既然你要用英文回，我也用英文回好了 The most fundamental concepts of KCL and KVL is: Maxwell Equation It CAN be rigorously proved using physical theorems The concept of V is based on the electric field, and the concept of I is based on the charge ALL simulator circuits uses KVL and KCL for calculation, theorem-wise there is no difference between lumped circuit components (L, R, C, etc.) and distributed circuits (transmission lines). The statement of KVL and KCL only work for lumped components is completely wrong. You just don't know how to express a transmission line with L, C components; a transmission line expressed with a limited number of L, C components is a low pass model, and increases in complexity as operating frequency increases. A proper approach is to use an s-parameter expression. You probably are not aware that there are many other circuit simulators other than spice, such as ADS, MWO, and Designer. All of them handle distributed components with ease. SPICE simulator is a time-domain simulator, therefore it has problems dealing with distributed components. You can also refer to any microwave engineering text, to see that KVL and KCL holds for any circuit. Even for multi-mode behavior in waveguides, an equivalent circuit in terms of s-parameters can be described; in this case, there are no DC current or DC voltage, but KVL and KCL still holds. KVL and KCL works because it is derived from Maxwell equations, not because we want it to work. The fundamental meaning of KVL and KCL is conservation of energy and conservation of charge. Any reputable textbook on electromagnetics, fundamental circuit theory will back my statement. ※ 引述《pow (体脂肪35%)》之铭言： : Just to share some idea... : I think the most fundamental assumption of using KCL and KVL is... : you are using abstract concept to describe things. : The beauty of EE is that we can use abstract V-I so well instead of : direct dealing with E-M...so well that we all feel : V-I is the real quantity, and embeded them into our everyday live. E-M is : from Physics department but we have our own way to live with it. : This V-I concept happened to work with the "pattern recognition" technique. : By "pattern recognition" I mean the topology. Top line is VDD. Bottom line : is ground. Left is input. And right is output. We borrow the concept of : topology from Math department, do some modification and apply to our EE world. : With the two weapons in hand, we can put complex calculation away and start : to "model" things. However, there's always limitation that our method : won't work anymore. Berkeley SPICE was developed to help us calculate complex : circuit, but still, they can only do V-I and topology. It's just that... : over time we get used to "back of the envelope" calculation (and SPICE.) : So we figure out some other way to "model" something that's not supposed to : work with V-I (instead, they should work with E-M). Probably it's our nature, : if the model fit data at certain range under certain environment, : we are happy. : So if you are claiming KCL and KVL will work as long as there's model : for it, this arguement is kind of loose, don't you think? It works because : we let it be that way, not because it works by itself. : ※ 引述《cpt (post blue)》之铭言： : : 我想做的只是提供一个我觉得合理的思考方式 : : 这当然和严谨的推导不同 : : epsilon0 和 mu0 是物理常数 : : 把它们设为零, 就好像把光速设为无限大一样 : : 在适当的尺度下, 这样的假设是合理的, 我并没有要做过多的延伸 : : 但是把 epsilon0 或 mu0 设为零, 在根本上当然与事实不符 : : 这也就是为什麽 KCL/KVL 在根本上也是与事实不符 : : 只不过这个误差在大部分情况之下可以忽略 : : KVL 之所以成立, 是因为我们额外去 model 了这个"电感" : : 但事实上电感无所不在, 如果有任何一小段没有 model 进去 : : KVL 就无法反映实际的物理现象 : : 因此, 与其说是"没有时变磁场", 为什麽不能想成"忽略时变磁场"呢? : : 这跟把 mu0 设成 0 不是一样的意思吗? : : (这不是在反驳你, 我是真的想知道为什麽这样不对) --

※ 发信站: 批踢踢实业坊(ptt.cc) ◆ From: 140.112.19.162