Drill through beam, pull through conduit, or what?

[endecotp]

Photo shows my bathroom ceiling. Note the old ceiling rose location, fan, lights above mirror (reflected in the mirror), and huge boxed-in steel beam. This is a flat; I have no real access to the room above. The ceiling is plaster on expanded-metal lath, with awful “deafening material” above.

I am trying to get individual control of the fan, mirror lights and main lights. I posted previously about pulling additional wires through the conduit from the switch, which I have now done.

Currently there is steel conduit from the ceiling rose to somewhere behind the fan, containing 3+E. From the fan, discrete surface-mounted trunking takes twin+earth to the above-mirror lights.

To separately control the above-mirror lights, my options seem to be:

1. Feed an additional single down the conduit from the rose to the fan. I’ve attempted this but failed.
2. Use the existing 3+E to pull through four singles. Access to the fan end of the conduit could be tricky, I think I may need to remove some wall tiles.
3. Use the existing 3+E to pull through a 5-core flex. Ditto.
4. Run a cable by a new route from the rose to the lights. I guess I would drill through the web of the beam! Could I do this through an easily-repaired “keyhole” in the boxing, with a long drill, or would I need to remove enough boxing to get up close to the beam?
5. As an electronic engineer by training, I am tempted to multiplex the fan and light control over the available wires using some collection of diodes and relays. I am discouraged by the fact that some of it would end up inaccessible.
6. Something involving “smart” devices.

What would you do? Any other options? In particular, what do you think about routing a new cable through the boxing-in?

Thanks.

 

[JohnW2]

..... 4. Run a cable by a new route from the rose to the lights. I guess I would drill through the web of the beam! Could I do this through an easily-repaired “keyhole” in the boxing, with a long drill, or would I need to remove enough boxing to get up close to the beam?

I think you would find that drilling a hole through a "huge" RSJ would be a major challenge even if you established very good access (lots of 'destruction' and repair), hence quite probably near-impossible through a 'keyhole'.
RSJs not uncommonly some with holes in their webs, but you might have to undertake massive de-construction to see if there are any (and might well find none ).

5. As an electronic engineer by training, I am tempted to multiplex the fan and light control over the available wires using some collection of diodes and relays.

One can do all sorts of clever things like that with DC. Indeed, with some electronics one can theoretically control an unlimited number of things over just two wire (or 'no wires', as in [6] below!). However, I assume you are talking about 'mains', which would be much more challenging (unless you created some DC for the 'controlling), probably having to use lots of relays (in lieu of diodes!). In any event, as you go on to say ....

I am discouraged by the fact that some of it would end up inaccessible.

I really think that any such solutions would have to be 'accessible', so perhaps the need for 'access panels/hatches' or whatever!

6. Something involving “smart” devices.

Much as I often voice reservations about some of the 'new-fangled' technologies, there must be some such solutions . Have you considered using Quinetic (or similar) switches? That approach would mean that only L+N (and E) would have to go in cables, the control all being 'wireless'.

 

[jj4091]

As John W2 says perhaps you could use these sort of things mounted discretely behind the beam but I'm just wondering if all the steel work (RSJ and expanded metal ) might affect their operation.

 

[endecotp]

5. As an electronic engineer by training, I am tempted to multiplex the fan and light control over the available wires using some collection of diodes and relays. I am discouraged by the fact that some of it would end up inaccessible.

Having wasted my Easter Sunday trying and failing to run an additional wire, I am now contemplating this scheme. The idea is that each switch controls one half cycle. Then the requirement is that the relay coils are happy with a half-wave-rectified input. For a mechanical relay, half wave rectified 240 V is about 170 V RMS, which I guess is probably enough, but with some odd harmonics; I wonder if it would buzz. For a solid-state relay, I wonder if a half-wave-rectified input would result in a half-wave-rectified output? I.e. if the opto-isolator only turns on for one half cycle, the triac would only fire for one half cycle. Anyone know? Maybe I’d be better off with a DC-input solid state relay with a capacitor on the input.

For the fan, which has a permanent live input, if the switched live input is just a control signal then it might work directly from the half-wave signal.

Feel free to tell me I’m crazy!

 

[flameport]

To separately control the above-mirror lights, my options seem to be:

Of those options the only real one is 2 - pull singles through the conduit.
Attempting to pull 5 core or any other multicore flex through a conduit is likely to fail due to it's large dimensions compared to single wires.

