If I am understanding correctly, you are working with a circuit that will be connected to AC power from the wall socket, and do some dc switching to energize downwind circuitry. You are asking specific questions, and these questions are easy to blurt out an answer to, but that answer could be based upon a misunderstanding of setup or application or your specific use case as there is the possibility that you are using the circuit differently that expected alongside a whole host of other things that could get any, some or all of us wrapped around the axle.
I think that the best answers to your questions can be arrived at via full understanding of the circuit to begin with, aka schematic. Then from there, make sure you understand what portions are AC and what portions, if any, operate in DC domain. You see, mathematics can be applied, and the mathematics can be greatly simplified if you are able to say that the circuit will always operate at a given frequency (such as 60 Hz from power company). Given an input voltage and the complete circuit with all component R, L, C values, you can compute the expected DC voltage at every single node (electrical connection point) in the circuit. Hopefully its a lower frequency all around, like 60 Hz, so that you don’t have to worry about any portion of the design being distribuited such that you can no longer rely upon a board trace to be a single voltage at all times (for GHz and beyond, this breaks down, and each trace becomes a complex system on its own accord). So this node based math is how EEs first analyze circuits in a classroom, and we often do it this way in the industry for the sake of simplicity and to obtain a quick result or understanding of whatz up.
However, most circuits are complex and will give you a headache, not to mention that the math will spiral out of control in complexity. They quickly exceed a complexity level beyond what you can figure out in your head. AC circuits with networks of resistors, capacitors, inductors, switches, and transformers are such beasts. When we get to a system like this, we may analyze mathematically using the node voltage method, (or mesh current, a similar and slightly more abstract hand analysis method). But more likelly, an EE worth his salt will crank out an electrically correct computer model of the circuit in question. Provided it is electrically correct and has the proper values in it, it should yield the behaviors and voltages that your simulation deck, as it is termed, predicts. You can also answer the what-if questions once you have this because you can easily make a change to your deck without having to do a bunch of soldering and making a mess in the lab. However, once you know your answer from your deck, you will have to assemble the final topology to achieve the desired result, provided there are changes your simulation prescribed from your starting point circuit.
There is also this to consider: DC analysis yields what is called the quiescent point or Q point of your analysis. This tells you what DC voltages each node of your mixed-signal project will eventually settle at a sufficient time after energizing the system with power from turning it on. A simulation analysis of the system immediately upon “power on” is called transient analysis, and should also be considered and modeled in cases where the spikes from switching power on can create issues and/or your circuit operates using an intentionally created resonance such as a switching supply or boost converter. Finally, once your Q point is known, you can do AC analysis and/or signal analysis. This would be overly complex to do by hand, and I highly recommend a computer simulation package for this sort of work.
If you want a recommendation about software you can use on the cheap/free, National Instruments Ultiboard/Multiboard/Circuit Design Suite can be downloaded and used for educational purposes for free (with caveat that if you use them to make money, you are required to purchase, and they are not inexpensive). There is a place on ni.com that will allow you to certify that you are in school or learning phase (does not have to be formal education, self-edu is ok) and also you agree to pay for the software should you begin to use commerically).
There are many different realms of such computer modeling pertaining to circuit design, some of which are for sure not the starting point for someone who just wants to understand basically what is going on with their simpler circuit. One such realm is RF, and NI purchased AWR/Microwave Office, which was the premier RF modeling enviiornment in the industry prior to the purchase, and remains so with further bolstered brand recognition courtesy of engineering powerhouse NI. There are also modelling enviornments for Signal Integrity (High Speed Digital Design) and Power Delivery. All of these in this paragraph are likely of no concern to you, but I included so you could see why the NI software is a fantastic starting point…it leads to powerful realms with experience… I spent years in the singal integrity simulation world, whcih satisfied my curiosity about advanced PCB design. Basically, with any of these, they are for the considerations at either very high frequency (RF, Si) or the finer aspects of delivering power to chips runnihg at very high frequency. The designer here has to take minute details into consideration that you would otherwise neglect, such as capacitance of a pin on a tiny chip. Knowing what to consider and what can be neglected for each of these sciences is the art that fills consultant’s bank accounts, and can generally only be learned through years of experience going back and forth from simulation workstation to the lab comparing simulated results with properly measured lab results. If the models you were supplied by your parts manufacturers are close enough, and you knew what to include in the sim deck, you can predict what will happen if, and the world is your oyster.
But those very high frequency simulation environments are far beyond what is needed here. The vast majority of answers for cirtuitry can be obtained from the classic simulation enviornments based upon some form of SPICE. These consider all nodes to be lumped and each component has minimal parasitic effects (these are the assumptions that break down when you enter the high frequency twilight zone). Your case requires a simple spice style sim deck considering only the basics, R, L, Cs, etc while considering all nodes as lumped (not distributed networks). Good options here include the NI stuff, or LTSpice (free from Linear Technology), or maybe even PSPICE (free years ago last time I used) or HSPICE (if you have access to a license).
In this manner, you can arrive at the answers you need, and you no longer need an expert, because you just became him/her. For me to arrive at correct answers I can give with confidence, I have to do exactly this, which takes time (lots), energy and communication to get all of the pertinent specifics. I think that it makes a lot of sense for me to pass on the best of what the industry taught me so that folks at this great makerspace can become extremely adept at incorporating the very best electronics aspects to the things that you make. Please note that usually the cool stuff you and your kids want for christmas have electronics fully integrated with them, whatever it is.
