Bipolar Layout/SiGe Bipolar Layout Issues

[Puppet123]

Hello,

I am doing Bipolar layout using two difference bipolar processes - one allows me to set the emitter area and lengths to whatever I want (there is a maximum) while another only has pre-determined npn bipolar cells that I have to use and then use multiples (ie. mult=n, where n is whatever emitter area I need).

What is the reason for the difference in these two PDKs/processes ? Is it because the npn in each PDK is modeled differently and I have to use multiples in one kit and not in the other ?

Thank you.

 

[dick_freebird]

The first sounds like what I was used to, the second more like
what I've seen from a SiGe foundry more recently. I believe it
has to do with the much fussier HF expectations and modeling
issues for SiGe applications. Get a 700MHz fT BJT slightly wrong
and maybe your 10MHz GBW op amp comes out 9MHz, you can
still sell some to somebody. But get 2.2GHz BW out of your SiGe
PA, and no 2.4GHz ISM band applications for you. So had best
be accurate, and delivering accuracy on a single fixed geometry
BJT is hard enough - make it 2D parametric w/ geometry, and it's
a mess for modeling group to fit. Such finesse might be "left for
later" and later never comes.