Buffer Tanks - 2 or 4 port connections

[Worcester]

OK, A nice technical one for the Bank Holiday Weekend

Buffer tanks - 2 port connection or 4 ?

Which is best 2 or 4 port, in What situation and why?

What factors e.g. does it depend upon the heating load types (e.g. underfloor / simple zone controlled rads)

So under what circumstances would a 2 port connection be best and when would 4 port connections be best.

Heat Pumps or Biomass Boilers?

Adding to that in each scenario which is the best way to send the call for heat to the Heat Source (HP or Biomass Boiler?

Finally when to use a loss loss header in preference to a buffer tank?

Example 1 Biomass boiler, long time to come on, like to run for a long time, long time for boiler to cool down - i.e. a large hysteresis, as such boiler manages buffer tank based on top and bottom temperatures and hysteresis on both. Heat load considers buffer as an inexhaustible supply of heat. Call for heat from rads / underfloor merely turns on circulation pump, so 90% of installations recommended as 4 port, though I have seen a few controlled this way though plumbed as 2 port, so with a proper back end management system means when boiler fired up that heat goes to heating circuits first.

Example 2 Ground Source Heat pump (non inverter) supplying mix of underfloor heating and rads, as all heat demands could be shut off manufacturers recommendation has been install buffer with a 2 port connection so that when heat load is less than production HP will heat up buffer tank and also meets the minimum flow requirements so doesn't get High pressure faults. - Call for heat from heating system

Example 3 Air Source Heat Pump (Inverter controlled) supplying mix of undefloor heating, fan coil and radiant panels, where load could vary from 600W to 10KW,
a) Supplier 1 recommended low loss header with call for heat from heating system
b) Supplier 2 recommends 100 Litre 4 port buffer tank with call for heat from heating system.
c) Supplier 3 recommends 100 Litre 4 port buffer tank with call for heat to HP from buffer tank with single sensor and 6°C hysteresis.

So question is Why the different recommendations re heat pumps, whether its GS or AS or Biomass - What's good and bad about each option?

Discussed it with various manufacturers and they simply maintain their position with no explanation as to why.

 

[Nostrum]

No idea why manufacturers recommendations vary as they are all very similar products.

LLH option for air source would depend on total water content of system I would imagine to prevent short cycling.

Froling recommend 2 port connection for buffer tank so that heating doesn't have to wait for buffer to be charged, I've always piped them in a 4 port configuration as always controlled them as you mentioned, boiler to buffer temp, radiators controls circulation pump only.

I can't really see how it makes any odds unless you intend switching the heat source off before buffer is satisfied to help reduce heat loss through buffer? If you could control the flow to buffer depending on flow rate as opposed to temperature, it would make more sense.

 

[IDCHAPPY]

I wish i was clevererer

 

[Worcester]

LLH option for air source would depend on total water content of system I would imagine to prevent short cycling.

I don't see how it does that (prevent short cycling) I can see how it prevents the low flow and hence HP faults on the heat Pumps.

Froling recommend 2 port connection for buffer tank so that heating doesn't have to wait for buffer to be charged

That's the only reason I can see for that as well.

I suppose there is one other scenario which is where the boiler is undersized for short term peak demand, so the boiler and buffer tank are providing heat directly, that may work better in a 2 port configuration...

So which is better for overall efficiency and hence fuel economy?

 

[Nostrum]

It's to do with minimum water content I believe.

The hysteresis set by the controller will also probably effect the required minimum water volume in order for the heat pump to run for a set amount of time. If the hysteresis is small then a small volume of water would take very little time to return at the set temperature. Although the controllers generally have anti short cycling settings as well. This is my interpretation of it anyway.

As for the biomass buffers and under sizing, Your looking at a much larger volume of water required then though, an accumulator, unless your set point is considerably greater than your flow temperature.

