Syntax in Hspice for measuring the settling time of the circuit

[nosrej]

Is there a specific syntax in Hspice for measuring the settling time of the circuit. . . What are the requirements in measuring the settling time of an IC? Please elaborate your explanations I'm just a newbie in using Hspice.
Thank you. . .

 

[crutschow]

Re: Netlist for Settling time

I don't think there are any special commands for measuring settling time. You just perform a transient response and measure the time from the output waveform as referenced to the circuit input waveform.

The requirements for measuring settling time are whatever you make them. It's usually to within some small percent of the final value (such as 0.1% or 0.01%, etc.).

 

[nosrej]

Re: Netlist for Settling time

how do i measure the output waveform using a netlist syntax? or should it be done in avanwaves using hspice?

 

[crutschow]

Re: Netlist for Settling time

how do i measure the output waveform using a netlist syntax? or should it be done in avanwaves using hspice?

I don't understand your question about measuring with a netlist syntax. :?: You do a transient simulation, however that's done in Hspice.

I'm not that familiar with Hspice so I don't know what "avanwaves" is.

 

[erikl]

Re: Netlist for Settling time

how do i measure the output waveform using a netlist syntax?

Use Hspice's .MEASURE (FIND and WHEN) syntax: View attachment hspice_command_ref_p_2-104--2-106.pdf

 

[nosrej]

.meas tran tsettling when I(mnload) = '0.01*I(mnload)'

I'm measuring the current from a sensor then it is mirrored by a current mirror. Is this syntax correct? or it is incomplete?

settling= failed
is the syntax incomplete?

 

[erikl]

.meas tran tsettling when I(mnload) = '0.01*I(mnload)'
...
is the syntax incomplete?

Yes, but I think you're getting close:

1. To measure settling time, you have to measure 2 event times: the begin (stimulus) and the end (time of satisfactory result); settling time is the difference between these 2 events = time points.

2. You can't ask it for an event time with an inequation: I(mnload) = '0.01*I(mnload)'
If I(mnload) is the current parameter to find the end time point, you first have to define a parameter for -- say -- a max. of this load current, e.g.

.param Imnloadmax = 1.0e-3 A   # and then:
.meas tran t_result when I(mnload) = '0.01*Imnloadmax'

3. If there are more than 1 transitions in your transient analysis, you have to tell the .meas statement, at which edge (e.g. RISE=1 ; FALL=2) it should take the measurement.

HTH! erikl

 

[nosrej]

Thanks I helped a lot. . . I could also try to measure it in avanwaves. Another question is how can I determine which is the point where the output reached its final value based from the graph below. The graphed below is the mirrored current from the sensor.

 

[erikl]

From such a chaotic curve behavior you can't define a reasonable final value, and so no settling time - IMHO.

 

[nosrej]

Is a pulse input really required to do settling time in transient analysis? Can't I just use vdd and vss with dc value

 

[keith1200rs]

A settling time is normally defined as the time it takes a signal to 'settle' to within a specified percentage (usually a very small percentage) of its steady state value after a step change if a specific amount (usually a large value). So, unless you have a different definition, you need to be applying a large step input.

Keith

 

[nosrej]

Does that mean that I have to replace the vdd source by a pulse step input? or where will I connect the step input in the circuit.

 

[keith1200rs]

If you are measuring the settling time of a circuit from power-on then yes, use a pulse for the VDD source. You will also need to define the power supply rise time.

Keith.

 

[nosrej]

what if I'm measuring the settling time of a mirrored current from the sensor? will I still use pulse as VDD source? or should I connect it somewhere else in the circuit.

 

[keith1200rs]

In which case you would put a step in the input current source, I would suggest.

Keith.

 

[nosrej]

Based from this graph is there still no settling time?