22
LTC1923
1923f
APPLICATIO S I FOR ATIO
WUUU
Noise and Slew Rate Control
One disadvantage of switching regulators is that the
switching creates wideband harmonic energy. The high
frequency content can pose problems to associated cir-
cuitry. To combat this issue, the LTC1923 offers a pin
called R
SLEW
that controls the slew rate of the output drive
waveforms. Slowing down the transition interval reduces
the harmonic frequency content by spreading out the
energy over a longer time period. The additional transition
time causes some efficiency loss (on the order of 2% to
3%) but significantly improves the high frequency noise
reflected onto the input supply.
Slew rate control is engaged by placing a resistor from
R
SLEW
to AGND. If slew rate control is not desired, the R
SLEW
pin should be tied to V
DD
allowing the output drivers to
transition at their fastest rate. The resistor value should be
set between 10k (fastest transition) and 300k (slowest tran-
sition). This provides about a 10:1 slew rate range to op-
timize noise performance. The “break-before-make” time
may need to be increased if slew control is implemented,
especially for slower transition rates. Adjustment can be
done by increasing the value of R
T
(C
T
can be reduced to
maintain the same frequency of operation), to ensure that
the bridge MOSFETs receive nonoverlapping drive.
Power MOSFET Selection
Four external MOSFETs must be selected for use with the
LTC1923; a pair of N-channel MOSFETs for the bottom of
the bridge and a pair of P-channel MOSFETs for the top
diagonals of the bridge. The MOSFETs should be selected
for their R
DS(ON)
, gate charge and maximum V
DS,
V
GS
ratings. A maximum V
DS
rating of 20V is more than
sufficient for 5V and 12V bridge applications, but as
mentioned in the High Voltage Application section, a 12V
maximum V
GS
rating is insufficient and higher voltage
MOSFETs must be selected. There is a trade-off between
R
DS(ON)
and gate charge. The R
DS(ON)
affects the conduc-
tion losses (I
TEC
2
• R
DS(ON)
), while gate charge is a
dominant contributor to switching losses. A higher R
DS(ON)
MOSFET typically has a smaller gate capacitance and thus
requires less current to charge the gate for the same
BV
DSS
. For 1A TEC applications, the Si9801DY or Si9928DY
complimentary N- and P-channel MOSFETs provide a
good trade-off between switching and conduction losses.
Above this TEC current level the MOSFETs selected should
have lower R
DS(ON)
to maintain the high end efficiency.
Efficiency Considerations
Unlike typical voltage regulators, where the output voltage
is fixed, independent of load current, the output voltage of
this regulator changes with load current. This is because
the TEC appears resistive and the current through the TEC
sets the voltage. The output power of the regulator is
defined as:
P
OUT
= I
TEC
2
• R
TEC
The efficiency of a switching regulator is equal to the
output power divided by the input power times 100%.
Often it is useful to analyze individual losses to determine
what is limiting the efficiency and which change would
produce the most significant improvement. Efficiency can
be expressed as:
Efficiency = 100% – (L1 + L2 + L3 + …)
where L1, L2, etc. are the individual losses as a percentage
of input power.
For this application, the main efficiency concern is typi-
cally at the high end of output power. A higher power loss
translates into a greater system temperature rise, result-
ing in the need for heat sinking, increasing both the system
size and cost.
There are three main sources which usually account for
most of the losses in the application shown on the front
page of the data sheet: Input supply current, MOSFET
switching losses and I
2
R losses.
1) The input supply current is comprised of the quiescent
current draw from the LTC1658, LTC2053, LTC1923 and
any additional circuitry added. The total maximum supply
current for these devices is on the order of 5mA, which
gives a total power dissipation of 25mW. This power loss
is independent of TEC current.
2) The MOSFET driver current results from switching the
gate capacitance of the power MOSFETs. Each time a gate
is switched from low to high to low again, a packet of
charge dQ moves from V
DD
to ground. The gate charging
current, I
GATECHG
= 2 • f • (Q
P
+ Q
N
), where Q
P
and Q
N
are
23
LTC1923
1923f
APPLICATIO S I FOR ATIO
WUUU
the total gate charges of the NMOS and PMOS on one side
of the bridge, and f is the oscillator frequency. The factor
of 2 arises from there being two sets of MOSFETs that
make up the full bridge. Note that increasing the switching
frequency will increase the dynamic current and therefore
power dissipation by the same factor. This power loss is
independent of TEC current.
Example: Q
N
= 10nC max, Q
P
= 15nC max, f = 225kHz,
V
DD
= 5V
Power loss = 2 • f • (Q
P
+ Q
N
) • V
DD
= 56mW
3) The DC resistances of the external bridge MOSFETs,
filter inductors and sense resistor are typically the domi-
nant loss mechanism at the high end TEC current. The
conduction path of the current includes one NMOS, one
PMOS, two inductors and the sense resistor so the DC
resistances associated with the components dissipate
power.
Example:
R
DS(ON)NMOS
at 5V = 0.055 max
R
DS(ON)PMOS
at 5V = 0.08 max
R
S
= 0.1
R
L
= 0.1,
I
TEC
= 1A
R
TEC
= 2.5
Total series resistance = 0.055 + 0.08 + 2 • 0.1 + 0.1
= 0.435
Power Loss = (1A)
2
• 0.435 = 0.435W
Output Power = (1A)
2
• 2.5 = 2.5W
This represents a 17% efficiency loss due to conduction
losses. The other two power loss mechanisms comprise
a little more than a 3% efficiency loss at this output power
level. This may sound alarming if electrical efficiency is the
primary concern and can be easily improved by choosing
lower R
DS(ON)
MOSFETs, lower series resistance induc-
tors and a smaller valued sense resistor. If temperature
rise is the primary concern, this power dissipation may be
acceptable. At higher current levels, this example does
illustrate that lower resistance components should be
selected.
