Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronic Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

CST Simulation running too long ...

Status
Not open for further replies.

Pushhead

Full Member level 4
Joined
Jan 19, 2005
Messages
199
Helped
20
Reputation
40
Reaction score
10
Trophy points
1,298
Activity points
1,870
Hi All,

I'm trying to simulate a simple solenoid (Something like a verification of biot-savart law) with N turns etc. The purpose is to measure H inside the device.

The problem :
-----------------
I'm getting the following msg :
======================================================================
Solver cycle: 1
======================================================================

Stimulation at discrete port 1


--------------------------------------------------------------------------------
Units settings:

Dimensions: cm
Frequency: MHz
Time: s

--------------------------------------------------------------------------------
Mesh summary:

< 0.5 :0.6 %
< 0.25 :0.4 %
< 0.1 :0.1 %
< 0.05 :0.0 %
< 0.02 :0.0 %

Minimal filling: 3.201 %

--------------------------------------------------------------------------------
*** Error ***

The current settings will lead to an extremely long simulation time.

You may improve the simulation speed by:

1. checking the frequency range and units settings
2. extending the simulation frequency range
3. checking the mesh to avoid very small mesh steps

--------------------------------------------------------------------------------

The means i've taken :
a. I've arranged a reasonable mesh (honestly)
b. The frequency bandwidth is 400% of what should be according to the signal's FFT.
c. as you can c in the units, i'm simulating a large one - ~60cm diameter.

The result :
Simulation time is tooooooooooooo long !
the timestep is (ns*1e-6).... I need Help !

(BTW: It would b nice if someone has a solenoid model to spare... )

Desperatly,
P.
 

Kit-the-great

Full Member level 3
Joined
Mar 30, 2002
Messages
151
Helped
5
Reputation
10
Reaction score
3
Trophy points
1,298
Location
Russia
Activity points
1,318
I'm not understand, what size of Your structure in wavelengthes? You can effectively simulate only the electrically small structures. For structures such coils You can try to use the mesh subgridding, if You use the CST version 5 or later.

Regards,
Kit
 

    Pushhead

    points: 2
    Helpful Answer Positive Rating

fekete

Member level 5
Joined
Jan 20, 2005
Messages
92
Helped
13
Reputation
26
Reaction score
6
Trophy points
1,288
Activity points
1,488
It seems that you are simulating a low frequency inductor. When you say CST which one of their tools are you using ? MWS, EMS, Mafia?
In case you are using MWS, it is not the right tool for low frequencies (when you can neglect the displacement current) but the other two would be.
 
  • Like
Reactions: nvt088

    Pushhead

    points: 2
    Helpful Answer Positive Rating

    nvt088

    points: 2
    Helpful Answer Positive Rating

plasma

Advanced Member level 2
Joined
Dec 20, 2001
Messages
536
Helped
79
Reputation
158
Reaction score
40
Trophy points
1,308
Activity points
3,854
Hi Pushhead

If You are using non linear dimension element.
You get this list that what it means

1. Checking the frequency range and units settings (always do it)
2. Extending the simulation frequency range (for getting smaller cells No. of cell depend on Wave Length of High Freq.)
3. Checking the mesh to avoid very small mesh steps (you have very small mesh steps in the simulation)

If you have circle try to segment it instead of using perfect circle.
also diagonal lines with very low angle.
PL
 

    Pushhead

    points: 2
    Helpful Answer Positive Rating

Pushhead

Full Member level 4
Joined
Jan 19, 2005
Messages
199
Helped
20
Reputation
40
Reaction score
10
Trophy points
1,298
Activity points
1,870
Thanks alot Plasma,

I've done it even b4 I asked the forum's help.
actually the solenoid is "square-o-noid".

The interesting part is that i've succeeeded in reducing the calculation time X10 by inflating the surrounding box volume.

Can you explain what is the meaning of expanding the box, or why it reduced the calc. time ?
 
  • Like
Reactions: nvt088

    nvt088

    points: 2
    Helpful Answer Positive Rating

plasma

Advanced Member level 2
Joined
Dec 20, 2001
Messages
536
Helped
79
Reputation
158
Reaction score
40
Trophy points
1,308
Activity points
3,854
I think that when the box is smaller it's not absorb the all radiating
energy and it behave like resonator and that expand the time for the pulse
to decay.
So when you increase the volume more energy absorbs and
it took you less time.

RGZ


Because you are working with low freq. the mesh cell are big so check if you have less then 3 cells between structure and boundary.
 
  • Like
Reactions: nvt088

    nvt088

    points: 2
    Helpful Answer Positive Rating

cplia

Member level 2
Joined
Jan 26, 2005
Messages
47
Helped
9
Reputation
18
Reaction score
3
Trophy points
1,288
Location
Europe
Activity points
433
Also, check the value of your lower frequency range.
For low-Q structures, such as your antenna here, reducing
the Fmin can influence positively (reduce) the simulation time.

I realised that when i was simulating one of my antennas,
between 1.5-3 GHz, the simulation time reduced when i tried
0-3GHz. I didnt even have to adjust the size of the bounding
box at all !

Also, from my experience, the meshline ratio limit plays a
dominant role, even when you have low (e.g. 10) lines/wavelength
setting. For my lower frequency antennas, e.g 400MHz, i
retain 10-14lines/λ, but i increase a lot the meshline ratio. This, i think,
can improve the accuracy in the meshing, without increasing
number of meshnodes in an exponential manner.

Let us know your outcome!
 
  • Like
Reactions: nvt088

    nvt088

    points: 2
    Helpful Answer Positive Rating

nvt088

Member level 5
Joined
Dec 24, 2013
Messages
81
Helped
10
Reputation
20
Reaction score
10
Trophy points
8
Location
Jupiter
Activity points
521
It seems that you are simulating a low frequency inductor. When you say CST which one of their tools are you using ? MWS, EMS, Mafia?
In case you are using MWS, it is not the right tool for low frequencies (when you can neglect the displacement current) but the other two would be.
Hi. Now I am simulating a Wireless Power Transfer System with CST 2013.

Working frequency is 6.7 MHz (wavelength=50m). What module in CST should I use? MWS, EMS...???
The size of coil is 60cm. What is the size of air box should I select? Because when I simulated, size of air box impacted alot on resonant frequency of system.
Many thanks

- - - Updated - - -

Also, check the value of your lower frequency range.
For low-Q structures, such as your antenna here, reducing
the Fmin can influence positively (reduce) the simulation time.

I realised that when i was simulating one of my antennas,
between 1.5-3 GHz, the simulation time reduced when i tried
0-3GHz. I didnt even have to adjust the size of the bounding
box at all !

Also, from my experience, the meshline ratio limit plays a
dominant role, even when you have low (e.g. 10) lines/wavelength
setting. For my lower frequency antennas, e.g 400MHz, i
retain 10-14lines/λ, but i increase a lot the meshline ratio. This, i think,
can improve the accuracy in the meshing, without increasing
number of meshnodes in an exponential manner.

Let us know your outcome!
I don't understand your idea. Please can you explain about the role of meshline ratio limit and lines/wavelength?
Many thanks!
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Top