1

I need to sweep Tox parameter in ngspice (BSIM level-54), but I found that there is no such parameter as Tox. Instead I need to play with Toxe, Toxp and Toxm. I tried to control these in my code, but it gives error: no such parameter.

e.g. 

Mp1 vout va vdd vdd pmos l={len} w={wpmos} toxe = {oxidelen}

Error on line 49 : mp1 vout va vdd vdd pmos l=     4.50000000000000059e-08   w=    2.50000000000000020e-06   toxe=    1.00000000000000002e-08 
unknown parameter (toxe)

I have also gone through the previous post "Dependency of gate oxide thickness in leakage" in which the author had changed Toxe, Toxp and Toxm in the model card itself.

Is there any way by which I can sweep Tox directly from my code?

user2943160
  • 2,878
  • 1
  • 17
  • 32
ssinghal
  • 21
  • 1
  • 2

1 Answers1

1

If you look at section 1.1 in the BSIM4 manual, there is a description of the oxide model. You just cannot "sweep" the model and get anything close to real behavior, but DTOX can be swept.

At some point, you'll start to tunnel when the oxide becomes thin. Also, depending on the process, you'll see "gate leakage" due to the dirty etch from the drain terminal to the gate. This looks like gate leakage, but you can mitigate it by changing your etch and moving out your contacts.

b degnan
  • 3,167
  • 2
  • 17
  • 36
  • But, if I change DTOX in the model card, that means I will be changing the oxide width of all the nmos/pmos in the design. What shall I do if I need to change the oxide width of only few selected transistors? – ssinghal Sep 01 '16 at 10:24
  • copy the card and then sweep that one model. you'll need to do some renaming. – b degnan Sep 01 '16 at 10:43
  • Got it. I will do the renaming part. – ssinghal Sep 01 '16 at 11:21
  • In order to see the effect of leakage current w.r.t Tox, I will be performing following steps: 1) Find leakage current in standby mode with TOX1 (or DTOX1) . 2) Find leakage current in standby mode with TOX2 (or DTOX2). I am assuming that the difference in leakage current in the above two steps is approximately due to change in oxide tunneling currents. Is that fine or some other leakages (like subthreshold) will play a signficant role due to the changes in TOX ? – ssinghal Sep 01 '16 at 11:22
  • Firstly, subthreshold is not "leakage", oxides cause leakage. it is an operating range. Changing the oxide thickness changes the voltage divider so it increases subvt current. Please see the images in this answer: http://electronics.stackexchange.com/questions/240747/simulation-model-for-floating-gate/240758#240758 – b degnan Sep 01 '16 at 12:04