Low Voltage Requirements
All components shown on the front page of this data sheet
will operate with a 2.7V input supply. Minor modifications
are required to guarantee correct operation. The voltage
on the REF input of the LTC2053 should be at least 1V
below V
DD
. Figure 14 shows how to implement this. By
dividing down the 2.5V reference with 500 of imped-
ance, feeding this to the REF input of the LTC2053 and the
integrating resistor of the LTC1923 error amplifier, any
common mode issues will be avoided.
REF
V
OUT
LTC1658
10k
NTC
A = 10
100k
1923 F14
4.7µF
10M
REF
10k
1µF
V
DD
2.7V TO 3.3V
CNTRL V
REF
V
DD
EAOUT
FB
V
SET
V
THRM
250
250
1µF
+
LTC2053
LTC1923GN
Figure 14. Low Input Supply Voltage Circuit
24
LTC1923
1923f
APPLICATIO S I FOR ATIO
WUUU
Higher Voltage Applications
A bank of TECs can be wired in series to minimize board
real estate utilized by the application. A higher voltage
supply may be required depending upon how many TECs
are placed in series and what their maximum voltage
drop␣ is. In other applications, only one high current
supply may be available, with the output voltage of this
supply being greater than the LTC1923’s absolute maxi-
mum voltage rating. The absolute maximum input voltage
for the LTC1923 is 6V. Since the current drawn by the
LTC1923 is small, it can be powered from a low current,
5V (or less) supply. A 12V application for driving the full
bridge is shown in Figure 15. Two LTC1693-1 high speed
dual MOSFET drivers are used to step up the lower voltage
produced by the LTC1923 drivers to the higher voltage
levels required to drive the full bridge. The LTC1693
requires proper bypassing and grounding due to its high
switching speed and large AC currents. Mount the low
ESR bypass capacitors as close to the pins as possible,
shortening the leads as much as possible to reduce
inductance. Refer to the LTC1693 data sheet for more
information. Since the LTC1693-1 low-to-high and high-
to-low propagation delays are almost identical (typically
35ns), there is minimal skew introduced by the addition
of these drivers. Sufficient dead time (typically 50ns)
between one leg of the bridge shutting off and the other
turning on, as set up by the LTC1923, will be maintained.
If this dead time is insufficient, the resistor tied to the R
T
pin can be increased to increase this time.
Care must be taken to ensure that the external MOSFETs
are properly selected based on the maximum drain-source
voltage, V
DS
, gate-source voltage, V
GS
, and R
DS(ON)
. Many
MOSFETs that have an absolute maximum V
DS
of 20V
have a maximum V
GS
of only 12V, which is insufficient for
12V applications. Even the 14V maximum V
GS
rating of the
Si9801DY may not provide adequate margin for a 12V
bridge supply voltage. Refer to Efficiency Considerations
for more discussion about selecting a MOSFET with
R
DS(ON).
Two pairs of resistors, R
T1
and R
T2,
must be added to
ensure that the absolute maximum input voltage is not
exceeded on the TEC
+
and TEC
inputs. The maximum
voltage on TEC
+
and TEC
must be less than the V
DD
input
supply to the LTC1923 which, for this example, is 5V. The
following equation will guarantee this:
V
R
R
R
k
V
BRIDGE
T
T
T
DD
1
100
1
2
1
++
<
where V
BRIDGE
is the supply voltage to the external bridge
circuitry and V
DD
is the input supply to the LTC1923.
These additional level shifting resistors affect some pa-
rameters in the data sheet. The direction comparator
thresholds are increased to:
(1 + R
T1
/R
T2
+ R
T1
/100k) • 50mV and
(1 + R
T1
/R
T2
+ R
T1
/100k) • –50mV
The output voltage on the V
TEC
pin represents the voltage
across the TEC (V
TECOOLER
) reduced by a factor of
(1 + R
T1
/R
T2
+ R
T1
/100k) or:
V
VTEC
= V
TECOOLER
/(1 + R
T1
/R
T2
+ R
T1
/100k)
The term containing 100k is the loading error introduced
by the input impedance of the differential amplifier. Typi-
cally this value will be 100k, but can vary due to normal
process tolerances and temperature (up to ±30%). Due to
this variability, it may be desirable to minimize the loading
effect to try to keep a tight tolerance on the TEC clamp
voltage. Although it will increase quiescent current draw,
this can be accomplished by making the value of R
T1
as
small as possible.
As a result of this level shifting, the TEC voltage necessary
to activate the clamp is raised. The voltage across the TEC
where the voltage clamp activates will be:
V
TECOOLER
= (1 + R
T1
/R
T2
+ R
T1
/100k) • 2.5V
One drawback with using the LTC1693 MOSFET drivers is
the inability to adjust the slew rate of the output drivers to
reduce system noise.

LTC1923EGN#PBF

Mfr. #:
Manufacturer:
Analog Devices / Linear Technology
Description:
Power Management Specialized - PMIC Hi Eff Thermoelectric Cooler Cntr
Lifecycle:
New from this manufacturer.
Delivery:
DHL FedEx Ups TNT EMS
Payment:
T/T Paypal Visa MoneyGram Western Union