A
DVANCED
L
INEAR
D
EVICES,
I
NC.
e
SAB MOSFET ADVANTAGES
TM
EPAD
E
N
®
AB
LE
D
ALD8100XX/ALD9100XX FAMILY of SUPERCAPACITOR
AUTO BALANCING (SAB
™
) MOSFET ARRAYS
GENERAL DESCRIPTION
The ALD8100xx/ALD9100xx family of Supercapacitor Auto Balanc-
ing MOSFET Arrays, or SAB™ MOSFETs, are designed to address
voltage and leakage current balancing of supercapacitors connected
in series. Supercapacitors, also known as ultracapacitors or
supercaps, connected in series can be balanced with single or
multiple ALD8100xx/ALD9100xx packages. These SAB MOSFETs
are built with ALD production proven EPAD
®
MOSFET technology.
SAB MOSFETs have unique electrical characteristics for superior
active continuous leakage current regulation and self-balancing of
stacked series-connected supercapacitors while dissipating near
zero leakage currents, practically eliminating extra power consump-
tion. For many applications, SAB MOSFET automatic charge bal-
ancing offers a simple, economical and effective method to bal-
ance and regulate supercapacitor voltages. With SAB MOSFETs,
each supercapacitor in a series-connected stack is continuously
monitored and automatically controlled for precise, effective bal-
ancing of its voltage and leakage current.
The SAB MOSFET regulates the voltage across a supercapacitor
cell by increasing its drain current exponentially across the
supercapacitor when its voltage increases, and by decreasing its
drain current exponentially across the supercapacitor when its volt-
age decreases. When a supercapacitor cell is charged to a volt-
age less than 90% of the desired voltage limit, the SAB MOSFET
across the supercap is turned off and there is zero leakage current
contribution from the SAB MOSFET. On the other hand, when the
voltage across the supercapacitor is over the desired voltage limit,
the SAB MOSFET is turned on to increase its drain currents to
keep the supercapacitor voltage from rising. Simultaneously, the
voltages and leakages of other supercapacitors in the series stack
are lowered to result in a near zero net increase in leakage current.
The ALD8100xx/ALD9100xx SAB MOSFET family offers a selec-
tion of different threshold devices for various supercapacitor maxi-
mum operating voltage values and desired leakage balancing char-
acteristics as well as different temperature range environments. A
list of the available ALD part numbers can be found in the tables on
pages 6 and 7 of this document. For individual datasheets and
specifications, please visit www.aldinc.com under “SAB MOSFET”.
SUPERCAPACITORS
Supercapacitors are typically rated with a nominal recommended
working voltage established for long life at their maximum rated
operating temperature. When a supercapacitor cell voltage ex-
ceeds its rated voltage for a prolonged time period, it experiences
reduced lifetime and eventual rupture and catastrophic failure. To
prevent such an occurrence, in most applications having two or
more supercapacitors connected in series, a means of automati-
cally monitoring and adjusting charge-balancing at their maximum
operating voltages is required. This is due to different internal leak-
age currents in each specific cell.
The supercapacitor’s leakage current is a variable function of many
parameters such as aging, initial leakage current at zero input volt-
age, material and construction of the supercapactor, its chemistry
composition, its leakage as a function of the charging voltage and
the charging current and temperature, operating temperature range,
and the rate of change of many of these parameters. Supercapacitor
balancing must correct for these changing effects automatically,
with minimal added leakage currents or power consumption.
