On Wed, 03 Nov 2004 18:32:03 +0000, Andy Hall wrote:

On Wed, 3 Nov 2004 16:24:51 -0000, "Christian McArdle"
wrote:

Yes, sorry if this was unclear. Some boilers have a facility to
continuously
vary their output..

Indeed, but they don't really need a continuously variable input from the
hot water cylinder. The additional benefit of knowing this information to a
control system is marginal at best.

Actually there is some value.

In the past I've used bimetal type thermostats, more recently
thermostats with temperature sensor and electronic control and
currently an arrangement with temperature sensor inside the cylinder
in a pocket.

The bimetal type I found to have quite poor response and taking a
while before a rapid temperature drop is detected and hence firing up
the boiler. Moreover, the hysteresis was around +/- 5 degrees - 10 in
tota - for slow changes in temperature.. So, with a nominal set point
of 60 degrees, when a shower was started or the bath run, with
hysteresis plus poor response, the boiler wouldn't come on until the
cylinder surface temperature was at just over 50, then on the way up,
the boiler wouldn't go off until around 67 or 68 degrees. Accounting
for all of these things the response window is close to 20 degrees.

An electronic type(e.g. TP75) does rather better because although it
is a switch on/off type, the hysteresis was smaller (+/- 3 degrees,
IIRC) and the probe gives a faster and more accurate response as well.

In my current system, the sensor is wired directly back to the boiler
and the controller measures the temperature (and displays it).
The HW temperature is set to a nominal 60 degrees as before.
I've monitored the behaviour with a PC connected to the boiler and the
response is quite intelligent.

If the water temperature has been drifting slowly down through small
amounts being used, nothing happens until it drops below 55 degrees -
i.e. 5 degrees below set point. At that stage, a hot water cycle is
initiated with the boiler being wound up to full power. As the
temperature approaches the 60 degree set point, the burner and pump
are modulated down so that there is no overshoot of temperature. The
controller appears to learn the characteristics of the cylinder load
because the behaviour improves over a few cycles following a reset.

If I start a bath etc. and there is a more rapid drop in temperature,
then the HW cycle is started earlier - typically as the temperature is
falling through about 58.5 degrees.

The result is that even with some stratification, the water
temperature is maintained within a narrower temperature range than
with the simpler thermostat.


Due to stratification, the tank doesn't so much heat up uniformly, as divide
into a hot and a cold zone, the boundary between of which moves down as
heating progresses. A single thermometer will show a very rapid move from
cold to hot as this boundary passes its installed location. This is why it
is difficult to provide good hysterysis from a single cylinder thermostat.
You actually need to sense the temperature at different heights to get a
feel for the amount of hot water available, not know the temperature at one
height.

I've found the results to be remarkably consistent with the analogue
sensing by the boiler. It is monitoring temperature accurately as
well as rate of change so seems to do a good job of overall control


I'm guessing but it looks like the boiler control system may well
implement a PID (or even just PD) type controller.

--
Ed Sirett - Property maintainer and registered gas fitter.
The FAQ for uk.diy is at www.diyfaq.org.uk
Gas fitting FAQ http://www.makewrite.demon.co.uk/GasFitting.html
Sealed CH FAQ http://www.makewrite.demon.co.uk/SealedCH.html