I was talking to the warmflow rep the other day who says their pellet boiler doesn't need a buffer in most circumstances. I've always wondered why the buffer sizes are so large for biomass, other than to cover as you say, the slow start up times etc. Most gas boilers have a similar turn down ratio.

I think buffers are much more common on the continent and so a lot of the manufacturers don't feel they need to explain their uses to anyone. We are looking at the Heliotherm range of ground source heat pumps which have full step less modulation, yet they still require (according to the manufacturer) large buffer vessels for maximum efficiency.

As for efficiency, I'm always wondering myself. It's a double edged sword, larger buffers will prevent as many start ups and allow longer run times at a lower temperature rise, but larger buffer will suffer from greater standing losses at the same time. The later isn't really calculated for, which seems like a bit of an oversight to me?

I am glad your on actually, as I wanted to ask a question I'd like you to answer! I'll post a thread.

 

[missplumb]

the only reason i can see for using a 2 port tank would be with a biomass and using it as an accumulator not a buffer. We would always suggest a buffer with any biomass boiler to reduce cycling. went to look at an install done pre 2008 which had a 90kw eta hack boiler, no buffer and he was told by the installer that due to the size of the house and 2x 250ltr unvented cylinders, there is enough system for the heat to dissipate, only problem is the boiler is connected directly by either the cylinder stats or a room stat. i was there for an hour and the boiler fired at least 6 times.

a lot of the biomass manufacturers recommend using the tank as an accumulator so the heat is drawn from there while the boiler is firing round its shunt circuit, once its upto temp and the mixer is open then the boiler takes over the heating and the tank remains uncharged until the heating demand is gone.

We would only really fit a lowloss header if the system had various different heating systems ie. ufh, rads, air blowers etc, or the temperature and flowrates needed to be different for the different circuits. Fitted quite a few LLH as we do alot of commercial buildings.

 

[Worcester]

So in the mixed emitter environment would you fit a llh in conjunction with a buffer tank? - Assuming the buffer is 'remote' from the heating distribution location so you aren't able take advantage of the multiple connection points on the buffer ?

 

[missplumb]

Absolutely, we had an installation in a garden centre where plant room was well away from the ditribution system. Ran pumped pipework from buffer round to lowloss which then fed a hot water circuit, underfloor heating circuit, and 2 overhead fan heater circuits. All the circuits needed different flow rates so the lowloss was perfect, combined with commissioning valves and variable speed pumps.

 

[missplumb]

this is our low loss header

 

[Worcester]

Gerberit Mapress or Yorkshire Pegler Xpress?

We tend to use Xpress as our local BSS keep 10x as much of that as the Mapress ... 1000's of connections made and No leaks (apart from the joint we forgot to press ) - we like that

 

[missplumb]

We use xpress as much as we can. We have had a couple of leaks but I'm sure there would be allot more if we had soldered them all

 

[ShaunCorbs]

Very poor practise, you need to leave the soldering to the men

Why it's handy to have esp when there's water in the pipe or when you can't turn the water off and have to put a valve in

 

[missplumb]

i guess i missed that post - obviously stuck in the dark ages!!

 

[Worcester]

Had a interesting discussions today with the UK Froling distributor (Not the BG subsidiary) wrt buffer and low loss headers.

One of the biggest issues on biomass boiler efficiency is to get the return temperature to the buffer / boiler as low as possible. - the boiler will always raise it if need with the return riser, it can;t lower it.

They are strongly recommending full weather comp with mixing valves on the heating circuits (not HDW). That will dramatically lower the boiler / buffer tank return temperatures for 95% of the year, so will help the boiler to modulate properly and therefore better fuel efficiency.

One component that would destroy that reduced return flow temperature is of course a low loss header... so they strongly recommend to design them out whenever possible. - Seems to make sense to me.

Part of the goal therefore of getting as low a return temperatrure as possible therefore points to using a two port buffer connection for the heating circuits (take DHW of the top of the buffer tank) Interesting to note that the 'official' Froling buffer tanks only have one column of four connection points. (plus top)

Thoughts?

