There's no field in the GUI to mark which pins are supply and ground. For RAM/logic parts the tester clearly powers the right pins on its own, but with a type=6 part (an 18-pin MSM6242B RTC, VDD on pin 18, GND on pin 9) I'm measuring pin 18 floating — average wanders between ~0.9 and 2 V, never a steady rail. So: how does Vetus decide which pins get VDD/GND for "other"? Is it derived from pin count, from type, or is power simply not driven for "other" unless specified somewhere?
I'll go over how it works. Also, I wouldn't leave a floating pin. All unconnected pins should be set to HI-Z. Power is given just like a HI/LO pin. It was easier that way because it doesn't add any new opcodes. Here is how the setup works.
MCU --> shift registers --> 0 or 1 sent to individual level shifter/buffer. So all signals are really buffered through all those little 6 pin components. the STM32 has no control over the voltage. That is all on the 5v supply and the 74LVC1T45s All of the capacitors are there to steady the signal on that side and hopefully stabilize before going through the ribbon cable. So if you leave it floating who knows what you could get out of the buffer. it might be turning off and on thus giving you the fluctuations.
What happens is when you change a pin setting. The unit sends the proper binary pattern through the shift registers to set the pins where you want them. Before doing so it sets all pins to HI-Z then shifts them, that way there is never any chance of a short. That being the case, there are only a few possible culprits. But for sure leaving a pin floating is generally not good for this.