The ALD8100xx/ALD9100xx family of SAB MOSFETs are designed
for automatic supercapacitor balancing. They are replacements for
many other passive or active supercapacitor balancing methods
where cost, board space, efficiency, simplicity and power dissipa-
tion are important design considerations. For example, in applica-
tions where supercapacitors require minimum long-term power dis-
sipation (internal leakage currents) as a primary goal, ALD8100xx/
ALD9100xx SAB MOSFETs are simpler and more effective in per-
forming the leakage balancing function, using significantly less
board space and contributing no additional charge loss beyond the
supercapacitor’s own leakages. Other common methods of charge
PIN CONFIGURATIONS
ALD8100xx
IC*
1
M1
M2
16
15
14
V-
V-
13
12
M4
M3
11
10
V-
V-
9
IC*
D
N2
G
N2
S
N2
V+
D
N3
G
N3
S
N3
D
N1
2
G
N1
3
S
N1
4
V-
5
D
N4
6
G
N4
7
S
N4
8
SCL PACKAGES
ALD9100xx
IC*
1
V-
8 V+
7 G
N2
6 D
N2
5 S
N2,
V-
G
N1
2
D
N1
3
S
N1
4
SAL PACKAGES
*IC pins are internally connected, connect to V-
©2014 Advanced Linear Devices, Inc., Vers. 2.0
www.aldinc.com
1 of 17
ALD8100XX/ALD9100XX FAMILY GENERAL DESCRIPTION (cont.)
balancing generally contribute additional continuous power dissi-
pation due to linear currents at all supercapacitor voltage levels,
whereas SAB MOSFET leakages decrease exponentially with de-
creased supercapacitor voltages. In many cases, the additional
leakage charge loss is near zero.
UNDERSTANDING SUPERCAPACITOR AUTO BALANCING
USING SAB™ MOSFETS
The principle behind the SAB MOSFET in balancing
supercapacitors is simple. It is based on the natural threshold char-
acteristics of a MOSFET device. The threshold voltage of a
MOSFET is the voltage at which a MOSFET turns on and starts to
conduct a current. The drain current of the MOSFET, at or below
its threshold voltage, is an exponential function of its gate voltage.
Hence, for small changes in the MOSFET’s gate voltage, its drain-
source on-current can vary greatly, by orders of magnitude. ALD’s
SAB MOSFETs are designed to take advantage of this fundamen-
tal device characteristic.
SAB MOSFETs are connected in the V
t
mode, meaning that the
Gate-to-Source and the Drain-to-Source terminals of each MOSFET
are always connected. In this mode V
GS
is always equal to V
DS
and when this joint terminal is connected across a supercapacitor,
it is also referred to as an Input Voltage, V
IN
. Each SAB MOSFET
has a well defined Drain-to-Source Current, I
DS(ON)
, for different
values of V
IN
Voltages. This current is also referred to as the Out-
put Current, I
OUT
, of the SAB MOSFET.
SAB MOSFETs can be connected in parallel or in series, to suit the
desired leakage current characteristics, in order to charge-balance
an array of supercapacitors connected in series. The array of com-
bined SAB MOSFETs and supercapacitors would be automatically
self-regulating with various leakage mismatches and environmen-
tal temperature changes. The SAB MOSFETs can also be used
entirely in the subthreshold mode, meaning the SAB MOSFET is
used at min., nominal and max. operating voltages in voltage ranges
below its specified threshold voltage.
With the ALD8100xx/ALD9100xx family, the threshold voltage V
t
of an SAB MOSFET is defined as its drain-gate source voltage at a
drain-source ON current, I
DS(ON)
= 1µA, when its gate and drain
terminals are connected together (V
GS
= V
DS
). This voltage is
specified as xx, where the threshold voltage is in 0.10V increments.
Two examples are: the ALD810025 features a precise threshold
voltage of V
t
= 2.500V at I
DS(ON)
= 1µA and the AL ºD810017 has
V
t
= 1.700V at I
DS(ON)
= 1µA.
As all ALD8100xx/ALD9100xx devices operate similarly, with lin-
ear voltage shifts, an ALD810025 is used as an illustration of its
characteristics. At voltages below its threshold voltage, the
ALD810025 rapidly turns off at a rate of approximately one decade
of current per 104mV of voltage drop. Hence, at V
IN
= 2.396V, the
ALD810025 I
OUT
is 0.1µA. At V
IN
= 2.292V, its I
OUT
becomes
0.01µA. When V
IN
drops further to 2.188V, its I
OUT
becomes
0.001µA. It should be apparent that at V
IN
≤
2.10V, the ALD810025
I
OUT
≤
0.00014µA, which is near zero when compared to 1µA at
V
IN
= 2.50V. At V
IN
below 1.9V, the SAB MOSFET Output Current,
I
OUT
, goes to essentially zero (~70pA). The I
OUT
≤
0.00014µA is
controlled and repeatable for different units from various produc-
tion batches.
