Hi.
I have to combine two 240W power bricks into one to power ISDT Q8 Max LiPo charger. I stumbled on the article (1) mentioning LTC4370 chip, and started searching.
I found this thing (2) on eBay and could not identify the chip and grasp how does it work without ground wires. Anyone has a clue, please?
(1)
https://slpower.com/archives/11891(2)
https://www.ebay.com.au/itm/402629555195
Operation is described in part 7 of LM74610 datasheet, particularly figure 9 and surrounding text.
It periodically turns off for a moment and uses the 0.7V drop of MOSFET's body diode as a power source.
This may or may not be acceptable, depending on your needs.
The concept of a ground in an electric or electronic circuit is only a way of allowing us, the poor human designers of such circuits, to think about what is going on without getting a headache. As such, a ground is, at least at the roots of the concept, literally a connection to the actual planet earth, which is assumed to be at a neutral or zero Voltage. We then look at everything else in the circuit in relation to that arbitrary point. And our limited minds can understand how the circuit works.
But charges and holes and whatever in actual circuits will flow as they will flow. Some, dare I say most circuits will function in relation to this assumed to be neutral Voltage. This is because WE designed them to work that way, not because it is any kind of magic or necessary thing that must be used in the design process. If you look at the equations that relate the ins and outs of things like resistors, capacitors, or transistors, you will not see any mention of a ground. There is no term for a ground in Ohm's law. It only speaks about relative Voltages and currents and the values of resistance. There is no 0 Volt term in it. Ditto for any and all of the other basic equations in electricity and electronics.
So, some circuits are not designed with a ground reference in mind. They simply depend on the current through a device or the Voltage difference across two or more of it's terminals. I have seen transistors where the emitter is at an elevated Voltage of 1000 Volts or more. And it conducts if the base to emitter current is present. And it does not conduct if that current is absent. The emitter could be connected to an AC Voltage that varies +/- 300 Volts. If the base is at the same Voltage, it will not conduct. If the base of an NPN transistor with it's emitter connected to that AC Voltage also varies by, perhaps +301 Volts to -300 Volts, then a varying b-e current will flow and it will conduct for part of the cycle. It is the b-e current that determines conduction, not any relation to any assumed reference point, like the earth's Voltage.
A case in point is a simple dimmer for a building light. These dimmers often, but not always designed to be inserted in series with the AC source and the light bulb with no connection to earth ground being needed.
I have had to read and understand schematics for devices that were many, many pages long. Believe me when I say that the concept of a single reference Voltage for the entirety of such a device is almost a total necessity if it is to be understood. But for smaller, simpler circuits, there are sometimes reasons for abandoning the concept of a single reference Voltage and working with a circuit that is a bit harder to understand but perhaps less expensive to produce. Engineering is always the art of managing compromises.
The concept of a ground in an electric or electronic circuit is only a way of allowing us,
Sure, and you add a lot of irrelevant crud, but the IC has to get it's power from somewhere. The datasheet of the posted lm74610 is much more enlightening. Using a single PN junction voltage drop to power an IC, then pump it up to over 6V and use that as a gate drive to keep it open for 98% of the time is a quite clever idea.
I read something about such "ideal diode" circuits before, but never as well explained as in this TI datasheet.
As far as I can tell, the only advantage of this over a P-channel MOSFET is it uses an N-channel MOSFET, which gives lower RdsOn for a given size.
May also have slight benefits where the supply voltage range is wide, but otherwise seems a rather expensive solution.
Magic had already talked about the datasheet. I was trying to talk about the idea that, while most circuits are designed around one or more Voltage rails and ground, that is NOT the only way that a circuit can be thought about or designed.
Sorry if you thought it was "irrelevant crud". Perhaps someone got something better from it.
The concept of a ground in an electric or electronic circuit is only a way of allowing us,
Sure, and you add a lot of irrelevant crud, but the IC has to get it's power from somewhere. The datasheet of the posted lm74610 is much more enlightening. Using a single PN junction voltage drop to power an IC, then pump it up to over 6V and use that as a gate drive to keep it open for 98% of the time is a quite clever idea.
I read something about such "ideal diode" circuits before, but never as well explained as in this TI datasheet.
Sorry if you thought it was "irrelevant crud".
It was "irrelevant crud" 
Not really . EPAIII Is trying to explain that anytime theirs a potential voltage difference across any element (the MOSFET in this case), current will flow. Current always flows from the highest potential to the lowest. This is why it is possible to "float" low voltage circuits on very high voltages.The concept also applies to your ideal diodes. All that is required is a charge pump that multiplies the potential voltage difference enough for the circuit to operate. No ground is required.