LTC6409
13
6409fa
Finally, noise figure can be obtained as:
NF = 10log 1+
e
no
2
e
no
2
(RS)
Figure 7 specifies the measured total output noise (e
no
),
excluding the noise contribution of source resistance, and
noise figure (NF) of LTC6409 configured at closed loop
gains (A
V
= R
F
/R
I
) of 1V/V, 2V/V and 5V/V. The circuits in
the left column use termination resistors and transform-
ers to match to the 50Ω source resistance, while the
circuits in the right column do not have such matching.
For simplicity, DC-blocking and bypass capacitors have
not been shown in the circuits, as they do not affect the
noise results.
Relationship Between Different Linearity Metrics
Linearity is, of course, an important consideration in
many amplifier applications. This section relates the inter-
modulation distortion of fully differential amplifiers to
other linearity metrics commonly used in RF style blocks.
Intercept points are specifications that have long been used
as key design criteria in the RF communications world as
a metric for the intermodulation distortion performance of
a device in the signal chain (e.g., amplifiers, mixers, etc.).
Intercept points, like noise figures, can be easily cascaded
back and forth through a signal chain to determine the
overall performance of a receiver chain, thus resulting
in simpler system-level calculations. Traditionally, these
systems use primarily single-ended RF amplifiers as gain
blocks designed to operate in a 50Ω environment, just like
the rest of the receiver chain. Since intercept points are
given in dBm, this implies an associated impedance of 50Ω.
However, for LTC6409 as a differential feedback amplifier
with low output impedance, a 50Ω resistive load is not re-
quired (unlike an RF amplifier). This distinction is important
when evaluating the intercept point for LTC6409. In fact,
the LTC6409 yields optimum distortion performance when
loaded with 200Ω to 1kΩ (at each output), very similar to
the input impedance of an ADC. As a result, terminating
applicaTions inForMaTion
Figure 6. A More General Noise Model Including
Source and Termination Resistors
calculation, a termination resistor (R
T
) is included and its
noise contribution is taken into account.
Now, the total output noise power (excluding the noise
contribution of R
S
) is calculated as:
e
no
2
= e
ni
1+
R
F
R
I
+
R
T
||R
S
2
2
+ 2 i
n
R
F
( )
2
+
2 e
nRI
R
F
R
I
+
R
T
||R
S
2
2
+ 2 e
nRF
2
+
e
nRT
R
F
R
I
2R
I
||R
S
R
T
+ 2R
I
||R
S
( )
2
Meanwhile, the output noise power due to noise of R
S
is
given by:
e
no
2
(RS)
= e
nRS
R
F
R
I
2R
I
||R
T
R
S
+ 2R
I
||R
T
( )
2
+
e
no
2
R
F
V
OCM
e
nRI
2
R
F
R
I
R
I
R
S
e
nRF
2
e
nRS
2
R
T
e
nRT
2
e
nRI
2
e
ni
2
e
nRF
2
i
n
+
2
i
n
2
6409 F06
LTC6409
14
6409fa
applicaTions inForMaTion
the input of the ADC to 50Ω can actually be detrimental
to system performance.
The definition of 3rd order intermodulation distortion
(IMD3) is shown in Figure 8. Also, a graphical repre-
sentation of how to relate IMD3 to output/input 3rd
order intercept points (OIP3/IIP3) has been depicted in
Figure 9. Based on this figure, Equation (4) gives the
definition of the intercept point, relative to the intermodu-
lation distortion.
