15M pump required, one large Vs 2 small in series?

[Last Plumber]

Some of the Schools I work in only have 4 meter head pumps installed.
There is more to this than I think you realise.
The Magna 3 is way oversized for your needs in my opinion.

My advice would be to ask a Heating Engineer to calculate this for you. You will need to pay for this spec obviously as it takes time, effort, knowledge and skill but it will work out cheaper than buying a commercial pump.
Don't make the mistake of thinking that oversized is best. It does not work like that and you will end up regretting your decision, an oversized pump will cause issues and is a waste of your hard earned cash.

I hope this helps!

 

[doitmyself]

used figure in literature for boiler, again very small.

Which literature were you looking at? There's nothing about heat exchanger resistance in the 40CDi Regular Installation Manual (I assume you are talking about the open vent version, not the system one). Did you take it from Technical Bulletin 48? If so the resistance of a 40kW CDi hex is about 3.5 metres, which isn't exactly small.

The 30CDi (which should be more than enough, including HW) has a resistance of approx 1.75m at full output.

What temperature differential did you assume when calculating the system resistance?

 

[Michael Groves]

I used the default figure in the sizing resistance in copper tube. It was tiny. 3.5M is much bigger. Yes looking at 40CDi regular, but with Robo kit. I thought 35CDi system was just too tight. In that case I would need 8.3M head.

I think 30kW would be too small, I also have 500L water cylinder with 20kW heat exchanger.

 

[Last Plumber]

I used the default figure in the sizing resistance in copper tube. It was tiny. 3.5M is much bigger. Yes looking at 40CDi regular, but with Robo kit. I thought 35CDi system was just too tight. In that case I would need 8.3M head.

I think 30kW would be too small, I also have 500L water cylinder with 20kW heat exchanger.

500 litres?
Why?
What do you have in the way of Hot water?

 

[Michael Groves]

It's a HMO, 11 tenants.
Currently only 3 showers, but every room has WHB. There is one communal kitchen. I plan to install maybe another 5 private showers in the larger rooms. Ultimately they'll still be 11 tenants max, but you could get a larger simultaneous draw off.
I must admit I've struggled with the capacity for this, how does it sound now with a little more background info?

 

[Last Plumber]

I think it sounds a lot to be honest with you.
I still think you would be wise to get a Heating Engineer to spec this job for you. Boiler size, Pump size and Domestic Hot water storage, any Cold water storage that may be necessary. You would benefit from the design and the advice they will be able to give. It will cost you a few hundred pounds but I think you will save a lot more. Plus you will have the peace of mind that it will be correct, the knowledge that the system is a as efficient as possible and someone to blame if it goes wrong.

 

[doitmyself]

I used the default figure in the sizing resistance in copper tube. It was tiny. 3.5M is much bigger. Yes looking at 40CDi regular, but with Robo kit. I thought 35CDi system was just too tight. In that case I would need 8.3M head.

By "default figure" do you mean 11 deg C as on page 2 (46 in the formula for flow rate on same page)? If so, that is incorrect for your boiler which is designed for a 20 deg C differential. 46 then becomes 84. If you redo your flow rates and resistance using 84, you will find that the head is about 30% of what you originally calculated. That's because resistance varies with flow rate squared.

What was "tiny"? Don't understand your reference to 3.5M and 8.3M.

You also need to take the 20C differential and lower operating temperatures (needed to ensure condensing) when selecting your rads, as rad output will be reduced.

It would have been helpful if you had mentioned HMO in your opening post.

Have you thought about having two boilers, a 17kW for the heating and a 20kW for the cylinder? A smaller heating boiler will modulate much lower and will be condensing more often - if the system is correctly designed. You could also include weather compensation.

The HW boiler needs to run with a flow temp of 70-75C and a 10 deg differential, otherwise the cylinder will not heat up to 60C quickly enough. This means the boiler will not condense.

 

[Michael Groves]

You make some very valid points, I think I need to wind this back to make sure I'm getting correct results.
Calculating head is complicated, I'm now thinking about a single boiler using W-Plan. If I also have a 400L cylinder the heat exchanger is 17kW. This way I can use a single 18kW boiler to run both.
Worst scenario if the boiler can't cope I could buy a second 18kW boiler further down the road and have a dedicated boiler for each.
I recall being told by Worchester a while back you want your boiler running at 100% for best efficiency?

Thoughts?

 

[ShaunCorbs]

Think it will be more than 17kw as normal 300l ones have between 20-25kw coils

Also you just got the one cylinder??

 

[Michael Groves]

Seems to be various heat exchangers, Gledhill 300L has 30kW heat exchanger !!!, I'm looking at Grahams Therma.

I'll just go for a single cylinder.

