[OpenTRV-dev] TRV control - my problem, and musings so far

Tue Apr 2 10:53:43 BST 2013

Hi Guys,

Having introduced myself I thought it might be prudent to put down the
"problem" that brought me here and the thoughts I've had prior to arriving
here as I've been kicking this can down the road (in my head - no hardware
here yet!).

I've had a quick browse through the list archive and I'm sure a lot of
this has probably been covered time and again. Apologies for that, but
it's offered in the hope that there might be the odd pearl of wisdom
waiting to be extracted and added to the wish list!

My "Problem":

In short, we are a couple living in a house with 14 radiators. The system
is pretty conventional. House is about 15 years old and the heating system
comprises a non-condensing boiler, HW / CH zone valves, Drayton TRV4s on
all but one radiator and room and hot water tank thermostats, fed by a
conventional programmer.

Being that there are only two of us we are wasting a fair proportion of
the heat we are using, since a maximum of two, and usually only one, room
will be occupied. Sometimes we'll spend the evening in the lounge watching
TV, sometimes most of it in the dining room having a meal, or in the
kitchen cooking it, often just in the study on our PCs. Upstairs, no room
is generally occupied until we go to bed. Sometimes, we're out all
evening. Despite this, all rooms are heated regardless, morning and

There are predictable elements to this. In the mornings only the bedroom,
bathroom, hall, landing and kitchen are used. In the evening, the upstairs
needn't be heated until late so it's warm by bedtime. There are also
unpredictable elements, such as which particular downstairs room we occupy
in the evening, if we're not out.

I'm also conscious that, when the system shuts down, there is a lot of
heat in the pipework and boiler that might be usefully directed to a part
of the house that would benefit from it or into the hot water tank, rather
than left to dissipate, as there is no pump over-run of any kind, other
than if the boiler overrun thermostat trips, which circulates water
through the bathroom towel radiators only as the zone valves are closed by

So, that's the problem I see. A couple of solutions I've considered:

1) Defeat unwanted zones on a time basis.

I've yet to test this theory, but I reckon a small heat source attached to
a TRV will cause it to close, or at least reduce the set temperature
significantly enough to be useful. The TRV4 has a large metal bulb as its'
sensing element, meaning that a resistor could be coupled with it quite
easily and very, very cheaply to apply a little heat. Feed this from a
timeswitch and we have a cheap and easy if crude solution to, for example,
avoid heating the lounge in the morning when it will not be used until the
evening. How to achieve this with a good WAF might need further study!

2) Control zones using an in-room "smart" device

Pretty much what's proposed here. I came across the Eberle TS 5.11 in
Conrad. It's a thermoelectric replacement TRV head operated by mains power
(although I believe lower voltage variants are around). Each room would
have temperature and occupancy detection and a heating schedule, with
set-back or "comfort" temperatures selected accordingly. I have mains
power near every radiator I want to control, so energy use of the head is
not an issue. I'd mused on intercepting the burglar alarm PIR circuits to
glean occupancy information. Obviously OK on a one-off basis but not
likely to appeal for a "consumer" product!

3) Replacement programmer

Something to replace the existing "dumb" heating programmer, perhaps using
occupancy information to determine whether to skip a CH and HW programming
period entirely, giving a web interface to allow monitoring and
programming changes to be made when around the house or even when away
from home. Control zone valves, pump and boiler independently to allow
intelligent use of heat after boiler shutdown. I've got both an AVR board
with Ethernet interface and a Raspberry Pi which I've bought as potential
controllers here.

4) Integration and control strategy

I'd mused that each "zone" could transmit a figure each time it "woke",
perhaps based on a comparison between current temperature and set point,
indicating the local level of "desperation" for heat. The programmer would
sum these and use it as a criteria to determine if it's worth lighting the
boiler. In this way, a single zone dropping a fraction below target while
the rest are happy wouldn't result in a potentially wasteful short cycle
of the boiler, yet a number of zones in such a state would, as would one
zone that's well below target or where the set point has recently been
increased. Each zone would receive a response from the programmer,
indicating if heat is being supplied, and also if a pump overrun condition
exists, allowing zones to open and take "waste heat" based on some

Anyway, that's a brain-dump of my ideas so far. As said, probably mostly
old news, but maybe an idea or two that might be useful?



More information about the OpenTRV-dev mailing list