Designer’s Guide Community

‹ Previous topic | Next topic ›

Pages: 1

The high temperature spice model (Read 592 times)

dandelion

Community Member

Offline

Posts: 98

The high temperature spice model
Mar 19^th, 2008, 7:36pm

Hi,
I am designing a circuit which contains a overheat block. The detection thershold is 160'C and if the 160'C is reached, the block will be shutdown.

But I checked the spice model and found the model is only for -40'C-125'C. These is no model descriptions above 125'C. Communicated with the Foundry, they said the above 125 model is not measured.

But for the overheat sub-block, i must simulated it wit 160'C. I wonder is there any risks in it? The actural silicon will be much descranpancy from the design?

Pls. share with your experienc.
Thanks

IP Logged

Geoffrey_Coram

Senior Fellow

Offline

Posts: 1999
Massachusetts, USA

Re: The high temperature spice model
Reply #1 - Mar 20^th, 2008, 6:22am

I'd suggest you do some plots of all the currents and capacitances as you step the temperature from 100 to 200 in 10-degree steps. Make sure they look reasonable.

Some models have linear temperature effects, and you might find that eg VSAT becomes negative at 190 degrees, which would be a strong indication that you won't get the right answer at 160. BSIM4 has VSAT_T = VSAT - AT * (T/Tref - 1);

If at first you do succeed, STOP, raise your standards, and stop wasting your time.

IP Logged

HdrChopper

Community Fellow

Offline

Posts: 493

Re: The high temperature spice model
Reply #2 - Mar 20^th, 2008, 7:10am

Certainly leakage currents will be a headache if not properly modeled. Usually, somewhere between 110 to 130C is where the leakage currents tend to play a role in the total error you get for several analog circuits.
I usually simulate up to 165C and at such temperature a common concern is balancing leakage currents in precision analog circuits.

You might want to consider this too.

Tosei

Keep it simple

IP Logged

Pages: 1

‹ Previous topic | Next topic ›