Transient simulation in ngspice with xschem.
# analog-design
n
Hi, I am designing a MOS biasing circuit in xschem. During transient simulation the results for output voltage are different for different time step. Two simulation results for "tran 100u 1" and "tran 10u 1" are attached below. In the first condition, the circuit is showing proper output voltage of around 880mV. In the second condition, the circuit is unstable. Can anyone please tell me why a comparatively small difference in time step is leading to such different results? Is circuit the problem or I am missing something regarding transient simulation?
pic_2.png
l
Show us your schematic, so we can see which biasing circuit you're simulating. Most problems with biasing circuits are caused by bistability, so they need start-up circuits. Maybe that is your problem.
n
Hi Luis Henrique Rodovalho , Thank you so much for your reply. I was just simulating a supply independent reference voltage generator to bias an opamp. I have not designed the startup circuit yet. I am guessing that is the issue. Please see the attached circuit. THank you so much for your help.
l
Yep. You will need a start up circuit. This kind of self biased current source is normally CTAT. You should stick with a constant gm one if you want a PTAT one. Depending on the sizing of the differential pair, you will need a PTAT current to keep its own gm constant too.
n
Ok. Thank you so much for your help.