the emulated
behaviour.
While we could have more closely
emulated this with, say, four linear
sections, we decided not to do that.
We found that a constantly changing
voltage during use interfered with
monitoring the device’s operation. In
other words, we have sacrificed real
ity for usability. Our simple voltage
Fig.2: our emulated cell voltage curve is much simpler than
that seen in many coin cell data sheets, but it still mimics
the cell going flat. Otherwise, we prefer to manually adjust
the voltage and observe what happens.
Practical Electronics | November | 2024
curve provides a voltage that behaves
very predictably.
It does omit the higher voltage at the
start, but that can easily be emulated
manually by initially setting the voltage
to 3.2V, observing the operation, then
manually dropping the voltage to 3V.
Another well-known aspect is that
a cell’s apparent capacity (in mAh) is
reduced if it needs to supply a heavier
load. The manufacturers also provide
graphs to characterise this behaviour.
One typical graph we saw showed
that a nominally 240mAh cell provides
only 150mAh with a continuous dis
charge of 3mA, nearly halving its ef
fective capacity.
We found quite a few curves that
demonstrate this behaviour. The data
varied quite a bit, but it was clearly
some form of polynomial relationship.
A good technique for finding the order
of polynomial relationships is to take a
plot of the logarithms of the variables in
question. The order of the polynomial
is related to the slope of this graph.
Consider the quadratic equation y
= x2. The value of log(x2) is equal to
2log(x), for positive values of x, so the
graph of log(y) or 2log(x) against log(x)
Fig.3: the reduction in useful capacity is modelled as a
straightforward quadratic curve. It’s a compromise between
simplicity and accuracy.
55
Constructional Project
The Coin Cell Emulator
shown at actual size, along
with the wire added to the back of the
PCB (right). This increases the thickness
of the PCB to bring it nearer to that of a
CR2032 cell (3.2mm thick vs 1.5-1.6mm thick for
the PCB). You’ll need to apply a bit of heat to get the solder
to take to the large copper area.
would have a gradient of two, suggest
ing a quadratic equation of some sort
(a quadratic is a second-order poly
nomial).
We found that the slopes of these
log/log plots were just over two. So we
modelled this with a quadratic equa
tion and found that it fit quite well to
the manufacturer data and was simple
enough for the 8-bit micro to calculate.
We didn’t see any charts that show
behaviour much above 5mA but this
model also allows us to extrapolate.
This extrapolation suggests severely
degraded capacity as the current enters
this region. Our experience is that coin
cells discharge very quickly if you draw
much more than 5mA from them, so
this makes sense.
Our model takes a parameter equal
to the current at which the cell capac
ity is halved. We have used a default
value of 3.5mA, which matches the
CR2032 data sheets we examined. It
also makes it easier to match your
Emulator to a specific type of cell if
that is required.
If this value is set to zero, then there
is no modelling and the Emulator will
show the same capacity no matter what
current is drawn. Fig.3 shows the graph
of the model against typical data from
a cell data sheet.
Regarding the short circuit behaviour
noted earlier, it should be apparent that,
like a real coin cell, the Emulator will
quickly ‘go flat’, effectively ending the
short-circuit condition.
Firmware
For the most part, the microcontroller
allows the user to set the output volt
age, although it can modify that based
on the discharge modelling. It moni
tors the voltages around the circuit
and calculates and sums the currents
in the two measuring shunts.
Fig.4: the rise time of the output is limited by the capacity of
the circuit to supply the current to charge the 1µF capacitor
at its output (the timebase is in µs here). The DAC that
controls the voltage has a settling time of around 10µs.
56
A timer keeps track of time inter
vals and allows the current to be ac
cumulated over time for the charge
and capacity calculations. The meas
ured charge (in mAh) is taken from the
actual value, while the SoC calculation
is based on the modified behaviour at
higher currents.
All this information is displayed on
the OLED screen. There are modes to
allow a test to be started and paused.
These tests turn on the output voltage,
start the timer and start the charge ac
cumulator. The test can be ended man
ually or automatically at a previously
set endpoint voltage.
Alternatively, the Emulator can
simply be used as a power supply that
can monitor the current consumed by
the circuit under test.
A settings screen can be used to trim
the parameters used to set the output
voltage. Since the Emulator can meas
ure its output, a calibration routine
can set these automatically. You can
also trim the resistance values of the
shunt resistors and adjust numerous
parameters that control the coin cell
emulation.
Since the PIC16F18146 has an in
ternal EEPROM memory (which can
withstand more write cycles than flash
memory), the calibration and setup
parameters are immediately stored in
EEPROM when modified.
