Temperature simulation in SPICE

[evilguy]

as we know in SPICE there is command for setting the simulation temperature. and the command is:

.TEMP

actually how this command affect the simulation? that this command is similar when we varies TNOM in SPICE parameter? for example if we set

.TEMP 30

and do dc operation point analysis, did we will get the same result when we do the same simulation with TNOM set to 30?
thank you.

 

[renwl]

it's the same.
TNOM is only the normal temperature setting.

 

[Guest]

Hi
in HSpice, most of options have by default values, the TNOM is also nominal temperature that circuit is simulated in it! (=25C deg) if you don't set your desired temp.
but when you use .Temp statement you may change this nominal value, but really this is chip temprature!

Regards,
SAZ

 

[evilguy]

from what i've read, TNOM is the temperature which we extract our transistor model. If i simply change the TNOM value, i should also consider other parameter as well such as VTHO, TOX and so on. am i right? when i simulate using

.TEMP 30

i dont get similar result when i change my TNOM to 30. that is the issue here. what is the correct method to simulate circuit in order to observe the effect of temperature to circuit performance? using

.TEMP 30

or using or simply changing TNOM=30

 

[renwl]

It's strange. I try it with spectre, and the results are the same. In spectre, the temp is the real temperature.

 

[pbs681]

from what i've read, TNOM is the temperature which we extract our transistor model. If i simply change the TNOM value, i should also consider other parameter as well such as VTHO, TOX and so on. am i right? when i simulate using

.TEMP 30

i dont get similar result when i change my TNOM to 30. that is the issue here. what is the correct method to simulate circuit in order to observe the effect of temperature to circuit performance? using

.TEMP 30

or using or simply changing TNOM=30

You are right, TNOM is the temperature at which we do the measurement for model extraction..... You cannot simply change the TNOM..... This is model parameter..... If you want to do temperature simulation, use .TEMP option..... say .TEMP 125

 

[evilguy]

It's strange. I try it with spectre, and the results are the same. In spectre, the temp is the real temperature.

i'm not familiar with spectre. is spectre also used similar model for transistor? i mean using BSIM model couse TNOM is one of parameter in BSIM model.

from what i've read, TNOM is the temperature which we extract our transistor model. If i simply change the TNOM value, i should also consider other parameter as well such as VTHO, TOX and so on. am i right? when i simulate using

.TEMP 30

i dont get similar result when i change my TNOM to 30. that is the issue here. what is the correct method to simulate circuit in order to observe the effect of temperature to circuit performance? using

.TEMP 30

or using or simply changing TNOM=30

You are right, TNOM is the temperature at which we do the measurement for model extraction..... You cannot simply change the TNOM..... This is model parameter..... If you want to do temperature simulation, use .TEMP option..... say .TEMP 125

thanks... i think i agree with you too. .TEMP is the correct way to simulate circuit in different temperature and i've prove it as my simulation is identical to one of paper that i've read. changing TNOM is wrong because it is suppose to be fix.