TP Report

Dimensionnement Charge Utile

Submitted by:
Rupali Mishra
Juan Valverde
Master ESECA M2
 Lab -1 TWTA characteristic exploitation and simulation

In this lab work, we have to propose a model of TWTA. Simulation will be issued at
12100MHz and performed between -20 and +6dB IBO.

TWTA characteristic calculation

From Technical data sheets and txt file of AM/AM characteristic (4 columns:
IBO (dB); OBO (dB), Phase shift (deg), DC Consumption (W)) have been already
provided to calculate the other parameters.

1. Define the saturation point Pin / Pout.

The saturation point is the point at which the compression is worth 1dB. As the input
power continues to increase, at some point the gain begins to decrease. The amplifier
goes into compression where no further output increases occur for an input increase. The
gain starts flattening and at high signal levels, the amplifier becomes saturated. As a
result of saturation it’s response becomes non-linear and produces signal distortion,
harmonics, and potentially intermodulation products.

2. Deduce Pin and Pout scale in W and in dBm from IBO and OBO (We draw your
attention on the convention stating that saturation ouput power is at +0dB OBO but
max measured RF power is at +0.03dB OBO).

It is given that Pin_ref = -7.83 dBm and Pout_ref = 51.38 dBm. So we calculated Pin
and Pout as follows:

Pin=IBO-7.83dB

Pout=OBO+51.38dB

All the values of Pin and Pout in Watt and in dBm can be observed in table below:
 Table-1: Pin and Pout in W and dBm

3. Deduce G(dB) evolution vs IBO (dB).

G in dB=10∗log[ Pout(W )/Pin(W ) ]

Figure-1: Gain (dB) vs IBO(dB)

4. Calculate TWT dissipation and retrieve consumption and dissipation curves.
We can calculate the dissipation as follows:
Pdiss = Pin + Pcons(w) – Pout

Table-2: Values of Power dissipation

Figure-2: Consumption and dissipation curves

5. Trace AM / AM characteristic in dB as a function of IBO (dB).

Figure-3: AM/AM transfer curve

AM/AM curves basically depict the relation between power in vs power out for TWTA
amplifiers. We have used deformalized value of IBO for plotting this curve.

6. Trace Pout vs Pin characteristic in dBm.

The curve Pout vs Pin has been plotted below in answer 7 along with the linear evolution.

7. Estimate a linear evolution of Pout over Pin range considering the gain in the
linear region (15dB IBO) and plot it on the same Pout/Pin graph.

Figure-4: Linear evolution of Pout over Pin

8. Considering an IP3 point at +58.05dBm (output power), estimate linear
and non linear IM3 evolution as well as C/I3. Plot them in the same graph.

C/I3 and IP3 characteristic allows us to determine the linearity of an amplifier with two
input carrier. Considering an IP3 point at +58.05dBm, IM3 and C/I3 can be calculated by:

(C/I3)dB = (Pout)dBw – (IM3)dBw = 2(IP3-Pout) dB

Table-3: Calculation of IM3, C/I3

 Figure-5:IM3 evolution Curves

9. Evaluate 1dB compression point and Gain compression at saturation. Estimate

and plot compression Gain evolution (dB) in function of Pin.

Table-4: 1 dB compression gain calculations

 In the above table we have calculated the 1 dB gain compression along with the IM3 ,
C/I3 considering the IP3 point as 58.05dBm. I dB compression gain point has been
highlighted in the above table.

Figure-6: Gain compression evolution curve

Behavioural model of the TWTA : Simulation 1 tone

In this part, In order to do the simulations with ADS, a behavioural model of the
TWTA has been established through a s2d file using s2d amplifier model from System-
Data Models library. We referred to Agilents documentation for information on s2d file
format.

We have written two s2d files with GCOMP6 and GCOMP7 options separately and fed
them in to the amplifier.

1. Compare corresponding AM/AM responses using Harmonic Balance of the

following scheme.

2. Compare the two methods (GCOMP6 and GCOMP7).

GC0MP6
This model is created from the S parameters at each frequency, the P1dB, the IP3 and the
point of saturation. We have considered -15dB as the saturation point.

