I think that you can't assign AC models because of the license that you are using. Nevertheless, you only need DC models and VRM models for DC drop analysis. Go ahead and run your analysis.
I see, the DC Drop Analysis was probably the wrong analysis for my problem.
Maybe you can help me out here. I got the problem that my loads, which are pulsed with PWM, influence my Power Plane.
As a result my sensors, which are powered with the same plane, deliver wrong measurements.
I believe I got some disturbance on my power plane, due to the on and off switching of my loads.
These disturbances directly influence the measurement of my sensors.
How can I simulate the impact of the on and off switching of my loads on the power plane? With Plane Noise Analysis?
I'm pretty new to the whole simulation world, so I would really appreciate your help here.
Yes, Plane Noise Analysis is the function that you want to use. You can assign AC models to the power pins of the of noisy device, assign models to decoupling capacitors and power supply pins, and then when you run the Plane Noise Analysis you observe the voltage noise at the pins of the sensor. From there you can try changing the decoupling capacitor values, but that probably won't have a significant impact on the noise at the sensor. You can also export the PDN nets to LineSim to experiment with moving or adding capacitors, or adjusting the design of the power and return nets.
This goes back to your initial question about assigning AC model to IC power pins. It looks like you need access to a more comprehensive PI license. You can coordinate with your license administrator and local Mentor AE to see if you need to adjust your environment to access an existing license or if you need to buy a different license.
Thanks Weston, your answer really helped me. I was able to do my Plane Noise Analysis. Although I was a bit surprised about the limited options for the configuration of the AC Models (e.g. Period time and pulse time are at max. 200ns). How would you suggest to model a load which is pulsed with 20kHz and therefore has a period time of 50us?
Or is this simulation simply not made for such low frequencies since an AC Model is meant to simulate an IC Power pin?
Is there any other way to model my load?
I expect that if you simulated the exact stimulus at the 50us period, you would simulate a whole lot of nothing. The noise should settle within 10's of nanoseconds, and then your simulation would show flat voltages. The purpose of the simulation is to study the interesting parts where noise happens. For the time when there is no noise, you can do a simple DC simulation.
If my assumptions are not correct then you might have a design that is outside the design space for which HyperLynx is optimized to work. Remember that every model is wrong, but some are useful. I try to find the useful model whenever I can.
Thanks for the fast response and the explanation. It makes sense to only watch the interesting parts (rising edge). I think I will restart the simulation and focus more on that.