Relationship between Vdssat and Vov

[jordan76]

vov mosfet

Hi,

As we know, drain-source saturation voltage Vdssat is not the same as the over-drive voltage Vov(=Vgs-Vth). But is there any relationship between them? How can we derive the Vdssat from simulations?

Thanks in advance!

regards,
jordan76

 

[ambreesh]

vdssat

Hi Jordan76,

I hope we all agree that VDsat (min), is the minimum Vds required to keep the transistor in saturation.
You always want Vds (min)>Vov.
And you can find Vdsat(min) without simulation, look into 2 stage opamp design example in Allen holberg.
I am sure you must have observed it that for a simle common source amp having an active current source as load, one cannot give the Vds across it before the simulation. For a tail current source of a diffamp we can surely say what would be its vds as long as input voltage range is know

Hope ti helps

 

[jordan76]

mosfet vov

Hi ambreesh,

Thanks,but I am afraid you may beat the wrong bush:?:

Actually my question is the relationship between Vdssat and Vov instead of Vds and Vdssat.

One friend of mine tole me that for long channel MOSFET transistor, Vdssat=Vov; but for short channel MOSFET transistor, this does not hold any more. So it brings forth my question about the relationship between Vdssat and Vov.

regards,
jordan76

 

[wwfeda]

vov voltage

Vdssat is when it reaches pinch off at the channel. Vgs-Vt is just the potential needed at the gate to inverse the channel under the oxide. This is usually a function of the doping under the oxide and substrate potential.

For short channel, you have a much earlier Vov compared to long channel. This is because the distance between the drain and source is much nearer and thus it creates a much higher electric field. This eletric field causes the velocity to max out at a lower Vds, and that inturn causes pinch off.

 

[ambreesh]

vov vdsat

Hi Jordan76,

As per your question , VDSsat is drain to source voltage for a MOS transistor. Now that is VDS and not VDSsat as I understand.
Vov= Vgs-Vth
As VDSsat(min) is observed lower than Vov (e.g by a 10 to 20 mv) and transistor is shown in saturation. Kindly observe the gds of the transistor for that Vds (<Vov). And compare the gds when it Vds is more than Vov we have designed for. We design for MOS transistors as good current sources.

VDSsat(min) we all find at times less then Vov for which we have designed the transistor cannot use Level 49 equations to do hand calculations.

Short channels do suffer from quick velocity saturation, but for pinchoff one still needs Vds>=Vgs-Vth

 

[jordan76]

gds vs vdsat

Thanks,guys.

Hi ambreesh,

And you can find Vdsat(min) without simulation, look into 2 stage opamp design example in Allen holberg.

Sorry, I still can not get it. Could you explain more a little bit?

Then I have another question for hand calculation: should we use Vov or Vdssat(min) to calculate Ids by square law?

BTW, I have clicked the helped me button^_^

regards,
jordan76

 

[ambreesh]

vdsat and vod

Hi Jordan76,

We always use Vov to calculate the W/L.
You would observe when you see the DC operating point parameters that VDSsat(min) is very close to the Vov what you have designed for, once the MOST is in saturation.
Thanks

 

[analogartist]

relation between vdsat and vds

Ambreesh,

Could you explain relation between gds ,vov(vgs-vth) and vdsat a little more.

If i have a device which has vov>vdsat ( by 100mv) and vds>vov ( by 500mv) and the device is shown to be in saturation. what will looking at gds tell me? What is the implication of gds(resistor for proper current source) in this case .

If I have a device where vov<vdsat (by 100mv) and vds>vod( by 500mv) and the device is reported to be in saturation. What will gds imply in this case?.

Also If I want to find the margin each device has, as said by you previous answer, I will have to look at vds-vov=margin?

-AA

 

[gpwu]

vov transistor

For long channel, Vov = Vdsat, IV curve looks like sqare law.
If same Vov as above for short channel device, Vdsat < Vov because higher field & IV curve looks like linear.

 

[amandeep_pec]

vdsat vgs-vt difference

I agree that Vov=Vdsat only for long channel devices...but what is the relation for short channel devices?

 

[broccoli]

transistors vov

it's what i want to konw, too.

 

[prince_123]

difference between vod and vds

Well I guess the need to search for those relation ship is to design your circuit in a more predictable manner.
Because of velocity saturation in submicron devices vdsat < vov, the transistor enters the saturation much earlier.

I suggest you to plot vgs-vth vs vdsat curve and do the curve fitting over it.
(It can be done by sweeping the vgs and performing DC analysis over the circuit and saving vdsat with every sweep point)
You will end up with the relation ship with some constants . These constants are technology dependent and will be different even for the same tech node of different foundries.

Please share your result with us all

 

[berryfan]

nmos vdsat derivation

I don't think this question is really serious.as long as we know whether the mos is in saturation or not is enough.

Added after 1 minutes:

long channel device in saturation is VGS-VT=VDSAT.

 

[prince_123]

vov vds

Well I think its important,
It might help to get more insight into the design, or lets say prior to spicing your circuit we know where the mos is.

 

[airboss]

vov vgs

i've forgot the formal answer to this question but i know if you wanna receive a complete answer to this, please read the text book "operation and modelling of mos transistor" by Y. Tsividus.

if you just need a 1-st order evaluation, then worry only about vov.
if you're running simulation, please compare vds and vdsat.
if you want to calcuate the output resistance, use the reciprocal of gds.
(the latter 2 points were what i learnt from my boss. i don't remember the reasons anymore.)