 

[Nostrum]

Zero ?

 

[Worcester]

Zero ?

Yep

 

[mutley racers]

Blimey

 

[Nostrum]

Interesting and yet a completely different answer to the one I received when I asked the same question a week or so back. I actually meant to post on here but didn't get around to it.

Who did you speak to out of interest?

I also raised the question when I did the commissioning and controller course but this wasn't mentioned at all.

Surely if the buffer is charged though the return will always be fairly high? What sort of temperature where they suggesting was optimal, 60?

 

[missplumb]

Was it zero carbon future?. So are they suggesting we design to use the tanks as accumulators rather than buffers? Surely if you have the return from the system designed to say 50oC And the boiler needed the mixer to be at 60 , then I would imagine the boiler would cycle more keeping the shunt circuit heated!!! And therefore using more fuel. We were told by ETA that the ideal buffer setting would be top target of 80 and bottom off at 70.

 

[Worcester]

@nostrum, i was discussing it with Finian
@missplumb, yep, we are getting the different stories from eta and froling

What froling (zcf) are saying is that the eta (innasol) way will cause more full power cycling and claim that their method will enable the boiler to modulate properly.

The buffer will still be charged as appropriately as DHW still requires full temp flow, and is the only consumer (optionally) directly connected to the buffer.

if the boiler is "stealing" flow to top up the return flow temperature, then the balance of the flow (production) required by the consumers (heating system) will come from the buffer, so the flow in that length of pipework will reverse (depleting the buffer), as it's only got a single column of connections that has to happen.

I can see the logic of both arguments, perhaps they are both correct for their particular boilers? Maybe, to an extent, it may therefore depend upon the boiler control software / systems?

 

[Nostrum]

I'm confused if I'm honest. Surely the load will be the same in either situation, therefore unless the load exceeds the output the boiler will function the same.

In both circumstances the boiler is triggered by the buffer sensor dropping below the set point (ignoring any hysteresis)

In both scenarios the buffer will 'give up' heat until the boiler fires, the only difference being which pipe it uses to export heat from the buffer, 4 port will have a dedicated export flow, 2 port will use the boiler flow in reverse?

Once the boiler fires the boiler pump will pull water from the bottom of the buffer and mix with the flow to desired temperature. The output energy will then be distributed through the heating load and into the buffer simultaneously until the buffer reaches set point again?

Where's the difference?! :skep: or what am I missing?

 

[Nostrum]

Just to add, I can see the benefit when using weather compensation and I guess in that scenario piping in the Froling configuration would reduce the return temperature to the boiler more directly.

 

[missplumb]

I'm gonna have to get my books out as that drawing doesn't show any mixing valve or shunt circuit, it's just using the boiler to service the heating and charge a heat store just for the hot water, in which case what the point of the buffer?

 

[Worcester]

I'm confused too!

The mixing valves are shown of the heating circuits - they're controlled by the weather comp program in the boiler.

From recollection the P4's are odd beasts as well - the flow comes from the bottom of the boiler (see the example diagram) and the return enters the top of the boiler and is fed through an internal coil in the water jacket that then terminates in the water jacket, so the return gets warmed up before entering the jacket it doesn't have a return riser.

Assuming this is using the froling controller, then from recollection, the boiler is also triggered by a call for heat from the heating system, i.e. the heating system and buffer tank are managed separately yet together.

 

[Nostrum]

As far as I am aware the 3200 controller doesn't have weather comp built in as standard although they will almost definitely have a module you can add to activate it. I've not seen any option for weather comp curves whilst setting up.

The smaller P1 and P4s use a modulating pump controlled by the controller to control the pump speed to maintain the return temperature to the required temp. This may be what is on the drawing.

On the last quote we got a couple of weeks ago they added an export pack which was not on the quote before. I queried what it was (listed as CB mixer) thinking they'd specced two variants for the shunt pump as I couldn't understand how they could spec a pump without asking for hydraulic spec of system.

