@iancalderbank 

Re-reading the thread on buildhub I refer to above I think I am now both confused and a little embarrassed.

I thought that the conclusion had been reached on the buildhub thread referred to that heat pumps can be set up with buffer tanks in a way that avoids material mixing, and that the control strategy outlined above was one way to do that (provided that the buffer tank is well designed etc).  I started the discussion on buildhub believing that this was not the case, but was thoroughly convinced by the arguments made.  I am now not sure!

I still think it may still be the case, and indeed it may even be possible without a buffer tank sensor, but I'm less convinced that I was.  I also may have jumped to the conclusion, based on the what ReedRichards argued in Jan 2023, that heat pumps with a buffer tank sensor and some sort of buffer tank control mode (which at least some, including Mitsubishi and Chofu have) do 'the right thing'.  Perhaps they don't!

Having said that I think its still possible that, even without a buffer tank probe, mixing could be fairly minimal.  The thermocline may drift up or down the tank, and will at times reach the top or bottom, at which point mixing will occur.  But that will (should?) (at least according to the arguments made, fairly convincingly i thought, by ReedRichards) trigger the heat pump to turn on/off pushing it back down/up, so mixing occurs only for a short period of time each 'cycle'.

So in summary, whilst I remain convinced it can work, I now am not sure what happens in practice unless you go to a lot of trouble.  Like you I cant see that buffers have a valid function in most domestic scenarios, so I didn't explore the matter extensively.  

This doesn't change my basic understanding namely that

1. (4 port) buffers provide an advantage only where there is a very concrete reason to fit them (which is rare in a domestic situation) - and are likely to give a disadvantage - so basically avoid unless there is a really good reason
2. it is in principle possible to make a sufficiently well designed (4 port) buffer work with zero to minimal mixing

However I think it makes (2) more difficult to achieve in practice than I had thought 

Although a little embarrassed by the final point, I take comfort in the knowledge that I always supported those who argue against buffers (in most circumstances).

@jamespa

Hi James.

Sorry for the delay in replying, I took a break from the forum at the weekend while the improvements were being made, only to find probably over 200 messages awaiting when I received the message that everything was functioning correctly. I am gradually working my way through them, though some people keep adding more. 🙄 

So the subject this week is buffer tanks and LLH's.

I fully agree with both you and Ian that buffer tanks and LLH's should not be installed unless absolutely necessary.

Can we first of all agree to use the term flowrate rather than pump speed, since different sized pumps will probably have different flowrates even if they are operating at the same speed.

I was surprised to read that the Ecodan can perform buffer tank control, so you can teach this 'old dog' new tricks. Having now had a closer look at the manual, this is how I think buffer tank control can be achieved. Obviously all the necessary thermistors need to be present and the water pumps controlled by the Ecodan controller.

It would appear that for buffer tank control to work correctly, the flowrate through the secondary water pump, or pumps, would need to be greater than the flowrate through the primary pump. This of course would cause mixing within the buffer tank which is the situation we are trying to avoid.

I cannot be 100% certain, but I think that this is how the system would operate.

With the primary pump running and the secondary pump stopped, warm water from the heat pump would start to heat the buffer tank from the top down.

When the buffer tank thermistor THW10 senses that the water temperature has reached a value close to the Zone 1 flow water temperature thermistor THW6, the controller starts the secondary water pump.

Because the secondary flowrate is greater than the primary flowrate, the warm water is being drawn out of the upper part of the buffer tank at a rate faster than warm water is being pushed in by the primary pump. A 'slug' (technical term 😋 ) of warm water is therefore being supplied to the heat emitters.

The secondary water pump will probably continue to operate until the temperature of the 'buffer water' sensed by THW10 is approaching the value of the Zone 1 return water temperature thermistor THW7, though there is mention of a 'Re-judgement of the pump on/off time', whatever that means. 🙄 At this point the secondary pump will be stopped.

I would then expect the above to repeat, filling the upper section of the buffer tank with warm water, then pumping this warm water to the heat emitters.

