Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

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.

simulate temperature rise due to current flow in copper track

Please login with a confirmed email address before reporting spam

Hi all

I am new to COMSOL and I would like to simulate the temperature rised of a short and straight copper track on PCB due to current flow. How should I start with? I have a fixed voltage and current to apply on it.

Does using solely the joule heating physics sufficient to simulate this? I can see that there is electric current -> electric potential for me to enter the voltage but no where to enter the current. How am I suppose to simulate this?

Thanks
Ron

9 Replies Last Post 15.02.2012, 02:00 GMT-5

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 10.02.2012, 13:20 GMT-5
Try this as an example model file

www.comsol.com/showroom/gallery/8577/


It will at least get you started.
Try this as an example model file http://www.comsol.com/showroom/gallery/8577/ It will at least get you started.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 11.02.2012, 07:34 GMT-5
I have come to here so far and could not get it work any further. I would like to apply 0.5V and 4.15A to the copper. How can I continue it? Attached is the model.

Thanks
Ron
I have come to here so far and could not get it work any further. I would like to apply 0.5V and 4.15A to the copper. How can I continue it? Attached is the model. Thanks Ron

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 11.02.2012, 08:09 GMT-5
Hi

Your model is slightly strange ;) first of all your air is underneath the copper ?, this is probably not correct, I would have not aded any "air" and replace it by a surface convective heat loss of some 2-4 [W/m^2/K].

Have you noticed you have 3 domains, as your Cu block and FR4 blocs intersects

Then if you define the resistivity (in particular with respect to T variantions) you cannot impose both current and voltage, you fix one typically a gnd and a voltage input region, or a current density boundary for a cte I supply, and the other is being calculated from Ohms law.

And the the FR4 is not really heating by joule effect so you might as well exclude it by defining a linear elastic material node

--
Good luck
Ivar
Hi Your model is slightly strange ;) first of all your air is underneath the copper ?, this is probably not correct, I would have not aded any "air" and replace it by a surface convective heat loss of some 2-4 [W/m^2/K]. Have you noticed you have 3 domains, as your Cu block and FR4 blocs intersects Then if you define the resistivity (in particular with respect to T variantions) you cannot impose both current and voltage, you fix one typically a gnd and a voltage input region, or a current density boundary for a cte I supply, and the other is being calculated from Ohms law. And the the FR4 is not really heating by joule effect so you might as well exclude it by defining a linear elastic material node -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 11.02.2012, 11:47 GMT-5
Hi Ivar

I am not too sure what you meant by only one source can be applied but anyhow, I just try my best to make the changes according to your advice and I have got it doing some simulation. However it is not something that I am expecting as usually the centre region of a 2 ends conductor will be the hottest part and the result gives such a high temperature which is quite abnormal. Of course, I understand that it is due to the boundary conditions that I have selected. Can you please point out which part have I made a mistake or need to be changed?

Thank you
Ron
Hi Ivar I am not too sure what you meant by only one source can be applied but anyhow, I just try my best to make the changes according to your advice and I have got it doing some simulation. However it is not something that I am expecting as usually the centre region of a 2 ends conductor will be the hottest part and the result gives such a high temperature which is quite abnormal. Of course, I understand that it is due to the boundary conditions that I have selected. Can you please point out which part have I made a mistake or need to be changed? Thank you Ron

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 13.02.2012, 01:49 GMT-5
Hi

you model is still somewhat weird ;)

1) get your two volumes to be in contact over a common boundary, now your copper is floating above the FR4, that is certainly not "physical, as no contact means no heat exchange use Z position offset 1.6175 mm
2) remove the Boundary Heat noce, your joule heating will heat the copper via the electromagnetic power set in
3) Add a Heat transfer in Solid node and select your FR4, that one si not conducting electricity hence will not heat up by Joule heating but by conduction from the Copper.
4) set a convecting cooling on the outside surface of the FR4 (top) and the Copper, but not the copper-FR4 interface (and you can forget the GND and Normal current density boundaries
5) then a current density of 467 A/mm^2 is very high, normal wire power leads are rated to max < 10 A/mm^2 So I'm not astonished your copper is fusing off
--
Good luck
Ivar
Hi you model is still somewhat weird ;) 1) get your two volumes to be in contact over a common boundary, now your copper is floating above the FR4, that is certainly not "physical, as no contact means no heat exchange use Z position offset 1.6175 mm 2) remove the Boundary Heat noce, your joule heating will heat the copper via the electromagnetic power set in 3) Add a Heat transfer in Solid node and select your FR4, that one si not conducting electricity hence will not heat up by Joule heating but by conduction from the Copper. 4) set a convecting cooling on the outside surface of the FR4 (top) and the Copper, but not the copper-FR4 interface (and you can forget the GND and Normal current density boundaries 5) then a current density of 467 A/mm^2 is very high, normal wire power leads are rated to max < 10 A/mm^2 So I'm not astonished your copper is fusing off -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 13.02.2012, 06:24 GMT-5
Hi Ivar

Thanks for your guidance. Howeer, I still have a bit questions. Why do you say the boundary conditions of GND and Current can be ignored? I thought they are needed as I am applying a constant current flow through the copper. Also, how is the current density determined? I just divide the current that I used over the cross sectional area of the copper. Is it wrong?