If they had said it was weather compensated mixing it would have made sense but they didn't, they simply said it was an export pump and mixing valve set. I've a feeling this has all only recently come about.

 

[Ermintrude]

.

 

[Nostrum]

,

As far as I am aware the 3200 controller doesn't have weather comp built in as standard although they will almost definitely have a module you can add to activate it. I've not seen any option for weather comp curves whilst setting up.

The smaller P1 and P4s use a modulating pump controlled by the controller to control the pump speed to maintain the return temperature to the required temp. This may be what is on the drawing.

On the last quote we got a couple of weeks ago they added an export pack which was not on the quote before. I queried what it was (listed as CB mixer) thinking they'd specced two variants for the shunt pump as I couldn't understand how they could spec a pump without asking for hydraulic spec of system.

If they had said it was weather compensated mixing it would have made sense but they didn't, they simply said it was an export pump and mixing valve set. I've a feeling this has all only recently come about.

I've just looked through the controller manual. It would seem as if when you plug a froling room sensor in the weather comp option becomes available

 

[Ermintrude]

That's a very small accumulator in test center...... I did my get as

 

[missplumb]

Looks like a SH 60kw ish log/pellet boiler with a 1800 ltr buffer.

 

[Ermintrude]

Etta

 

[missplumb]

Yeah ETA SH

 

[Nostrum]

Looking at the schematic a few times today I can see where this may differ to a 4 port buffer.

In this configuration the return temperature will fluctuate in line with the heating load. In a 4 port buffer configuration the temperature difference between flow and return will be relatively constant throughout the heating cycle until the buffer set point is reached, regardless of heating load.

In this configuration, the boiler cannot heat the buffer as the flow is reversed. The boiler will receive a return directly from the heating load which I presume will be a much lower water volume and from mixed circuits.

I don't quite understand how the flow from boiler and flow from buffer is balanced to feed the heating circuit ? Any ideas?

Only once the heating circuit is satisfied will the boiler reheat the buffer which would probably mean the boiler ramps up assuming the boiler modulates depending on return temperature rise. EG: a slower rise will cause the boiler to increase output to achieve set point.

Thoughts?

 

[missplumb]

The flow from the boiler and buffer would be created by the export pump, but that is what I don't quite get as the boiler pump would only run when the buffer needs heat. I've been to several systems piped like this and because the heat loads have not been calculated, the size of the boiler has been well undersized. This means the boiler is constantly running and buffers sit there cold.

 

[Nostrum]

Yes and what is stopping cold water being pulled through the boiler in reverse until the buffer temp drops (which will then trigger boiler to fire)

Thinking about it the system is self balancing, any short fall from the boiler will be made up through the buffer.

The buffer will provide instant heat when heating is called on without the boiler firing until the buffer temp drops, the only pump running would be the mixer pump.

At the end of the heating demand the mixer pump will stop and the buffer will be loaded as normal using the boiler pump.

 

[missplumb]

Yes providing the boiler is sized correctly.

 

[Nostrum]

Yes providing the boiler is sized correctly.

That should go without saying and would apply to any system :biggrin5:

I'm wondering, there must be the need for the buffer/boiler set point to be higher than the heating flow temp in order for the mixing set to work properly and allow enough time for the boiler to fire.

 

[missplumb]

I don't see how the system without the mixing valve would work properly as you couldn't have the boiler running and the system being serviced by the buffer. If you have the mixer valve, that provides the mechanical separation between the system and buffer, until the boiler has reached its operating temp.
I think I'll stick to simple boiler - buffer - heatload.

 

[Nostrum]

Once the heating pump is running the boiler and buffer are effectively working in cascade. The buffer temp and boiler output should presumably equalise although I can't quite get my head around how as the only input to the buffer is the return temp. Surely the buffer would theoretically become exhausted at some point? Or will the fixed flow rate from the boiler pump mean that the only flow through the buffer is excess to what the boiler pump can provide?

