Which is safer/safest pump style in a garden 'splasher style' swimming pool?

The RCD and AC 240v option however, has an earth at the motor, so theoretically, a failure will cause an earth fault and trip the RCD, but, if for some design reason, the earth isn't affected by the mechanical fault and water gets to the live, the pool is then electrified to 240v and then RCD does not trip. Typically, this is accounted for by bonding the pool water to the earth, though it's not as I understand it so trivial, since if you do that, you expose the water to potential fluctuations in the earth voltage (TNCS supply). To avoid that, you'd then convert the supply to TT setup.

"Bonding" (or Earthing) water isn't simple. The basic issue is that relatively clean fresh water has considerable resistance - even a few cm of 15mm plastic pipe can easily come in at several tens of kΩ. The size of the pool might in effect give a much larger c.s.a. and hence lower resistances, but often the contact area - either to the bond, or to the fault - is limited - which tends to mean that there's a considerable resistance that can't be avoided. Even at 230V, 7.67kΩ would be enough to mean a 30mA RCD wouldn't reliably trip.

Then there's some fundaments of ADS to worry about - the usual disconnection times are only expected to protect around 95% of the population under normal dry conditions - the remaining 5% may have lower body resistances and so be more vulnerable to a given voltage for a given duration. Immersed in water, probably a lot more of us would fall into that category. RCDs are usually faster than minimum requirements - but not necessarily that much faster under all circumstances, even when working to specification. Then there's the issue of RCD reliability - a number of studies suggest that around 7% RCDs in service are faulty (either fail to trip at all or don't trip within the specified time).

As Mike suggests, the water resistance effect can be used to advantage though - e.g. put the pump in a separate container away from the main body of the pool, connect it using relatively long and thin insulating tubing, and the water resistance then helps rather than hinders (both in terms of voltages from local faults and any PME earthing).

   - Andy.

Thank you, that's a really informative response. I wasn't aware of the figures for RCDs and ADS though I understand that RCD performance varies between seemingly identical devices; due to manufacturing tolerances I guess.

With regards to your mention of the length of the pipes and csa impacting the RCD trip times. I'm interested because if I were to bond the pool, and I don't want to, I'd bond it (in this scenario) with either a short length of earth wire trailing through the return pipe, or perhaps a ring of copper sheet snuggly inside the end of the pipe, entering immediately after the pump, that way resistance between pump and the earth wire is minimal. This way my logic would be that the rcd should trip before the current gets to the main body of pool water. Being really really reluctant to bond the water and convert to a TT supply with grounding and the associated testing etc, I've shyed away from this idea, not investigating just how long the pipework would need to be before and after the pump in order that an RCD would certainly stop the current before it moved into the pool.

Thank you, that's a really informative response. I wasn't aware of the figures for RCDs and ADS though I understand that RCD performance varies between seemingly identical devices; due to manufacturing tolerances I guess.

With regards to your mention of the length of the pipes and csa impacting the RCD trip times. I'm interested because if I were to bond the pool, and I don't want to, I'd bond it (in this scenario) with either a short length of earth wire trailing through the return pipe, or perhaps a ring of copper sheet snuggly inside the end of the pipe, entering immediately after the pump, that way resistance between pump and the earth wire is minimal. This way my logic would be that the rcd should trip before the current gets to the main body of pool water. Being really really reluctant to bond the water and convert to a TT supply with grounding and the associated testing etc, I've shyed away from this idea, not investigating just how long the pipework would need to be before and after the pump in order that an RCD would certainly stop the current before it moved into the pool.