Learning Forum

[Dracarys99]

I want to do EMI EMC analysis on my PCB of a ZCS buck converter. I am using SIWave for this. When I have to select current and voltage sources do I have to mention all the components using the voltage (say +18V) as a voltage source or just the pin header that is supplying it? Let's say I have a 18 V to 5 V regulator ic. What should I assign this IC as? Also, does PWM generator ic like TI's SG3524 be assigned a current source?n

[manny]

I'm going to provide a high-level answer to your question that should get you up and running. First off this is an advanced simulation that you are after. The workflow for conducted emissions analysis of a buck converter is to start with SIwave, export to Q3D, build the circuit, run the analysis and view the emissions results. In SIwave, set up pin groups from the components under ICs and I/Os. Assign Sources/Sinks. The sink/source assignments can be done on the Export to Q3D dialog. In general, power consuming active devices are set to current sinks. VRMs are set to voltage sources. The export to Q3D generates a broadband model. Now set up a circuit simulation where you can assign SPICE/IBIS models or generic sources and generate the waveform. Good luck with your simulation! n

[Dracarys99]

Hello Array , thank you for your response. I have many doubts, but I'll ask a few first. I thought conducted and radiated emissions were the Near and Far field solvers in SIwave. It appears not from your answer, so what are those for? When you say pin groups, do you mean this : Say I have a PWM IC SG3524 on my board. So from the create pin groups option I choose the corresponding RefDes and select all nets and then hit create pin groups ? I am not clear about the purpose of this, I think it is to have a better analysis once I have assigned current and voltage sources. nAlso, could you suggest/tell me what kind of a circuit have I got to set up for EMI analysis? Do you mean something like the LISN model? Sorry if this question sounds trivial, I have just begun venturing in this domain ! nThank you, your answer has been very helpful.n

[manny]

As I mentioned to you before this is an advanced simulation. No, this is not a trivial question You're absolutely right - you'll need to use the LISN model in your circuit design. nYou are right about the pin groups. Assume you select an I/O device such as a controller. Selecting the controller from the Components tree in SIwave will select all the associated pins. Use the Create/Manage Pin Groups... under the Tools menu. Choose the part name and all the ref designators. And select all nets and create Pin Group(s). n

[Dracarys99]

ArrayYour comments have been really helpful, it is starting to make some sense to me. I've proceeded a bit since then, and I have some questions:-n1) When I simulate any solver, (I started off with the simplest (seemingly?) DC IR drop analysis,) I get Capacitors ignored, Inductors ignored resistors ignored messages. How do I correct this? Is this because it cannot map the capacitors etc. to the actual models? n2) Can I put any of the current sinks' GND terminal as the Ideal Ground Node (0V) which comes up just before simulating DC IR? n3) When I try to compute resonant modes for my 2-layer PCB of a soft-switching buck converter, I get an error Could not compute resonant mode voltage plots. Could not locate requested layer pair in mesh file. Then I checked my layer stack up and I found that it shows the elevation of the bottom layer = 0mm. I suspect this is why I cannot get the voltage differences between top and bottom layers. How do I correct this? Is it more to do with my PCB design than in SIwave? n4) Are Header pins also supposed to be assigned a voltage source apart from regulator ICs?nThank you very much.n

[manny]

1) The passive devices are not attached to the nets that you are simulating in your DCIR analysis so SIwave generates warnings. These are nothing to worry about. I'll get back to your other questions later.n

[Dracarys99]

nCan someone please clear my doubts? Thank youn

[manny]

Hi Dracarys99,nQ2. Yes. You can choose any source's negative reference node, and it will be used as the ground reference for voltage plots.nQ3. Yes, it is more to do with your PCB design. Please check that your 2-layer printed circuit board has both a power and ground plane. This is because you need to have at least two reference planes to compute the resonant modes.nQ4. I don't have an answer because I don't know what a header pin is. nHope this helps.nBest,nMannynn

[Dracarys99]

ArrayBy header pin I mean a pin header or an electrical connector. I've used 2 of them in my board to get external power supplies of 18 V and 24 V, so they should also be assigned a voltage source? I'm confused since the power source is external (i.e not present on the board).nAlso, I have two 2-layer PCBs which were actually bolted together to make a 4-layer PCB (crude method). Now the PCB files are thus two separate ones for the two boards. Is there any way I could combine the two to study the actual fabricated 4-layer board in SIwave? I have the Gerber files and I know the components, so if I can somehow assign the footprints to the components in SIwave, then I can do my analysis.nnThank you.n

[manny]

Hi Dracary,nYes, you can assign a voltage source. Please watch this video on DCIR there is a similar connector to supply power from off board - to that we assigned a voltage source. The full video series is here.nWe can combine PCBs to make a multi-zone PCB in SIwave. Here's a video on merging PCBs in SIwave.nBest,nMannynn

[Dracarys99]

nThank you for your reply.n1) I tried to export to Q3D in SIWave 2017.2. The first time I clicked on this option I got to the Ansys Electronics Desktop - Nexxim simulator automatically. But now for a different board, when I click this option it takes me to Q3D modeler and not the NExxim environment where I can see ports, and make my circuit for LISN and consequently EMI analysis. How do I get to the Nexxim software?n2) For this board, I have an AGND polygon pour in the bottom left corner. But this is not being recognized in the DC-IR analysis, since I can see that there is no triangular mesh here. Why and how can I rectify this?nn

[Dracarys99]

n

[manny]

Hi Dracarys99,n It's possible that you missed to select the Invoke Schematic upon Completion option in the Q3D export dialog on SIwave. Be sure to also select the AC Inductance/Resistance checkbox at the bottom of the dialog to avail the option of invoking the schematic. Please note in Electronics Desktop you can insert a circuit design type, create your circuit schematic and use the Nexxim solver once you configure your setup. To use a circuit schematic select Insert Circuit Design from the Project menu of the Electronics Desktop.nDC current and voltage drop analysis performs adaptive mesh refinement. In your design, I think the reason AGND is not being recognized by the mesher is because it is possible that AGND is not in the conductive path or in the return current path from the voltage sources to the current sinks in the terminal components.nBest,nMannyn

[Dracarys99]

Array I have an asymmetric half bridge converter circuit PCB which I need to do DC-IR analysis on. The power supply is external to the board (300 V). The board has mainly IGBTs and diodes for the power circuit, and Diode snubbers and switch snubbers which seem to be the relatively large power consuming parts.nIn this case, should I be assigning the pin header / board connector through which power supply is coming as a voltage source of 300 V? nHow do I assign current sinks in this case? There are some current sensor ICs, but they consume negligible power I think compared to snubbers.nAlso, in this board (2 layer) there is no GND plane at all, it is all routed. And my initial (though wrong) DC-IR analysis looks largely mesh-less, like the bottom left corner of the PCB I posted in this thread earlier. It's going to be like that for J and I in DC-IR but what about Power P?.Array Array Array Array n

[Dracarys99]

Hello,nIn addition to the doubts in my previous comment I have another problem. Here is the board in DipTrace, where there are 8 pads on the left side of the board marked B+ G+ R+ Y+ etc. nBut when I am converting this into ODB++ format and importing it into SIWave, I do not see these left side pads.nThese are pads that will be used to merge another board with this one for analysing it altogether. How do I correct this?.Also, this is basically the lower 2 layers of a 4 layer PCB (which has been made as a 2+2 layer board). In this case will I be able to simulate using REsonant Mode solver? None of the individual boards have a power plane it has been done by traces.n

[Dracarys99]

Kindly help with the above queriesn

[Dracarys99]

n