To be honest I'm struggling a bit to reconcile your sketch where 9 ends got to neutral and 3 ends go to each phase with what appears in  the photo,  I presume you have used meter and convinced yourself it is OK.  Obviously the neutral is not needed unless during the operation of the SCR controls, one phase may be cut off while the other 2 remain on - but for proportional controls where there needs to be a part way on as well as a fully on and an off,  this is quite possible so the neutral may well be needed.

However, to set your mind at rest, 4mm2 copper will be good for a lot more current than normal 37A or so that is the PVC covered rating, when insulated like this. The down side is that what it connects to may not be that keen on the heat . Ceramic choc bloc may be your friend here where you transition to larger cable but in something less heat resistant, though that example I linked to may not be quite enough.

You are in the right ball park with 80-88 amps per phase when running, so you need  16mm or more cable for the feeder and 100A or higher breaker. I presume you have gone up a bit for length and voltage drop or routing environment  reasons.

How hot is this wiring box expected to get during operation ? - I assume they have not used high temperature wire for fun, so you need to do your connections either in something equivalent, or use those tails and have a cooler junction box nearby but safely  out of the worst of the heat.

More questions than answers I think but keep us updated.

Mike

I too am a little confused by the photo - I suspect I'm being mislead by the assumption that the capped terminals are line and the uncapped ones N - if it were the case that L1 was uncapped and L2, L3 and N capped it almost makes sense - but without knowing for use where the ends of each element appear, it's difficult to be sure.

If we can assume that the 4mm² only need fault protection (i.e. the individual elements can't overload, or short to case in a way that'll be detected and disconnected - e.g. by upstream RCD) then it's just a matter of checking the cable withstand against the let-through of the protective device. The devil will be in the detail for that one (possibly unusual values for 'k' as well as needing manufacturer's data for the circuit breaker) but it may well be OK.

The upstream SCRs will likely alter the earth fault loop impedance - so it might be worth double checking how you're providing shock protection downstream of them. (In practice it may well be by disconnection by magic smoke escape from the electronics) but a more formal proof might be needed.

   - Andy.

The 'safe' current handling for silicone covered wires is movable feast. If we want to keep the copper below 70C the normal PVC tables apply, but if you do not mind running the copper at 150C or so, then the rating is considerably higher as the attached table shows.

The chart is based on US wire sizes and the highlighted line is not the 4mm one, but for example shows 35A for a wire of a touch under 1.5mm2 - something we would not normally consider with PVC or XPLE the wire nearest to 4mm2 is rated at 110A. For the wire to be damaged by a fusing event it probably needs to rise considerably higher still.

Mike.

Thanks both for sharing your thoughts.

so yes my sketch is a little busy and confusing at the bottom of the page  (I was rushing!) 

I tried to show that I will have 

3 x heater L1’s connecting to my main supply 35mm2 L1.

3 x heater L2’s connected to my 35mm2 L2

3 x heater L3’s connected to my 35mm2 L3

3x heater N’s. connected to my 35mm2  N

(And yes, if all 3 elements are equally balanced then 0 Amps in the neutral.) 

and I’ve had to go to 35mm2 for my main supply cable due to the nature of this particular installation (ambient, grouping etc etc) 

I have finally had feedback from the heater manufacturer who confirm everything in their terminal box (the photo) needs to be rated to minimum 125C as that’s how hot it can get in there.

I’m struggling to find suitable connectors. thanks for the link to the ceramic ones but I’ve decided to bite the bullet and replace the existing wires with longer heat resistant wires that I can then terminate in an adjacent terminal box at a much cooler temperature 

my concerns were based on having 3 sets of heater wires at only 4mm2 where my upstream MCB is rated to protect the 35mm2 supply SWA… I’m effectively creating 3 new circuits with smaller wires. Yes the current in each cable is 1/3rd so I’m not concerned about the wire current rating but nevertheless the MCB is still 125A! I just don’t know if that is good practice? I guess the 4mn2 cables are  very short. I thought maybe I should be putting in additional branch protection eg 3 x 32Amp breakers in the junction box but that seems crazy. 

Yes it took me a while to study the photo as I was also very confused! Oddly they have ceramic caps on L1, L2 and N but not on the L3s …?? … Maybe it is a space constraint thing.