Figure-7: Simulation circuit with GCOMP6 s2d file

GCOMP6 s2d file

 Figure-8

As we have selected -15 dB as the saturation point, In the graph also we can observe the
saturation around -15dB which has the value 49.257dBm.

We plotted the graph for linear gain and gain compression with the help of Gain
equations as follows.

Figure-9: Gain compression

The difference between the values of marker m1 and m2 is 1dB. We can clearly observe 1
dB gain compression in the above graph.
GC0MP7
We repeated the same procedure as GC0MP6 for GC0MP7 but with the value of gain at
each value of power.

Figure-10: Schematic for GC0MP7 method

GC0MP7 File
 Figure-11

Figure-12: Gain Curve

In the above graph, the difference between values of two markers clearly shows the diffrence
of 1 dB compression.

We have observed that the model created from the GCOMP7 file is more accurate. Because in this
model , for each value of the input power a value of the gain is provided. But we can also see that the
file GCOMP6 makes it possible to obtain good results knowing only 3 parameters of the amplifier.
Third order intermodulation products: Simulation 2 tones
We have proposed a circuit to evaluate GCOMP7 model file and to evaluate third order
intermodulation products levels with 2 equi level carriers.

Figure-13: Schematic for simulation 2 tones

Figure-14: Simulation results for IM3 level

If we compare the initial theoritical results to the simulation results, we have found that
the value for IM3 level are almost the same. The value in theoritical curve is 58dBm and
in simulation results it has been observed as 57.536dBm. That means both the results rae
close enough.
 Lab 2 : Model of a payload Input section
Gain and Noise figure budgets

This lab session aims at modelling with ADS an input section of a communication payload
from payloads input port until TWT input with respectively :

- Model of a receiver : Calculation of Gain and NF budget of a receiver

- Model of the input section : Stating from calculation of attenuators values on a gain
budget on the input section
- Use of the CAMP in ALC mode
-
For this purpose, Simulation Budget library of ADS will be used to model the input section of
a communication payload for BSS channels (17,3GHz<=f<=18,1GHz).

We may consider a 36 MHz channel centered around 17.7 GHz (uplink frequency) with a
frequency conversion of 5600MHz as an example.

17-12 GHz Receiver model (Budget_NF.dsn)

From the receiver model of Figure 1, parameter correctly the model of receiver (by affecting
values to the variables of the design) and tune both Gain and noise figure of the first amplifier
to obtain following performances for the receiver:

 Gain : 58.5 dB
 NF : 1.9dB
Use the Budget_NF data display for exploitation of the simulation results.
NF_RefIn_NoImage_dB is to be used to monitor NF of the RF chain.
 Figure-15: Schematic for Budget simulations

The last figure shows the tuner values about the Gain and the Noise Figure, we obtain the
next behavior:

Figure-16:Output Gain of the input

 Figure-17:Noise figure of the input

Input section model

Considering uplink flux values are included between (-93 and –72 dBW/m²) and considering
a constant Rx antenna gain value of 28dB.

1. Calculate Pin dynamic (Pin min and Pin max) at the input port of the payload for the all
BSS frequency dynamic. We recall that Pin can be estimated from input fluxes from :

In order to modelizer the -0,82 dB input losses of the payload (from input port until receiver
input), we introduce an attenuator Input_loss before the receiver.

Reminder: an input filter is commonly used between input port and the receiver.
Payload input losses up to the receiver input (-0.82dB) are considered as an
approximation at central frequency of the channel for input losses of first band pass
filter and RF harness.

Where 𝐺𝑅 = 𝑓𝑙𝑢𝑥 ∙ 𝐺, we can specify that the effective area is 𝐴𝑒 = 𝐺 ∙ 𝐴𝑖 where Ai is the
isotropic area.

