Buffers and ASHPs.............again

In the last 12 months following advice gathered from here and additional research in the industry we have not installed ASHP’s without a buffer.

Wherever possible we err towards Scandinavian units like Danfoss and Nibe, but costs and coverage sometimes mean that we do end up specifying inverter driven units, usually from Daikin. We have monitoring via in line meters in place on 3 units at present, 2 Daikins and a NIBE, and at present the 2 Daikins are outperforming their simulations and it’s too early to tell with the NIBE given the “wetness” of the build. We are adding more meters to installs in the new year.

Ignoring the argument about whether inverters are really required on ASHP’s we are left in a situation where in order to be competitive we have to be able to offer an inverter product due to their low cost in comparison with the Swedish alternatives but won’t install without a buffer.

However this can be an issue when using a buffer, because Daikin and Mitsubishi do not have thermistors to control the space heating side that can be situated in the buffer to weather compensate into it.

The Mitsubishi can have a buffer but the HP side pump and house side pump have to be interlocked via the air stat to be both called on at the same time as the heat pump. The air stat then switches them all off when satisfied so the buffer only heats during a demand from space heating.

Daikin’s can set be up so that the HP is independent from the house side of the buffer and pump, but has to sample via engaging the pump for something like 3 mins in every 5.

Our concern is that on the Mitsubishi the buffer is really only there to improve defrost and guarantee flow rates as opposed to max compressor run times, although the extra mass/volume will help to a degree.

On the Daikin the sampling pulls from the lowest/coldest part of the buffer and therefore does not benefit from being held off for longer as would be the case in a Scandinavian unit with a thermistor in the centre of the buffer. In addition there is a risk that the pump sampling will upset the stratification anyway, by mixing the contents. However if we work with 30 litres a kw the volume should reduce this impact to a degree.

Neither Mits nor Daikin are keen on the idea of relocating their control (flow) thermistors to the buffers.

There is little we can do to change the Mitsubishi set up, so our quandary is do we fit a small buffer (say 100 litres) to assist defrost and flow rate, on the basis that there is a remote chance that the air stat may switch out before the buffer reached temp if it was sized at (20-30 litres per kw).

On the Daikin, it would seem that the sampling via the return does mean that the HP should run for longer because it has to switch out via flow temp being met via the return from the base of the buffer, but it will then cut back in sooner because the sampling comes back direct from the base of the buffer/house. Conversely a Scandinavian unit will run shorter to satisfy the centre point thermistor but hold off for longer.

We may be obsessing too much about this (it wouldn’t be the first time), but would be interested in others views of how to integrate buffers best with inverter units. Ideally what we are looking to achieve is lowest ramp down, longest run time and biggest off periods, but suspect we will have to achieve a compromise somewhere!

One idea we have thought about is setting up the Daikin to sample from maybe the bottom third of the buffer either by speccing a bespoke buffer or tapping a Gledhill with an Essex flange, we would still have the mixing issue to consider though.

We are going to look at the Panasonics in the New Year and also NIBE’s new inverter although the latter is unlikely to be at a price point that makes it worth speccing ahead of a Daikin or Mits.

There is a very real risk that I am too close to this issue to see flaws in my arguments so please feel free to put me straight, we just want to get the units set up as best we can!