Can anybody find my error? (wireless RF modeling)

G

Ginu

Guest
Hello,

I'm simulating a network of multiple wireless nodes that can transmit
to one another using finite battery resources. I'm getting some
really
strange results in my testing of some power calculations. It's not a
complicated problem - the problem exists in finding the bounds on my
transmit power before even doing anything with it.


There is a technical term for the minimum received power that a node
must receive for it to be able to decipher a message, known as the
minimum input receiver sensitivity (MIRS). Also, because each node as
a finite amount of battery power, there is a maximum transmit power
that can be achieved to ensure that the power dissipated for a
transmission does not exceed the amount of battery power that a node
has. With these two pieces of information, I form these constraints
on
my transmit power:


Pt >= max(0, P_MIRS) -> Lower Bound


Pt <= min(0, P_max) -> Upper Bound


Now the first constraint is formed by observing that the minimum
transmit power must be the maximum of 0 watts and the transmit power
required to reach the MIRS at the receiver. I calculate P_MIRS as:


MIRSdBm = -92;
pathLossExponent = 1.8;
Txpower = MIRSdBm - Gt - Gr + 20*log10(4*pi* (freq * 1E6) / (3E8)) +
10*pathLossExponent*log10(distance);
P_MIRS = 10^((Txpower-30)/10);


Now, with Gt = Gr = 0 dBi, freq = 2400 (I'm modeling frequency 2.4
GHz
or 2400 MHz) and distance = 2m, my Txpower = -46.5dBm and P_MIRS =
2.2205e-008 Watts. This is my Lower Bound.


For the second constraint, P_max is calculated by considering the
remaining battery power and power dissipation. The power dissipation
of a transmission is calculated using q*R: q is the energy
consumption/
bit of the wireless technology used and R is the transmission rate in
bits/sec. R = w*log2 (1 + (P_max*G) / (w * N + Interference) is
the physical representation of the transmission rate called the
Shannon Capacity. Multiplying the two quantities gives energy
consumption/sec
which gives units of power. Now, the quantity q is a term that can be
found in the literature for a technology, but since it is considered
for the transmission as a whole (including a transmitter and
receiver), to find the power dissipation at the transmitter only, I
assume a 50-50% split for power dissipation between transmitter and
receiver. So I form the constraint below to find P_max by stating
that
the battery power must be greater than or equal to the maximum power
dissipation for a transmission that has used P_max of transmit power
in Watts:


P_battery >= (q/2)*R >= (q/2)*w*log2 (1 + (P_max*G) / (w * N +
Interference);


Rearranging this formula gives:
P_max= [2^(P_battery/(w*q/2)) - 1]*[(w * N + Interference)]/G


Now, I have P_battery = 1 Watt, w = 1 MHz = 1E6 bandwidth, q = 0.5E-6
joules/bit (for the Zigbee technology), N is the noise @ center
frequency 2400 MHz around bandwidth w, Interference is the
interference @ center frequency 2400 MHz around bandwidth w, and G is
the channel gain @ center frequency 2400 MHz calculated as


Gdb = Gr + Gt - 20*log10(4*pi*(freq * 1E6) *d0/c) -
10*pathLossExponent*log10(distance/d0);
G = 10^(G/10);


Again, Gr = Gt = 0 dBi, freq = 2400, d0 = 1m, c = 3.0E8,
pathLossExponent = 1.8 and distance = 2m. So I get values for Gdb =
-45.46 dB and G = 2.8415e-005.


Now this gives me a P_max of 1.9479e-009 Watts. For this calculation
my N = kT = (1.23E-23)*(300 Kelvin) and Interference = 0. This P_max
=
1.9479e-009 Watts is my Upper Bound.


So after all of this without even any interference, my upper bound <
lower bound. Also, as you can see, with 1 whole Watt of battery power
and no
interference, I can't even transmit a distance of 2m away. There's
definitely something wrong.


I hope someone can show me what I've done wrong.


Thanks a bunch!
Omar
 
On Aug 2, 5:11 pm, Ginu <osheik...@gmail.com> wrote:
Hello,

I'm simulating a network of multiple wireless nodes that can transmit
to one another using finite battery resources. I'm getting some
really
strange results in my testing of some power calculations. It's not a
complicated problem - the problem exists in finding the bounds on my
transmit power before even doing anything with it.