This is the point where you all throw your arms in the air and say "but there is no continuous flow of water through the heat emitters, so they must not be receiving sufficient thermal energy".

As James correctly pointed out, the heat emitters don't care if they receive the thermal energy as a continuous flow or in 'dollops' (another technical term 😋 ), provided the dollops are sufficiently frequent and it is the same quantity of thermal energy.

Consider the following example:-

A 50 litre buffer tank being supplied from a heat pump at a flowrate of 20 lpm. To fill the upper section of the buffer tank with warm water should take in the region of 1.25 minutes, at which point the secondary pump is started. If the flowrate through the secondary pump is 21 lpm, then it will probably take 20 minutes or so to take the warm water from the buffer tank before the secondary pump is stopped. The quantity of warm water being pumped into the buffer tank and pumped out to the heat emitters is approximately the same, it is just the time taken that differs slightly.

Of course the above would not work with a LLH, because there would be insufficient volume.

Let's now consider what would happen if the primary water flowrate is greater than the secondary flowrate.

Because the flow in, is greater than the flow out, the thermocline will move down the buffer tank and start the secondary pump when THW10 reaches the required temperature. The secondary pump will continue to operate because the thermocline will not move back up, and some of this warm water will mix with the cooler water returning from the heat emitters.

The return water to the heat pump will therefore be warmer, so will tend to cause the LWT to increase, which in turn will cause the heat pump controller to reduce the compressor speed to keep the LWT at the required value.

The system is therefore self-regulating, with the heat pump supplying the quantity of thermal energy required to meet the demand presented by the heat emitters.

I hope this makes sense, but please feel free to ask questions.

@derek-m leaving aside the efficiency debate which we've done to death.  have you ever seen system discussed on here where there a buffer tank thermistor controls things?  I can't recall seeing a picture of a buffer with anything other than just 4 pipes.

Posted by: @derek-m

↑

I was surprised to read that the Ecodan can perform buffer tank control, so you can teach this 'old dog' new tricks. Having now had a closer look at the manual, this is how I think buffer tank control can be achieved. Obviously all the necessary thermistors need to be present and the water pumps controlled by the Ecodan controller.

I have to stress that I don't actually know what it does as I don't have one, only that it does something.  It looks like you have found some detail from the manual which I didn't find.

Posted by: @derek-m

↑

It would appear that for buffer tank control to work correctly, the flowrate through the secondary water pump, or pumps, would need to be greater than the flowrate through the primary pump. This of course would cause mixing within the buffer tank which is the situation we are trying to avoid.

I think the key point from the discussion on the buildhub forum, which I believe still stands, is that it doesn't matter in a stratified tank, if the instantaneous flow rates differ, because all that happens is that the thermocline moves up/down (rather like what happens when you draw off water from a DHW cylinder).  What matters is that on average they are the same so that the thermocline never reaches the top or bottom (or if it does, does so for only a small fraction of the operation time.

Posted by: @iancalderbank

↑

@derek-m leaving aside the efficiency debate which we've done to death.  have you ever seen system discussed on here where there a buffer tank thermistor controls things?  I can't recall seeing a picture of a buffer with anything other than just 4 pipes.

No I haven't, but I would expect the thermistor is just a sensor going into a pocket in the side of the buffer tank.

Posted by: @jamespa

↑

Posted by: @derek-m

↑

I was surprised to read that the Ecodan can perform buffer tank control, so you can teach this 'old dog' new tricks. Having now had a closer look at the manual, this is how I think buffer tank control can be achieved. Obviously all the necessary thermistors need to be present and the water pumps controlled by the Ecodan controller.

I have to stress that I don't actually know what it does as I don't have one, only that it does something.  It looks like you have found some detail from the manual which I didn't find.

Posted by: @derek-m

↑

It would appear that for buffer tank control to work correctly, the flowrate through the secondary water pump, or pumps, would need to be greater than the flowrate through the primary pump. This of course would cause mixing within the buffer tank which is the situation we are trying to avoid.

