the concept of input power

Status
Not open for further replies.

matin-kh

Member level 3
Joined
Nov 9, 2013
Messages
67
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,286
Location
iran
Visit site
Activity points
1,955
Hi every one,
I'm working on AGCs. I can not understand a concept in a paper. as you see in the pictures the gain and the Linear-in-dB error are plotting versus input power(dBm). i can not understand what is the input power and how can we measure this things?





with best regards
Matin
 

Power input is standard definition of power.
viz : v^2 / R, or I^2 x R, or V x I etc.

Thus (for example) if the input impedance of your AGC block is 50 ohms, and the pk-pk input voltage is 100 mV, then the input power would be :

Vrms = 35.36 mV rms
and input power is V^2 / R = 2.5 uW

This translates into ~ -16 dBm input

Now your AGC graph shows that for a -16dBm input, the gain is roughly 5 dB.
Hence output power would be -16+5 = -11dBm

This reverse translates to 79.4 uW output.

If your load at output of AGC is again 50 ohm, then you will see a pk-pk voltage swing of around 178 mV.

(i think all my above calcs are correct. I leave the detailed workings to you)
 
Linear-in-dB error curves are result of nonlinearity of AGC curves at different temperatures. Ideal AGC curve will have 0dB error for full range of input power or straight horizontal line at 0dB. Ideal AGC curve will have constant slope or gradient of -1.
 
dBm is the power in decibels referenced to 1mW. This is calculated as 10 log (P / 1mW) where P is the measured power. Thus 0dBm is equal to 1mW and with a 50Ω source/load the voltage would be v = √(50 * 1mw) = 223.6mV rms.
 
thank you very much for your help.
now I understood the concept of power in AGC now I have two more questions:
1-if i want to have a curve like above(Gain versus input power) should i use power source and use ac analysis in cadence which input power is varying?
2-how can I measure linear-in-dB error like above in cadence?
Thanks
 

Status
Not open for further replies.

Similar threads

Cookies are required to use this site. You must accept them to continue using the site. Learn more…