There is a technical term for the minimum received power that a node
must receive for it to be able to decipher a message, known as the
minimum input receiver sensitivity (MIRS). Also, because each node as
a finite amount of battery power, there is a maximum transmit power
that can be achieved to ensure that the power dissipated for a
transmission does not exceed the amount of battery power that a node
has. With these two pieces of information, I form these constraints
on
my transmit power:

Pt >= max(0, P_MIRS)        -> Lower Bound

Pt <= min(0, P_max)            -> Upper Bound

Now the first constraint is formed by observing that the minimum
transmit power must be the maximum of 0 watts and the transmit power
required to reach the MIRS at the receiver. I calculate P_MIRS as:

MIRSdBm = -92;
pathLossExponent = 1.8;
Txpower = MIRSdBm - Gt - Gr + 20*log10(4*pi* (freq * 1E6) / (3E8)) +
10*pathLossExponent*log10(distance);
P_MIRS = 10^((Txpower-30)/10);

Now, with Gt = Gr = 0 dBi, freq = 2400 (I'm modeling frequency 2.4
GHz
or 2400 MHz) and distance = 2m, my Txpower = -46.5dBm and P_MIRS > 2.2205e-008 Watts. This is my Lower Bound.

For the second constraint, P_max is calculated by considering the
remaining battery power and power dissipation. The power dissipation
of a transmission is calculated using q*R: q is the energy
consumption/
bit of the wireless technology used and R is the transmission rate in
bits/sec. R = w*log2 (1 + (P_max*G) / (w * N + Interference) is
the physical representation of the transmission rate called the
Shannon Capacity. Multiplying the two quantities gives energy
consumption/sec
which gives units of power. Now, the quantity q is a term that can be
found in the literature for a technology, but since it is considered
for the transmission as a whole (including a transmitter and
receiver), to find the power dissipation at the transmitter only, I
assume a 50-50% split for power dissipation between transmitter and
receiver. So I form the constraint below to find P_max by stating
that
the battery power must be greater than or equal to the maximum power
dissipation for a transmission that has used P_max of transmit power
in Watts:

P_battery >= (q/2)*R >= (q/2)*w*log2 (1 + (P_max*G) / (w * N +
Interference);

Rearranging this formula gives:
P_max= [2^(P_battery/(w*q/2)) - 1]*[(w * N + Interference)]/G

Now, I have P_battery = 1 Watt, w = 1 MHz = 1E6 bandwidth, q = 0.5E-6
joules/bit (for the Zigbee technology), N is the noise @ center
frequency 2400 MHz around bandwidth w, Interference is the
interference @ center frequency 2400 MHz around bandwidth w, and G is
the channel gain @ center frequency 2400 MHz calculated as

Gdb = Gr + Gt - 20*log10(4*pi*(freq * 1E6) *d0/c) -
10*pathLossExponent*log10(distance/d0);
G = 10^(G/10);

Again, Gr = Gt = 0 dBi, freq = 2400, d0 = 1m, c = 3.0E8,
pathLossExponent = 1.8 and distance = 2m. So I get values for Gdb > -45.46 dB and G = 2.8415e-005.

Now this gives me a P_max of 1.9479e-009 Watts. For this calculation
my N = kT = (1.23E-23)*(300 Kelvin) and Interference = 0. This P_max
1.9479e-009 Watts is my Upper Bound.

So after all of this without even any interference, my upper bound
lower bound. Also, as you can see, with 1 whole Watt of battery power
and no
interference, I can't even transmit a distance of 2m away. There's
definitely something wrong.

I hope someone can show me what I've done wrong.

Thanks a bunch!
Omar

Sorry,

Gdb = Gr + Gt - 20*log10(4*pi*(freq * 1E6) *d0/c) -
10*pathLossExponent*log10(distance/d0);
G = 10^(G/10);

should read
Gdb = Gr + Gt - 20*log10(4*pi*(freq * 1E6) *d0/c) -
10*pathLossExponent*log10(distance/d0);
G = 10^(Gdb/10);

Is it how I'm calculating my channel gain power? I'm assuming that
it's alright to use G in my R calculation where P_max*G is my maximum
received signal power. If this isn't it, I have no clue what I've done
wrong.
 

Welcome to EDABoard.com

Sponsor

Back
Top