3 ms·
Yep, you are completely right. It is a triangle of size, efficiency and cost. Simplest flybacks, and half bridges can't compare to topologies with extra magne
by baybal2 7y ago
Yep, you are completely right.
It is a triangle of size, efficiency and cost.
Simplest flybacks, and half bridges can't compare to topologies with extra magnetics on efficiency, but you need to find space for extra xformers.
The appeal of high end 3-5 switches is that you can get high efficiencies even in simplest topologies, and employ no extra magnetics.
My engineers too hit icebergs few times when we tried making power supplies "the smart way."
PZT xformers are smaller than iron, but their output may not be completely sinusoidal.
Fancy vitroperm, finemet, and other similar magnetics can have rather low tolerances for magnetic properties, and some even drift over time as their metallurgy changes at relatively low temperatures.
GaN switches can be easily damaged without carefully made gate driver.
Capacitive isolation is not foolproof, and substandard safety caps are everywhere.
SiC switches are said to be as reliable as galvanic iso, are not so reliable in practice.
- kens 7y ago> It is a triangle of size, efficiency and cost. It's interesting to consider that the reason there are tradeoffs is because uniformly bad technologies get discarded, so the remaining choices have tradeoffs. For chargers, consider the 1990s-style line transformer and linear regulator (big wall warts). These are larger, less efficient, and cost more than a modern switching power supply, so essentially nobody uses them. So it may seem like you're stuck in a triangle of choices that are bad in some way, but in the historical picture, these choices are all very good.
- milesvp 7y agoYeah, you’re sort of describing engineering in a nutshell. It’s all about tradeoffs, and there’s so much min-maxing going on. Then eventually you discover some property or design where the tradeoffs aren’t as bad while still mostly maxing the desired property and it becomes a common solution. What’s interesting, is that periodically you can go back to what looked like a bad solution, and because of advancements in related designs, this solution can actually end up being superior in some ways.