An intuitive explanation of the multiphase Buck Trans-Inductor voltage regulator (TLVR)

[@esijal]

Great and insightful presentation 🙏

[@sambenyaakov]

Thanks

[@josemariamolinagarcia5562]

Thanks Sam! Looking forward to the new version of TLVR presented at APEC

[@sambenyaakov]

Will see

[@michaelbyrnes1822]

Thank's sam great time topost a video

[@sambenyaakov]

👍

[@two_number_nines]

I wish I could find an old motherboard with ring shaped toroid inductors so I can wind a secondary like this on them. Also, wouldn't these transitional current spikes cause the lower side transistors to blow up if they were already at their current/power limits in the no-tlvr configuration?

[@ats89117]

Interesting as always, but what are the implications if all the transformers are wound on a single core?

[@two_number_nines]

there would be no filtration, as the core would see pure DC and there would be no magnetic field collapse to drive the flyback cycle

[@ats89117]

@@two_number_nines Good point.

[@sambenyaakov]

This will have no series to parallel amplification

[@martinmartinmartin2996]

Very complex to analyze. so many factors to consider; such huge currents make one wonder the possible diameter of the windings.

[@sambenyaakov]

see www.tdcommons.org/cgi/viewcontent.cgi?article=6147&context=dpubs_series

[@brandonhicks7549]

Interesting concept. Using the same magnetic cores for the transformers, you can either double the wire cross section with the same turns, keeping inductance the same which should halve the resistance, or double the turns quadrupling the inductance. The later is a bad idea, so relatively speaking the ESR loss from the transformer regulator will be twice as high in the primary as the conventional regulator, however the secondary will also have losses. It would seem that in the worst case the secondary current will be approximately 1/N times the current carried in the primary, so this can be minimized with a large number of phases. These sorts of regulators usually suffer from large switching losses in the half-bridges, to keep the switching speeds high enough to satisfy the di/dt requirements. Effectively doubling(or N times, in your theoretical “nonlinear controller”) the di/dt using this technique might mean lower switching frequency (and higher inductor core losses) or fewer phases (and higher conduction losses) I can see where the trade offs may pay to use this method, but it’s not trivial to compare without a detailed optimization for both designs.

[@tamaseduard5145]

👍🙏❤

[@sambenyaakov]

😊👍🙏

[@WhittierSteven-w6z]

Kuvalis Stravenue

[@HoraceFrancis-f2m]

Ullrich Squares

[@DevenRandall-p9j]

Daniel Shoal

[@LeonardMack-z4h]

Noelia Tunnel

[@DonDean-e7r]

Torphy Coves

[@RonaldColeman-f3z]

Dickens Mission

[@samlau2852]

hi, what is your comments about those (5+ yr ago ) youtubers claiming V= --L l'(t)? like this one { kzbin.info/www/bejne/eGmmgXqfgJpsnJo }. Is the -ve sign wrong??

[@sambenyaakov]

I think it is a matter of convention. The sign is meaningless until you define and mark the positive polarity of the terminal.

[@ClaraMonica-v9w]

Alexandria Mountains