Resistive power dividers

Hi folks
Here is a paper by G Adams on resistive power dividers. Can anyone please explain to me table 1 ?
The paper says if u get 1.7 db loss from port one to port 2 , u'll get 20 db loss from port 1 to port 3

can anyone derive this?

bode plots i guess

---------- Post added at 13:58 ---------- Previous post was at 13:56 ----------

the capacitor resistor combination may make a filter in this case that gives u a 20dB loss in power

the capacitor resistor combination may make a filter

Click to expand...

No capacitors are involved in the design.

can anyone derive this?

Click to expand...

How about applying ohms law and calculating the output voltages?

Here are my analysis [Producing results which DO NOT match the values given in the table ]

Lets assume the loss from port 1 to port 2 is 1.72db . Since its a loss , I take it as -1.72db .
Using (db)=20 log V2/V1 we get V2/V1 =0.82 or V2=0.82V1

Now , this implies for a 2 way unequal power divider , V3= (1-0.82) V1 i.e. V3= 0.18V1

Converting this 0.18 into db gives -14db

Thus for a 2 port unequal resistive power divider , if S21= -1.72 db , then S31 = -14db

***The paper says S31=-20db
and also the table values do not match

nnayak82 said:

Here are my analysis [Producing results which DO NOT match the values given in the table ]

Lets assume the loss from port 1 to port 2 is 1.72db . Since its a loss , I take it as -1.72db .
Using (db)=20 log V2/V1 we get V2/V1 =0.82 or V2=0.82V1

Now , this implies for a 2 way unequal power divider , V3= (1-0.82) V1 i.e. V3= 0.18V1

Converting this 0.18 into db gives -14db

Thus for a 2 port unequal resistive power divider , if S21= -1.72 db , then S31 = -14db

***The paper says S31=-20db
and also the table values do not match

Click to expand...

It's Power Divider not Voltage Divider

nnayak82 said:

Here is a paper by G Adams on resistive power dividers. Can anyone please explain to me table 1 ?
The paper says if u get 1.7 db loss from port one to port 2 , u'll get 20 db loss from port 1 to port 3

can anyone derive this?

Click to expand...

Hai
Based on equations in article i get the following values of resistors:
Rs = 0.1 Zo
Rp = 5 Zo
Rt = 4.45 Zo
Ru = 1.25 Zo
So S21 would be 20log10(0.82) = -1.73 dB, as it should be,
S31 = 20log10(0.9*0.55/5) = -20 dB
So everything is correct.

Lets keep resistor values aside . Lets consider the power divider to be 3 port with port one as input port.
I was referring to a simple unequal resistive power divider and I would like to derive the value of attenuation[in db] at port 3 when we know the attenuation at port 2 .As Bigboss pointed out we are dealing with power here so my mathematical analysis are as follows :::

****************************************************************************

The author says::" 1.72 db loss at port 2 from port 1 and then we will have 20 db loss from port 1 to port 3. "
****************************************************************************

My derivation gives Since we have -1.72 db loss at port 2 from port 1 , then P2/P1 = 10^(-1.72/10) = 0.67
or P2=0.67 *P1
Assuming input power =1
1-0.67=0.33 and hence P3=0.33*P1
Or converting the ratio of P3/P1 into db yeilds -4 db and NOT 20 db as given by author
I know I might be making some mistake in above calculations and hence would like to know it .....

Assuming input power =1
1-0.67=0.33 and hence P3=0.33*P1

Click to expand...

The obviously erronous assumption is a lossless power splitter. But by using resistors, a power divider must have losses. The amount of dissipated input power depends on the particular circuit and can't be derived from the attenuation numbers. Thus all calculations based on it are void.

You better should calculate the actual transmission factors of the given divider circuit with usual equations for resistive networks. Then you'll be able to reproduce the given numbers, as all other contributors already did. (I also checked it and could confirm).

pavel_adameyko said:

Hai
Based on equations in article i get the following values of resistors:
Rs = 0.1 Zo
Rp = 5 Zo
Rt = 4.45 Zo
Ru = 1.25 Zo
So S21 would be 20log10(0.82) = -1.73 dB, as it should be,
S31 = 20log10(0.9*0.55/5) = -20 dB
So everything is correct.

Click to expand...

I still did not understand how you calculated S31
Please explain