This exponential relationship between the SAB MOSFET’s V
IN
and
I
OUT
can be an important consideration in replacing certain
supercapacitor charge balancing applications currently using fixed
resistors, operational amplifier circuits or other forms of charge bal-
ancing. These other conventional charge-balancing methods con-
tinue to dissipate a significant amount of current, even after the
voltage across the supercapacitors has dropped, because the cur-
rent dissipated is a linear function, rather than an exponential func-
tion, of the supercapacitor voltage (I = V/R). For supercapacitor
series stacks with more than two cells, the challenge of leakage
balancing becomes even more onerous.
With most other passive or active circuits that offer charge balanc-
ing, active power is still being consumed even if the supercapacitor
voltage falls much below its operating voltage. For a four-cell
supercapacitor stack, for example, this translates into a 2.0V x 4
~= 8.0V power supply for an IC charge-balancing circuit. As the
number of cells increase, adding components to the charge bal-
ancing circuit, increased circuit complexity and power dissipation
becomes a greater challenge. A supercapacitor stack using the SAB
MOSFET charge-balancing method, on the other hand, would not
cause extra power dissipation when the number of cells increase.
This method provides an exponentially decreasing amount of charge
loss with time, and helps preserve, by far, the greatest amount of
charge on each of the supercapacitors.
If V
IN
of the ALD810025 is greater than its V
t
threshold voltage, its
Output Current, I
OUT
, behavior has the opposite near-exponential
effect. At V
IN
= 2.60V, for example, the ALD810025 I
OUT
increases
tenfold to 10µA. Similarly, I
OUT
becomes 100µA at a V
IN
of 2.74V,
and 300µA at V
IN
of 2.84V. (See Table 1.)
As I
OUT
changes rapidly with the applied V
IN
, the SAB MOSFET
device acts like a voltage limiting regulator with self-adjusting cur-
rent levels. When the SAB MOSFET is connected across a
supercapacitor cell, the total leakage current equals the two in com-
bination automatically compensate and correct for each other.
Consider the case when two supercapacitor cells are connected in
series, each with an SAB MOSFET connected across it, charged
by a power supply to a voltage equal to 2 x V
S
.
If the top supercapacitor has a higher internal leakage current than
the bottom supercapacitor, the voltage V
S(top)
across it tends to
drop lower than that of the bottom supercapacitor. The SAB
MOSFET I
OUT
across the top supercapacitor, sensing this voltage
drop, drops off rapidly. Meanwhile, the bottom supercapacitor
V
S(bottom)
voltage tends to rise, as V
S(bottom)
= (2 x V
S
) - V
S(top)
.
This tendency for the voltage rise also increases V
IN
voltage of the
SAB MOSFET across the bottom supercapacitor. This increased
V
IN
voltage would cause the I
OUT
of the bottom SAB MOSFET to
increase rapidly as well. The excess leakage current of the top
supercapacitor would now leak across the bottom SAB MOSFET,
reducing the voltage rise tendency of the lower supercapacitor. With
this automatic self-regulating mechanism, the top supercapacitor
voltage tends to rise while the bottom supercapacitor voltage tends
to drop, creating simultaneously opposing actions to the
supercapacitor leakage currents.
With appropriate design and selection of a specific SAB MOSFET
device for a given pair of supercapacitors, it is now possible to
regulate and balance two series-connected supercapacitors with
essentially no extra leakage current, since the SAB MOSFET only
conducts the difference in leakage current between the two
supercapacitors.
Likewise, the case of the bottom supercapacitor having a higher
leakage current than that of the top supercapacitor works in similar
fashion, where the bottom supercapacitor voltage tends to drop,
compensated by the tendency of the top supercap voltage to drop
as well, effected by the top SAB MOSFET. This SAB MOSFET
charge balancing scheme also works with four, eight or more
supercapacitors in series by using an equivalent number of SAB
MOSFETs in multiple package(s).
