lumped components
If I would like to make a Wilkinson microstrip power divider, is it possible for MWO to somehow include the isolation resistor. Either in the EM-simulation orafterwards by incorporating the layout with the schematic.
I have done this in ADS and then applied internal ports where the resistor is placed. Afterwards the EM-simulation, it is possible to import the simulated layout into the schematic where I can use a lumped resistor.
Please advise
Regards
Full disclosure: I work for Sonnet.
We have had a long and fruitful relationship with AWR. I do not know about their EMSight for internal ports, however SonnetLite interfaces with Microwave Office very nicely. In fact it uses the same COM based interface (we actually helped AWR set it up). A couple menu picks and your layout is in Sonnet. Exit Sonnet after analysis and the results are back in MWO. Really nice.
You can get the free SonnetLite (www.sonnetsoftware.com) and you can use up to 4 ports and you can include one ideal internal component (in this case a resistor). If you use the full version of Sonnet (available at most universities, if you do not have it at your university, contact us immediately), you can include internal S-parameter blocks and/or co-calibrated ports. All this interfaces with MWO very easily. SonnetLite even includes the interface to MWO.
Actually, the same is true for ADS too. The ADS/Sonnet interface is free in SonnetLite too. If you use Sonnet's co-calibrated ports (only in the full version of Sonnet), something special happens. The Sonnet EM data automatically goes back to ADS as a layout-lookalike schematic element. You can send a layout to Sonnet from ADS schematic or from ADS layout as you like. Just a couple menu selections and you are there. I recommend all engineers learn both ADS and MWO. You will become a more valuable engineer. The interoperability both ADS and MWO have with 3rd parties, like Sonnet, make them especially valuable frameworks. I recommend avoiding any frameworks that lack that kind of interoperability because they limit the solutions available to you.
Now, what is really special about Sonnet's internal ports (or, as we call them, "co-calibrated ports"), is that they are perfectly calibrated (i.e., to within numerical precision, provided your port connecting lines are not over-moded). No kidding. No one else can do that. The mathematics was quite an effort, but it has all been published. This can be done only in a shielded analysis. It can not be done in an unshielded analysis.
Perfect calibration is not an issue for low frequency work. But it is really important at higher frequencies. I can discuss this more if anyone is interested.
[quote="rautio"]Full disclosure: I work for Sonnet.
Hi rautio, since u said u have a fruitful experience with mwo, may i know whether u know about using mwo to design filter?Is it it must be in circuit schematic rather than in EM structure?
I am using mwo 2006 to design my final year project and having a bit problems because not familiar with the software and there is no anybody can help me here.
I usually use Sonnet with MWO as the front end. Make up a schematic and layout of your divider and verify it works using standard microstrip elements. Try to choose the line widths and and spacings so that everything falls on grid, and has at least 3 grid units over the width of the line (or more if you have curves, which Sonnet can handle better). If the geometry is set up nicely and snapped to an EM grid then you can use an "EM Extraction" (see the manual) to analyze the circuit. It will add edge ports for the I/O and via ports where the resistor attaches. You can then analyze this in either MWO or Sonnet. You may right-click on the circuit and choose "add extraction" to preview the EM extraction before simulation, which you should do to verify the ports and geometry look good.The MWO via ports are not de-embedded (the via inductance is not removed from the calculation) so you may want to EM simulate just the via then de-embed this manually to see if it affects your result. I believe the Sonnet via port does de-embed the inductance but maybe Dr. Rautio can verify that. The Sonnet co-calibrated port can probably be used by "editing in native editor" when using MWO as the front-end but I have not verified that yet.
You can subsection your circuit by assigning elements to different extraction groups the same way you would split up your circuit with subsectioning in Sonnet. This is handy since the junctions are probably more critical than the lines (unless they are tight and curvy) and simulating the whole circuit may be excessive.
Check out these "webinars" for more info on using MWO and doing EM extraction. Pretty much everything I mentioned about MWO is clearly explained here. It will take several hours to watch them but they are well worth it.:
http://web.appwave.com/News_&_Events...d_Webinars.php
Also make sure you have a good resistor model that includes the pads. Read appnotes from Modelithics for the full story. The shunt capacitance can burn you when making a Wilksinson. However you can come up with a rough lumped model given pad and resistor dimensions to see if it affects your circuit.
http://www.modelithics.com/
I think the "internal port" in EMSight of MWOffice is not calibrated for good reason: there is simply no way to ensure or confirm that there is only one "mode" propagating across an "internal port". Most of the times, there is no well defined "mode(s)" at an internal port, how can we determine it is not over-moded and can be "perfectly" calibrated?
Hi Guys -- Since no other EM analysis in this universe has co-calibrated ports, you do have to go into Sonnet's editor, "native editior" in AWR MWO, then add the co-cal ports in Sonnet, then you terminate and go back to AWR and launch a Sonnet analysis, or stay in Sonnet and launch a Sonnet analysis. Co-cal ports are something completely new in microwave design, so be sure to read the Sonnet manual about how to use them. It would be a good idea to learn about co-cal ports because they are going to completely change how we do microwave design.
