[mindslight]

But keep in mind many times electrical conduit is not used as a bona fide conduit, but rather more of a physical convenience (note how the bit you quoted is talking about conductors not cables. So if you bury a length of PVC pipe (grey or white) with a cat6 network cable plus a UF power cable (required for the wet location, regardless of the pipe), you're fine.

Although residential in the US rarely (never?) goes above 170 volts (peak), which is more forgiving than the ~325V of most of the world.

[sokoloff]

I believe that a grey non-metallic conduit piece would meet the definition of a raceway in article 100 (to which the insulation rating rules would apply). A white PVC pipe is not “expressly designed” for that use (and would therefore clearly not be a raceway).

Article 100 defines a raceway as “An enclosed channel designed expressly for holding wires, cables, or bus bars, with additional functions as permitted in this Code”

[jaclaz]

At least in Italy, that would apply as well.

I.e. you can have in same conduit a "protected" (double insulation) mains cable (the type is called FG16 now, it was FG7) and a low voltage/signal cable.

The idea is that the FG cable in itself, having the external insulation besides the single wire insulation and being suitable to "unprotected" installation can coexist with signal cables, the conduit in this case is only an added mechanical protection.

The type of certified 400 V insulated cables I mentioned earlier is instead allowed to coexist with "normal" single wires (this is the normal way electricity is distributed in buildings) inside a conduit.

The issue using FG cables (besides the fact that it is way harder to be inserted in small conduits) might be that it would be suitable to power (say) a mains receptacle (live+neutral+earth) but wouldn't be suitable for (still say) a diverter or a reverse switch due to the colour coding of the wires.

[cptskippy]

A cable is at least one wire which is a conductor, they use the term conductor because the standard applies to more than wires and cables.

Claiming a conduit is just a convenience doesn't change anything.

Where did you get the 170/325 numbers? In residential applications it's 220-240v for most of the world. North America uses split phase for most outlets but major appliances still run on 220-240v.

[mindslight]

I think it's going to come down to the AHJ, or your own judgement if there is no inspection. The way I've seen it explained on electrician forums is that you're not using the piece of PVC for its properties as conduit under the electric code (like say outer damage protection/containment for running single THHNs), so it isn't considered a conduit. But it would obviously be wrong to apply this argument to things like conduit fill or abrasion protection - regardless whether the pipe was manufactured intended for electrical conduit or not.

While they call mains voltage 120V, that's actually RMS voltage and not peak voltage. The peak is 170V, which I think is a better gauge for thinking about insulation and safety (by the time power is being resistively dissipated, you've already lost). Residential "240V" circuits use opposing legs of a split phase, so nothing is more than 170V from ground, which is what matters for insulation and most failure modes. You're not going to get any more of a shock from a "240V" residential circuit than a "120V" one unless you manage to touch both ungrounded conductors at the same time - it's the arc flash risk that gets worse.

Although now that I'm really thinking about it, maybe mixing insulation types is not foolproof at the extremes. If there is a wire with 170V inside sitting in free space, the worst case assumption is that it can have 170V on the outside of its insulation as well (due to the parasitic resistance and capacitance across the insulation, and ignoring the parasitics to ground. If this sounds strange to you, think about how those non-contact voltage testers can work). So if you put a grounded conductor with a low insulation rating right next to it, that insulation could have voltage across it higher than its own rating (depending on its parasitic R/C), which may cause it to break down over time. The possibility is likely moot with real world values, but still.