[HFSS] Inconsistent Results
时间：03-31
整理：3721RD
点击：
So, I have been using HFSS to simulate resonating circuits for about a month now. However, I am starting to check my old calculations and I have come across some rather troubling issues.
The main problem is that it seems running the same simulation multiple times will give me radically different results. The resonance will be off by up to 40% of its previous value, and the quality factors change by up to a factor of 5.
I have tried making stricter converging requirements, but that doesnt seem to help. In fact, if I create 2 different setups with the exact same convergence criteria and adaptive sweeps, I will still get different results between the two.
Does anybody know of what would likely be causing this?
Note- To analyze the resonance frequencies of my circuits, I have been using entirely discrete frequency sweeps where the fields are being saved. I start with a search between 1 and 4 GHz, then find a blip in the s-parameter plot. I then zoom in slowly to find the resonance frequency to 6 decimal points (The quality factors are on the order of 50,000). Now usually, I do not create a new setup everytime I zoom in, but rather just change the numbers in the setup to match where I should focus. I usually end up using a previous point as the new adaptive frequency (I constantly change the adaptive frequency to the max value of each frequency sweep)? Could this be creating a problem?
I am also curious about the meshing. Could it be changing each time?
The main problem is that it seems running the same simulation multiple times will give me radically different results. The resonance will be off by up to 40% of its previous value, and the quality factors change by up to a factor of 5.
I have tried making stricter converging requirements, but that doesnt seem to help. In fact, if I create 2 different setups with the exact same convergence criteria and adaptive sweeps, I will still get different results between the two.
Does anybody know of what would likely be causing this?
Note- To analyze the resonance frequencies of my circuits, I have been using entirely discrete frequency sweeps where the fields are being saved. I start with a search between 1 and 4 GHz, then find a blip in the s-parameter plot. I then zoom in slowly to find the resonance frequency to 6 decimal points (The quality factors are on the order of 50,000). Now usually, I do not create a new setup everytime I zoom in, but rather just change the numbers in the setup to match where I should focus. I usually end up using a previous point as the new adaptive frequency (I constantly change the adaptive frequency to the max value of each frequency sweep)? Could this be creating a problem?
I am also curious about the meshing. Could it be changing each time?