You should not rely on any version of the USB protocol to protect something from happening. It might be possible to implement what you want but that is an out-of-spec use of the protocol and personally I would never want to rely on it because it was never meant to be capable of that.
I think you are misunderstanding a lot of what I am trying to accomplish here. I am trying to get confirmation of what current the USB spec is allowing me to draw, which I will control ourselves to ensure the board is always compliant. I don't really see where you are viewing USB itself as being a protection device anywhere in this plan. USB itself serves as no protection in any of what I've described, in fact a great deal of the board is to protect the Host pc's usb ports/circuitry specifically with the idea that there isnt inherent protection of a level that I am looking for.
If you want to build something that will be safe to use in a lab context then you must account for that environment yourself. There is only one way to cope with such a harsh environement: total isolation. There is a reason why all those fancy T&M tools from HP; Gossen; Rigol and many more have a galvanic isolation on the communications bus (USB, ETH, RS-xxx; GPIB; etc... ).
If you look over my initial post, you'll see that the entire point is complete galvanic isolation, and indeed that is the reason for the 4 main isolating components described, exactly for that purposes. The Hi-speed USB isolator, is the key component of this device, so I am once again confused because that is quite literally what the plan is: To have complete isolation on one side of the board to the other side of the board.
I will write it again: the simplest, don't care and no more worries solution is to use an additional cheap second hand computer.
I feel I have already sufficiently explained why this suggestion is not relevant to my project as what you are suggesting is to simply not do my project and continue to risk hardware rather than build a safety device once that improves my workflow significantly by not simply having to use low quality hardware due to a lack of proper isolation.
In your context an MCU is going to be one of your cheapest components. The very capable ESP32-C3 is 1€ and you can find other MCU as low as 10cent. Unless you are going to build millions of devices and the price of the device is sensitive to a couple cents, it is a no brainer decision to include an capable MCU especially if your device is something that will be used to make debugging of firmware/hardware easier, that MCU will bring a lot of capabilities for a ridiculous cost that otherwise would not be possible. Chances are that your potential customers don't care about 5€ more or less but they do about that extra functionality.
I don't quite think you are understanding the goal or purposes of the device I am aiming to build. The ESP32 would be completely wasted other than to configure USB hub controllers, and would necessitate extra, unnecessary complexity. I have many ESP32s laying around, this isn't like Im not familiar with or am unaware of how mcus work or how cheap they are, it is that I
am and do not want the added complexity, especially when it comes to a device I want to be fairly bullet proof safety wise.
I don't have potential customers. Im building something that I want, for me, to improve my workflow.
You still have a 'modern' use case for the FT2232HQ? I'm genuinely curious about what you want to do with it, because I think of it as a legacy interface with old fashioned serial buses. It works perfectly well but is expensive, very expensive.
This is the most confusing part of your comment. These are completely inexpensive. They cost at most 7 dollars Canadian. They also have all of the capabilities I care about with regards to debugging such as compatibility with JTAG, SPI, I2C, UART, and SWD (through software) using either open source libraries and software (ex. OpenOCD) or their driver set.
They are a perfect fit for my use case, and I can't quite see why I would degrade my experience by having to bit bang the JTAG protocol manually through code as I would need to do with something like an esp32.
They have 480mbps bandwidth which is more than enough bandwidth for dual channel communication for any of the specified protocols, have reasonable software support, and deal with all of the GPIO worries and most of the protocols I care about internally with their MPSSE modules.
All in all I think there have been a lot of crossed wires here, so let me reiterate:
The goal is a debugger board with isolation that allows me to comfortably debug things without adding friction to my workflow in a protected manner.
To do this I will be putting a FTDI FT2232HQ behind complete galvanic isolation just as described in my initial post.
To power the isolated supplies however, power is required, and that is where this post comes into play, with my seeking to find a method to best identify how much current I can safely pull from the other side of the isolation wall.
Just one more time, I want to reiterate, a major part of this boards aims, involve being galvanically isolated, hence so much of my confusion at much of your reply. Hopefully that clears things up.