Response time
Figs.4 and 5 show the rise and fall
times of the output voltage in response
to a change in the setpoint. These charts
Fig.5: the longer fall time of the Emulator output is almost
entirely due to the 1ms time constant of the 1kW/1µF RC
combination. After about 4ms (four time constants), the
voltage settles near its 0V endpoint.
Practical Electronics | November | 2024
Assembly
The Emulator is built on a small PCB
with surface-mounting components.
They are the typical range of SOIC,
SOT-23 and M3216/1206 parts that are
fairly easy to solder. Fig.6 is the PCB
overlay diagram; you can also refer to
the photo of the PCB before the OLED
module is attached.
We recommend using a fine-tipped
soldering iron, solder flux paste, thin
solder wire, tweezers, a magnifier and
good lighting. Solder wicking braid
is helpful for removing bridges and
excess solder. Work outside if you
don’t have good ventilation or fume
extraction.
Start with the mini-USB socket,
CON1. Apply flux to all its pads and
rest the part on top. Its locating pegs
should lock into holes in the PCB,
aligning it.
100nF
IC2
MAX4239
D1
K
CON3
PIC16F18146
1
10 m F
100nF
100nF
+
10kW
BC817
–
+
CON4
–
1kW Q1100nF10k 1kW 1mF 22W 10k 100nF
Clean your iron’s tip and add a small
amount of fresh solder, then touch it
to where the pins meet the PCB pads.
After that, apply a generous amount of
solder to the four larger pads that affix
the connector’s shell.
If you have bridges between the
pins, add some extra flux and press
some fresh braid against the bridge
with the iron.
When the braid has taken up solder,
slowly draw both away together. If the
part is flat against the PCB, surface
tension should leave enough solder
to form a solid joint.
Fit Q1 next by spreading flux on its
PCB pads and resting it in place, being
sure to align the body with the silk
screen printing. Tack one lead, ensure
the part is flat and aligned within all
pads, then solder the remaining leads.
Solder the two ICs next, using a sim
ilar process, starting with one lead to
locate the part. Both ICs should have
a small pin 1 divot in one corner, so
1 double-sided PCB coded 18101231, 78 × 44mm
1 Mini-USB SMD connector (CON1)
1 5-way right-angle male header, 2.54mm pitch (CON2; optional, for ICSP)
1 1.3in I2C blue OLED module (MOD1)
3 2-pin SMD tactile switches (S1-S3)
4 small self-adhesive rubber feet
Semiconductors
1 PIC16F18146-I/SO microcontroller programmed with 1810123A.HEX,
wide SOIC-20 (IC1)
1 MAX4238 or MAX4239 low-offset op amp, SOIC-8 (IC2)
1 BC817-40 NPN transistor, SOT-23 (Q1)
1 LL4148 SMD diode, SOD-80/MiniMELF (D1)
Silicon Chip Coin Cell
Emulator Kit
Capacitors (all SMD M3216/1206 X7R)
SC6823 (~£18 + P&P)
1 10μF 10V
1 1μF 16V
https://pemag.au/link/AC1J
6 100nF 50V
Resistors (all SMD M3216/1206 1% ¼W unless noted)
5 10kW (code 1002 or 103)
2 1kW (code 1001 or 102)
1 22W ½W (code 22R0 or 22R)
IC1
100nF
CON1
Parts List – Coin Cell Emulator
Practical Electronics | November | 2024
GND VCC SCL SDA
4148
MOD1
10kW LL4148
CON2
10kW
Fig.6: assembling the PCB
mainly involves fitting
SOIC and M3216/1206
SMD parts. Take care
with the orientation
of the two ICs and D1.
‘Mousebites’ are provided
so you can separate the
PCB between CON3 and
CON4; the two halves can
be rejoined with some
light-duty figure-8 wire.
ICSP
were taken in an unloaded state (al
though the Emulator accurately indi
cated the expected 0.3µA draw from
the 10MW scope probe at 3V!).
As expected, the rise time is short,
about 20µs from 0V to 3V. About half
of this is due to the 10µs settling time
of the DAC, with the other half being
the time to charge the 1µF capacitor
with the 200mA available.
The fall time is dominated by the
1ms time constant of the 1kW/1µF
pair and takes about 4ms to settle near
its final value. An external load will
speed this up.
RST
Coin Cell Emulator
S1
S2
S3
align that with the PCB markings.
For IC2, this might be a notch at the
pin 1 end.
For diode D1, ensure its cathode
stripe aligns with the ‘K’ marking on
the PCB. After this, none of the com
ponents are polarised. The capacitors
will not be marked, so be careful not
to get them mixed up. The resistors
will be marked with codes, as shown
in the parts list.
The PCB will now need a thorough
cleaning to remove flux residue. At
the minuscule currents the Emula
tor measures, any contaminants can
cause leakage and interfere with meas
urements.
