Learning Forum

[Saransh]

Hi All,

I'm trying to do a very simple electrical simulation in Maxwell which is shown in the picture attachment.

This is a simple copper bar (10um long, 1um wide and 0.6um thick) with 1mA current being sent from the top and collected from the bottom. The cavity in the middle is created so that there can be identical currents (500uA) on both the sides of the cavity. When I simulate this in Maxwell, I never get identical currents when I integrate over the metal area on both the sides of the cavity. It varies a lot in uA region where I get 504uA on one side and 500uA on the other (these are typical numbers to show 4uA variation). As you see, this doesn't even add up 1mA!

For coarse simulations, this might have been okay. But the task I am dealing with requires me to observe changes in currents in these two halves in the presence of external magnetic field (. The B would cause slightly more current in 1 half and less in the other, as per the Hall effect. This current difference would be of the order of nA if B is mT range. There is no way I can observe nA current deflection if I have uA current inconsistency over the two sides.

I have tried making the mesh finer (100nm max element size) and reducing the error percentage in simulation setup to 0.0001 but despite that, I get uA range difference. This is becoming a serious problem for me and I would really appreciate any help on how to remove this inconsistency.

The way I am observing the current is by integrating the current density (J) over the two side planes separately and getting their I value. I don't know if this is the best way of doing it or whether this might be causing the inconsistency.

Any help would be very much appreciated!

 

[mbeliaev]

Hello,

 

I think this exceeds software limitation. What I may suggest to try is to scale current and external field in your design and then scale back your output to nA range

 

--

Maksim

[Saransh]

Hi Maksim,

Thank you for your response. I tried this and got the following results:

Before increasing the current and field, I made a slight modification to the design - I removed the cavity from the bottom which was creating the two separate terminals for current. Instead, I just have a metal plate now (literally a rectangular plate!) in which I send current from the top and integrate the current densities on the bottom - this integration I do for the left half of bottom and right half so the two surface integrals give me the received current in both the halves.   

With this new very simple rectangular plate, I get absolutely identical current in the two halves up to the 15th decimal place, meaning fA range (when the DC current in plate is uA range). This occasionally changes to nA range when I simulate more, all in absence of magnetic field. 

Now, as per your suggestion, I changed the field to very high - around 5T in the plate region. And the current in the plate is 1A, again very high. A very surprising thing happens now - I get these two currents in the two halves: (Right) 0.487500000000004, (Left) 0.487499999999999

This is very surprising for me because it seems the two currents are very very identical! This can not happen if the Hall effect manifests at all. I get exactly the same current distribution in absence of the magnetic field which essentially means that the crazy high 5T field has no change on current distribution!

This makes me doubt whether Maxwell can even simulate the change of current in the presence of external magnetic field? Can it simulate for the Lorentz force which acts on the charge carriers when they're exposed to external magnetic field? Somebody please clarify!

P.S: I am using the magnetostatic mode of solution, I wonder if that has got something to do with the lack of Lorentz force's effect on the current.

 

[mbeliaev]

can you please send me your model and I will have a look

[Saransh]

Hi Maksim,

Thank you for your kind offer to look into the file. I have uploaded my complete Maxwell simulation file on google drive and here is the link:

https://drive.google.com/file/d/1FTnODPbUV2YMpCYPNx3zCalYuQ8K6o9a/view?usp=sharing

When you download and run it, you'll want to see these two in the results section from calculator:  

Integrate(Surface(Left_Section1), Mag_J)
Integrate(Surface(Right_Section1), Mag_J)

These are the two current values that should have a difference in presence of magnetic field.

Please contact me over my email (ssharma3@caltech.edu) in case of any clarification. Thank you very much for your kind help!

[Saransh]

 Hi Maksim,

 

Looking forward to hearing from you soon.

 

Thank you.

[Saransh]

I have tried both Magnetostatic and Transient simulations but neither provides any Hall current deflection, no matter how high the applied magnetic field or DC bias current.

Seems like Maxwell can not simulate for Lorentz force, which is quite surprising since I thought it uses Maxwell's Electromagnetic equations which very well take into account this type of force. 

I have no option but to get rid of our Maxwell license since I don't seem to be getting any help from any community member.