I think the key point from the discussion on the buildhub forum, which I believe still stands, is that it doesn't matter in a stratified tank, if the instantaneous flow rates differ, because all that happens is that the thermocline moves up/down (rather like what happens when you draw off water from a DHW cylinder).  What matters is that on average they are the same so that the thermocline never reaches the top or bottom (or if it does, does so for only a small fraction of the operation time.

That is basically what I was saying, the buffer tank thermistor senses the water temperature and starts and stops the secondary water pump to prevent the thermocline from reaching the top or the bottom.

If the primary flowrate is greater than the secondary, then the secondary pump will run at the same time as the primary pump, and the heat pump will self-regulate the quantity of thermal energy.

It is when the secondary flowrate is greater than the primary flowrate that the thermistor temperature variations cause the secondary pump to be started and stopped. Variations in flowrate will cause the timing of the on and off periods of the secondary pump to also vary.

I have just received a quote for a variable speed secondary pump to replace the existing variable speed pump which had no PWM control supplied. Once fitted, I intended to retain the buffer tank and vary the secondary pump speed/ flow and try to optimise the system by comparing overall performance at different speeds. I have been most interested in this topic and have the utmost respect for the contributors, but it has made me reconsider what would be the optimum way to improve performance. I should note that I have no thermistors fitted in the tank and I doubt if I could find a reliable contractor, here in the Algarve, to properly connect one.

So, which way is this discussion going? Is there a new view forming that, indeed, a buffer tank system can be made to work rather better than previously thought and that it might be the preferred option?

In my case, do I fit the new pump or do I ditch the buffer tank and connect the Ecodan directly to the radiator system?

At age 87 I don't have a lot of time to experiment!  

Posted by: @davidalgarve

↑

So, which way is this discussion going? Is there a new view forming that, indeed, a buffer tank system can be made to work rather better than previously thought and that it might be the preferred option?

I think where we are at is that a buffer tank can be made to work (but it's by no means guaranteed that it will have been set up properly) but its not the preferred option unless there is a very good reason for existence.  I don't think this conclusion is likely to change (this discussion hasn't actually changed it, merely refined the understanding of the possibilities for controlling buffer tanks correctly)

Posted by: @davidalgarve

↑

In my case, do I fit the new pump or do I ditch the buffer tank and connect the Ecodan directly to the radiator system?

To answer this question we need to answer why is the buffer tank there in the first place.  Do you know.  Have you got any of:

A backup boiler

Ufh AND radiators

Something else which makes your system complex (ie it's something other than a hot water source feeding rads or ufh, plus dhw.)

If the answer is yes to the question above then maybe there is a case to keep the buffer 

If it's no then either remove it and the secondary pump or plumb the buffer as a 2 port volumiser in the return (still removing the secondary pump)

Posted by: @jamespa

↑

I think the key point from the discussion on the buildhub forum, which I believe still stands, is that it doesn't matter in a stratified tank, if the instantaneous flow rates differ, because all that happens is that the thermocline moves up/down (rather like what happens when you draw off water from a DHW cylinder).  What matters is that on average they are the same so that the thermocline never reaches the top or bottom (or if it does, does so for only a small fraction of the operation time.

.... I think where we are at is that a buffer tank can be made to work .....

so I get that with a good control and a tall stratified buffer, mixing could be avoided. however : If the sides are still not at the same flow rate : then if the 1ary is higher, then it will have to modulate on/off. Meaning some loss of efficiency when the HP cycles. If the 2ary is higher, then that will have to modulate on off. Meaning some period when there is no heat moving on the 2ary, so the water in the "slugs" that move when it does move, has to be hotter to keep the same average temp on the 2ary side, meaning  the HP side has to run at a higher WT in order to meet the heat loss. Either way you do it, even though you avoided mixing, you still lost some efficiency. Only if the heat source is one that doesn't lose efficiency at higher temperatures or when it cycles (i.e. its not a HP), does the modulating-through-a-stratified-buffer approach not lose efficiency.