Ambient temperature increases cause supercapacitor leakage cur-
rents to increase. The SAB MOSFET threshold voltage is reduced
ALD8100XX/ALD9100XX SUPERCAPACITOR
Advanced Linear Devices, Inc.
AUTO BALANCING (SAB) MOSFET ARRAY FAMILY
2 of 17
ALD8100XX/ALD9100XX FAMILY GENERAL DESCRIPTION (cont.)
with temperature increase, which causes its I
OUT
to increase with
temperature as well. This current increase compensates for the
leakage current increase within the supercapacitor, reducing the
overall supercapacitor temperature leakage effect and preserving
charge balancing effectiveness. This temperature compensation
assumes that all supercapacitors and SAB MOSFETs operate in
the same temperature environments.
SAB MOSFET LIMITATIONS
2) Determine the max. leakage current of each supercapacitor.
During supercapacitor charging, consideration must be paid to limit
the rate of charging so that excessive voltage and current does not
build up across any two pins of the SAB MOSFETs, even momen-
tarily, to exceed their absolute maximum ratings in voltage, operat-
ing current, and power dissipation. In most cases though, this is
not an issue, as other design constraints elsewhere limit the rate of
charging or discharging of the supercapacitors. For many applica-
tions, no further action, other than checking the voltage and cur-
rent excursions, or including a simple current-limiting charging re-
sistor, is necessary.
For each SAB MOSFET, its V+ pin must be connected to the most
positive voltage and its V- and IC pins to the most negative voltage
within the package. SAB MOSFETs have numerous pins required
for its manufacturing process, which must be connected to the
supercapacitors when in use, for proper circuit operation. Multiple
packages can be cascaded for higher system voltages as long as
absolute maximum ratings are observed for each individual pack-
age.
Note that each Drain pin of a SAB MOSFET has an internal re-
verse biased diode to its Source pin, which can become forward-
biased if the Drain voltage should become negative relative to its
Source voltage. This forward-biased diode clamps the Drain volt-
age to limit the negative voltage relative to its Source voltage, and
is limited to a 80mA max. rated current between any two pins.
Each Gate pin also has a reverse biased diode to V-. When for-
ward biased, the maximum diode current must be within the abso-
lute maximum ratings. All other pins must have voltages within V+
and V- voltage limits. Standard ESD protection facilities and han-
dling procedures for static sensitive devices must also be followed
before the SAB MOSFETs are installed. Once the SAB MOSFET
is permanently connected to the supercapacitors, ESD concerns
are relieved because any extraneous electrostatic charge would
be absorbed by the supercapacitor and would not cause exessive
voltage increase.
EXTENDED TEMPERATURE RANGE OPERATION
SAB MOSFETs are built with solid state integrated circuit tehcnology.
They are available for operation over a wide temperature range,
with appropriate derating, screening and packaging. Standard com-
mercial grade devices are rated for operation at 0
°
C to +70
°
C.
Industrial temperature range (“I” suffix) units are rated for -40
°
C to
+85
°
C. Custom versions are also available for military temperature
ranges (“M” suffix), -55
°
C to +125
°
C.
MATCHING SAB™ MOSFETS TO SUPERCAPACITORS
Figure 1 shows a basic connection diagram of two SAB MOSFETs
connected across two supercapacitors, powered by a V+ power
supply with an external (or internal) resistor, with basic equations
of SAB MOSFET and supercapacitor voltages and currents.
3) Determine the desired nominal operating voltage of the
supercapacitor.
4) Determine the maximum operating voltage rating of the
supercapacitor.
5) Calculate or measure the maximum leakage current of the
supercapacitor at its maximum rated operating voltage.
6) Determine the operating temperature range of the supercapacitor.
7) Determine any additional level of operating leakage current in
the system.
Next, determine the normalized I
OUT
of an SAB MOSFET at a pre-
selected V
IN
operating voltage.
