The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> mosfet noise
https://designers-guide.org/forum/YaBB.pl?num=1319098905

Message started by summi on Oct 20^th, 2011, 1:21am

Title: mosfet noise
Post by summi on Oct 20^th, 2011, 1:21am

dear all,
can any one tell me why mosfet gives more noise than a resister of same conductance and same bias current.

br,
summi.

Title: Re: mosfet noise
Post by rfidea on Oct 20^th, 2011, 3:51am

Hi,

I can think of three reasons.

1. At low frequencies you have more 1/f noise in the MOS device.
2. The MOS device is in the saturated region and not in the linear region where it behaves as a pure resistor.
3. There is filtering effects inside the MOS due to parasitic capitance.

Title: Re: mosfet noise
Post by raja.cedt on Oct 20^th, 2011, 4:13am

hello rf,
i agree with you, but let us assume there is no 1/f noise and parasitic and only thermal noise and operating in saturation.

So in this case i feel mos gives more noise because resister noise=4kT*(i/v) but for mos 4kT*gamma*(2*I/Vov), so i feel for the same current mos has less Vov, hence more noise..agree?

Thanks,
raj.

Title: Re: mosfet noise
Post by rajkumar palwai on Oct 20^th, 2011, 10:25am

@Raja,
summi took an assumption that the conductance and currents are same in resistor and mosfet. Ideally the mosfet noise current is 4kt(gamma)(gds@vds=0). But as gds@vds=0 is same as gm, we use typical equation of 4ktgm. So in the case of mosfet & resistor conductances being same, both of them produce same amount of noise current.

@summi
But if we see the noise voltages, the resistor noise voltage is 4kt/r*r^2 = 4ktr. Where as the noise voltage at mosfet o/p is 4ktgm*ro^2 , which is much greater than resistor noise voltage.

However, as you may be aware that all these equations hold true only when the mos is in saturation. If mosfet is in linear region than all the noise currents and voltages are same for both mosfet and resistor.

-rajkumar

Title: Re: mosfet noise
Post by raja.cedt on Oct 20^th, 2011, 11:25am

@rajkumar...yes you are correct if gm and I is same. But for me it seems it is very difficult unless you have too much overdrive (assume you have 1k res with 1ma bias current, now to get gm of 1ms @ 1mA current you need around 2v Vov). So i feel same current assumption is not fare under same gm.

@ every one: for the same condactance Mos gives more noise due to many reasons, but if from simple expression about noise in mos, there extra term 2/3, but is this correct to conclude mosgives less noise? Please check 'Design-Oriented Estimation of Thermal Noise in
Switched-Capacitor Circuits' paper ---foot note 9

Thanks,
raj.

Title: Re: mosfet noise
Post by rfidea on Oct 20^th, 2011, 12:54pm

Hi guys!

I was assuming the comparison of the noise was for the mos in its linear region when the noise should be about the same. In the saturation region the noise for the mos will be larger because of the arguments you have described.