BSIM4 MOSFET Model - RF Cafe Forums

 Post subject: BSIM4 MOSFET Model

Posted: Wed Jul 11, 2007 9:14 am 

I'm trying to simulate a circuit containing a MOSFET at a frequency of around 2GHz. The problem I'm facing is that I can't seem to find a MOSFET that operates in that frequency range (It needs to be a switching MOSFET, which means the transistor operating frequency should be around 20GHz for it to have a small enough capacitance). I tried creating my own BSIM4 MOSFET model, but the problem is that I'm not sure about the values of the model parameters for it to be physically realizable. Does anyone know of any place where I can get detailed actual MOSFET model parameters for high frequencies?

 Post subject:

Posted: Wed Jul 11, 2007 10:50 am 

Hi Quasar,

Depending on what type of MOSFET you are needing and what your voltage levels are will determine what kind of model parameters you will be needing. For instance, if it is for large voltages (>5 V) you will be needing a different type of MOS with different parameters than one which is less than 5 V.

I don't know much about higher voltage FETs, but for <5 V applications an okay place to start is through the MOSIS web site (www.mosis.org). They keep a detailed list of SPICE parameters for different size technologies (from 1.2um CMOS down to 90nm) with different runs so that you can see the process variations. As far as needing an ft of greater than 20 GHz, you will need to use at least 0.35um CMOS. Even then, it may be pushing the limits.

If you can't find the information you need on the MOSIS site, you can always check foundry web sites. Sometimes they post their SPICE parameters there for older processes and you can see if they are in the ballpark that you are in.

Hope this helps.

-J

 Post subject:

Posted: Wed Jul 11, 2007 12:49 pm 

Well, fortunately I don't need a high voltage MOSFET. I did check out the MOSIS website and they do have good details but the problem is that there are certain parameters of the BSIM3 model that are not mentioned. I think that's due to the fact that they test wafers, not transistors. Information such as the contact resistance of the gate, drain, and source, the width and length of the channel, the area and perimeter of the drain, and the source. All of these and some more are critical for the design and are not mentioned. So I ended up getting weird results on the simulator. I guess I'll have to try to find a way to guess these values, and I'll try using the 0.25um or the 0.18um CMOS. Thank you for your help.