For example, the ALD810025 has a rated Drain Current of 1µA at
applied V
IN
of 2.50V. If the desired normalized I
OUT
is 0.01µA,
then the ALD810025 would give a bias V
IN
voltage of approximately
2.3V at that current, which produces an equivalent ON resistance
of 2.3V/0.01µA ~= 230MΩ (using the rule of thumb: one decade of
I
OUT
change per 0.10V of V
IN
change).
The proess of selecting SAB MOSFETs to match specific models
of supercapacitors begins by analyzing the parameters and the re-
quirements of a given set of supercapacitors:
1) For better leakage current matching results, pick the same make
and model of supercapacitors to be connected in a series. If pos-
sible, select supercapacitors from the same production batch. (Note:
SAB MOSFETs are precisely set at the factory and specified such
that their unit-to-unit variation is not a concern.)
FIGURE 1
V+
RSRC
+
VIN1
+
VC1
IC1
VSUPPLY
-
IOUT1
+
VIN2
IOUT2
-
+
VC2
-
IC2
-
Basic Equations:
V+ = VC1 + VC2 + RSRC (IC1 + IOUT1)
V+ = VC1 + VC2 + RSRC (IC2 + IOUT2)
VC1 = VIN1
VC2 = VIN2
IC1 + IOUT1 = IC2 + IOUT2
VSUPPLY = VC1 + VC2
ALD8100XX/ALD9100XX SUPERCAPACITOR
Advanced Linear Devices, Inc.
AUTO BALANCING (SAB) MOSFET ARRAY FAMILY
3 of 17
ALD8100XX/ALD9100XX FAMILY GENERAL DESCRIPTION (cont.)
CHOOSING A SPECIFIC SAB MOSFET
In choosing SAB MOSFETs for a specific application, go to the
SAB MOSFET selection table, (Table 1 for ALD8100xx devices,
Table 2 for ALD9100xx devices) where each SAB MOSFET Part
Number and its respective parameters are listed. First, select an
SAB MOSFET I
OUT
Current horizontally across the top row of the
Table(s). Next, look down that column to the row that contains the
maximum desired V
IN
voltage. The appropriate ALD part number
is in the first column of that row. The part number of an SAB
MOSFET references its rated threshold voltage, but that is not nec-
essarily the desired operating voltage where the auto-balancing
supercapacitor operates. Generally, the recommended maximum
supercapacitor I
OUT
auto-balancing for the ALD8100xx/ALD9100xx
family is about 1mA. When supercapacitor leakage current exceeds
1mA, the effectiveness of the SAB MOSFET auto-balancing gradu-
ally diminishes and there is additional leakage current contribution
from the SAB MOSFET itself as its V
IN
increases. Please contact
techsupport@aldinc.com for more information or technical assis-
tance.
A DESIGN EXAMPLE
A single 5V power supply using two 2.7V rated supercapacitors
connected in series and a single ALD810026 SAB MOSFET array
package (using two of the four devices in the package).
For a supercapacitor with:
1) max. operating voltage = 2.70V and
2) max. leakage current = 10µA at 70
°
C.
3) At 2.50V, the supercapacitor max. leakage current = 2.5µA at
25
°
C.
Next, pick ALD810026, a SAB MOSFET with V
t
= 2.60V. For this
device, at V
IN
= 2.60V, the nominal I
OUT
= 1µA. See Table 1, at
V
IN
= 2.50V, I
OUT
~= 0.1µA.
At a nominal V
IN
of 2.50V, the additional leakage current contribu-
tion by the ALD810026 is therefore ~= 0.1µA. The total leakage
current for the supercapacitor and the SAB MOSFET = 2.5µA +
0.1µA ~= 2.6µA @ 2.50V operating voltage. When operating volt-
age becomes 2.40V, additional ALD810026 leakage current con-
tribution decreases to about 0.01µA.
At V
IN
of 2.70V across the ALD810026 SAB MOSFET, I
OUT
= 10µA.