 

[doitmyself]

W Plan is either heating or hot water; you can't have both at the same time. If it's a HMO, you won't have any control over when anyone wants to wash, bath or shower. So there may be up to eight people showering early in the morning when the heating needs to be on. With W plan hot water takes priority; everyone freezes as the heating is not on. You can time the hot water to recover when heating is not required, i.e. overnight or when most of the inhabitants are at work; but this can't be guaranteed to work all the time.

Have you had the water pressure and flow rate checked to make sure it is sufficient for eight simultaneous showers?

You haven't answered my earlier questions, so here they are again:

1. Where did you get the idea from that the CDi 40 heat exchanger was low resistance (tiny?)?
2. Where does the 3.5m head come from?
3. Where does the 8.3m head come from?

 

[Michael Groves]

Water Pressure is very good in the property, they'll be no problem.
Agreed it's one or the other with W-Plan, it'll take some tweaking to get it right. But remember with HMO, they're not all rushing out the door to get to work, showers are spread throughout the day. For those that wash anyway!
1. I used a default figure from calculating copper pipe resistance for average boiler, this was proven to be inaccurate. My bad.
2. You said from Technical Bulletin 48 it was 3.5M ?
3. I calculated 4.8M using calculations supplied previously. I simply added your 3.5M and got 8.3M

 

[Michael Groves]

Just been reading modern boilers prefer a W or S-Plan where each circuit gets 100% of boiler. Most 300 & 400 cylinders have a reheat time of 38-45 minutes. Therefore giving Hot Water 100% for 45 minutes every few hours should not affect Central heating that much.
I was also reading if you give hot water preference, you can reheat as soon as it drops, so maybe only need 15 minutes whenever someone takes a shower, so central heating should not be affected unless you get several showers in a row on a cold night.
They were also saying you want the biggest heat loss to stop boiler cycling, so again W or S-Plan so boiler always running at max.

 

[doitmyself]

1. I used a default figure from calculating copper pipe resistance for average boiler, this was proven to be inaccurate. My bad.
2. You said from Technical Bulletin 48 it was 3.5M ?
3. I calculated 4.8M using calculations supplied previously. I simply added your 3.5M and got 8.3M

1. Modern heat exchangers have very narrow passageways for the water to pass through, so the resistance is very high. Also, the resistance is not fixed, it varies with the flow rate.
2. The figure of 3.5m only applies when the heat exchanger is running at max output with a flow-return differential of 20C. The same hex running at 20kW, with the same differential, has a resistance of approx 0.8m. This is because the flow rate is halved. See attached chart for the CDi hex.
   
3. The resistance of 4.8m assumed a differential of 11C. Resistance is proportional to flow². The flow rate for a 20C drop will be 55% if the flow for a 11C drop (11/20) so the resistance will be 0.55² = 0.30 x 4.8 = 1.45m.
4. The available head on WB system boilers is 2m, which gives a margin of 0.55 metres.

 

[doitmyself]

Just been reading modern boilers prefer a W or S-Plan where each circuit gets 100% of boiler. Most 300 & 400 cylinders have a reheat time of 38-45 minutes. Therefore giving Hot Water 100% for 45 minutes every few hours should not affect Central heating that much.

Only W Plan can guarantee HW priority. That's because it uses a diverter valve (similar to Y plan but without a mid-position).

Most cylinder thermostats have a differential of about 10C. If the hot water is stored at 60C and incoming cold is at 20C, the cylinder temperature will drop to 50C when 25% of the hot water has been used. That's equivalent to two ten-minute showers for a 400 litre cylinder.

They were also saying you want the biggest heat loss to stop boiler cycling, so again W or S-Plan so boiler always running at max.

That doesn't make sense. Can you post a link to this?

 

[Michael Groves]

Sorry I can't find the article, what I meant was the return temp needs to be much lower than supply temp.

 

[Chuck]

Most cylinder thermostats have a differential of about 10C.

3°C is typical for a domestic mechanical cylinder thermostat.

If the hot water is stored at 60C and incoming cold is at 20C, the cylinder temperature will drop to 50C when 25% of the hot water has been used.

This would only be true if the incoming cold mixed completely with the hot water. It doesn't; cylinders are designed to prevent incoming cold disrupting the stratification so the transition between hot and cold as the tank runs out is fairly abrupt.

That's equivalent to two ten-minute showers for a 400 litre cylinder.

A 400 litre (nominal capacity) cylinder at 60°C will be able to supply about 440 litres of mixed water at 50°C, which is a painfully hot shower. A ten minute shower uses between 60 (eco head) and 150 (power shower) litres of mixed water so you should get between 3 and 8 showers out of a tank depending on personal preferences.

 

[doitmyself]

3°C is typical for a domestic mechanical cylinder thermostat.

