Under LTSpice, for the VBIC (level=4 or level=9) and MEXTRAM (level=504) Bipolar Transistor Options, as some people pointed me, I dont know how to simulate an e.g. 2N3904 under VBIC or MEXTRAM?

The single reference, suggested by Dietmar Warning (if he is the real) is in sourceforge.net from 2010.

IS the situation still the same?. Or there is a better reference?


I did some BJT modelling previously, from which someone posted a link to the full document on the MEXTRAM model, and also the VBIC model paper may help you.

Here is my condensed advice: don't use the default models from LTspice, write your own .txt file that explicitly defines the parameters that you want.

Writing your own LTspice model

For example when I was trying to model a germanium BJT I defined a text file called bjtmod.txt, which contained only one line

.model IDEAL_GERM_NPN NPN(IS=1E-6, BF=120, BR=5)

You can see here first the statement is to define a .model, called IDEAL_GERM_NPN, which is an NPN BJT with parameters IS=1E-6, BF=120, BR=5.

To use this in LTspice you must place it in the same folder as the schematic you are working on. Including it to the schematic is done by writing a SPICE statement:

.include "bjtmod.txt"

To set your BJT to use this model make sure that the Value field is set to IDEAL_GERM_NPN.

You can then customize each parameter of the model, including avalanche terms. From the reference you provided:

.model 2n3904L4 NPN(level=4, rcx = 2.65711, rci = 0.0, rbx = 0.634251, rbi = 5.203349, re = 0.0001, is = 1.26532e-10, nf = 1.5, nr = 2.89609, fc = 0.512134, cje = 4.64214e-12, pe = 0.4, me = 0.256227, cjc = 3.015688e-12, cjep = 7.53922e-13, pc   = 0.4, mc = 0.238109, cjcp = 0, ps   = 0.75, ms   = 0.5, ibei = 6.13333850374693e-13, nei = 1.5, iben = 2.30771e-09, nen  = 3.31052, ibci = 6.13333850374693e-12, nci = 2.89609, ibcn = 2.30771e-09, ncn = 1.9876, vef  = 566.483710419377, ver  = 8.69028854282359, ikf = 0.0272221, ikr  = 0.272221, tf   = 4.19578e-10, xtf  = 0.906167, vtf = 8.75418, itf  = 0.0105823, tr  = 6.82023e-08, td   = 0, ea = 1.05, eaie = 1.05, eaic = 1.05, eane = 1.05, eanc = 1.05, xis  = 1, xii = 0.9, xin = 0.668948, kfn  = 0, afn  = 1, avc1 = 0.02, avc2 = 110)

Lastly, don't get fixated on whether it resembles a real BJT. It won't. Some of these parameters vary hugely between devices. Extraction procedures are phenomenally complex, take that from someone who wrote one for only around 12 BJT parameters.

  • \$\begingroup\$ Thanks you for your answer. Do you know if there is some conversion procedure from SGP models (which most manufacturers actually provide) into VBIC or MEXTRAM? Should i start a new question with that? \$\endgroup\$ – Brethlosze Jun 29 '18 at 14:35
  • 1
    \$\begingroup\$ I wouldn't keep asking questions for each small task. Perhaps see if you can find a contact at a semiconductor manufacturer and see if they would be willing to share a MEXTRAM model with you. \$\endgroup\$ – loudnoises Jun 29 '18 at 15:00
  • \$\begingroup\$ Yes of course i dont want to keep you tied to this question. I just want to clarify my issues. Thanks. \$\endgroup\$ – Brethlosze Jun 29 '18 at 16:05
  • 1
    \$\begingroup\$ Oh I didn't mean to be rude, I meant for the site. It is better if you can come up with questions that deal with the root of the issue rather than the trivial aspects. That way you will receive better answers. \$\endgroup\$ – loudnoises Jun 29 '18 at 16:53

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.