There is also an experience level factor here that affects the validity of the simulation performed. Experience shows the engineer what to do to get his/her deck to match reality, and so either with or without this experience, it is advisable to verify your sim by comparing to measured reality for the sake of sanity checking, and I would say its simply the most important requirement for someone without a lot of experience. You need to know that your deck is spitting out coherent information as opposed to straight nonsense, which is a common result as it is easy to set a weird parameter that might, say, accidentally ground nodes intended to settle to Q point and carry signal.
So guru is a set of big shoes to wear. But I am EE with significant industry experience and I hang around DMS. When it comes to answering electronics questions that nobody knows the answer to for sure or with a great deal of confidence, is in dispute, or is outside the experience of pretty much everyone, this is how to get the answers. What i have outlined here is the brute force approach that anyone can use to figure things out for themselves.
Yes, it is quite a bit of work and will take time to get confident with the tool you select. And I have seen folks working in the simulation/computer modeling industry that could not match a real circuit in simulation, hence would be unable to predict behavior of a new circuit prior to expending the dollars and time to build it. But should you get good at this, you are quite employable and you are likley to discover companies may go so far as to fight over your services, which can be steered into salary/compensation bidding war to your benefit. In my decades on this planet, I have yet to see a slowing in need for electronics/software/hardware//firmware, dating all the way back to my getting my first electronic toy as a kid, Merlin. (anyone remember that one?) Its like the damand for the stuff has only accelerated.
MY VERY BEST ADVICE TO ALL INTERESTED IN UNDERSTANDING HOW TO MAKE ELECTRONICS WORK FOR YOU IS TO LEARN ABOUT SIMULATION/MODELING AND HOW YOU CAN GET YOUR OWN ANSWERS AND HAVE CONFIDENCE IN WHAT YOU ARE DOING…THIS IS THE PATH TO GURUDOM, AS IT SERVES AS THE BOUNDLESS FOUNTAIN OF INFORMATION AND PREDICTION FROM WHICH ALL THE VERY BEST GURUS DRINK PRIOR TO PUBLICLY SPRINKLING PIXIE DUST UPON THE AILING CIRCUIT IN QUESTION, FILLING THE ROOM WITH THE SOUNDS OF AWE, SHOCK, AND DISBELIEF AT THE BLACK ART THAT WAS CLEARLY DEMONSTRATED BEFORE ALL EYES.
And finally, yes I know I wrote you a book, I’m famous for excessively lengthy talk posts, However, this was a great interlude for this suggestion. Further, I would like to put down a challenge for you, should you decide to take it. I would get all giddy should questions on talk about electronics started sounding very much like this: “I was not sure what was important here, so I made the following sim deck (see attached). I got this result when I did (transient/AC/DC/Monte Carlo, etc) analysis, and while my Q points looked good and were verified here (see attached), I have concerns about this portion of the analysis and whether this aspect is modeled properly. What do you think, and are there things I can do to check this aspect or have a degree of confidence as to whether or not it is a good prediction?”
Should I see questions in electronics posed like this, then I know DMS electronics is leading the maker movement electronic front. I’m thrilled to chip in from the experience set I managed to accumulate in the industry. In the meantime, I suggest starting with LTSpice (easiest to acquire and install), model a simple DC voltage divider (to understand quiescent modeling in this tool), then a low pass filter setup/test bed (simple RC circuit that should have 20db/decade rolloff to test AC analysis), then move on to attempting to model the circuit discussed above. You should be able to find an LTSpice quick video tutorial on Linear Tech website, or youtube (the greatest information exchange so far in history of mankind) that can get you agile enough in the tool to do some simple analyses.
And maybe I should do some classes on computer modeling of circuitry? This would benefit the space more than anything else I can think of to offer.
For a great number of components out there, the manufacturers provide models that can be brought into the simulation environment for this exact purpose. Some devices make no sense to model, as chiefly it is analog effects that are simulated. Digital drivers can often be modeled for sub GHz simulation by creating a voltage source with source resistance and maybe pin capacitance if its even a factor at all (this is for cases where your analog circuit or system is stimulated by a digital line from some chip…happens all the time) You could also determine the thevenin equivalent current source and model it that way if you just want to be abstract. If you want suggestions on how to model various items, feel free to send me a list of the specific components you have questions about, datasheets included is best so we both can see the analog characteristics. I have had my best success by taking care not to overcomplicate things by including 6th order parasitcs in a 20 kHz simulation setup. You will develop deep understanding simulating as you will have to fully understand how to set things up, how to stimulate to get the parameter of interest, and you can easily probe DC currents in every branch and also the signal at each node. You can quickly add something you wanted to try out, then that and see the effect and how it changes things. Like I said, you will develop DEEP understanding of the circuit in question. You can also, perhaps with some help initially, drive a transmission line (cable or board trace), SE or differential, with some serial standard driving so many Mbps or even Gbps. If you do need to solve a problem like this, be sure to contact me as I spent a good decade on these board/cable related sim/reality concerns. So you bet I can also help with less speed related, more typical analog what-ifs.
Finally, I would like to point out that the VirtualBench in the lab is a great tool for correlating your simulation to reality. It has the ability to stimulate your circuit, controlled by software, then measure back the result using ADC lines (analog in), and can trigger iteself appropriately with ease since all that functionality is in one box, no need for GPIB converters, usb drivers, and trigger lines between instruments, alongside all the headaches presented trying to get all that to work…
Yes I know its a lot of work I just prescribed, but this is where those confident answers get them when it comes to the nuances of electronic ciruitry, specifically analog concerns.