If that's the case then the buffer really isn't used much during operation in most circumstances (providing good design)

This would make sense as the buffer only becomes active to satisfy the boilers run time requirements and to provide heat when the boiler is starting up, or to make up any shortfall in boiler output for short periods of time.

 

[Nostrum]

Which will actually work in the way I was suggesting at the beginning of the thread. It just uses basic principles to achieve it as opposed to any controlled intervention.

Quite a clever set up if I'm correct, if I'm not then I've just wasted half my day thinking up a load of tosh. Where's Worcester for his opinion?

If you could control the flow to buffer depending on flow rate as opposed to temperature, it would make more sense.

 

[Worcester]

Worcester's been having a day with family

I think we're there, and having gone round in circles as we've gotten our heads around it a combination of MissPlumb's and Nostrum's responses answer it.

i.e assuming the heating circuits use a mixing valve (room stat, external temp sensor, weather comp) and are managed by the boilers inbuilt systems, then the buffer provides instant heat the boiler then fires up only working at the level need to maintain heat demand, it also monitors the buffer, so if there is also a hot water demand or if the buffer tank is cooling - part of the heating return has to go that way if flow is coming from it, the boiler will ramp up and then after heating demand ceases, it will continue to heat the buffer up until it's replenished.

Needs a good boiler controller, so won't work properly with some of the simpler boilers out there.

ETA show basically the same configuration, so it works when the boiler has a comprehensive control system - this schematic presumes using the inbuilt eta controls.

If you're NOT using the built in controls and / or the boiler has simplistic management, or a cheap buffer that doesn't stratify well/properly, then I can see possible problems with plumbing it the 2 port way: increased cycling, buffer tanks not heated properly, poor performance of the heating sytem (client perception).

So what do Windhager / HDG recommend? - I know that Okofen reckon that they don't need a buffer on smaller systems perhaps for similar reasons.

So:
Using Froling/ETA built in heating controls, using indoor and outdoor stats, mixing valves on the heating circuits and their weather comp system, the boilers should run more efficiently. Reducing the flow through the buffer also helps to maintain good stratification.
If you're not using their control systems then a 4 port buffer is a 'safer' configuration as far as ensuring heat is always available, though may trigger greater full power cycling as opposed to modulation; subject to: programming the correct buffer tank control, by using multiple stats - as a minimum top and bottom stats, and now looking at the hysteresis programming of those stats. Once again the "simpler" boilers may not be able to handle those types of inputs and modulate properly.

I think we've finally answered my original question!

For the simpler heat pumps we have developed our own buffer controllers specifically to address this using two stats and hysteresis management. - The controller is less than £100, needed a bit of lateral thinking to find it though

So next up is how do you configure the buffer tank control on / off points, what temperatures, what hysteresis?

 

[Nostrum]

My thoughts would be that the buffer set point would have to be at least 10 degrees higher than the design flow temp and the hysterisis would be maybe 5 degrees below set point to prevent any short cycling on the buffer reheat.

Realistically though it will depend on a lot if you want to be spot on, buffer size, boiler output, heating load.

Here's another thought, I'm contemplating using a buffer with hot water coil in one of the upcoming jobs so that we can use the space of the existing oil boiler and cylinder to house the buffer internally, rather than externally and have to export heat through underground pipe work constantly which will increase losses.

In the 2 port configuration, with heating on in depths of winter and customer runs a bath or has a long shower. Buffer temp will drop fairly quickly, heating pump will be on so flow to buffer will be reversed, how's the buffer going to remain charged?

 

[Worcester]

In the 2 port configuration, with heating on in depths of winter and customer runs a bath or has a long shower. Buffer temp will drop fairly quickly, heating pump will be on so flow to buffer will be reversed, how's the buffer going to remain charged?

Same as with any 'small' buffer that could suffer depletion, in the 2 port case you would effectively have heating priority as opposed to in the 4 port case the possibility of reduced temperature in buffer, hence reduced flow temp to both buffer and dhw.

