Physically it'll be fine - modern RCDs work by detecting an imbalance between L and N currents and that'll be the same whether the escaping current finds its way back via the TN earthing system, a TT electrode or even directly to the general mass of the Earth.

Regs compliance wise it's less clear. There is (to me a rather odd) requirement that earthing systems in TT systems should 'correspond' to protective devices (RCDs):

411.5.1 Every exposed-conductive-part which is to be protected by a single protective device shall be
connected, via the main earthing terminal, to a common earth electrode. However, if two or more protective devices
are in series, the exposed-conductive-parts may be connected to separate earth electrodes corresponding to each
protective device.

which makes it sound like your setup might not comply (but still it's not entirely clear as half of what's protected by your RCD will be TN rather than TT). I suspect that 411.5.1 hails from the days of voltage-operated ELCBs where such an arragement makes a lot more sense.

   - Andy.

Thanks for the reply AJewsbury, I do speak and read English but often the regs (such as the 411.5.1) really confuse me how they are written. I read 411.5.1 and think well, there wont be any exposed conductive parts in this installation, the only conductive parts I can think of are the SWA glands which will be covered by their shroud, and possibly the water heating element which is not exposed ! There is the earth rod itself and its connection to the earth cable but this will be in a pit with a lid only removable by use of a tool; all the switchgear & equipment is class 2.

I had a hunch that applying a single RCD to both TN and TT systems wouldn't sit well with the regs, maybe 411.5.1 confirms that but im not sure how to decipher it. I wonder if installing an additional RCD main switch at the outdoor enclosure would be better in this instance, even though it would be in series with the upstream RCD and offer no discrimination. As I said a trip would not necessarily cause any major inconvenience anyway. 

Easily sorted. Just install another  new 30mA R.C.D. in a suitable enclosure near to the hot tub location. Job done. You then have two additional  lines of defence against electrocution. It has been suggested that R.C.D.s can fail in use, and not operate correctly when really needed.

Z.

In this case the exposed-conductive-part is the conductive hot tub water.

Z.

I do speak and read English but often the regs (such as the 411.5.1) really confuse me how they are written.

You're certainly not alone there! The regs have been around in some form since 1882 - so we're having to contend with about 140 years worth of 'amendement by committee' as well as attempts to merge in requirements from other standards (e.g. IEC ones) that might have been originally written from a different standpoint or using a different style of language. Often definitions don't keep up with changes in technology or accepted practice so we end up with an increasing amount of 'interpretation' to make sense of things.

Take the definition of an exposed-conductive-part for instance - on the face of it it would seem only to apply to things than can be easily touched - yet we normally consider things like flush steel back boxes or the steel armour of SWA cables to be exposed-conductive-parts as their connection to Earth is fundamental to providing protection against electric shock by automatic disconnection of supply. With SWA for instance, an armoured cable is normally selected because it's in a situation where damage to the cable is foreseeable - mechanically the armour alone isn't sufficient to keep things safe from all damage - nails or digging tools can easily displace the individual wires and penetrate through to the live conductors - so we earth the armour and arrange the protective devices so that if something conductive does penetrate the cable then the supply is cut off. Even then the armour itself isn't often directly exposed but the nail or garden fork that's in contact with it will be - so from a shock prevention point of view the result is the same as if the armour could be touched directly. Likewise with flush back boxes - if the box was allowed to remain live after a fault, then the (possibly slightly damp) plastered wall a few mm away will be at a similar hazardous voltage (often referred to as "live wall") - so pose a risk of shock in the same way as if the box itself could be touched.

But, perhaps paradoxically, we do rely on the outer plastic sheath of SWA cables and the boots over glands to keep those earthed parts out of reach in normal conditions - e.g. when we're running a TN cable into a TT zone - where otherwise simultaneously accessible parts would be connected to different earthing systems. So perhaps we should have some sympathy for the difficulty in coming up with a simple definition that covers all situations nicely.

As for a practical solution, I'd agree with Zoomup - adding an extra RCD as the transition point from TN to TT avoids any doubt, they're not particularly expensive these days and the duplication has the benefit of reducing the risk from a sticky/faulty RCD considerably. There's no loss of discrimination since nothing would loose power as a result of the new RCD tripping that wouldn't have lost power anyway when the original RCD tripped, if anything the situation might be slightly improved since there is a small possibility of the new RCD tripping first, leaving the upstream RCD still on.

The transition from TN to TT does require some care however - a fault from a pre-RCD line conductor to the TT earth won't usually result in disconnection but will raise the TT earthing system to 230V and that dangerous situation can persist indefinitely. So where there's no upstream RCD, or you prefer not to reply on it, then all the live wiring upstream of the RCD within the TT area needs to be protected by double or reinforced insulation (or equivalent).

   - Andy.

Thanks for the thoughtful and in depth response guys, im soaking up the info like a sponge so its not fallen on deaf ears. Yep my line of thought without really going in depth was to fit an RCD at the TN - TT transition point, and at least my reasoning has been confirmed, which is good. Thanks agin