10µA is also the max. leakage current design margin, the differ-
ence between top and bottom supercapacitor leakage currents that
can be compensated.
If a higher max. leakage current margin is desired, then SAB
MOSFET selection may need to go to the next SAB MOSFET part
down in Table 1, which is ALD810025. For ALD810025 operating
at a max. rated voltage of 2.70V, the max. leakage current margin
is ~= 50µA. For this device, I
OUT
at 2.50V is ~= 1µA, which is the
average current consumption for the series-connected stack. The
total current for the supercapacitor and the SAB MOSFET is = 2.5µA
+ 1µA ~= 3.5µA @ 2.50V operating voltage.
Because an SAB MOSFET is always active and always in “on”
mode, there is no circuit switching or sleep mode involved. This
may become an important factor when the time interval between
the supercap discharging or recharging, and other events happen-
ing in the application, is long, unknown or variable. The circuit op-
eration is also greatly simplified.
In real life situations, the actual circuit behavior is a little different,
further reducing overall leakage currents from both supercapacitors
and SAB MOSFETs, due to the automatic compensation for differ-
ent leakage currents from the supercapacitors by themselves and
in combination with the SAB MOSFETs. Take an example of two
supercapacitors in series, assuming that the top supercapacitor is
leaking 10µA and the bottom one is leaking 4µA (both at the rated
2.7V max.) while the power supply remains at 5V DC. The actual
voltage across the top supercapacitor tends to be less than 2.5V
(50% of 5.0V), due to its higher internal leakage current, and re-
sults in a lowered current level than 10µA because the current tends
to be lower at less than 2.7V. As the total voltage across both
supercapacitors is still 5.0V, each supercapacitor would experience
a lowered voltage than its maximum rated voltage of 2.7V, thereby
resulting in reduced overall leakage currents in each of the two
supercapacitors.
These leakage currents are then further regulated by the SAB
MOSFETs connected across each of the supercapacitors. The end
result is a compensated condition where, for example, the top
supercapacitor has V
IN
of ~2.4V across it and the bottom
supercapacitor has V
IN
of ~2.6V. The excess leakage current of
the top supercapacitor is bypassed across the bottom SAB
MOSFET. Meanwhile the top SAB MOSFET, with ~2.4V across it,
is biased to conduct very little I
OUT
. Note also that the top
supercapacitor is now biased at ~2.4V and, therefore, would expe-
rience less current leakage than when it is at 2.7V. The key point
here is that this process of leakage current balancing is fully auto-
matic and works for a variety of supercapacitors, each with its own
different leakage current characteristic profile.
A second factor to note is that with ~2.4V and ~2.6V across the two
supercapacitors, as in this example, the actual current level differ-
ence between the top and the bottom SAB MOSFETs is at about a
100:1 ratio (~2 orders of magnitude). The net additional leakage
current contributed by the ALD810026 in the design example above
would, therefore, be approximately 0.01µA. In this case, leakage
currents between the two supercapacitors can be at a ratio of 100:1
and still experience charge balancing and voltage regulation. If a
range of supercapacitor leakge currents can be determined or se-
lected for a particular model of supercapacitor across different pro-
duction batches, then a SAB MOSFET part can be specified that
further minimizes any SAB leakage currents and still maintains bal-
anced supercapacitor voltages within a narrow range.
The dynamic response of a SAB MOSFET circuit is very fast, and
the typical response time is determined by the RC time constant of
the equivalent ON resistance value R
ON
of the SAB MOSFET and
the capacitance value C of the supercapacitor. In many cases the
R
ON
value is small initially, responding rapidly to a large voltage
transient by having a smaller R
ON
C time constant. As the volt-
ages settle down, the equivalent R
ON
increases. As these R
ON
and C values can become very large, it can take a long time for the
voltages across the supercaps to settle down to steady state lev-
els. The direction of the voltage movements across the
supercapacitor, however, can indicate that the supercapacitor volt-
ages are moving away from the voltage limits.