To be sure you are not overmoded on the port connecting lines, make sure there is no radiation, and make sure the port connecting lines are << 1/2 wavelength in width, and make sure the ground return current is << 1/4 wavelength away from the signal line. This is just good engineering design practice. If the co-cal ports fail, it is likely that your design will fail too. One area that is still a little difficult is if you have multiple levels and there is circuitry on levels immediately above or below the desired co-cal ports, but we can work that situation too.
Groups of co-cal port should not be too close together either. This is because fringing fields between groups is not removed. Typically, this takes the form of mutual inductance between each group's automatically added ground vias. A way around this is to simply use a floating ground for each group. Again, this is a brand new concept in microwave design that has never before been available. Read the Sonnet manual about it. This is a very powerful technique and is especially useful in things like Si RFIC and LTCC where "ground" is not obvious.
Loucy -- I had similar feelings to what you describe when I started working on calibrating internal ports. That is part of why it took me a full 30 days of total immersion (i.e., I literally did nothing but eat, sleep, and work the problem, day and night) to find a solution. But the problem is solved and you can read all about it:
Deembedding the effect of a local ground plane in electromagnetic analysis
Rautio, J.C.;
Microwave Theory and Techniques, IEEE Transactions on
Volume 53, Issue 2, Feb 2005 Page(s):770 - 776
Digital Object Identifier 10.1109/TMTT.2004.840576
To see that it works, and it works really really well, read:
EM-Component-Based Design of Planar Circuits
Rautio, J.C.;
Microwave Magazine, IEEE
Volume 8, Issue 4, Aug. 2007 Page(s):79 - 90
Digital Object Identifier 10.1109/MMW.2007.383305
And my latest pride and joy, which absolutely requires perfect port calibration, and introduces a brand new area of compact modeling (it works almost like magic!), read:
Synthesis of Compact Lumped Models From Electromagnetic Analysis Results
Rautio, J. C.;
Microwave Theory and Techniques, IEEE Transactions on
Volume 55, Issue 12, Part 1, Dec. 2007 Page(s):2548 - 2554
Digital Object Identifier 10.1109/TMTT.2007.909141
You can get these on IEEE Xplore, or email me and I will reply with pdfs of the desired papers. I am not posting the papers because they are copyrighted.
Hi rautio,
May i know how to get the student full version of Sonnet?I need it to do my final year project.
Thanks,
Chai Ling
Hi Rautio,
I am very interested in the articles you mention. Could you email me these at: tyassin@live.com
regards
Hi Tyassin -- Three papers emailed in three emails.
Hi Chailingleow -- Send an email to sales at sonnetsoftware.com. There is a one time cost, for a full university copy, but if you have a special situation, let us know. Mention that you talked to me on EDAboard. -- Jim Rautio
Hi Rautio,
There is not much students here that design antenna.They dont require to use sonnet.I am just the only one interested in it because i cant simulate my filter design in mwo because it consume hours and sonnet lite dun have enough features to simulate it.So i just request that i want to use the full trial version for 30 days.Thanks for your offer.
Regards,
Chai Ling
Sounds good, Chai. Please send an email to sales(at)sonnetsoftware.com. Sonnet and MWO Emsight are based on the same theory. Sonnet is much more advanced because we do EM only, and we have done that for 25 years. So, Sonnet should be faster, but if it is not fast enough, there are many things that can be done to get faster results. I suggest you make a posting on our forum, at www.sonnetsoftware.com if you need further help.
Hi Dr Rautio,
email to sonnectsoftware?
No need to pay?This all must wait till i get the evaluation copy then I will carry it on
After i check it out, just only i knew ur positions :p
Very happy that i found this forum here, throughout my studies there is rarely any peoples that can help me around here. But i really get some helping hand here in EDAboard :D
Use Agilent Momentum for accurate resuts.& try it on your design.
Hi Dr. Rautio,
I use sonnet lite for microwave planar designs. I have seen some of the papers of yours regarding co-calibrated ports. May you send these papers to me.
Deembedding the effect of a local ground plane in electromagnetic analysis
Rautio, J.C.;
Microwave Theory and Techniques, IEEE Transactions on
Volume 53, Issue 2, Feb 2005 Page(s):770 - 776
Digital Object Identifier 10.1109/TMTT.2004.840576
To see that it works, and it works really really well, read:
EM-Component-Based Design of Planar Circuits
Rautio, J.C.;
Microwave Magazine, IEEE
Volume 8, Issue 4, Aug. 2007 Page(s):79 - 90
Digital Object Identifier 10.1109/MMW.2007.383305
Synthesis of Compact Lumped Models From Electromagnetic Analysis Results
Rautio, J. C.;
Microwave Theory and Techniques, IEEE Transactions on
Volume 55, Issue 12, Part 1, Dec. 2007 Page(s):2548 - 2554
Digital Object Identifier 10.1109/TMTT.2007.909141
Sonnet released V12.52 this past April and one of the new features in the Sonnet AWR Microwave Office Interface is to automatically convert the via ports created with em extraction into co-calibrated ports. At this point, the ports are ungrouped but they are much better than using via ports when connecting a component.
Mike