Honeywell L641A = approx 10C
Drayton HTS3 = approx 8C
Danfoss ATC = 6-10C

Only mechanical thermostats with a working anticipator have a differential around 3C; and cylinder thermostats do not have anticipators

How do you design a straight sided cylinder to prevent incoming cold disrupting the stratification?

I wasn't suggesting that showers were normally taken at 50C. The two 10 minute showers assumed that the shower temperature was 40C; i.e. 50% cold @20C and 50% hot at 60C.

 

[Michael Groves]

Apparently modern condensing boilers are able to carry a higher temperature than older boilers, so a 15mm pipe is able to carry 9kw, a 22mm pipe 24kW and 28mm pipe 70kW.
On that bases if my heat load is only 17kW I only need run 22mm pipe, I don’t need to run any 28mm?

 

[fezster]

Grundfos advised 1st pump should be configured with autoAdapt, with second pump fixed rate. This way AutoAdapt will adjust accordingly.
But Grundfos said 2nd pump would always be running at 100% which would reduce life and ties in with your comment.

I've got some more realistic prices, difference is about £200

I'm still leaning towards 2 in series.

Interesting discussion about 2 pumps in series. All the literature I've read says that 2 in series increases the head and 2 in parallel increases the flow. What I couldn't quite understand is how 2 pumps in series with different specs work in tandem? Eg. I have a 25-80 with a 15-60 in series - what exactly happens in that scenario, does anyone know?

 

[Last Plumber]

Interesting discussion about 2 pumps in series. All the literature I've read says that 2 in series increases the head and 2 in parallel increases the flow. What I couldn't quite understand is how 2 pumps in series with different specs work in tandem? Eg. I have a 25-80 with a 15-60 in series - what exactly happens in that scenario, does anyone know?

How far apart are they?

 

[fezster]

Approximately 4-5m of 28mm pipework. There is also a zone valve in between the 2.

One pump is serving the entire system, the other only a single zone. This zone has the index rad on it and the pump in series is legacy before the 25-80 was fitted. The plan was to remove it but given it was working, it was left in.

 

[Last Plumber]

Oh I see.
In all honesty, I have only ever sized pumps correctly and fitted the right one. I can't say I have ever experimented with pumps in tandem. I have seen it done but only in a situation where the pump was undersized and someone added another to the circuit further along at the point where the pressure and flow from the original was exhausted. I can see how this may work but it is not something I personally would do.

There is no reason why pipes and pumps cannot be correctly sized for the job they are intended to do.

 

[doitmyself]

Apparently modern condensing boilers are able to carry a higher temperature than older boilers

It's nothing to do with higher temperature, e.g. 80C compared to 70C, but with the greater difference between flow and return temperatures, i.e 20C compared to 11C. The larger difference is achieved by a lower flow rate - 14.35 litres/min @ 20C compared to 26.1 lpm @ 11C for a 20kW boiler.

The lower flow rate (lmp) means that the water velocity (metres/sec) is reduced proportionally. Experiments have shown that the water velocity needs to be above 0.3 metres/sec (to reduce the chance of sludge settling in horizontal pipes) and below 1.5 metres/sec (to reduce the noise caused by the water passing through the pipe.

The pdf "Small bore heating systems", which I attached earlier, has more information.

 

[fezster]

Oh I see.
In all honesty, I have only ever sized pumps correctly and fitted the right one. I can't say I have ever experimented with pumps in tandem. I have seen it done but only in a situation where the pump was undersized and someone added another to the circuit further along at the point where the pressure and flow from the original was exhausted. I can see how this may work but it is not something I personally would do.

There is no reason why pipes and pumps cannot be correctly sized for the job they are intended to do.

It's a bit of a mish mash to be honest. The 25-80 should be more than sufficient for that zone so the second circulator needs to come out at some point. Probably when the Low Loss Header goes in

I was just intrigued as to how it worked as logically it doesnt make sense to me that a second slower pump can work in tandem with one which is capable of much higher flow rates. But it does work, so not sure.

 

[Last Plumber]

It's a bit of a mish mash to be honest. The 25-80 should be more than sufficient for that zone so the second circulator needs to come out at some point. Probably when the Low Loss Header goes in

I was just intrigued as to how it worked as logically it doesnt make sense to me that a second slower pump can work in tandem with one which is capable of much higher flow rates. But it does work, so not sure.

It sounds like Ballcocks to me but like I say, it's not an experiment I would get involved in. What's the point?

 

[Chuck]

How do you design a straight sided cylinder to prevent incoming cold disrupting the stratification?

You get a reasonable degree of stratification without doing anything special, but good quality cylinders, e.g. a Megaflow, have a carefully designed diffuser on the inlet to minimise mixing.

