That doesn't work - you vapourise the sharpie/substrate but the copper remains. The sharpie doesn't spend enough time getting hot before it's turned to vapour.

Laser cutting doesn't work (BTDT) because copper is too good a reflector of the laser light and absorber of heat. That was with a 40W CO2 laser anyway - no doubt you can vapourise it with sufficient power.

Indeed, pen fault detection devices ( example ) monitor the L-N voltage to be 230V +/- 10% which drives a 3-pole contactor to disconnect all three terminals if the voltage is out of spec.

It seems unlikely they would be used for testing the LCD output directly, as there's no pins to read the segments on each digit. But they could be the MCU's programming interface like JTAG or similar (pull a pin high/low at power on time to enable programming mode) and the pins are GPIOs in normal use. They could also enable a debugging / calibration mode in firmware by having a routine to detect a specific pattern on them when the firmware starts.

To note there are MP3 player chips like the VS1053B which can play MP3 audio piped to them. That takes care of having to do the MP3 codecs in a microcontroller and the audio output. You might be able to pair this with an Arduino or similar Bluetooth project to reuse existing software components.

The transformer in circuit 2 is drawn with a larger secondary just to fit on all the different windings without having them unreadably close together. Circuit diagrams are about conveying information, and here they have to make the secondary bigger to fit it all on. That trumps anything about relative sizes of the windings.

Another option is an active two-port analogue switch chip (which are available in versions designed for specific protocol signalling, eg USB, PCIe). The switch is controlled by a pin on the plug-in connector. Pulled high by default, it goes straight on. Grounded it goes to the connector.

@martij7 Looks like I can see BS EN60898-1:2019 (track changes version). There's no mention of Z-curve or K-curve circuit breakers in the standard like there are B-curve, C-curve and D-curve. (The standard calls them 'B-type' instead of 'B-curve'). Section 5.3.5 lists the different types and their instantaneous trip characteristics.

One of the J2 pads is already grounded. If you mix up your 'to raspberry' connections you'll connect the ground of your circuit to the pwr_on pin. Assuming your circuit is grounded elsewhere it will force pwr_on constantly to ground and be like holding down the power button all the time.

How 'high voltage' is HV? In the supply industry 240V is 'low voltage'. If you were to have a positive rail short to the case, would the PSU(s) shut down safely (ie standard short circuit protection in the PSU)? If so, I can't see a safety problem. There are also various questions about extraneous current paths (ie return current coming via the earth and not the ground rail, making a larger inductive pickup and collecting interference).