Datasheet

LT3748
10
3748fb
For more information www.linear.com/LT3748
applications inForMation
Selecting Actual R
REF
, R
FB
and R
TC
Resistor Values
The preceding equations define how the LT3748 would
regulate the output voltage if the system had no time delays
and no error sources. However, there are a number of
repeatable delays and parasitics in each application which
will affect the output voltage and force a re-evaluation of
the R
FB
and R
TC
component values. The following approach
is the best method for selecting the correct values.
The expression for V
OUT,
developed in the Operation sec-
tion, can be rearranged to yield the following expression
for R
FB
:
R
FB
=
R
REF
• N
PS
V
OUT
+ V
F
( )
+ V
TC




V
BG
where:
V
OUT
= Output voltage
V
F
= Output diode forward voltage
N
PS
= Effective primary-to-secondary turns ratio
V
TC
= 0.55V
The equation assumes the temperature coefficients of the
output diode and V
TC
are equal and substitutes R
FB
/N
PS
for
the value of R
TC
. This is a good first order approximation
but will be revisited later.
First, the value of R
REF
should be approximately 6.04k
since the LT3748 is trimmed and specified using this
value. If the impedance of R
REF
varies considerably from
6.04k, additional errors will result. However, a variation in
R
REF
of several percent is acceptable. This yields a bit of
freedom in selecting standard 1% resistor values to yield
nominal R
FB
/R
REF
ratios.
With starting values for R
FB
and R
TC
, an initial iteration
of the application should be built with final selections of
all external components (transformer, diode, MOSFET,
etc.). The resulting V
OUT
should be measured and used
to re-evaluate the value of R
FB
due to non-idealities in the
sampling system:
R
FB(NEW)
=
V
OUT(DESIRED)
V
OUT(MEASURED)
• R
FB(OLD)
With a new value of R
FB
selected, the temperature co-
efficient of the output diode in the application can be
tested
to verify the nominal R
TC
value. The R
TC
resistor
should be removed from the circuit under test (this will
cause V
OUT
to increase for this step) and V
OUT
should
be measured over temperature at a desired target output
load. It is very important for this evaluation that uniform
temperature be applied to both the output diode and the
LT3748—if freeze spray or a heat gun is used there can
be a significant mismatch in temperature between the
two devices that causes significant error. Attempting to
extrapolate the data from a diode datasheet or assuming
the nominal R
TC
value may yield a better result if there is
no method to apply uniform heat or cooling such as an
oven. With at least two data points (although more data
points from hot to cold are recommended), the change
in V/°C can be determined by:
∆V
OUT
∆TEMP
=
V
OUT1
– V
OUT2
TEMP1– TEMP2
Using the measured V
OUT
temperature coefficient, an exact
R
TC
value can be selected using the following equation:
R
TC
=
R
FB
N
PS
•
1.85mV/°C
∆V
OUT
∆TEMP
If the value of R
TC
has changed significantly, which can
happen with the use of some output diodes that have
a very low forward drop, the R
FB
value may need to be
changed to restore V
OUT
to the desired value. As in the
previous iteration, after measuring V
OUT
, a new R
FB
can
once again be selected using:
R
FB(NEW)
=
V
OUT(DESIRED)
V
OUT(MEASURED)
• R
FB(OLD)
Once the values of R
FB
and R
TC
are selected, the regulation
accuracy from board to board for a given application will be
very consistent, typically under ±5% when including device
variation of all the components in the system (assuming
resistor tolerances and transformer windings matching
of 1% or better). However, if the transformer, the output
diode or MOSFET switch are changed or the layout is
dramatically altered, there may be some change in V
OUT
.
Downloaded from Arrow.com.Downloaded from Arrow.com.Downloaded from Arrow.com.Downloaded from Arrow.com.Downloaded from Arrow.com.Downloaded from Arrow.com.Downloaded from Arrow.com.Downloaded from Arrow.com.Downloaded from Arrow.com.Downloaded from Arrow.com.