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.

SAW Device Impulse Response in the Time Domain

Please login with a confirmed email address before reporting spam

Hi,

I have built a 2D model of a SAW delay line device in the time domain, where I set up a voltage pulse on one set of IDTs, which creates a surface wave that travels across the surface and is detected by the other set of IDTs. This works and gives me the results that I expect with my set up. I designed this for a frequency of 20MHz and had placed the IDTs reasonable close together. When looking at the potential plot over time, I can see the surface wave travelling across the surface.

I then took the same model (copied it) and adjust it for a different frequency, 5MHz which alters the dimensions. I built the model in a way where the dimensions of the domains were dependant on the frequency so they adjusted automatically with one change. Additionally, I change one of the dimension to have a greater distance between the IDTs.

This updated model now reponds differently, instead of a surface wave moving across the boundary of the domain, I now have bulk wave that travels through the bulk of the domain and then reflects causing a different response at my output IDTs.

Currently I'm thinking perhaps it could be due to the meshing, as I think I am using the same mesh parameters for the both, and for a 5 MHz design this is probably finer than needed. Resulting in a greater simulation time.


3 Replies Last Post Apr 21, 2019, 10:32 p.m. EDT

Please login with a confirmed email address before reporting spam

Posted: 5 years ago Nov 27, 2018, 6:25 a.m. EST

Just an update, I adjusted the meshing which improved the simulation time, ~17 hrs to ~3.5 hrs. The meshing I had not set up to adjust with updates of the models dimensions.

A thought a collegue of mine had was that COMSOL is solving for the more efficient modes of vibration, and with my increase in seperation between the input and output IDTS, the surface wave is now seen as less efficient than the bulk wave as the thickness of the material is considerably less than the distance the surface wave would travel. So in an attempt to remedy this, I am trying with an increased substrate thickness, comparing with my previous model's IDTs seperation vs substrate thickness ratio.

Just an update, I adjusted the meshing which improved the simulation time, ~17 hrs to ~3.5 hrs. The meshing I had not set up to adjust with updates of the models dimensions. A thought a collegue of mine had was that COMSOL is solving for the more efficient modes of vibration, and with my increase in seperation between the input and output IDTS, the surface wave is now seen as less efficient than the bulk wave as the thickness of the material is considerably less than the distance the surface wave would travel. So in an attempt to remedy this, I am trying with an increased substrate thickness, comparing with my previous model's IDTs seperation vs substrate thickness ratio.

Please login with a confirmed email address before reporting spam

Posted: 5 years ago Nov 27, 2018, 10:46 a.m. EST

Update: Increasing the thickness of the substrate layer appears to improve my results for the suface wave propagation. An FFT of the output signal shows clearer peaks at the frequencies I expect.

My thought is that this thickness increase, minimises the bulk reflected waves produced in the simulation and/or at least increases the propagation distance they need to travel through the bulk of the material and so cause less interference.

I am running models with another increase in thickness, as I think this should produce a clearer output surface wave produced signal.

Another question this has raised for me is that differenct cuts of piezoelectic material have different properties, some more prone to bulk wave propagation while others are less so, does COMSOL take this into account with its predefined material library?

I am only trying 2D models in an effort to save time, and have difficulty understanding how to orientate my material for the correct cut in 2D, I can do it 3D based on the Euler angles, but the 2D rotation set up is different.

Update: Increasing the thickness of the substrate layer appears to improve my results for the suface wave propagation. An FFT of the output signal shows clearer peaks at the frequencies I expect. My thought is that this thickness increase, minimises the bulk reflected waves produced in the simulation and/or at least increases the propagation distance they need to travel through the bulk of the material and so cause less interference. I am running models with another increase in thickness, as I think this should produce a clearer output surface wave produced signal. Another question this has raised for me is that differenct cuts of piezoelectic material have different properties, some more prone to bulk wave propagation while others are less so, does COMSOL take this into account with its predefined material library? I am only trying 2D models in an effort to save time, and have difficulty understanding how to orientate my material for the correct cut in 2D, I can do it 3D based on the Euler angles, but the 2D rotation set up is different.

Please login with a confirmed email address before reporting spam

Posted: 5 years ago Apr 21, 2019, 10:32 p.m. EDT

Hi,David Czerski I'm a student,and studying the" SAW Device Impulse Response In The Time Domain".Can I have a look at this model? Thank you !

Hi,David Czerski I'm a student,and studying the" SAW Device Impulse Response In The Time Domain".Can I have a look at this model? Thank you !

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.