sensitivity in amplifiers

Status
Not open for further replies.
H

hauser

Newbie level 3
Joined
Jan 1, 2015
Messages
4
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
29
Hello,

I’m looking for a parameter called „sensitivity“ in the datasheet of the amplifier LMH6882:https://www.ti.com/product/LMH6882/technicaldocuments, but this doesn’t appear to be there. Question: should this parameter be provided by an amplifier? I just want to know how strong the incoming signal must be to get an output.

Hauser
 


Nope.

The thing specifies a gain (variable in this case) and a noise level (More or less 10dB at highest gain getting worse as you turn down), how much signal input you need to get above the noise depends on your working bandwidth and any processing gain in the obvious way.

Sensitivity is a property of a system not a component.

Regards, Dan.
 

Its adjustable voltage gain is 26dB (20 times) to 6dB (2 times). Then its sensitivity can be anything you want if you keep the input signal level low enough to prevent the output from clipping. Like any amplifier, its noise level will probably limit its useable sensitivity.
 


The correct way to express amplifier or receiver sensitivity is by its noise "floor", above which a small input signal can cause an output different from noise.

The noise power at amplifier input is

Pn = kTB in Watts, or,

Pn (dB) = -174 + NF +10 log B,

where B is the bandwidth in Hz, NF is noise figure in dB and "-174" is the log of the Boltzmann constant, k, multiplied by ambient temperature, T, in Kelvins.

Most amplifiers have specified their noise figure, NF , in dB. In variable-gain amplifiers, NF sometimes varies and is lowest for the maximum gain.

Pn defines the "noise floor" which means any signal power lower than it will not cause an output higher than the amplified noise input.
In systems, we then use the S/N ratio to define its capability to operate at low signal levels. Good systems have S/N > 10...20 dB. Lower S/N values often cause problems in demodulators.
 


Pn is a value which is independent of the component. Can man expect the incoming signal to be amplified as long as this signal is greater than Pn?

Hauser
 

hauser said:
Pn is a value which is independent of the component. Can man expect the incoming signal to be amplified as long as this signal is greater than Pn?

Hauser
Click to expand...

Pn is dependent on component noise figure, NF (dB).
I forgot to put it in the "linear" equation for Pn (W).

Any signal can be amplifier along with input noise but at the output we can only use a signal exceeding the nose level, thus we use S/N ratio.

Certain techniques like synchronous detection or correlation are used to extract signals from noise. , also if signal power is e.g. 60 dB lower than noise power. But signal parameters must be known in advance so such procedure can work.
 


I think the above mentioned dimension shall be dBm, not dB, see attachment.

In my case, the NF of LMH6882 is 9.7 dB = 39.7 dBm, the bandwidth is 600 MHz, so Pn = -174 + 39.7 + 10log10(600 MHz) = -174 + 39.7 + 87.8 = -46.5 dBm.

If the signal is stronger than -46.5 dBm, there will be amplified signal on the output. Is it correct?

Hauser
 


Thank you for the correction, yes, the Pn comes in dBm. And yes, any signal above Pn will appear above noise at amplifier output. For signal and noise simply add amplifier gain in dB.

I only do not see how from NF = 9.7 dB you made 39.7 dBm. This would be almost ten watts. Use only NF-9.7 dB, please
 



1 W = 1000 mW, so 1 dBW = 30 dBmW, that’s why I made 9.7 dB = 39.7 dBm. Maybe it’s wrong?

Hauser
 

hauser said:
1 W = 1000 mW, so 1 dBW = 30 dBmW, that’s why I made 9.7 dB = 39.7 dBm. Maybe it’s wrong?

Hauser
Click to expand...

The equation for noise power with "-174" gives the power in "dBm", therefore related to one milliwatt.
Noise figure an bandwidth are given in dB, or ratios. So NF is 9.7 dB, no relation to power level but by itself it is a power ratio.
The "-174" constant comes in dBm, it is the log of kT, Boltzmann constant (Joules/Kelvin) times ambient temperature (Kelvins). Joule is Watt x sec, seconds are canceled out later by multiplying by 10 log BW (bandwidth in Hertz).
Noise figure comes in dB, then you cannot equate it to dBm. It is only a ratio in dB by which the noise power in dBm grows to a higher power level in dBm, related to one milliwatt.
 

Status
Not open for further replies.

Similar threads

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