I do not really need too much isolation and I can just use an op-amp. And given that it only adds a few milliohms, I think it is the better choice. Plus with such large currents there may be other magnetic fields that would give inaccurate readings. And would it really be that accurate? I am just pretty skeptical.

I can think of a few methods to measure current: The current transformer (AC only); Hall effect; High side sense; low side sense; there is a new method I read about *Magnetic"

You have to pick one. A high side monitor will require some effort, so it makes sense to buy. See: https://www.maximintegrated.com/en/app-notes/index.mvp/id/746 I'm not advocating any part. Sometimes you don;t have much of a choice with full scale voltage.

High current connections are usually done with "bus bars". On the 45 A power supplies I bought, two copper rectangular bars with threaded holes were where you attached your wires.

Some high current lugs are here: https://www.platt.com/platt-electri...+ground+lug&sectionid=4&groupid=61&CatID=1254

Connections to welding cables are normally to a lug where the wire is secured with a screw. For stranded cable, usually some sort of plate (wire guard) is pushed against the stranded wire. You do need to look at suitability.
Some connections, like crimp are not suitable for solid wire. Crimp lugs for very large diameter cables might require a hydraulic crimper.

One work gizmo needed 6V AC at 3000 Amps (yes three thousand). The connection to the transformer was a bar probably about 8" wide x 3/8" thick.

A wire chart is here: https://www.powerstream.com/Wire_Size.htm

I used 8 AWG for transmission about 6 feet maximum in my application in the 40-45 A range.

The soldering isn't the big deal. It's the flux vapors. Solder typically isn't used with high current connections.
Silver plated tubing is actually used. Silver soldering or brazing is a very solid connection. When you actually get really picky about solder and accuracy, every solder joint is a thermocouple. Those effects are usually neglected.

I think that covers about 1/2 of post #58.

 

Ok. So I think I'll just make a complete schematic for all the power components and then connect them all to a piece of acrylic or some other heat resistant non-conductive material. I'll just have to figure out the exact layout and get the connectors. Then maybe I'll have a PCB or something for the control stuff (esp32, op-amps, control ICs, all that stuff).

So then is what I suggested in #58 for the switches and filter practical and efficient? Or are there certain ICs that you would recommend that would be much more accurate and better suited for what I want?

 

Also for constant current could I just apply a varying voltage (probably 0-5V) to a BJT? Or is that incredibly impractical for such large currents? Is keeping it simple stupid or is simple better here? And my understanding is that Ic=Ib*B, Ic being the max collector current, Ib being the base current. But I have heard that beta can change based on temperature and other conditions. So is this a good, reliable choice? Or is there no escaping the complex control systems?

The major control element choices are the BJT, the MOSFET and the IGBT and, of course, parallel combinations. I'd so a search for IGBT vs MOSFETS and see what you get.

Most major variations in control depend on your REFERENCE stability.

And if that is a good choice, can you place the collectors and emitters of an IGBT parallel to it? They would not be operated at the same time, but would crossover or them being connected like that be a problem?

Current sharing is always a problem. For BJT's small emitter resistors are used as well as Beta matching.
High power audio amplifiers use this technique. With MOSFETS, you try to match Vgs and transconductance.
A match might be within 10%.

One last question about the filter part. Do I just need a capacitor in parallel and an inductor with flyback diode in series? Is there something else missing here, or is the inductor unnecessary? I want to have very low ripple (10s of mVs) even with more power drawn, but I want it to be able to change the voltage within a millisecond or two. I do not want it supplying overcurrent for 10s of milliseconds. So I need to figure out the right capacitor and inductor value. Also, I do not want any LC resonance going on. I am thinking 75-300 uF ultra-low ESR 60+V for the capacitor and 200-400 uH 4 or less mOhm ESR 50+A 60+V for the inductor. What do you recommend here?

These
http://www.ti.com/lit/an/slva549/slva549.pdf
https://www.digikey.com/-/media/Ima...locker/APEC_Ripple_BlockerFINALWhitePaper.pdf

https://www.eevblog.com/forum/chat/possible-to-reduce-output-voltage-ripple-below-1mv/

https://www.edn.com/design/analog/4350857/Capacitor-amplifier-reduces-ripple-without-dc-loss


offers some techniques to reduce ripple.