 @davidalgarve if you already have a buffer tank, and you are of a bent to setup your own PWM controls for pumps to equalise the flow rates, then that is probably the "least plumbing change" option. there are others already working on this as well. take a look on openenergymonitor forum, worth touching base with some of them.

But if you there is no other reason to keep it (warranty has also been mentioned by some, not a complex system, no other heat source) and you want to strive for the absolute best efficiency, you are sure the flow rates are ok... then take it out.

Hello JamesPa. The contractor I employed, seemed to know what he was doing (and I was desperately trying to escape from LPG at a time when the price was skyrocketing), but it soon became clear that he didn't and subsequent Mitsubishi support here in Portugal has been almost non existent.

I suppose that the buffer tank made more sense as the contractor left it, with the system running at constant temperature i.e. 55C during the first winter, but it was very expensive to run, hence my educating myself rather more and now running much more efficiently and comfortably, on weather compensation curve.

I do not have a backup boiler or UFH but do have 15 aluminium radiators (mostly with SRV's set 2-3Deg above normal required room temperature) and 3 towel rails with normal thermostatic valves set high.

The ASHP also heats a more recent DHW tank via a 3 way valve.

Does that suggest to you that I should remove the tank and secondary pump?

Posted by: @davidalgarve

↑

I have just received a quote for a variable speed secondary pump to replace the existing variable speed pump which had no PWM control supplied. Once fitted, I intended to retain the buffer tank and vary the secondary pump speed/ flow and try to optimise the system by comparing overall performance at different speeds. I have been most interested in this topic and have the utmost respect for the contributors, but it has made me reconsider what would be the optimum way to improve performance. I should note that I have no thermistors fitted in the tank and I doubt if I could find a reliable contractor, here in the Algarve, to properly connect one.

So, which way is this discussion going? Is there a new view forming that, indeed, a buffer tank system can be made to work rather better than previously thought and that it might be the preferred option?

In my case, do I fit the new pump or do I ditch the buffer tank and connect the Ecodan directly to the radiator system?

At age 87 I don't have a lot of time to experiment!  

It is always difficult knowing what to do for the best, particularly when the advice being provided may differ slightly.

It is equally difficult trying to provide advice when all the relevant information may not be available.

In your situation I would suggest the following:-

Re-pipe the system so that the buffer tank becomes a volumiser. Remove the secondary water pump and replace it with a piece of copper pipe, so that if necessary the pump can be replaced. Then operate the system and see how it performs.

The above should create the minimum disruption and also allow the system to be returned to the original if the modification fails to perform adequately.

Posted by: @davidalgarve

↑

Hello JamesPa. The contractor I employed, seemed to know what he was doing (and I was desperately trying to escape from LPG at a time when the price was skyrocketing), but it soon became clear that he didn't and subsequent Mitsubishi support here in Portugal has been almost non existent.

I suppose that the buffer tank made more sense as the contractor left it, with the system running at constant temperature i.e. 55C during the first winter, but it was very expensive to run, hence my educating myself rather more and now running much more efficiently and comfortably, on weather compensation curve.

I do not have a backup boiler or UFH but do have 15 aluminium radiators (mostly with SRV's set 2-3Deg above normal required room temperature) and 3 towel rails with normal thermostatic valves set high.

The ASHP also heats a more recent DHW tank via a 3 way valve.

Does that suggest to you that I should remove the tank and secondary pump?

 

Posted by: @derek-m

↑

It is always difficult knowing what to do for the best, particularly when the advice being provided may differ slightly.

It is equally difficult trying to provide advice when all the relevant information may not be available.

In your situation I would suggest the following:-

Re-pipe the system so that the buffer tank becomes a volumiser. Remove the secondary water pump and replace it with a piece of copper pipe, so that if necessary the pump can be replaced. Then operate the system and see how it performs.

The above should create the minimum disruption and also allow the system to be returned to the original if the modification fails to perform adequately.

Having read your follow up question and the answer from @derek-m , it suggests you follow the route which both @derek-m and I suggested.  Replumb buffer as volumiser (in the return not the flow), bypass the pump.  Male sure its not possible for most of the emitters to shut down simultaneously.