hi:
There is one question.
We can only use the fixed-size model in the design of BJT ic ?
Or we can zoom in (zoom out) the emitter area of the model which is provided by foundry?And whether it will affect the precision of simulation?
Because i know if the width or length of the mos exceed the limition of the model , the result of the simulation will be inaccurate.
Thanks.
Usually, the foundry should provide several kinds of BJT models, these models have different emitter area.
If you need to change the emitter area, you can use different model or parallel several same models. For example, a model area is 'A', parallel two models, the area is '2A', three models is '3A'...
For a accurate simulation result it is better to use the models provided by the founadary . For BJT it is very important to select the proper models , otherwise it will result in lot of mismatch in current values
Similar to MOS the scaling is not exacly proportional. In MOS-Models that is at lower frequencies implemented within a single model. For bipolar you could use only subcircuit extented models or HiCUM. The fabs monitor actual the derivation from the scaling law but know that the EDA infrastructure is missing the scaleable support for bipolar. Therefore most fab external kits and models use singular devices.