ADL5336-EVALZ AD [Analog Devices], ADL5336-EVALZ Datasheet - Page 23

• Manufacturer Part Number: ADL5336-EVALZ
• Description: Cascadable IF VGAs
• Manufacturer: AD [Analog Devices]

Data Sheet
EFFECT OF SETPOINT ON EVM
While in AGC mode, the EVM can degrade depending on the
output setpoint each VGA is set to. There is a strong relationship
between the output setpoint of VGA2 and EVM performance
while the output setpoint of VGA1 is held constant. Conversely,
the EVM does not change much while the output setpoint of the
VGA2 is held constant and the output setpoint of VGA1 is
changed. This behavior can be seen in Figure 66 where several
different setpoints of both VGAs were tested. This example uses
a 16 QAM modulated signal at 4.5 Msym/sec using a pulse
shaping filter and an alpha of 0.35.The frequency used was 140
MHz and C
codes of 11.
Table 9. Total Cascaded Output Noise
V
176
176
176
i
Figure 67. EVM vs. RF Input Power While VGA1 Setpoint Held Constant to
(mV rms)
–10
–15
–20
–25
–30
–35
–40
–45
–50
–5
250 mV rms and VGA2 Setpoint Swept; VGA1/VGA2 Gain Code = 11
0
–65
–10
–15
–20
–25
–30
–35
–40
–45
–50
–5
0
–65
Figure 66. EVM vs. RF Input Power over Several Setpoints
–55
VGA2 88mV rms
VGA2 125mV rms
VGA2 176 mV rms
VGA2 250mV rms
VGA2 500mV rms
VGAIN2 250/88
VGAIN2 250/125
VGAIN2 250/176
VGAIN2 250/250
VGAIN2 250/500
AGC
VGA1 88mV rms, VGA2 250mV rms
VGA1 250mV rms, VGA2 250mV rms
VGA1 707mV rms, VGA2 250mV rms
VGA1 250mV rms, VGA2 88mV rms
VGA1 250mV rms, VGA2 500mV rms
VGA1 250mV rms, VGA2 125mV rms
VGA1 250mV rms, VGA2 176mV rms
–55
= 0.1 μF. Both VGAs were set to maximum gain
–45
–45
RF INPUT POWER (dBm)
–35
A
0
6
12
RF INPUT POWER (dBm)
–35
V1
–25
(dB)
–25
–15
–15
V
GAIN1
–5
–5
V
176
353
707
O1
5
(mV rms)
5
15
15
25
2.5
2.0
1.5
1.0
0.5
0
25
Rev. B | Page 23 of 32
A
+6
0
−6
V2
(dB)
CASCADED VGA/AGC PERFORMANCE
The
Cascading VGAs decreases the overall noise figure by keeping
as much gain as possible before the final gain stage/noise source. A
single X-AMP has constant output referred noise. For an 8 dB
NF amplifier, with 36 dB maximum gain, in a 200 Ω matched
system, output referred noise V
noise contribution from the source, is the constant source noise
multiplied by the gain (as the gain is reduced, the noise contribution
from the source decreases). Measuring noise figure as 20 × log10
(total noise/noise from source), the dB-for-dB degradation in
NF typical of this architecture can be seen.
When the gain is partitioned into two VGAs, consider 18 dB
each. If each has an 8 dB NF, then each has an RTO noise of
18 nV/√Hz, including the source noise, and 16.5 nV/√Hz,
excluding the source noise. At maximum gain, the total RTO
noise is 145 nV/√Hz. As overall gain is decreased, the gain of
VGA2 is decreased first. When the gain of VGA2 is decreased
by 6 dB, the noise contributions from the source and VGA1
both decrease by 6 dB for an overall RTO noise of the system
that falls to 74 nV/√Hz.
When VGA1 and VGA2 are cascaded and operating in AGC
mode, setpoint programming affects dynamic range. The noise
measured at the output of VGA1 is relatively constant across
gain, which is a feature common to X-AMP VGAs. However,
measured at the output of VGA2, the noise contribution from
VGA2 is constant, but the noise contribution from VGA1 depends
on the gain of VGA2. For a given overall gain (VGA1 and VGA2),
the gain partitioning between VGA1 and VGA2 controls total
RTO noise and distortion.
To illustrate, consider the case where both VGAs are programmed
to a maximum gain of 14 dB and the setpoint of VGA2 is 101,
or 353 mV rms. Gain and signal levels can also be looked at
when the setpoint of VGA1 is programmed to 011, 101, and
111, 176 mV rms, 353 mV rms, and 707 mV rms (see Table 9).
ADL5336
V
353
353
353
O
is designed for easy cascading of the two VGAs.
(mV rms)
N, RTO
= 144 nV/√Hz. RTO, the
n
20
10
5
1
n
10
10
10
2
ADL5336
n
22.4
14.1
11.2
TOTAL