Feasible Temperature Drop across a Radiator

[LeadByExample]

Hi all,

This topic is sort of the result of another post started by me.

My question is as follows:

In a domestic central heating system, what temperature drop across a radiator is taken as standard and what is the maximum feasible drop?

Or in other words what drop could be achieved.

If the answer is conditional, add the condition (i.e. if it is only valid for Combi Boilers or very large homes in isolated areas etc. etc.).

I have read 11°C and 12°C were 'standard' but I have also come across 10°C.

Regards

PS.
Please note this is the drop across the radiators, NOT the drop in temperature between flow and return at the boiler! Which should be (around) 20°C, I believe.

[SimonJohns]

Anywhere from 10-20c does me good

[doitmyself]

In a domestic central heating system, what temperature drop across a radiator is taken as standard and what is the maximum feasible drop?

BS EN 442 specifies a drop of 10°C; however installers tend to use 11°C. The reasons for that are historic, and mathematical.

When the UK lived in the dark ages and measured temperatures in °F, boilers were designed to have a differential of 20°F. on the Celsius scale this is 11.111111°C (recurring), which was rounded for simplicity to 11°C.

This is a bit of an anomaly as BS3528, which preceded BS EN442, specified flow 90°C and return 70°C, ie a 20°C differential!

What drop could be achieved.

This will depend on the design of the system. There is nothing to prevent you achieving a 20°C drop across the rads if the system has been designed with this in mind.

[steadyon]

I'm probably being thick, but I don't see how a temperature drop of (say) 11 degrees across each radiator in a system can give a boiler flow / return differential of 20 degrees.

[SimonJohns]

I'm probably being thick, but I don't see how a temperature drop of (say) 11 degrees across each radiator in a system can give a boiler flow / return differential of 20 degrees.

I questioned it for years in college. Never had a good answer so gave up on it

[doitmyself]

I'm probably being thick, but I don't see how a temperature drop of (say) 11 degrees across each radiator in a system can give a boiler flow / return differential of 20 degrees.

The problem is the different flow rates required. A 20kW boiler running with a 20C differential requires a flow rate of (20 x 60)/(4.18 x 20) =14.35 litres per minute. 20kW or rads running with a differential of 11C requires a flow rate of 26.1 lpm, which is nearly twice as fast.

If you connect them together directly you have a problem - how do you double/halve the flow rate? The answer is: you can't - unless you decouple the boiler circuit from the radiator circuit. You can do this either by using "closely spaced tees" or by using a "low-loss header". The boiler circuit is driven by one pump and the radiator circuit by a second pump.

[jonmanty]

The design temperature differential in a heating system between flow and return is normally the same at all points in the system unless something special is going on such as underfloor.

Conventional sizing uses a temperature drop of 10-12 degrees, however with condensing boilers this kind of design drop will not achieve maximum efficiency as the return temp will be too high to maintain the boiler in condensing mode during operation. Over the years there has been much debate on this subject and there is a move now towards a 20 degree differential with condensers - the main point here is that the flow rate significantly reduces with a higher differential which can then take advantage of smaller pipe sizes. On the negative side heat emitters need to have the correct correction factors applied which will result in a sizeable increase in size to take account of the lower flow rate.

[LeadByExample]

Hi all,

Thank you very much for your input, it was very informative.

This will depend on the design of the system. There is nothing to prevent you achieving a 20°C drop across the rads if the system has been designed with this in mind.

Conventional sizing uses a temperature drop of 10-12 degrees, however with condensing boilers this kind of design drop will not achieve maximum efficiency as the return temp will be too high to maintain the boiler in condensing mode during operation. Over the years there has been much debate on this subject and there is a move now towards a 20 degree differential with condensers - the main point here is that the flow rate significantly reduces with a higher differential which can then take advantage of smaller pipe sizes. (...)

This sounds hopeful, although I don't think a 20°C drop across the radiator is feasible, it would indicate I have a lot more margin to 'play' with. It indicates I could even 'downsize' our choice of combi boiler as a lower flow rate is required. Especially if the condensing only kicks in at a 20°C differential at the boiler.

Can I safely assume I could achieve a drop of 14.5°C without many problems?

To jonmanty

(...) the main point here is that the flow rate significantly reduces with a higher differential which can then take advantage of smaller pipe sizes. On the negative side heat emitters need to have the correct correction factors applied which will result in a sizeable increase in size to take account of the lower flow rate.

Could you elaborate on this, as I don't understand to which 'correction factors' you are referring?