Could always use one of the strap on DHW heat exchanger units - both Froling and ETA offer them, flows up to 40l/min. "Virtual Combi". There again as soon as the Buiffer tank is being deleted the boiler output should raise to match demand up to it's maximum. Should work fine IF using the built in controllers.

 

[Nostrum]

But how's the boiler able to up any flow to the buffer if the heating pump is on?

The hydraulics don't allow flow to heating and buffer simultaneously from what I can see. You'd have to turn off the heating pump.

 

[Nostrum]

Just looked at the schematic for this set up and it seems they use an additional buffer sensor half way down the tank, maybe this tells the controller to stop heating if it reaches a trigger point?

 

[Worcester]

Unless heat load = boiler output, there will be some spare capacity to heat the buffer. Hence you'll need to design boiler output in this case to cope full heat load plus an acceptable buffer recovery - actually that's no different from normal.

Practice says boiler size = heat load + 200W per person for DHW, if you had a property demanding full heat load, you'll still have a problem heating the DHW cylinder - it'll take a whiie. Plus boilers come in discreet sizes, not actual heat load size, so there should always be some spare capacity.

As you suggest it could also be charging to the middle sensor to ensure capacity for DHW.

 

[missplumb]

As Worcester said you could use the hot water module on the buffer and pipe the boiler direct to the buffer and come out the other side to your heating. You then add the middle sensor to help monitor the buffer charging status. Difference between the two schematics you've shown is the ETA shows a shunt circuit with mixer valve, the Froling does not.
We have always used the 4 port buffer system, we set top of buffer to 80 min with 5 deg hysteresis and bottom off at 70 deg. We find that the systems are never short of heat as the buffer is always already 70% charged and boiler only fires using buffer sensors only.

 

[Ermintrude]

The buffer works when the boiler isn't even awake.

Demand established :-

Buffer satisfies load whilst boiler gets to 55c and flue up to 100c

Boiler hits :- 55c and flue 100c 3 port and pump wind in ..... Buffer still providing load satisfaction

Boiler hot and 3 port swings in:-

Buffer almost depleted but boiler coming on line.

Load side pumps pull from easiest route which is boiler now that the 3 way is playing ball.

 

[Nostrum]

Won't the boiler and buffer be asking for two separate temperatures though? If the return temperature is coming back in excess of what the boiler is looking for, 20 degrees differential for example, the boiler will modulate to try and match the load?
so unless the buffer is happy to run at 60 also (which is lower than the desired temp at mixer) I can't see how sufficient flow will enter the buffer to raise the temp?

Surely in the example I gave above, you could end up with a buffer at 60 providing no additional input to the heating circuit until the heating mixer pump turns off, at that point the buffer could be loaded by the boiler?

Thats fine for heating, but hot water production is relying on the buffer.

It's a shame they don't explain the controller in a little more detail, maybe most people aren't interested but I'd prefer to know the capabilities of the product in full detail rather than just pipe it up off a schematic as told. It's my companies reputation at stake at the end of the day.

 

[Nostrum]

The buffer works when the boiler isn't even awake.

Demand established :-

Buffer satisfies load whilst boiler gets to 55c and flue up to 100c

Boiler hits :- 55c and flue 100c 3 port and pump wind in ..... Buffer still providing load satisfaction

Boiler hot and 3 port swings in:-

Buffer almost depleted but boiler coming on line.

Load side pumps pull from easiest route which is boiler now that the 3 way is playing ball.

Yep we've gathered how the set up works on start up, but I'd like to know how the controller can maintain the buffer temp whilst under load.

 

[missplumb]

It doesn't, the buffer is only there for peak load demand, once the boiler is up to temp it takes over the best demand and the buffer sits doing not allot as the least circuit of resistance is not through thousand of litres of water in the buffer. As Ermi said, once the demand has gone and heating mixers are diverting back to the boiler, then the buffer is re charged ready for the next peak demand.