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.

Questions about Gm/Id Methodology

Status
Not open for further replies.

gaom9

Full Member level 4
Full Member level 4
Joined
Oct 8, 2007
Messages
228
Helped
7
Reputation
14
Reaction score
5
Trophy points
1,298
Location
China
Activity points
3,294
gm/id methodology

Hi,
Recently, I begin to learn the Gm/Id Methodology for analog circuit design, and I find it is quite useful.
I have read some papers about it, and I think the most important things for it is to get the curves of Gm/Id versus Id/(W/L) and Gm/Id versus VA, and then make some tradeoff to the all requirments, such as noise, match, bandwith, gain, and so on.
But I have no idea to get the two curves and the Va (early voltage).
Can anyone tell me the way to get them through Cadence and hspice, please? And I want some samlpes about this methodology, too, please.

Thank you very much!
Best regards!
 

gm, gds, hspice

Hi gaom9,

Hope this helps!

.lib "/xxxxxxx/SPICEFILEFROMFOUNDRY" TT
.options post brief dccap
.param gs=1
vgs g 0 dc 'gs'
mn1 g g 0 0 nch L=0.18um W=10um

.op
.dc gs 0.2V 1.8V 10mV

.probe gm_id = par('gmo(mn1)/i(mn1)')
.probe id_w = par('i(mn1)/W(mn1)')

.end
 

    gaom9

    Points: 2
    Helpful Answer Positive Rating
gm id methodology

Hi jnuhope,
Thank you very much!
Now I can get the curves of Gm/Id vs Id/(W/L)
But I think it should be .probe id_w_l = par('i(mn1)*L(mn1)/W(mn1)') in your reply. And then I use the log scales for X. Ok the curve come out. Thank you very much!
And can you tell me the way to get the Va(Early Voltage) and Inversion Coefficient IC. Also how to get the Kbeta in Hspice or Cadence.
Thank you very much!

I have push the buttom 'helped me'.

Best regards!:D:D:D
 

cadence how to extract gm/id curves

Do u read this paper:


IC and n values aren't a part of operation point parameters for bsim3 models. So u need to know them before simulation. In paper u can find how estimate this values.
Early voltage u can define through Gds, 1/Gds=Va/Id.
P.S. Kbeta is also operation point parameter.
 

edaboard gm/id hspice

I think DenisMark is right. I personally don't simulate Inversion Coefficient and early voltages (it's cmos afterall, right?) but you just have an idea where your transistor is biased(weak inversion or strong inversion)

I simulate gm*ro as follows:

.....
vg g 0 dc 'vt+0.2'
vd d 0 dc 1
.op
.dc vd 0 1.8 10mV
.probe av =par('gm0(mn1)/gdso(mn1)')

**provide dc voltage to gate as 200mV higher than vt, vt can be assigned by a rough value, then sweep vd**
 

site:www.edaboard.com gaom9

How to measure!
Sweep over voltage source not correct way and more complex. One can use transconductor in closed loop connection, aka diode connected MOST. So one can make a sweep over current source Id.
This is correct way, we stabilize amplifier's operation point than calculate small signal parameters by breaking loop. In this case breaking the loop isn't required because we extract operation point parameters.
E.g. one have common source amplifier with ideal current source load and need to simulate AC gain and phase margin, one correct way how to do it is as i mentioned above.
 

gm versus id

Thank you very much!
I calculate the value of I0 and let n=1.4. to get the curves of IC.


Best regards!
 

top papers in analog gm/id

hi, what is the method for plotting gm/id vs IC in spectre?
 

how to get gm/id vs id curve

Hi;

I have a question regarding the BW consideration in gmid methodology.

What value should i set for ft, such that I meet the BW requirements. I was thinking of designing for 10*ft; is this a reasonable approximation. Also, is there any method to select the ft.

Thanks
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top