A HIGH LEAKAGE CURRENT DESIGN EXAMPLE
A nominal 12V DC power supply connects across a supercapacitor
series stack consisting of six 2.0V supercapacitor cells. Each cell
has a nominal operating voltage of 2.0V and is rated at 2.5V max.
Maximum voltage across the stack is 13.92V, which results in a
per-cell voltage of 2.32V. The max. leakage current for the
supercapacitor is rated at 1mA at 2.5V.
Next, we choose a maximum acceptable supercapacitor in-balance
stack voltage of 2.42V, which allows for temperature and aging ef-
fects, among other factors. When we look down the column of
1000µA (1mA) in Table 1 to locate a V
IN
voltage of 2.42V, we find
the corresponding ALD part number to be ALD810019.
ALD8100XX/ALD9100XX SUPERCAPACITOR
Advanced Linear Devices, Inc.
AUTO BALANCING (SAB) MOSFET ARRAY FAMILY
4 of 17
ALD8100XX/ALD9100XX FAMILY GENERAL DESCRIPTION (cont.)
In the graph titled “Input Voltage vs. Output Current”, locate the
V
IN
point as follows. First, find the V
t
of the ALD810019 from the
SAB MOSFET Selection Table, which is 1.90V. Next, subtract 1.90V
from 2.42V, which is 0.52V. Check the I
OUT
current variation and
voltage variation as a function of temperature. If the temperature
variation allowance is 60mV, then the maximum supercap inbalance
voltage is 2.48V (2.42V + 0.06V) across temperature.
In cases where the supercapacitor leakage current is 1mA max.,
the ALD810019 is suggested. In cases where supercapacitor leak-
age currents are up to 3mA, then a part such as the ALD81016
can be used, although this may cause increased leakage current
through the SAB MOSFET itself. Another way to reduce leakage
currents would be to parallel connect mulitple ALD810019 devices
to auto-balance leakage currents greater than 1mA.
A 4.2V SUPERCAPACITOR STACK DESIGN EXAMPLE
A supply voltage of 4.2V across two supercapacitors gives 2.1V
across each supercapacitor cell. With a maximum leakage current
of 100µA for each cell at 2.22V maximum V
IN
cell voltage, the cor-
responding ALD part number is ALD910020SAL, a dual 8L SOIC
package.
The ALD910020 would support an I
OUT
(supercapacitor leakage
current) of 300µA at V
IN
= 2.30V; 100µA at V
IN
= 2.22V; 10µA at
V
IN
= 2.10V and 1µA at V
IN
= 2.00V, respectively. An inbalance
leakage current ratio between two supercapacitor cell units of 100µA
to 1µA, a 100 to 1 ratio, would produce one cell voltage of 2.22V
and the other cell voltage of 1.98V, which adds up to 4.20V. Simi-
larly, a lower supply voltage than 4.2V would be divided between
the two supercapacitors corresponding to their respective leakage
currents. Consider the case when the supply voltage is 4.10V, each
with an ALD910020 connected to it. If the leakage current ratio
between the supercapacitors remains the same, then one cell would
be biased at 2.22V (100µA) and the other would be biased at 1.88V
(4.10V - 2.22V). This would cause the ALD910020 to have a max.
leakage current contribution of less than 0.1µA.
PARALLEL-CONNECTED AND SERIES-CONNECTED SAB
MOSFETS
In the first design example on the previous page, note that the
ALD810026 is a quad pack, with four SAB MOSFETs in a single
SOIC package. For applications where two supercapacitors are
connected in series, the ALD9100xx dual SAB MOSFET is recom-
mended for charge balancing. If a two-stack supercapacitor re-
quires charge balancing, then there is also an option to parallel-
connect two additional SAB MOSFETs of the quad ALD8100xx for
each of the two supercapacitors. Parallel-connection means that
the drain, gate and source terminals of each of the two SAB
MOSFETs are connected together to form a single MOSFET with
twice the output current and twice the output current sensitivity to
voltage change. In this case, at an operating V
IN
voltage of 2.50V,
the additional I
OUT
current contribution by the SAB MOSFET is
equal to 2 x 0.1µA = 0.2µA. The total current for the combined
supercapacitor and SAB MOSFET is = 2.5µA + 0.2µA ~= 2.7µA @
2.50V operating voltage. At max. voltage of 2.70V across the SAB
MOSFET, V
IN
= 2.70V results in a I
OUT
of 2 x 10µA = 20µA. So
this configuration would be chosen to increase max. supercapacitor
charge balancing leakage current at 2.70V to 20µA, at the expense
of an additional 0.1µA I
OUT
leakage at 2.50V.
