[help] CMOS design procedure

[speedracer]

rcaverly +pdf

hi!

i am trying to learn the design procedure for a mixer. could someone tell me how to determine the bias current and the bias voltage for the device, since the source of the device is at different potential as the substrate.

thanx!

speedracer

 

[khouly]

i think u can estimate the bais current from calculation of the gm required by this device
frist assume the source and bulk connnected and calculate the current then begin to etirate with the vbs , to get minimun error

 

[amitbhaiji]

Which type of mixer are you talking about....is it active gilbert cell mixer.
Amit

 

[speedracer]

thanx for the replies so far.

yes, i am studying the gilbert cell mixer. the professor has not been very clear about the whole procedure. i figured out the width of the transistor, but i became confused by the rest since there are too many variables.

so, i should figure out Ids next, right?

thanx!

speedracer

 

[khouly]

also u must be clear which transistor
there r 3 function of the transistors in gilbert cell mixer
the transconductors and switch and the active load

 

[speedracer]

hi! khouly,

you are absolutely right. i am planning to do the rf differential amp first, then the switch then the active load. i am using 0.35 um cmos technology.

i am totally confused by the next step after i fiugred out the dimension of the transistor. i know i am suppose to figure out the Id but the substrate and source are at different potential, so how can i accurately determine Id?

thanks for the help!

speedracer

 

[khouly]

hi
assume frist that the source and bulk in the same potential then begin to etirate to calculate the new VT , then put it back in current equations , and so on
until to get very samall error

in my graduation project one of my friends designed a mixer "gillbert cell " work around 2.4 Ghz , i will see the document . and i will try to hel u as far as i could

 

[speedracer]

hi! khouly,

my professor and his t.a. are definitely no help. my friends and i estimated the device width from the text. we are trying to find a more deterministic method to ensure an appropriate Ids to keep the transistor in saturation.

we were told that to simulate the transistor by itself, but the Ids vs. Vds curves could not tell us what is the 'best' Ids or Vgs. are we totally off base?

thanx!

speedracer
irinakuz@yahoo.com

 

[T Zul]

The IV curve is the best place to ensure that your circuit has the correct biasing. For example by running your transistor at a particular width and length will lead you to an IV curve in saturation mode.

T Zul

 

[boy]

This may help.

hxxp://eesof.tm.agilent.com/adsdemo/seminar4/extras.html#workbook

 

[uncle_urfi]

**broken link removed**

**broken link removed**

**broken link removed**

**broken link removed**

**broken link removed**

**broken link removed**

**broken link removed**


Also see the book by Stephen Maas on mixer design.

Have fun!

 

[layes]

gm
id
and
SR
settling time
....