The other option is to use the 3+E that's there, with the switched L for the mirror lights and add a pull cord to the fan such as https://www.edwardes.co.uk/products/ced-wmps2-2amp-240volt-mini-pull-cord-switch-white-

 

[JohnW2]

..... I am now contemplating this scheme. The idea is that each switch controls one half cycle. Then the requirement is that the relay coils are happy with a half-wave-rectified input. For a mechanical relay, half wave rectified 240 V is about 170 V RMS, which I guess is probably enough, but with some odd harmonics; I wonder if it would buzz. For a solid-state relay, I wonder if a half-wave-rectified input would result in a half-wave-rectified output?

Quite clever, and simpler than I has envisaged (because I hadn't contemplated running relays with half-wave rectified input!).

I suspect that a mechanical relay might well buzz with such an input and, like you am not sure what would happen with a solid state one (although I suspect that what you go on to suggest may well be right). Therefore ....

Maybe I’d be better off with a DC-input solid state relay with a capacitor on the input.

... I think I would be inclined to include smoothing capacitors and use (mechanical or solid state) DC-input relays.

I was initially concerned that there might be a problem if both switches were 'on' - but, on reflection, that ought to work as intended.

For the fan, which has a permanent live input, if the switched live input is just a control signal then it might work directly from the half-wave signal.

I imagine so, provided that you used the correct one of the half-wave rectified signals for the fan. I think that the most common arrangement is that the first negative half cycle of the AC input pulls down the voltage (to the 'forward' voltage across a zener) at the (joined) inputs to one gate of a CD4001 - and it takes only that half-cycle to trigger the timer.

Feel free to tell me I’m crazy!

Not at all. The one thing I would say is that, as bernard and myself sometimes say about some of the unconventional 'control systems' we utilise within our domestic electrical installations, it's all very well whilst we are the householders but, beyond that time, 'unconventional' things such as you are proposing (and bernard and myself have already implemented) would probably totally confuse (and/or mystify) a high proportion of electricians who might find themselves working on the installation in the future
Similarly, when choosing components for something such as you propose, one probably needs to consdier than an future electrician might find a way of IR-testing the circuitry with 500 V DC!

Kind Regards, John

 

[SUNRAY]

My experience of the proposal: AC relays will buzz and arc, use some modest smoothing capacitors and ideally DC relays.

 

[JohnW2]

My experience of the proposal: AC relays will buzz and arc, use some modest smoothing capacitors and ideally DC relays.

Indeed - as I wrote ...

I suspect that a mechanical relay might well buzz with such an input and, like you am not sure what would happen with a solid state one (although I suspect that what you go on to suggest may well be right). Therefore .... I think I would be inclined to include smoothing capacitors and use (mechanical or solid state) DC-input relays.

 

[SUNRAY]

 

[JohnW2]

View attachment 379558

How does the fan get its permanent live with that arrangement?

 

[SUNRAY]

How does the fan get its permanent live with that arrangement?

Having the fan individually switched it doesn't require a permanent live.
I assume this arramgement isn't showing one

 

[JohnW2]

Having the fan individually switched it doesn't require a permanent live. ... I assume this arramgement isn't showing one

Maybe I misunderstood, but I took it from what the OP had told us was that it was a timer fan triggered by the rooms lights (but not shown as such on his diagram)- i.e. as below. If that weren't the case, if he had two switches, two loads and three cores then, as you've illustrated, he would not have any need for anything 'clever', would he? ...

 

[endecotp]

Having the fan individually switched it doesn't require a permanent live.

I will still have a perm live to the fan. I had considered not doing so, but this particular fan doesn’t work well with perm and switched live joined together, and I don’t want to replace it.

 

[JohnW2]

I will still have a perm live to the fan. I had considered not doing so, but this particular fan doesn’t work well with perm and switched live joined together, and I don’t want to replace it.

I'm getting a little confused
If it has both a switched live and permanent live, then I presume I was right in assuming that it is a timer fan?
What controls (and/or what do you want to control) the fan - activated by the lights or separately manually switched?

What do you mean by saying that it doesn't 'work well' with perm and switched live joined? It is obviously only the perm live which can actually 'power' the fan - the S/L merely triggers the timer to start, and I can't see how there can be a difference from triggering it with, say, the switched supply tio the light and having it permanently connected to L (in which case the tiner will get triggered as soon as soon as the permanent live becomes 'live'.

 

[SUNRAY]

I was working from the origional post

I am trying to get individual control of the fan, mirror lights and main lights.

In which case I would have thought run on will not be required as it's not controlled by the light switch.



How is the fan going to be controlled as run on and individually?