[SOLVED] Source follower input/output swing

[mxm14]

I found in an article that in a source follower the output/input swing is done by

• Vin_pp = Vdd - ( Vt + 2.Vds_sat )

I tried to obtain that using Vds>Vdsat or Vgs>Vt but i cannot obtain this..maybe i have not clear the definiton of output/input swing..
The circuit is the classical source follower with an nmos having Vin as Vg, Vd connected to Vdd and Vs=Vout connected to the Ibias

just as reference:

 

[crutschow]

I tried to obtain that using Vds>Vdsat or Vgs>Vt but i cannot obtain this.

You have given little useful information to help solve your problem.
What is M1, what are the voltages you are applying and seeing, and what is I1?

 

[mxm14]

You have given little useful information to help solve your problem.
What is M1, what are the voltages you are applying and seeing, and what is I1?

M1 is the mosfet, vds voltage between drain- source, vddsat the overdrive voltage, vgs voltage between gate and source, I1=Ibias. The circuit is a classic source follower, so i put M1 in saturation, apply a signal on the gate, the output signal is on the source. What is not clear for me is this input-output swing definition Vinpp and its calculation

 

[barry]

I think you’re confused and you’re trying to confuse us.

First of all, you mention an article, but don’t reference it. Then you talk about “input/output swing” and “output/input swing”; I don’t think these things exist. There’s output swing and there’s input swing. Then you’ve got an equation for input swing (not “input/output swing”), and I think that’s wrong-why is Vdsat multiplied by 2?

And then this: “I tried to obtain that using Vds>Vdsat or Vgs>Vt but i cannot obtain this”. I don’t even know what that means.

Maybe you should try simulating this, that might give you more insight.

 

[mxm14]

The article says the same thing you can find here pag 271.
It s here that calls Vinpp input/output swing, so yes i m a bit confused..

Analog/RF and Mixed-Signal Circuit Systematic Design

Despite the fact that in the digital domain, designers can take full benefits of IPs and design automation tools to synthesize and design very complex systems, the analog designers’ task is still considered as a ‘handcraft’, cumbersome and very time consuming process. Thus, tremendous efforts...

books.google.it

"And then this: “I tried to obtain that using Vds>Vdsat or Vgs>Vt but i cannot obtain this”. I don’t even know what that means."-> I simply try to consider the conditions for having M1 on(Vgs>Vt) and M1 in saturation (Vds>Vdsat) in order to see if the disequation of the first post(Vinpp = Vdd - ( Vt + 2.Vdssat )) derives from this two conditions.

So i m just asking if somebody knows the derivation of this
Vinpp = Vdd - ( Vt + 2.Vdssat )

 

[FvM]

I agree with barry, you can't derive a formula without knowing what it describes exactly. "input/output swing" is no commonly understood technical term. I assume that the author has an idea what this means, but we can only guess.

 

[mxm14]

Understood ...let s see if somebody else knows what is referring to

 

[BradtheRad]

Isn't that a current source in the schematic? I think the circuit requires a different approach other than emitter follower.

The current source performs by adopting whatever voltage is necessary to produce desired Ampere flow.

V_out is that resulting voltage.

V_out doesn't necessarily match V_in. They may be related by formula depending on mosfet parameters:

a) gain
and
b) Amperes yanked through it by the current source.

I could be wrong.

 

[barry]

Isn't that a current source in the schematic? I think the circuit requires a different approach other than emitter follower.

The current source performs by adopting whatever voltage is necessary to produce desired Ampere flow.

V_out is that resulting voltage.

V_out doesn't necessarily match V_in. They may be related by formula depending on mosfet parameters:

a) gain
and
b) Amperes yanked through it by the current source.

I could be wrong.

The current will source (actually sink in this case) whatever current it’s configured for, regardless of the voltage at Vout. Think of it as adjusting its impedance to match the required current.

 

[davidrichard4748]

We can't give a formula if you don't know exactly what you're explaining. "Swing in/out" is not a common sense term. I think the author knows what that means, but we can only guess.

 

[mxm14]

I never heard this swing in/out too and this is the reason why I ask for..let's see if somebody else knows this definition or can derive it from the formula given by the author

 

[barry]

I never heard this swing in/out too and this is the reason why I ask for..let's see if somebody else knows this definition or can derive it from the formula given by the author

You keep saying "let's see if somebody knows about this." Nobody knows about this because it's not a real thing.

And the link you provided didn't work for me, so I can't see your source.

 

[andre_luis]

The first pages of the mentioned 'article' can be found here: https://link.springer.com/chapter/10.1007/978-3-642-36329-0_12

 

[stanislavc]

The article says the same thing you can find here pag 271.
It s here that calls Vinpp input/output swing, so yes i m a bit confused..
...

So i m just asking if somebody knows the derivation of this

Vinpp = Vdd - ( Vt + 2.Vdssat )

So this makes sense to me like so:
On top of p. 271 at your link, I see the equation you are referring to. Vinpp designates the max input swing you can have, while the source-follower's transfer function is still valid. Source follower's transfer function is Vout~=Vin+Voffset (it's ~= because there is some distortion as you swing from min to max, but in ideal sense, with an ideal current source, as in schematic, it's an exact equal '=')

Now regarding the derivation of the equation. Take a case where you have supplies Vdd on top and 0V on bottom for the full circuit. The highest point that the stage that drives Vin can reach is Vdd (let's assume you don't have a charge pump or other ways of setting Vin higher than Vdd by the previous stage). At this point Vout will be lower than Vdd by Vgs of the fet.
=> Vin_max=Vdd, Vout_max= Vdd - Vgs = Vdd- (Vth+Vov) = Vdd - (Vth+Vds_sat)
^ this also assumes that Vov = Vds_sat of a fet; which isn't exactly the case but is a good 1st order approximation in undergrad / grad circuit books)

The lowest voltage that Vout can reach without distorting the source follower operation is the smallest voltage that the bottom current source can support while providing a constant bias current. Since most current sources are implemented as a nfet (here) with a drain which sinks the current, the min voltage it can handle across it is again Vov, while staying in saturation.
=> Vout_min=Vov_currentSourceFet = Vds_sat_currSrcFet

Now we can calc the Vout_swingMax= Vout_max - Vout_min:
Vout_swingMax= Vdd-(Vth+Vds_sat)-Vds_sat = Vdd - (Vth+2*Vds_sat)
^ assumes that current source fet is biased similarly as the main fet so they have equal Vds_sat

We also know that Vout_swing_max = Vin_swing_max in a source follower (again, ignoring the distortion factor for a 1st order approximation).

=> Vin_peak-to-peak_swing = Vdd - (Vth + 2*Vds_sat)

 

[mxm14]

Thank you for your clear answer!