-
Categories
-
Platforms
-
Content
I need some query about cmos circuit
I am reading about cmos circuit, I wonder how designer think when they design new chip
cmos circuit specification
Input voltage
output voltage
input current
output current
transistor width
transistor length
fall time
raise time
- they determine the requirement of customer
answer- less delay, low power consumption, low voltage
( i made temporary specification for cmos inverter )
Input voltage 3.3 v
output voltage 3.0 v
input current 1.5 micro amp
output current 1.0 micro amp
transistor width 6 micron
transistor length 2 micron
fall time 2 ns
raise time 3 ns
how designer decide specification
I mean how they make chip that work fine
I just want to know how does designer think ,
I am reading about cmos circuit, I wonder how designer think when they design new chip
cmos circuit specification
Input voltage
output voltage
input current
output current
transistor width
transistor length
fall time
raise time
- they determine the requirement of customer
answer- less delay, low power consumption, low voltage
( i made temporary specification for cmos inverter )
Input voltage 3.3 v
output voltage 3.0 v
input current 1.5 micro amp
output current 1.0 micro amp
transistor width 6 micron
transistor length 2 micron
fall time 2 ns
raise time 3 ns
how designer decide specification
I mean how they make chip that work fine
I just want to know how does designer think ,
Unfortunately your specs for an inverter would produce a largely useless device. The inverters output would be insufficient to drive even a single input.
What you are defining is an incomplete set of input and output specs for a logic family. These would normally include 4 voltages for inputs and 4 for outputs defining the limits on valid voltages for a logic high and low. The inputs will typically accept a wider range than the outputs will produce.
Then you would specify an output that was capable of significant fan- out. In some logic families, fan-in can also be valid.
Then there's a plethora of other things including rise and fall times (perhaps min and max), propagation delay, power supply specs and thermal considerations.
None of that will define a perfect inverter. For a particular family, the spec will consist primarily of the relationship between the input and the output logic states. Perhaps the perfect inverter will meet the tightest end of all specs where there is a range, and well exceed the minimum where there is just a minimum spec (and similarly, but opposite, for those which just have a maximum).
Will that be the perfect inverter? No. A perfect inverter would be the first choice every time you wanted an inverter. Such a beast does not exist. For example you might want open drain or complementary outputs -- you cant have both. Or you might want the ability to source and sink a large current, or you might want a current limited output. a gain you can't have both. You might want to interface with ECL and such a requirement would seriously impact noise immunity if you also wanted to interface with CMOS. Maybe you want to drive long lines, with minimal crosstalk. This would impose limitations which would conflict with the ability to have a very small rise and fall time.
You may have an easier time describing a "perfect" woman.
What you are defining is an incomplete set of input and output specs for a logic family. These would normally include 4 voltages for inputs and 4 for outputs defining the limits on valid voltages for a logic high and low. The inputs will typically accept a wider range than the outputs will produce.
Then you would specify an output that was capable of significant fan- out. In some logic families, fan-in can also be valid.
Then there's a plethora of other things including rise and fall times (perhaps min and max), propagation delay, power supply specs and thermal considerations.
None of that will define a perfect inverter. For a particular family, the spec will consist primarily of the relationship between the input and the output logic states. Perhaps the perfect inverter will meet the tightest end of all specs where there is a range, and well exceed the minimum where there is just a minimum spec (and similarly, but opposite, for those which just have a maximum).
Will that be the perfect inverter? No. A perfect inverter would be the first choice every time you wanted an inverter. Such a beast does not exist. For example you might want open drain or complementary outputs -- you cant have both. Or you might want the ability to source and sink a large current, or you might want a current limited output. a gain you can't have both. You might want to interface with ECL and such a requirement would seriously impact noise immunity if you also wanted to interface with CMOS. Maybe you want to drive long lines, with minimal crosstalk. This would impose limitations which would conflict with the ability to have a very small rise and fall time.
You may have an easier time describing a "perfect" woman.
Steve, you know very well that there aint no such thing as a "perfect woman"
But can we describe (define) a "perfect woman"?Steve, you know very well that there aint no such thing as a "perfect woman"