OIP3=P
O
+
IMD3
2
(4)
P
O
is the output power of each of the two tones at which
IMD3 is measured, as shown in Figure 9. It is calculated
in dBm as:
P
O
=10log
V
2
PDIFF
2 R
L
10
3
(5)
where R
L
is the differential load resistance, and V
PDIFF
is
the differential peak voltage for a single tone. Normally,
intermodulation distortion is specified for a benchmark
composite differential peak of 2V
P-P
at the output of the
50Ω
V
IN
+
50Ω
V
IN
+
+
150Ω
1.3pF
150Ω
150Ω
150Ω
50Ω
1:4
6409 F07
1.3pF
V
OCM
V
IN
+
600Ω
e
no
= 4.70nV/√Hz
NF = 14.41dB
+
200Ω
1pF
200Ω
100Ω
100Ω
50Ω
1:4
1pF
V
OCM
V
IN
+
e
no
= 5.77nV/√Hz
NF = 10.43dB
+
500Ω
500Ω
100Ω
100Ω
50Ω
1:4
V
OCM
V
IN
+
e
no
= 11.69nV/√Hz
NF = 8.81dB
+
150Ω
1.3pF
150Ω
150Ω
150Ω
50Ω
1.3pF
V
OCM
V
IN
+
e
no
= 5.88nV/√Hz
NF = 17.59dB
+
200Ω
1pF
200Ω
100Ω
100Ω
1pF
0.4pF
0.4pF
0.8pF
0.8pF
V
OCM
e
no
= 9.76nV/√Hz
NF = 16.66dB
+
250Ω
250Ω
50Ω
50Ω
V
OCM
e
no
= 14.23nV/√Hz
NF = 13.56dB
Figure 7. LTC6409 Measured Output Noise and Noise Figure at Different Closed Loop Gains with and without Source Impedance Matching
LTC6409
15
6409fa
R
S
50Ω
V
S
R
F
LTC6409
100Ω
100Ω
R
F
C
F
R
I
R
I
R
T
R
L
50Ω
C
BA
6409 F10
1dB
LOSS
IDEAL
4:1
IDEAL
1:4
1dB
LOSS
+
C
F
R
T
applicaTions inForMaTion
Figure 8. Definition of IMD3
Figure 9. Graphical Representation of the
Relationship between IMD3 and OIP3
amplifier, implying that each single tone is 1V
P-P
, result-
ing in V
PDIFF
= 0.5V. Using R
L
= 50Ω as the associated
impedance, P
O
is calculated to be close to 4dBm.
As seen in Equation (5), when a higher impedance is used,
the same level of intermodulation distortion performance
results in a lower intercept point. Therefore, it is impor-
tant to consider the impedance seen by the output of the
LTC6409 when working with intercept points.
Comparing linearity specifications between different am-
plifier types becomes easier when a common impedance
level is assumed. For this reason, the intercept points
for LTC6409 are reported normalized to a 50Ω load im-
pedance. This is the reason why OIP3 in the Electrical
Characteristics table is 4dBm more than half the absolute
value of IMD3.
If the top half of the LTC6409 demo board (DC1591A,
shown in Figure 12) is used to measure IMD3 and OIP3,
one should make sure to properly convert the power seen
at the differential output of the amplifier to the power that
appears at the single-ended output of the demo board.
Figure 10 shows an equivalent representation of the top
half of the demo board. This view ignores the DC-blocking
and bypass capacitors, which do not affect the analysis
here. The transmission line transformers (used mainly
for impedance matching) are modeled here as ideal 4:1
impedance transformers together with a –1dB block. This
separates the insertion loss of the transformer from its
ideal behavior. The 100Ω resistors at the LTC6409 output
create a differential 200Ω resistance, which is an imped-
ance match for the reflected R
L
.
As previously mentioned, IMD3 is measured for 2V
P-P
dif-
ferential peak (i.e. 10dBm) at the output of the LTC6409,
corresponding to 1V
P-P
(i.e. 4dBm) at each output alone.
From LTC6409 output (location A in Figure 10) to the input
of the output transformer (location B), there is a voltage
attenuation of 1/2 (or –6dB) formed by the resistive divider
Figure 10. Equivalent Schematic of the Top Half of the LTC6409 Demo Board
P
S
POWER
FREQUENCY
IMD3 = P
S
– P
O
∆f = f2 – f1 = f1 – (2f1 – f2) = (2f2 – f1) – f2
6409 F08
P
S
P
O
P
O
2f1 – f2 2f2 – f1f1 f2
IMD3
1×
IIP3
P
O
OIP3
P
OUT
(dBm)
P
IN
(dBm)
6409 F10
P
S
3×

LTC6409CUDB#TRPBF

Mfr. #:
Manufacturer:
Analog Devices Inc.
Description:
High Speed Operational Amplifiers 10GHz GBW, 1.1nV/vHz Diff Amp/ADC Drvr
Lifecycle:
New from this manufacturer.
Delivery:
DHL FedEx Ups TNT EMS
Payment:
T/T Paypal Visa MoneyGram Western Union