BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:US/Michigan BEGIN:DAYLIGHT DTSTART:20260308T030000 TZOFFSETFROM:-0500 TZOFFSETTO:-0400 RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3 TZNAME:EDT END:DAYLIGHT BEGIN:STANDARD DTSTART:20261101T010000 TZOFFSETFROM:-0400 TZOFFSETTO:-0500 RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11 TZNAME:EST END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20260428T002718Z UID:E5802E76-7755-4AD1-B37C-F5B3823346BC DTSTART;TZID=US/Michigan:20260423T173000 DTEND;TZID=US/Michigan:20260423T193000 DESCRIPTION:Pizza Sponsor: [Cornucopia Technical Sales](https://www.cornuco piatechnicalsales.com/)\n\nVenue Sponsor: [Apex Test Labs](https://www.ape xtestlabs.com/)\n\n[][]\n\nKristof P. von Czarnowski\n\nSpeaker Bio:\n\nKr istof P. von Czarnowski is an RF and hardware engineer specializing in hig h-frequency and mixed-signal system design for automotive applications. At Lear Corporation since 2022\, he leads transceiver design\, implementatio n\, and validation\, while also contributing to low- and high-voltage hard ware platforms with a focus on signal integrity\, power integrity\, and el ectromagnetic susceptibility and emissions.\n\nHe serves as an electromagn etic compatibility (EMC) simulation subject matter expert\, working to shi ft EMC from reactive post-test mitigation to proactive design integration. His work aims to embed EMC constraints directly into the development proc ess through design guidelines\, automated design rule checks\, simulation workflows\, and libraries - pushing for earlier issue detection\, reducing validation risk\, and improving first-pass success rates across schematic design\, PCB layout\, and compliance validation.\n\nKristof holds undergr aduate and graduate degrees in engineering from Oakland University\, where he specialized in electromagnetics and wireless systems.\n\nHis technical interests include practical EMI/EMC simulation workflows\, SI/PI co-desig n\, and model validation for RF and power distribution networks. His recen t work explores AI-assisted engineering tools and machine learning approac hes to automate component characterization and accelerate system-level EMC prediction - ad\n\ndressing the industry's missing-model bottleneck that prevents full system-level compliance prediction.\n\nAbstract\n\nEvery ele ctromagnetic compatibility problem begins with Maxwell's equations. Yet th e path from fundamental physics to actionable EMC prediction remains opaqu e to many practicing engineers. This presentation aims to demystify this j ourney.\n\nWe start at the foundation: how Maxwell's equations in differen tial and integral form lead to fundamentally different numerical solver fa milies - volume-based methods (FEM\, FDTD) versus surface-based methods (M oM\, PEEC). The choice is not arbitrary\; it is dictated by your problem's physics\, materials\, and electrical size. Through practical examples\, w e demonstrate solver selection for automotive EMC: when to use full-wave 3 D versus quasi-static extraction\, how to bridge the MCAD/ECAD domain gap\ , and why near-field antenna geometry matters.\n\nA detailed biconical ant enna walkthrough demonstrates the complete workflow - from geometry defini tion exploiting axial and planar symmetries\, through material assignment strategies (PEC versus realistic copper)\, radiation boundary setup for CI SPR 25 test distances\, port excitation\, and convergence.\n\nBut an anten na model alone does not predict EMI. The critical step is integration: com bining the biconical antenna with a device harness in a unified simulation domain to capture mutual coupling\, then extracting S-parameters for the complete antenna-harness-LISN system. We demonstrate the EM-to-SPICE hando ff - Touchstone curve-fitting with passivity and causality enforcement\, i ntegration with nonlinear sources\, and time-domain-to-frequency-domain co nversion for emission prediction.