1. If I do this to extend the circuit 30 feet. should I use 10/3 with ground or 8/3 with ground?
10/3 is fine for a 24A load on a 30A circuit up to 200' one-way from the service panel.
2. What is the difference between A: "eliminating the current (10-30) receptacle" to extend the circuit and B: just buying a 10-30 plug and connecting it with the above wire (#1) to a 14-30R, making sure the 14-30R is properly grounded to a cold water supply pipe I have? (I know I can't use the neutral as a ground).
A) eliminates the "extension cord" aspects and will require you to do electrical construction per the NEC to deliver the receptacle to the final spot where the UMC plugs into the NEMA 14-30.
B) effectively uses an "extension cord", to use a comparison with standard 120V 5-15 outlets. NEC governs the code compliance for the electrical construction to the existing NEMA 10-30 receptacle (which I understand to be code compliant as existing installation), then you would need an appropriate, UL-listed/rated extension/adapter cord to provide your NEMA 14-30. By constructing one with a 10-30P to a 14-30R, you've created an extension cord.
The NEC contains some rules for using extension cords. Some of the rules state that extension cords are not permitted to be used as a substitute for the fixed wiring of a structure; run through holes in walls, structural ceilings, suspended ceilings, dropped ceilings, or floors; run through doorways, windows, or similar openings; or be concealed behind building walls, structural ceilings, suspended ceilings, dropped ceilings, or floors.
I understand your concern if I terminate with a 14-50R and a software update overrides my 30A charging setting. Am I correct in assuming if I use the 14-30 Tesla adapter on the Mobile Connector, the Tesla won't try and charge at 40A?
Correct. If you use a 30A adapter on the MC, it will charge at 24A maximum.
Thanks again for your valuable and expert advise. I know you are thinking of our best interest, but every house in my area was built around 1969, probably with the Zinsco box, and none have burned down yet. I did experience a problem once with a connection between a copper pigtail from a breaker to the aluminum house wiring. But all that happened was the joint corroded and wouldn't pass power. I trimmed the joint and coated it with Noalox and it's fine now.
I personally experienced an FPE panel failure in my California rental home in 1997 that nearly burned it, due to a breaker's failure to trip. It was a combination of failures: a crimp-splice (common in the 1960's and 1970's) had failed, creating a higher-resistance point which generated heat when an appliance was run (in my case, a microwave oven and some lighting), eventually creating a bigger, lower-resistance short when the insulation melted from the wires. The breaker failed to trip, causing the reaction to continue at extremely high currents, turning the entire circuit into a giant heater in my wall. I was probably an hour from seeing the house catch fire. I was lucky to be home and awake when it gave way or I would have lost everything.
I should also note that every house in my neighborhood (as it sounds yours is) was built in typical California "pattern-home" style, and they had not experienced any major fires that I was aware of while living there. Mine was the first in my neighborhood.
FPE and Zinsco panel failures are common and have a high rate of occurrence -- just search for them... Moreover, the type of failure that can happen is the worst type -- a breaker that trips once while testing (giving you some sense of security) may permanently jam and fail to trip again. With FPE and Zinsco panels, you really don't know when the failure will occur, but we do know that they have very high failure rates and there is a reason most insurance companies will refuse to cover a home with an FPE or Zinsco panel when they find out. There is a reason that Solar City refused to do that install for you. Hopefully your friend's son will be able to explain it to you as well.
A simple question: knowing that Zinsco panels are inherently dangerous, have many, MANY catalogged failures, why would you not replace it at the first possible moment? These are not "if it ain't broke, don't fix it" scenarios, because they
are broke by default.