plotting in Kelvin instead of Celsius for temperature sweep

Setting the units on the temperature form doesn't change what you're sweeping - it merely makes it easier to set a single temperature. The units in spectre are always Celsius.

You can easily achieve what you want in ADE L (assuming you're doing a DC sweep here) by doing:

1. Create a variable called "kelvin" with value (say) 300
2. Create a variable called "tempC" with value kelvin-273.152
3. On the Setup->Temperature form, set the temperature in degrees to VAR("tempC")
4. In the DC analysis, rather than sweeping temperature, sweep a variable and sweep kelvin from (say) 230 to 390.

Voila! (this example is just the current through a resistor with some made up temperature coefficients just to provoke some kind of interesting simulation results):

Andrew

Setting the units on the temperature form doesn't change what you're sweeping - it merely makes it easier to set a single temperature. The units in spectre are always Celsius.

You can easily achieve what you want in ADE L (assuming you're doing a DC sweep here) by doing:

1. Create a variable called "kelvin" with value (say) 300
2. Create a variable called "tempC" with value kelvin-273.152
3. On the Setup->Temperature form, set the temperature in degrees to VAR("tempC")
4. In the DC analysis, rather than sweeping temperature, sweep a variable and sweep kelvin from (say) 230 to 390.

Voila! (this example is just the current through a resistor with some made up temperature coefficients just to provoke some kind of interesting simulation results):

Andrew

thanks a lot Andrew for the cool ! trick.

Request one more suggestion from you. I tried sweeping starting from -273degC or even at -200degC and see DC convergence issues causing simulation errors. The circuit is basically an opamp and few diodes.

I have ++aps mode enabled

Any workarounds for this ?

regards

I've no idea what models you are using, but have your models been characterised at cryogenic temperatures? Most foundries don't characterise at such low temperatures (and a fraction of a degree above absolute zero is quite optimistic for reasonable results anyway). I expect the models become non-physical at such extreme temperatures, leading to convergence problems.

++aps is unlikely to have anything to do with it.

Andrew.

ok thanks Andrew