Your flux might recommend a sol
vent, but we find that isopropyl alco
hol works well (another great option is
Chemtools Kleanium G2). Wipe away
any excess solvent and allow the re
mainder to evaporate thoroughly.
Give the PCB a thorough check now
that it has been cleaned, as any prob
lems will be easier to spot and repair
before the OLED is fitted, as it covers
many of the components.
Now solder on the three tactile
switches, being sure to align them
within their silkscreen outlines and
keep them flat against the PCB. If you
need to program your microcontroller,
add the CON2 ICSP header.
Next, solder the OLED module in
place using its four-pin header, align
ing the pin markings and spacing it
above the other components on the
PCB. When you are happy with its lo
cation, solder stiff wires to the lower
corners of the OLED module and secure
them to the through-hole pads in the
PCB below.
Finally, attach the rubber feet to
the underside of the PCB so it won’t
scratch your work surface.
Programming
PICs supplied in kits or purchased
separately from the Silicon Chip Online
57
Constructional Project
Table 1 – Settings
Page – Parameter
Notes
Set Cap
The default allows for brief tests. It
can be set from 1-10mAh steps of
1mAh or up to 250mAh in steps of
5mAh.
Endpoint
It can be set from 0 to 3.4V in
0.1V steps, the same as the output
voltage.
Current Comp.
The current at which the effective
cell capacity is halved. It can be set
in steps of 0.1mA; if set to 0mA,
there is no compensation. 3.5mA is
typical for CR2032 cells.
Voltage Fall
Below this level, the cell voltage
setpoint is linearly decreased to
reach 0V at 0% SoC. If set to 0%,
then there is no decline in voltage.
Calibrate
Ensure the output is not connected
to any loads and press S1 to start.
This sets the Q1 Vbe and DAC span
automatically. Pressing S2 sets all
parameters back to their defaults.
Set Q1 Vbe
Set by the Calibrate step. If voltages
across the range are still too high,
increase this value. There is a slight
offset below 0.3V output; voltages
are not as accurate in that range.
Set DAC span
If the voltage offset increases
across the range, decrease this; if it
becomes lower, increase it.
Set R(hi) (22W)
It can be set in steps of 0.01W within
10% of 22W. 1% parts should not
need calibration.
Nominal emulated cell
capacity
Default = 10mAh
The voltage at which
tests stop
Default = 2V
Determines how cell
capacity is affected by
high currents
Default = 3.5mAh
SoC at which the
cell voltage starts to
decline
Default = 5%
Start automatic
calibration voltage
Transistor Q1 baseemitter junction
voltage
Default = 588mV
The nominal span of
the DAC output
Default = 4002mV
Actual value of 22W
resistor
Default = 22.00W
It can be set in steps of 1W within
10% of 10kW. 1% parts should not
need calibration.
Trim Timer
The Emulator’s 1Hz clock is available
at the RST pin (with respect to
ground). This can be measured to
help trim the timer. Each step will
change the frequency by about 0.4%.
Exit Setup
All values are saved to EEPROM as
soon as any changes are made and
new settings are used immediately.
The displayed value is
the period of the timer
counter
Default = 243
Press S1 to return to
normal operation
58
Shop come programmed, so skip this
section if you have one of those.
The PIC16F18146 requires a PICkit 4,
PICkit 5 or Snap programmer. If you are
using a Snap (which does not provide
power), you can supply power using
a USB cable connected to CON1. You
might need to use some short exten
sion wires to prevent the Snap from
fouling the USB cable.
You can use the Microchip IPE to
program the 1810123A.HEX file. If you
don’t have the IPE installed, it can be
downloaded and installed for free as
part of the most recent MPLAB X IDE.
Once programmed, the startup OLED
screen should look like Screen 1.
Setup
The Coin Cell Emulator is usable
without calibration, but we recommend
doing it since it is easy and only needs
to be done once. Hold in S3 until the
screen goes blank, then release it to
enter SETUP mode.
Table 1 summarises the individual
setup pages you can cycle through by
pressing S3. In general, S1 decreases
a parameter while S2 increases it. On
some pages, they trigger specific ac
tions, such as starting the automatic
calibration process or returning to
normal operation from SETUP.
The first four SETUP screens relate
to the emulation settings and can be
skipped to reach the calibration set
tings.
We recommend just running the au
tomatic “Calibrate” step. If the Emula
tor’s other measurements are off, you
could consider changing other values,
such as the resistances or timer trim.
Cycle to the Exit Setup page and press
S1 to return to regular operation.