For 17.3 GHz:

  The minimum flux -93 dBW/m² then:
𝑐2 −93⁄ 28 (3𝑥108 )2
𝑃𝑖𝑛−𝑚𝑖𝑛 = 𝑓𝑙𝑢𝑥 ∙ 𝐺 ∙ = 10 10 ∙ 10 ⁄10 ∙ = 7.567𝑝𝑊
4 ∙ 𝜋 ∙ 𝑓2 4 ∙ 𝜋 ∙ (17.3𝐺 ) 2

𝑃𝑖𝑛−𝑚𝑖𝑛 (𝑑𝐵𝑚 ) = 10 log(7.567𝑝) + 30 = −81.2105

 The maximum flux -72 dBW/m² then:

𝑐2 −72⁄ 28 (3𝑥108 )2
𝑃𝑖𝑛−𝑚𝑎𝑥 = 𝑓𝑙𝑢𝑥 ∙ 𝐺 ∙ = 10 10 ∙ 10 ⁄10 ∙ = 952.665𝑝𝑊
4 ∙ 𝜋 ∙ 𝑓2 4 ∙ 𝜋 ∙ (17.3𝐺 )2

𝑃𝑖𝑛−𝑚𝑎𝑥 (𝑑𝐵𝑚 ) = 10 log(952.665𝑝) + 30 = −𝟔𝟎. 𝟐𝟏𝟎𝟓 it is the highest value.

For 18.1 GHz:

 The minimum flux -93 dBW/m² then:

𝑐2 −93⁄ 28⁄ (3𝑥108 )2
𝑃𝑖𝑛−𝑚𝑖𝑛 = 𝑓𝑙𝑢𝑥 ∙ 𝐺 ∙ = 10 10 ∙ 10 10 ∙ = 6.913𝑝𝑊
4 ∙ 𝜋 ∙ 𝑓2 4 ∙ 𝜋 ∙ (18.1𝐺 ) 2

𝑃𝑖𝑛−𝑚𝑖𝑛 (𝑑𝐵𝑚 ) = 10 log(6.913𝑝) + 30 = −𝟖𝟏. 𝟔𝟎𝟑𝟐 it is the lowest value.

 The maximum flux -72 dBW/m² then:

𝑐2 −72⁄ 28 (3𝑥108 )2
𝑃𝑖𝑛−𝑚𝑎𝑥 = 𝑓𝑙𝑢𝑥 ∙ 𝐺 ∙ = 10 10 ∙ 10 ⁄10 ∙ = 870.313𝑝𝑊
4 ∙ 𝜋 ∙ 𝑓2 4 ∙ 𝜋 ∙ (18.1𝐺 )2

𝑃𝑖𝑛−𝑚𝑎𝑥 (𝑑𝐵𝑚 ) = 10 log(870.313𝑝) + 30 = −60.6032

2. Estimate the input power dynamic to specify at the input of the receiver from input power
dynamic estimated at payloads input, A second attenuator Intraloss1 is modeling losses of
filters / IMUX and RF harness up to the Channel Amplifier input. Routing of RF harness and
IMUX performances may lead to discrepancies between power levels at CAMPs input for the
different channels.

In order to equilibrate those unbalanced losses, attenuators are included by design inside the
RF chain. Additional Sat1 is so a model of this sating attenuator(s) used for power balance
purpose on our RF chain.Term2 of the chain represents CAMP input
 Figure-18: Schematic for budget simulations

Figure-19

On the graph we can see that the two markers meet the calculated parameters because

M1 is for the lowest frequency and Pin_max = -60.211 dBm, then with losses we obtain -
61.031 dBm
M2 is for the biggest frequency and Pin_min = -81.603 dBm, then with losses we obtain -
82.423 dBm

That matches the calculations.

We need to take care about the frequency range because the P in min and the P in max
has a direct dependency.

Figure-20

On the graph we can see that the two markers meet the calculated parameters because

M1 is for the Pin_max = -60.211 dBm, then with input losses of -0.82 dB, Gain of 58.5 dB:

Out = -60.211 -0.82 + 58.5 = -2.531 dB

M2 is for the Pin_min = -81.603 dBm, then with input losses of -0.82 dB, Gain of 58.5 dB:

Out = -81.603 -0.82 + 58.5 = -23.923 dB

That matches the calculations.