I wasn't suggesting that showers were normally taken at 50C. The two 10 minute showers assumed that the shower temperature was 40C; i.e. 50% cold @20C and 50% hot at 60C.

No, you asserted that a 400 litre tank could only supply water for two 10-minute showers at 40°C. Since the V40 capacity of a 400 litre tank, i.e. the amount of water it can provide at 40°C, is typicially 450-500 litre you must have assumed that the showers were using 20-25 litres/minute.

 

[doitmyself]

No, you asserted that a 400 litre tank could only supply water for two 10-minute showers at 40°C.

I think we are talking about different things.

I was just saying that, after two 10 minute, 10litre/min, 40C showers (50% @ 20C, 50% @ 60C), the mean temperature of a 400 Litre cylinder will have dropped to 50C, at which temperature a typical cylinder stat will turn the boiler on.

Of course you can get more than two showers out of a 400 litre cylinder.

 

[snowhead]

I was just intrigued as to how it worked as logically it doesnt make sense to me that a second slower pump can work in tandem with one which is capable of much higher flow rates. But it does work, so not sure.

The second pump ADDS to the first pump's flow rate.

Much like running downhill, the hill being the first pump.

 

[Chuck]

The second pump ADDS to the first pump's flow rate.

Much like running downhill, the hill being the first pump.

There's a bit more to it than that. For pumps connected in series the flow rate through each must be equal and the total pressure drop across the pair, Ptot, will be determined by the resistance of the rest of the heating system. One must then determine the pressure across each pump by using the 'pump curves' (pressure vs flow charts) to a combination of flow rate for which the sum of the pressure drops matches Ptot.

Such arrangements may or may not have stable solutions.

 

[fezster]

There's a bit more to it than that. For pumps connected in series the flow rate through each must be equal and the total pressure drop across the pair, Ptot, will be determined by the resistance of the rest of the heating system. One must then determine the pressure across each pump by using the 'pump curves' (pressure vs flow charts) to a combination of flow rate for which the sum of the pressure drops matches Ptot.

Such arrangements may or may not have stable solutions.

I was looking at the pump curves for the 2 pumps earlier and determined that even the 15-60 pump can handle the flow rates my boiler expects, just at a much reduced head. This is actually OK in my scenario, though, as the 25-80 is providing adequate (but maybe marginal) head for my index circuit, so the slight extra head the 15-60 gives (roughly 1m at the flow rate Im looking at) is probably all that's needed. What you say corroborates this I think - and more importantly, the setup is working (and has been for ~5 years).

I'm still convinced the second pump can be taken out, but I'm loathed to call someone in just to do that, so will wait until I have other work carried out.

 

[tolly]

Whats the setup though, unless your using a LLH you need to add the resistance of the heat exchanger into your pressure loss calcs. In modern boilers this can be a big amount due to the small waterways inside for effeciency.

Ive always worked out pressure loss of index circuit and work out your flow rate then cross refeence those two figures with the pump curves.

 

[doitmyself]

Interesting discussion about 2 pumps in series. All the literature I've read says that 2 in series increases the head and 2 in parallel increases the flow. What I couldn't quite understand is how 2 pumps in series with different specs work in tandem? Eg. I have a 25-80 with a 15-60 in series - what exactly happens in that scenario, does anyone know?

As the pumps are in series, the flow must be the same through both pumps. You just add together the head for each pump, at the same flow rate, to get the head of the pair of pumps.

See Pumps in Parallel and Series.

 

[tolly]

I should probably read the thread first. Most of what i wrote has been said already!

The domestic heating design guide has a good bit on sizing pumps and working out your flow rates etc. I think its published by the CIBSE...the orange book. Worth a read if youve ever got a largish heating system to work out calcs on.

 

[fezster]

Whats the setup though, unless your using a LLH you need to add the resistance of the heat exchanger into your pressure loss calcs. In modern boilers this can be a big amount due to the small waterways inside for effeciency.

Ive always worked out pressure loss of index circuit and work out your flow rate then cross refeence those two figures with the pump curves.

Yes, the Vaillant 438 has a 4.05m head loss at 38KW/20 degree differential (minimum flow rate 0.45 l/s)! This is far higher than my CH index circuit which is around 2.5m. So total head loss of ~6.55m.

The grundfos 25-80 provides ~6.7m head at this flow rate (speed setting 3) and the 15/60 around 3.5m (speed 3 - ive reduced to speed 1, which is around 1.4m).

So (in theory) the second pump is now unnecessary. I'm hoping once I have a LLH fitted, I can use the 15/60 to pump my heating circuits, and use a 25-55 (I have spare) for the primary circuit. The 25-80 would probably be overkill looking at the pump charts, and a 15/60 too small to overcome the hex resistance.