For stacks of series-connected supercapacitors consisting of more
than three or four cells, it is possible to use a single SAB MOSFET
array for every supercapacitor stack (up to 4 cells) connected in
series. Multiple SAB MOSFET arrays can be arrayed across mul-
tiple supercapacitor stacks to operate at higher operating voltages.
It is only important to limit the voltage across any two pins within a
single SAB MOSFET array package to be less than its absolute
maximum voltage and current ratings.
LOW LEAKAGE ENERGY HARVESTING APPLICATIONS
Supercapacitors offer an important benefit in energy harvesting ap-
plications with a high impedance energy source, in buffering and
storing such energy to drive a higher power load.
For energy harvesting applications, supercapacitor leakage cur-
rents are a critical design parameter, as the average energy har-
vesting input charge must exceed the average supercapacitor in-
ternal leakage currents in order for any net energy to be harvested.
When the input energy is a variable, meaning that its input voltage
and current magnitude is not constant and dependent upon other
parameters such as the source energy availability (energy sensor
conversion efficiency, etc.), the energy harvested and stored must
supply and exceed the necessary leakage currents, which tend to
be steady DC currents.
In these types of applications, in order to minimize the amount of
energy loss due to leakage currents, it is essential to choose
supercapacitors with low leakage specifications and to use SAB
MOSFETs to balance them.
For the first 90% of the initial voltages of a supercapacitor used in
energy harvesting applications, supercapacitor charge loss is lower
than its maximum leakage rating, at less than its max. rated volt-
age. SAB MOSFETs, used for charge balancing, would be com-
pletely turned off, consuming zero leakage current while the
supercapacitor is being charged, maximizing any energy harvest-
ing gathering efforts. The SAB MOSFET would not become active
until the supercapacitor is already charged to over 90% of its max.
rated voltage. The trickle charging of supercapacitors with energy
harvesting techniques tends to work well with SAB MOSFETs as
charge balancing devices, as it is less likely to have high transient
energy spurts resulting in excessive voltage or current excursions.
If an energy harvesting source only provides a few
µA
of current,
the power budget would not allow wasting any of this current on
capacitor leakage currents, and on many other conventional charge
balancing methods. Resistors or operational amplifiers used as
charge-balancing circuits would dissipate far more energy than
desired. It may also be an important consideration to reduce long
term DC leakage currents as energy harvesting charging at low
levels may take up to many days.
In summary, in order for an energy harvesting application to be
successful, the input energy harvested must exceed all the energy
spent, due to the leakages of the supercapacitors and the charge-
balancing circuits, plus any load requirements. With their unique
balancing characteristics and near-zero charge loss, SAB MOSFETs
are ideal devices to use for supercapacitor charge-balancing within
energy harvesting applications.
LONG TERM BACKUP BATTERY APPLICATIONS
Similar to energy harvesting applications, any low leakage long-
term application, such as a long-term backup battery requiring
supercapacitors at the output to reduce output impedance and to
boost its output power, would benefit from SAB MOSFET deploy-
ment. Over a long time span, reducing leakge currents is an impor-
tant design parameter. For example, a low DC leakage current of
just 1µA over 5 years translates into 44.8mAhr of energy lost.
ALD8100XX/ALD9100XX SUPERCAPACITOR
Advanced Linear Devices, Inc.
AUTO BALANCING (SAB) MOSFET ARRAY FAMILY
5 of 17
京公网安备 11010802033920号
Copyright © 2005-2026 EEWORLD.com.cn, Inc. All rights reserved
电子工程世界
