Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Resistive voltage divider
Posted Apr 16, 2022, 3:54 a.m. EDT Electromagnetics Version 6.0 7 Replies
Please login with a confirmed email address before reporting spam
Hi!
I have received an mph file of the geometry of a voltage divider consisting of 10 high voltage resistors and 1 low voltage resistor all connected in series with each other. The goal is to add materials to all components, add ground and electrical potential limits, and then calculate the problem.
I'm very new to COMSOL and I do not know how to apply the electrical potential limit. The potential on the upper sphere of my divider should be 1000V. When I assign 1000 V as an electrical potential limit to the sphere and click calculate, I get zero voltage everywhere except the sphere.
how do you get the potential to go from the sphere down to earth and show me how stray capacitances look like between ground and each stage of the divider?
Note! I attached the mph-file if somebody want to check my potential boundary and give me some guidance. I deleted the mesh and the results of last run to compress the size of my file, you can just add a mesh and compute. All help is appreciated!
Attachments:
Please login with a confirmed email address before reporting spam
Tony,
I do not really understand your setup. You apply the 1000V potential on the top sphere and on the connections between the resistors. Shouldn't it only be applied to the sphere? You set up a time dependent study step without a time setting. Shouldn't this be a stationary study?
And in ec all domains must have at least a very small conductivity, also the air domain, otherwise the solver won't converge.
Cheers Edgar
-------------------Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
Please login with a confirmed email address before reporting spam
Hey Edgar, Thanks for your time and reply!
The truth is that I haven't designed niether the geometry nor the time dependency of the study but I am only told to add materials, boundaries and compute. My pure guess of why "time-dependent" is that I am asked to generate plots of the step and frequency response simulations of the divider which might need a time dependent study.
However, I applied 1000 V on the connections between the resistors because that made the solver converge in 2-3 seconds. When I only applied 1000 to the sphere, then the simulation took 30 mins to be completed and the results showed zero potential everywhere but the sphere. This delay could be due to the convergence problem that you have mentioned about the insulators having 0 condictivity.
Now, 1. I will remove the boundaries from the connections and introduce some condictivity to air and the plexiglass tube. 2. check the time setting 3. compute
Regards Tony
Please login with a confirmed email address before reporting spam
I recommend to do a stationary study first to check plausibility. A time dependent study may not converge without ramping up the potential BC.
-------------------Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
Please login with a confirmed email address before reporting spam
Hi Edgar,
I disabled the time dependent study and added a stationary study to my model. Still no electric potential is appearing in the potential plots. Could you please have a look at the attached model and give me some advice on how to proceed?
Regards Tony
Attachments:
Please login with a confirmed email address before reporting spam
There is no solver for the stationary step.
-------------------Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
Please login with a confirmed email address before reporting spam
Yeah, I completely missed that. The stationary case is now working well and I am getting a potential distribution as shown in the figure below. [drive.google.com/file/d/15IGwEW2AKxVwxwtsTgMC-dJ8bdOtWji-/view?usp=sharing]
For the time dependent study, I am using this function V =1000*(1-exp(-t/tau)) for my electric potential BC but the simulation is not converging. It took 30 mins to complete it and it is giving zero potential everywhere. Any idea why it is taking so long time ? I am not sure if the boundary conditions I have applied cause that kind of convergence problem. Could you please take a look at the attached model and see if the boundaries look good?
Attachments:
Please login with a confirmed email address before reporting spam
You don't have any output times in the time dependent step. If the stationary study converges the BCs should be ok. You will probably need to play with the time stepping to get the time dependent solver to run.
-------------------Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com