Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

the proper value of Vgs-Vth of the input MOST

Status
Not open for further replies.

swolf

Newbie level 6
Joined
Nov 27, 2007
Messages
12
Helped
1
Reputation
2
Reaction score
0
Trophy points
1,281
Activity points
1,377
I am designing a folded cascode OTA for pipelined ADC. It requires 400M Hz GBW with lowest power consumption. the load capacitance is fixed (i.e. 4 pF), so i need to design the input MOST to get the largest gm under the lowest id of the input MOST.
as we know, gm is inversely propotional to (Vgs-Vth), the question is: what is the proper Vgs-Vth?
i designed the mosts and simulation result is:
#(0.18um process) the simulation tool is Spectre.#

w=300u, L=0.18u, Id=1.023mA, Vgs=558mV, Vth=470mV, Vdsat=118mV, Vds=1.187mV, gm=14.6mS, ron=1.16K

is this value (Vgs-Vth=88mV) proper? would it course some problems in reality? or can i still decrease the Vgs for more larger gm? what is the proper value in reality?
Would you give me some hints? Thank you in advance!
 

swolf said:
as we know, gm is inversely propotional to (Vgs-Vth), the question is: what is the proper Vgs-Vth?
As I know, gm is proportional to (Vgs-Vth) directly.
swolf said:
is this value (Vgs-Vth=88mV) proper? would it course some problems in reality? or can i still decrease the Vgs for more larger gm? what is the proper value in reality?
Would you give me some hints? Thank you in advance!
As I know, in 0.18um process, Vov<100mV is not good
because MOSFET behavior in this Vov is similar to sub threshold.
regards
 

I think your device is in triode region.
Better to design VDS > VDSAT, so that the device is in sat region.
 

i am sorry that i typed a wrong value: Vds should be 1.187 V

gm=2*Id/(Vgs-Vth), when Id is fixed, we can obtain larger gm by decreasing vgs-vth, is that right?

hr_rezaee said:
As I know, in 0.18um process, Vov<100mV is not good
because MOSFET behavior in this Vov is similar to sub threshold.
so what will happen to the circuit when input mosts operates in the sub threshold?

thank you
 

gm=2*Id/(Vgs-Vth) is valid for strong inversion, which means your Vgs-Vt is larger than 2nVt. 2nVt is around 78mV and depends on what technology you use.
As I know, for low power design, input devices operate in the moderate inversion, which operate between the strong and weak inversion.
 

swolf said:
gm=2*Id/(Vgs-Vth), when Id is fixed, we can obtain larger gm by decreasing vgs-vth, is that right?
yes, it is true.
sorry.
 

GBW=Gm/Cload, where Gm-transconductance in input diff pair. Vgs-Vth=88mV ans Gm/Id=14.3 indicate moderate inversion level. U can increase Gm/Id up to about 20-22 in weak inversion region (Vgs-Vth<-50mV) simultaneously decrease Id (power consumption).
Also for diff pair in weak inversion u can use PTAT current reference to make GBW close to constant.
Weak and moderate inversion level is optimal region for amplifier. Strong inversion for diff pair is used in Gm-C filter where linearity in attention.
 

Very clear for me. I was still wondering how you get a rough number like 14.3 or 20-22 for moderate or strong inverison by using BSIM model. I know some people like Jesper or Binkley use EKV models and they can decide what regions the devices are in.
 

Yes, it is right.
Input pair works in weak inversion is OK.
Although most model is not very good in modeling ro in this region.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top