Connections
Set R(lo) (10kW)
Actual value of 10kW
resistor
Default = 10000W
Screen
CON3, CON4 and the circular pads
can all be used to connect to a circuit
under test. For most of our prototyp
ing, we simply used a header socket
for CON3 and ran jumper wires to
our circuit.
The circular section of the PCB is
designed to be slotted into the side
of a cell holder. The photo oppo
site shows the Emulator connected
to our Advanced Test Tweezers. It
probably won’t work with other cell
holder types where the cell is insert
ed from above.
Since the PCB is only 1.6mm thick,
it will not be a tight fit for holders
that expect a 3.2mm-high CR2032
Practical Electronics | November | 2024
Coin Cell Emulator
Screen 1: the initial screen seen when
the Emulator powers on allows the
output voltage setpoint to be changed
with pushbuttons S1 and S2. S3
switches to the other screens. Holding
S3 for three seconds enters the Setup
mode, shown in Table 1.
Screen 2: the output can be toggled
on and off when this screen is shown.
Note also the supply voltage display
at upper right. If this is flashing, the
supply is lower than 4.5V or higher
than 5.5V, and the Emulator may not
function correctly.
Screen 3: S1 and S2 start and
reset the stopwatch timer and
charge accumulator measurement,
respectively. If the timer is running,
this screen will show PAUSE instead,
with S1 pausing the timer if pressed.
cell, although many holders are de
signed to accept 1.6mm thick CR2016
cells. You could carefully bend the
cell holder’s tabs to add more tension.
We also added some thickness to the
Emulator by soldering on some pieces
of wire, as shown on page 56.
Another option is to carefully break
the PCB between CON3 and CON4
(there are ‘mouse bites’ in the PCB to
facilitate this). You could then run a
pair of wires between CON3 and CON4
to join them.
the measured current. It is in a larger
font as it is the most important pa
rameter to observe. If “I(lo)” is shown,
the reading is expected to be accurate
to 0.1µA as only the 10kW resistor is
being used as a shunt.
When “I(hi)” is shown, the Emulator
has switched to the higher range and
the 22W resistor comes into play. When
this happens depends on the output
voltage and supply voltage (which re
lates to IC2’s headroom). At 3V output,
it will occur at around 130μA.
The second-last line shows the stop
watch timer, which measures up to
999 days, or almost three years. The
text on this line indicates if the timer
is running and, if so, the charge meas
urements on the next line are also ac
cumulating.
The µAh reading on the last line
measures actual charge consump
tion (not adjusted). It can be used to
validate the total current consump
tion and estimate potential capacity
losses due to high current usage. The
SoC figure does take into account the
adjusted current.
Pressing S3 shows Screen 2, which
allows the output voltage to be switched
on and off; S1 switches it off, while S2
switches it on.
Screen 3 is reached by pressing S3
again; it allows the timer and charge
accumulator to be paused, started
and reset. S1 will start and pause
the timer, while pressing S2 resets
the timer and accumulator when the
timer is paused.
Press S3 again to reach Screen 4.
Pressing S1 (“GO”) on this screen
will switch on the output voltage and
start the timer and accumulator; S2
(“PAUSE”) will pause the timer and
switch the output off. Thus it can be
used to start and stop testing cycles.
Once you’ve started a test, the cur
rent draw will be shown, and the timer
and accumulator will go up while the
SoC goes down.
As the SoC passes 5%, the output
voltage will drop to simulate the cell
running flat. When the output volt
age reaches the endpoint, the test will
pause, as if S2 were pressed on this
screen, allowing the statistics to be
recorded.
Operation
Screen 1 shows the default Emulator
cell voltage of 3V, which can be changed
on that page. Other features on Screen
1 are common to the operating Screens.
The third line of text shows the status
of the output voltage; the first figure is
the setpoint (target) output voltage and
whether it is on or off.
The other voltages are the values up
stream and downstream of the shunt,
respectively. They can be considered
the internal cell voltage and external
‘terminal’ voltage, respectively. The
first should be very close to the set
voltage (when on), except if the emu
lated cell is nearly flat.
The fourth line (in larger text) shows
Conclusion
We’re already making good use of
the Coin Cell Emulator in designing
an upcoming project. It’s also coming
in handy as a general power supply
PE
for low voltages and currents.
The circular section of the PCB is designed to slot straight into the cell holder
we’ve used for various projects, including the Advanced Test Tweezers
shown here. In this case, testing would be easier if we separated the
PCB between CON3 and CON4 for a more flexible connection.
Screen 4: pressing S1 here starts the
timer and charge accumulator and
switches on the output voltage. S2
pauses the test, allowing the results
to be recorded. The test will be
automatically paused if the Emulator
reaches its endpoint voltage.
Practical Electronics | November | 2024
59
|