\n\nThe practical bottleneck is not Maxwe ll's equations - the physics is solid. The challenge is missing component models when modeling your DUT (e.g.\, PCB with SMD components). Vendor lib raries are incomplete or physically inconsistent (non-passive\, non-causal )\, and measuring every passive on a real BOM is not scalable. We conclude with a physics-informed machine learning approach that synthesizes broadb and capacitor models from part descriptions alone\, achieving accuracy suf ficient for design comparison without waiting for vendor data.\n\nSimulati on does not replace measurement. But it enables the critical capability ev ery EMC engineer needs: the ability to rank design alternatives before bui lding hardware. Layout variant A versus B. Filter topology trade-offs. Shi elding effectiveness. You don't need absolute dBμV accuracy to pick the b etter design - you need the ranking to be correct. And when combined with analytical design rule checking and validated component models\, computati onal electromagnetics delivers that predictive capability early in the des ign cycle\, where changes are inexpensive and design freedom is highest.\n \nAgenda: \n5:30 Pizza and networking\n\n6:00 Presentation\n\n7:30 End\n\n Apex Test Labs\, 815 N Opdyke Road\, Auburn Hills\, Michigan\, United Stat es\, 48326 LOCATION:Apex Test Labs\, 815 N Opdyke Road\, Auburn Hills\, Michigan\, Uni ted States\, 48326 ORGANIZER:scott@emcsociety.org SEQUENCE:21 SUMMARY:Computational Electromagnetics from Maxwells Equations URL;VALUE=URI:https://events.vtools.ieee.org/m/553583 X-ALT-DESC:Description: <br /><h2>&nbsp\;</h2>\n<h2><span style="color: rgb (224\, 62\, 45)\;">Pizza Sponsor: <a href="https://www.cornucopiatechnical sales.com/">Cornucopia Technical Sales</a></span></h2>\n<h2><span style="c olor: rgb(224\, 62\, 45)\;"><img src="https://static.wixstatic.com/media/1 a5db7_a98a42e26aad412194401a6e8efdbbb9~mv2.png/v1/fill/w_373\,h_96\,al_c\, lg_1\,q_85\,enc_avif\,quality_auto/1a5db7_a98a42e26aad412194401a6e8efdbbb9 ~mv2.png"></span></h2>\n<h2><span style="color: rgb(224\, 62\, 45)\;">&nbs p\;</span></h2>\n<h2>Venue Sponsor: <a href="https://www.apextestlabs.com/ ">Apex Test Labs</a></h2>\n<p>&nbsp\;</p>\n<p>&nbsp\;</p>\n<p><img src="ht tp://www.emcsociety.org/big_iee.jpg" alt="" width="100" height="100"><img src="http://www.emcsociety.org/emcsociety.jpg" alt="" width="200" height=" 100"></p>\n<p>&nbsp\;</p>\n<div>&nbsp\;</div>\n<div><strong id="docs-inter nal-guid-88e8ec44-7fff-345e-6008-f841b0f4039b"><img src="https://events.vt ools.ieee.org/vtools_ui/media/display/3840c9b7-7370-466f-b1a7-a44ec3ce3f77 " width="200px\;" height="200px\;"></strong></div>\n<div>&nbsp\;</div>\n<d iv>Kristof P. von Czarnowski</div>\n<div>&nbsp\;</div>\n<div>Speaker Bio:< /div>\n<div>\n<p dir="ltr">Kristof P. von Czarnowski is an RF and hardware engineer specializing in high-frequency and mixed-signal system design fo r automotive applications. At Lear Corporation since 2022\, he leads trans ceiver design\, implementation\, and validation\, while also contributing to low- and high-voltage hardware platforms with a focus on signal integri ty\, power integrity\, and electromagnetic susceptibility and emissions.</ p>\n<p dir="ltr">He serves as an electromagnetic compatibility (EMC) simul ation subject matter expert\, working to shift EMC from reactive post-test mitigation to proactive design integration. His work aims to embed EMC co nstraints directly into the development process through design guidelines\ , automated design rule checks\, simulation workflows\, and libraries - pu shing for earlier issue detection\, reducing validation risk\, and improvi ng first-pass success rates across schematic design\, PCB layout\, and com pliance validation.