Temperature Step in a .TRAN Simulation?

LTSpice or regular?

I'm pretty sure you can do it with a .STEP or a .TEMP command in LTSpice.

Simulating thermal distortion?

.STEP isn't the same as a step in temperature _during_ the simulation run.

.STEP simply applies a data change at the _beginning_ of a run.

I want, during a transient solution, to have a temperature change to test a temperature-related shutdown mechanism.

I suspect I will have to do a replacement in subcircuits...

TEMP => V(MyTemp)

and fudge it with a voltage step :-] ...Jim Thompson

Easy in LTspice.

.step temp -55 125 10

Cheers

Phil Hobbs

that will give you a simulation run for each static temperature, afaiu Jim wants the temperature to change during simulation

I don't think it is possible, I've tried it with all conceivable variations of trying to modulate temp/tnom etc.

it all fails with anything but a constant

-Lasse

So I gathered after I posted. You can set individual device temperatures to different constants, allegedly, so maybe putting some curly-brace function of time in there would work.

Cheers

Phil Hobbs

Yes, during simulation.

But what I'm working on is a behavioral model of a real temperature sensor that I've curve-fitted.

Since "TEMP" is a "reserved word" in Spice I think the temporary solution, for testing purposes, is to substitute V(MyTemp) for TEMP in the equations, then simply apply a voltage step or maybe a ramp would be better.

Hmmmm! Helps to discuss... here's the solution...

Do the substitution, as above and leave it permanently.

The user can either input a voltage on node MyTemp, or hang an EValue source on it...

Evalue MyTemp 0 VALUE {TEMP} ; >:-}

Kicking ideas around almost always produces solutions! ...Jim Thompson

I can't make temp work but tnom works, as long as it is a constant..

Version 4 SHEET 1 1852 680 WIRE 208 16 -144 16 WIRE 544 16 208 16 WIRE 208 32 208 16 WIRE -144 48 -144 16 WIRE 544 96 384 96 WIRE 208 112 48 112 WIRE 384 144 384 96 WIRE 480 144 384 144 WIRE 48 160 48 112 WIRE 144 160 48 160 FLAG 208 208 0 FLAG 544 192 0 FLAG -144 128 0 SYMBOL npn 144 112 R0 SYMATTR InstName Q1 SYMATTR Value 2N2222 SYMBOL voltage -144 32 R0 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName V1 SYMATTR Value 5 SYMBOL npn 480 96 R0 SYMATTR InstName Q2 SYMATTR Value 2N2222_1 SYMBOL res 192 16 R0 SYMATTR InstName R3 SYMATTR Value 1k SYMBOL res 528 0 R0 SYMATTR InstName R4 SYMATTR Value 1k TEXT -176 232 Left 2 !.tran 1 TEXT 688 160 Left 2 !.model 2n2222_1 ako:2n2222 tnom=120 TEXT 688 208 Left 2 !;.model 2n2222_1 ako:2n2222 tnom=27

-Lasse

Ruminating whilst shopping at the grocery... this behavioral model has an output pin with a _very_current-limited drive that is equivalent to the voltage that an LM94022 produces. Internally, the circuit views this voltage to determine circuit shutdown.

So a simple solution... back drive this pin with an equivalent to the LM94022 that is _not_ current-limited, and provide it with an input pin to load MyTemp into its equation. ...Jim Thompson

Hence my comment about thermal distortion. It is a classic case where you want the temperature to change dynamically during a simulation.

Ooh ouch. So ideally you'd like to make sure that your temperature detection/shut down works correctly as temperature is changing dynamically?

I think I'd separate the circuit into part A and part B, where part A is the temperature sensor and part B is the rest. Characterize part A over temperature, fake its behavior in a subcircuit, then make sure that part B responds correctly to the changes in part A at a variety of different temperatures.

It'll probably be valid, unless the temperature shows strong gradients across the chip or changes as rapidly enough to be close to the time constants of any RC pairs.

Yes. How I ended up fudging is sort of round-about, but it works...

In the actual circuit there is the equivalent of an LM94022 temperature sensor with GS0 and GS1 both set to 1's.

Internally to the overall circuit the output voltage of the LM94022 drives a comparator (with hysteresis) to effect the thermal shutdown.

Fortunately this voltage also comes out on a pin, so that the user can monitor temperature, if he wants (to me, not very useful :-)

Since I modeled the LM94022 for this application in the first place, I realized that the LM94022 has a weenie output, limited to +/-150uA.

So I made a test tool, an equivalent to an LM94022, but with a voltage source output, replaced every instance of "TEMP" in the equations with "V(MyTemp)", where "MyTemp" is an input pin.

The output of this test tool overdrives the internal LM94022 output, if I apply a voltage to pin "MyTemp" equal to the desired temperature.

So I can apply a ramp to "MyTemp" and watch the thermal shutdown function under transient conditions.

Works great!

...Jim Thompson