3. Propose a saturation value to equilibrate power level at the input of the CAMP knowing
that CAMP dynamic is tuned between -24,5 and -45,9 dBm. Power level evolution along the
RF chain can be visualized thanks to Out_Pwr_dBm plot.
We need to add to attenuators on the output, intraloss has a value of -18.7 dB and represent
the losses by the IMUX and RF harness, and saturation represent the losses to equilibrate the
power with an input = CAMP
For an Out = -2.531 dB, intraloss of -18.7 dB and a CAM = -24.5 dBm then:
Saturation = +2.531 +18.7 -24.5 = -3.269 dB
For an Out = -23.923 dB, intraloss of -18.7 dB and a CAM = -45.9 dBm then:
Saturation = +23.923 +18.7 -45.9 = -3.277 dB
We obtain a mean of -3.273 dB that correspond to the saturation value

Figure-21
In conclusion we can see that CAMP dynamic is tuned between -24.508 and -45.900 dBm.

4. Estimate the noise figure of this input section.

Figure-22

For this we must use the relationships:

And for the Cascaded Noise Temperature we need to use:

For this equation we work in a table and we are going to compute each term and after we can
add term by term in order to see the same simulation pattern:

In_loss AMP1 MIX1 AMP2 Intraloss Saturation

Gain [dB] -0.82 33.397 -5 30.103 -18.7 -3.28
Gain 0.827942 2186.251 0.316228 1024 0.01349 0.469894109
NF [dB] 0.82 1.895 3 3 18.7 3.28
NT [K] 60.26601 158.64 288.6261 288.6261 21208 327.1603233

NT Terms [K] 60.26601 191.6076 0.159454 0.504238 0.036183 0.041377155

Cascaded Cumulative NT [K] 60.26601 251.8736 252.0331 252.5373 252.5735 252.6148582

Cumulative NF [dB] 0.82 2.715 2.716278 2.720316 2.720606 2.720936838

Calculation of Noise Figure, noise temperature, and Cascaded Noise Temperature

We can see how the Cumulative Noise Figure (dB) is almost the same that in the simulation.

CAMP setting in ALC mode

In FGM mode, CAMP gain is the same for any input power level and CAMP output varies
linearly with input power. A gain-controlled amplifier may be used for modelization of this
FGM mode of the CAMP with a gain value adjustable with an FGM step TC (command
voltage of the amplifier).

But the AGC_AMP amplifier of ADS that can be used in power budgets is to be coupled with
an AGC_PwrControl element for power regulation through ALC loop. In ALC mode of the
CAMP, the CAMPs output power remains constant and the gain decreases linearly with
increasing input power.

In order to get an output power of 0.5 dBm to drive the TWT in saturation, we need to add a
gain control element, for the CAMP dynamic between -24,5 and -45,9 dBm we have:

For an CAM = -24.5 dBm, Out TWT = 0.5 dBm then:

CAMP Gain = + 24.5 + 0.5 = 25 dB

For an CAM = -45.9 dBm, Out TWT = 0.5 dBm then:

CAMP Gain = + 45.9 + 0.5 = 46.5 dB

In order to simulate this test, we need to add a fixed amplifier of 46.5 – 25 = 21.5 dB, and a
controlled amplifier of 0 to 25 dB with a reference of 0.5 dBm.
Figure-23

Figure-24
We can see the dynamic range on the index = 0, and after in the index = 7 and index = 8 the
constant gain to drive the TWT in saturation equivalent to 0.5 dBm

Propose an evolution of the previous scheme to have a model of CAMP in ALC mode with a
constant output power of 0,5dBm to drive the TWT in saturation (this output level in ALC
mode can be adjusted in reality).


Figure-25

For the analysis of the Noise figure we know that the last stages over the circuit, the
cumulative noise figure almost does not change because the weight on the last steps
are lower than the previous ones. We need to take care to choose the first element
with the lower Noise figure in order to have the lowest value on the last element. We
can see the next figure:
 
Figure-26