Joined: June 2006
Still here, Dave? Thanks for continuing the conversation.
|Any simulation that incorporates propagation delay is not strictly boolean anymore.|
Boolean means that there are two possible states. Do propagation delays add more possible states?
|Moreover, the mosfets only operate in two states, on or off.|
Interesting. And here I thought that MOSFETs had three operating modes, each continuous and roughly linear, making MOSFETs no more boolean than BJTs or JFETs. But you're the genius.
|Unlike older silicon transistors, mosfets ...|
And what are MOSFETs made of?
|...mosfets require no resistive or capacitive elements.|
More accurately, it's complementary logic that requires no resistors. MOSFETs aren't the only transistors that can be used in complementary logic, and MOSFETs require pull-up or pull-down resistors if they're not used in a complementary configuration.
And using MOSFETs rarely eliminates the need for capacitors. More often, their inherent capacitance is a hindrance rather than a help.
|I had assumed that I was talking with people who were sufficiently knowledgable to recognize that the difference between modeling a CMOS processor at the transistor level and the gate level is a quibble because the individual logic gates are composed of just a few simple on/off mosfet (transistor) switches.|
Like most EEs, I used SPICE, which most definitely does not treat MOSFETs as on/off switches. What transistor modelling tool did you use?
Oh, I almost forgot to give you an opportunity to solve a problem. Here's one, and it's about as easy as it gets: Given an NMOS transistor with I_dss = 16mA and V_p = -4V, what is the gate-to-source voltage required to establish I_d = 32mA?
"I wasn't aware that classical physics had established a position on whether intelligent agents exercising free were constrained by 2LOT into increasing entropy." -DaveScot