Thanks
Ron
Hi Ivar Thanks for your guidance. Howeer, I still have a bit questions. Why do you say the boundary conditions of GND and Current can be ignored? I thought they are needed as I am applying a constant current flow through the copper. Also, how is the current density determined? I just divide the current that I used over the cross sectional area of the copper. Is it wrong? Thanks Ron

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 13.02.2012, 08:12 GMT-5
Hi

it was for the convective loss you could ignore the two small faces.

Are you sure you have such a high current density ? that you didnt miss something. You can use COMSOL to calculate the area or give the value directly in current Amps. But I doubt you get 500A/mm^2 in DC through such a small Cu lead without some better cooling

--
Good luck
Ivar
Hi it was for the convective loss you could ignore the two small faces. Are you sure you have such a high current density ? that you didnt miss something. You can use COMSOL to calculate the area or give the value directly in current Amps. But I doubt you get 500A/mm^2 in DC through such a small Cu lead without some better cooling -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 14.02.2012, 17:33 GMT-5
Hi

I have come to this so far but something strange is I could not set the same boundary condition for both ends of the track. For example, I had added a through hole pad at the end with copper (wire) in it and solder on top. I can add convective cooling for only one solder but not both. Any reason?

Thanks
Ron
Hi I have come to this so far but something strange is I could not set the same boundary condition for both ends of the track. For example, I had added a through hole pad at the end with copper (wire) in it and solder on top. I can add convective cooling for only one solder but not both. Any reason? Thanks Ron

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 15.02.2012, 02:00 GMT-5
Hi

You have slightly too many BCs defined here ans a few are incompatible.

First of all ,with EC you are solving for the voltage and as U[V]=R[Ohm]*I[A] you define a source I[A] and a geometry with a material and a given resistivity + an input and output current flow boundary, which means you have defined R[ohm], or at least all what defines with some calculations the value R(x,y,z,t). Then you leave COMSOL resolve the voltage U.
Which means you can remove the initial condition V=0.13V applied everywhere
Then you define the input and output (GND) current boundaries. I would restrict myself to one boundary lets say 15 the lower solder surface, and 55 for the Terminal (it could be the top solder surfaces 12 and 45 , but certainly not all the pad surfaces

You have also a RED warning in the material section for your solder. The epsilon r is missing, I would propose to use "1" its a rather safe value if you do not have a known one for your material

Then something is wrong in the geometry not 100% sure what, but you could try to add a last "union" and select all Cu objects and combine them by unchecking both Keep input objects AND Keep interiour boundaries. by the way you canb also check the "Create Selection" and then change the name to "Union Cu", this will add a selection items allowing you to select the copper as 1 item in a quick and handy way (check the dynamic help "Union Selection"

You are using an external natural convection for your boundary, this gives a reasonable temperature (260°C) but corresponds to some h=14 W/m^2/K on average, up to 22 onthe hot parts (3D surface plot "jh.ccflux/(T-20[degC])"), a rather high value, for me, for natural convection, but might be correct, anyhow worth some simple validation and tests. You could add another heat convection loss on the lower side of the Fr4 plate, this would further reduce the temperature by some 40°C

--
Good luck
Ivar
Hi You have slightly too many BCs defined here ans a few are incompatible. First of all ,with EC you are solving for the voltage and as U[V]=R[Ohm]*I[A] you define a source I[A] and a geometry with a material and a given resistivity + an input and output current flow boundary, which means you have defined R[ohm], or at least all what defines with some calculations the value R(x,y,z,t). Then you leave COMSOL resolve the voltage U. Which means you can remove the initial condition V=0.13V applied everywhere Then you define the input and output (GND) current boundaries. I would restrict myself to one boundary lets say 15 the lower solder surface, and 55 for the Terminal (it could be the top solder surfaces 12 and 45 , but certainly not all the pad surfaces You have also a RED warning in the material section for your solder. The epsilon r is missing, I would propose to use "1" its a rather safe value if you do not have a known one for your material Then something is wrong in the geometry not 100% sure what, but you could try to add a last "union" and select all Cu objects and combine them by unchecking both Keep input objects AND Keep interiour boundaries. by the way you canb also check the "Create Selection" and then change the name to "Union Cu", this will add a selection items allowing you to select the copper as 1 item in a quick and handy way (check the dynamic help "Union Selection" You are using an external natural convection for your boundary, this gives a reasonable temperature (260°C) but corresponds to some h=14 W/m^2/K on average, up to 22 onthe hot parts (3D surface plot "jh.ccflux/(T-20[degC])"), a rather high value, for me, for natural convection, but might be correct, anyhow worth some simple validation and tests. You could add another heat convection loss on the lower side of the Fr4 plate, this would further reduce the temperature by some 40°C -- Good luck Ivar

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.