How would the lower flow rate require larger radiators?

Or in other words, assuming the mean difference between the water temperature and room temperature remains unchanged, the output (in watt) of the heat emitter remains the same. The only difference being the water releasing more energy per unit volume (i.e. a 14.5°C drop releases more energy than 11°C drop, of course). I understand it means the water has to remain longer in the radiator (by flowing slower) hence the lower flow rate needed, but I fail to see how that would mean larger radiators are required.

Could you give an example with figures to explain?

The design temperature differential in a heating system between flow and return is normally the same at all points in the system unless something special is going on such as underfloor.

I assume you are referring to the individual heat emitters, i.e. every one will have the same differential, e.g. all have a 14,5°C drop etc. If not, can you explain what it is you referring here?

To Doitmyself

When the UK lived in the dark ages and measured temperatures in °F, boilers were designed to have a differential of 20°F. on the Celsius scale this is 11.111111°C (recurring), which was rounded for simplicity to 11°C.

I always appreciate background info, I find it helps to understand better the more I know of how things came to be.

This is a bit of an anomaly as BS3528, which preceded BS EN442, specified flow 90°C and return 70°C, ie a 20°C differential!

Have a look at Radiator standard BSEN442. It appears the standard was changed to attain a higher degree of accuracy and to achieve a standardised set of test conditions in order to have the results comparable with other tests (as they will have been held under similar conditions).

However, of course, this does not mean these testing conditions have to be emulated exactly in domestic situations.

Regards

[jonmanty]

Not really sure what you are trying to achieve here.

When designing a standard system, first factor to determine before sizing radiators and pipework is system operating temperatures.

Conventional system - typically 80 flow, 70 return with 75 mean water temperature

Condensing boiler - if it's to operate in fully condensing mode during all periods of operation then return needs to be 55 or below - typically 75 flow, 55 return with 65 mean water temperature


Radiators are sized from room heat loss calculation normally with a %factor added on for intermittent heating. In order to properly size the radiator the mean water to room temperature needs to be calculated, if this is 50 degrees then the radiator can be selected straight from the catalogue if it's anything else then to size properly a correction factor from the manufacturer table is applied.

So if we've got a 3kW heat requirement to the room at 21 degrees -

a) the conventional system has a mean water to room air temperature differential of 75 minus 21 = 54 degrees and a correction factor of 1.106 from radiator tables (worked out by interpolation) can be applied to establish the radiator size from the manufacturer table

So 3kW/1.106 = 2.712kW selected output from the radiator catalogue

b) the condensing system has a mean water to room air temperature differential of 65 minus 21 = 44 degrees and a correction factor of 0.847 from radiator tables (worked out by interpolation) this can then be applied to establish the radiator size from the manufacturer table

So 3KW/0.847 = 3.541kW selected output from the radiator caalogue

The pipework is sized using the formula-

Heat requirement = mass flow rate x specific heat capacity x temp difference (flow to return)

a) so with the conventional system the pipework to the radiator has a mass flow rate of

3kW = M x 4.18 x (80-70)

3 = 41.8M

M = 0.071L/s

b) with the condensing system the pipework to the radiator has a mass flow rate of

3kW = M x 4.18 x (75-55)

3 = 83.6M

M = 0.036L/s

So in conclusion when designing flow and return temperatures are normally fixed at all points in the system usually based on boiler operating requirements, radiator and pipe sizes are then worked out based on those set conditions.

[doitmyself]

It indicates I could even 'downsize' our choice of combi boiler as a lower flow rate is required. Especially if the condensing only kicks in at a 20°C differential at the boiler.

Where do you get that idea from? The output of the boiler, in kW will not change.

Condensing occurs when the return temperature is below 55°C, not when the differential is 20°C.

Could you elaborate on this, as I don't understand to which 'correction factors' you are referring?

Look in the back of the Elite catalogue

assuming the mean difference between the water temperature and room temperature remains unchanged, the output (in watt) of the heat emitter remains the same.

But it does not remain the same. Take a nominal 1kW radiator, the outputs will be as follows (assuming 20°C room temperature)

Flow °C - Return°C - Watts
75 ------------- 65 ------- 1000 (Mean water temp 70°C)
80 ------------- 60 --------- 986 (Mean water temp 70°C)
75 ------------- 55 --------- 854 (20°C drop across rad, return at condensing point)
65 ------------- 55 --------- 741 (10°C drop across rad, return at condensing point)

Have a look at Radiator standard BSEN442.

Known this for years! Grandmothers and eggs again.