I'm in day 3 of learning LTSpice, finally dragging myself into the 90's, and I am wondering if there is a repository of Spice models for active devices (SITs JFETS, old MOSFETS etc.) that are used in the FW/PASS/ZEN etc. designs? I have looked around here but only find smatterings of devices in the last 20 years. If there isn't maybe we could start one?
Thanks!
Paul
Thanks!
Paul
It's worth asking, "Why should I believe this spice model is correct?" .
When someone attaches a model file to a Forum post, or puts it on a web page, where did they get it? Did they know it is correct? Or did they merely copy it from an earlier post or website? Did each person in the chain of spice model copiers, know and trust the person they got the model from?
It's also worth asking, "Why should I believe the spice models in the presupplied LTSPICE library are correct?" Those discrete transistors were never built or sold by Linear Technology / Analog Devices. LT and AD make integrated circuits, not discretes. LT's software team merely got those models from someone else, and (I believe) simply copied them into the LTSPICE library without verifying or testing them. Perhaps they made typos when copying. Perhaps the originator(s) did a sloppy job. Why should you blindly trust them?
A different way to ask the question might be, "How much do I trust this spice model?" Do I expect it to have 1 part per million inaccuracy? 1% inaccuracy? 10% inaccuracy?
One possible path out of this jungle of uncertainty, is to only trust models that were extracted and tested and verified by people you know and trust. Such as Bob Cordell, who measures devices himself, extracts spice models himself, and puts them on his Cordell Audio website where anybody can download them and use them. Then as a corollary: if necessary, use but don't trust, models that weren't extracted and tested and verified by people you know and trust.
When someone attaches a model file to a Forum post, or puts it on a web page, where did they get it? Did they know it is correct? Or did they merely copy it from an earlier post or website? Did each person in the chain of spice model copiers, know and trust the person they got the model from?
It's also worth asking, "Why should I believe the spice models in the presupplied LTSPICE library are correct?" Those discrete transistors were never built or sold by Linear Technology / Analog Devices. LT and AD make integrated circuits, not discretes. LT's software team merely got those models from someone else, and (I believe) simply copied them into the LTSPICE library without verifying or testing them. Perhaps they made typos when copying. Perhaps the originator(s) did a sloppy job. Why should you blindly trust them?
A different way to ask the question might be, "How much do I trust this spice model?" Do I expect it to have 1 part per million inaccuracy? 1% inaccuracy? 10% inaccuracy?
One possible path out of this jungle of uncertainty, is to only trust models that were extracted and tested and verified by people you know and trust. Such as Bob Cordell, who measures devices himself, extracts spice models himself, and puts them on his Cordell Audio website where anybody can download them and use them. Then as a corollary: if necessary, use but don't trust, models that weren't extracted and tested and verified by people you know and trust.