</p>\n<p dir="ltr">Kristof holds undergraduate and grad uate degrees in engineering from Oakland University\, where he specialized in electromagnetics and wireless systems.</p>\n<p dir="ltr">His technical interests include practical EMI/EMC simulation workflows\, SI/PI co-desig n\, and model validation for RF and power distribution networks. His recen t work explores AI-assisted engineering tools and machine learning approac hes to automate component characterization and accelerate system-level EMC prediction - ad</p>\n<p dir="ltr">dressing the industry's missing-model b ottleneck that prevents full system-level compliance prediction.&nbsp\;</p >\n<strong><strong id="docs-internal-guid-4b751bdb-7fff-ca3b-d1e3-abdcc89d 278d"></strong></strong></div>\n<div>&nbsp\;</div>\n<div><strong>Abstract< /strong></div>\n<div>\n<p dir="ltr">Every electromagnetic compatibility pr oblem begins with Maxwell's equations. Yet the path from fundamental physi cs to actionable EMC prediction remains opaque to many practicing engineer s. This presentation aims to demystify this journey.</p>\n<p dir="ltr">We start at the foundation: how Maxwell's equations in differential and integ ral form lead to fundamentally different numerical solver families - volum e-based methods (FEM\, FDTD) versus surface-based methods (MoM\, PEEC). Th e choice is not arbitrary\; it is dictated by your problem's physics\, mat erials\, and electrical size. Through practical examples\, we demonstrate solver selection for automotive EMC: when to use full-wave 3D versus quasi -static extraction\, how to bridge the MCAD/ECAD domain gap\, and why near -field antenna geometry matters.</p>\n<p dir="ltr">A detailed biconical an tenna walkthrough demonstrates the complete workflow - from geometry defin ition exploiting axial and planar symmetries\, through material assignment strategies (PEC versus realistic copper)\, radiation boundary setup for C ISPR 25 test distances\, port excitation\, and convergence.</p>\n<p dir="l tr">But an antenna model alone does not predict EMI. The critical step is integration: combining the biconical antenna with a device harness in a un ified simulation domain to capture mutual coupling\, then extracting S-par ameters for the complete antenna-harness-LISN system. We demonstrate the E M-to-SPICE handoff - Touchstone curve-fitting with passivity and causality enforcement\, integration with nonlinear sources\, and time-domain-to-fre quency-domain conversion for emission prediction.</p>\n<p dir="ltr">The pr actical bottleneck is not Maxwell's equations - the physics is solid. The challenge is missing component models when modeling your DUT (e.g.\, PCB w ith SMD components). Vendor libraries are incomplete or physically inconsi stent (non-passive\, non-causal)\, and measuring every passive on a real B OM is not scalable. We conclude with a physics-informed machine learning a pproach that synthesizes broadband capacitor models from part descriptions alone\, achieving accuracy sufficient for design comparison without waiti ng for vendor data.</p>\n<p dir="ltr">Simulation does not replace measurem ent. But it enables the critical capability every EMC engineer needs: the ability to rank design alternatives before building hardware. Layout varia nt A versus B. Filter topology trade-offs. Shielding effectiveness. You do n't need absolute dB&mu\;V accuracy to pick the better design - you need t he ranking to be correct. And when combined with analytical design rule ch ecking and validated component models\, computational electromagnetics del ivers that predictive capability early in the design cycle\, where changes are inexpensive and design freedom is highest.</p>\n<strong><strong><stro ng id="docs-internal-guid-9b3c40d8-7fff-4dcb-292b-104ad7d2d7c6"></strong>< /strong></strong></div><br /><br />Agenda: <br /><p>&nbsp\;</p>\n<p>5:30 P izza and networking</p>\n<p>6:00 Presentation</p>\n<p>7:30&nbsp\; End</p>\ n<p>&nbsp\;</p> END:VEVENT END:VCALENDAR