that sounds reasonable, and good that you've been able to measure it . how, if you don't mind me asking?

  getting 21lpm at 4.5M implies your pipework is doing ok. 56 as you say is a big target, you'd have to substantially repipe, or buy a monster head commercial pump. But 9M head pumps are not  ridiculously expensive.  this seller may have one UPMM left https://www.ebay.co.uk/itm/314885973099 . 

This version of the UPMM is PWM modulating meaning you need to send it a PWM signal to slow it down. otherwise its flat out all the time (its designed that way, if no control signal run full speed) . there is no "1 -2 -3" dial. 

this came up on another thread, with a buffer / LLH in play, in an ideal world flow rates both sides the same, in practice, secondary side slightly slower so that the FlowT will have least possible drop across the buffer. however as you say when you add the second heat pump in, then the flow rate on the primary side will be way higher, so you'll have lots of mixing, and the two heat pumps will each only be seeing a reduced DT and putting out a reduced output because of this.

to make this work properly I think you need to try to build circulating pump speed control on both sides. if you're running on one heat pump, you need its circulating pump going full speed, and your house circulating pump at ~24lpm to match. if you're running on 2 heat pumps then you need the house-side circulating pump flat out, and the two heat pump's circulating pumps dialed down somewhat, so that the  total of the two matches what the house side can do when flat out.

what pumps do you have on the primary side (serving your two heat pumps), and are they PWM capable? and or do they have 1-2-3 dials that give a usable tuning of flow rate?

however PWM is a much cleaner and controllable way of setting circulating pump speed vs "1-2-3" dials. hence my suggestion to get a PWM one for the house side if you're going shopping.

you say are not into coding or automation ?

you can buy pre-built modules like this https://www.ebay.co.uk/itm/275254025141 which will generate a 0-10v DC output in response to the dial setting. connect that to the PWM input of the circulating pump, you have a simple manually adjustable circulating pump speed controller.  don't wire this anywhere near the AC side , its purely a DC low power low voltage controller.

Hi @iancalderbank, and thanks for that. I'll have a look at my pump situation when I've got a bit of time later.

Just to answer your point about the calc though, I just used the heat transfer equation, Q = mst, to find mass flow rate. I took the Q Watts from the Samsung controller generation figure, and measured the delta-t across the secondary side as it leaves/returns to the buffer, typically this was around 5°. The theory is sound but it's a bit crude I know, principally because I'm relying on the Samsung metering being accurate and my own analogue temperature gauges which I know aren't accurate but are at least consistent.

I measured it several times under stable temperature conditions and it all kept giving me similar answers, and the flow rate seemed to make sense from a velocity point of view, so I thought fair enough. Really, I just wanted an indication and to give me something to have an informed discussion with the installer. 

I may yet have to get into coding and start sorting things for myself. Everything I've learned about my set-up so far has been from experience and reading stuff on here! 😊

Hi @iancalderbank, and sorry for the delay. Turns out there are more important things to do than fettle the heat pump 😂.

The primary side pumps are Wilo Para ku 25/7 - see pic and manual download attached. The manual seems to indicate they are PWM-enabled but there's no cable entry point (ref 11) as shown in the diagrams. Certainly they aren't operating in this mode and just pump out at a constant 26lpm.

I've not altered the pump speed at all (it's remained on setting III) but I did wonder about trialling it at II and seeing what impact that has. Presumably a reduction in heat output but maybe the system becomes better balanced and more efficient overall? I've read elsewhere that these HPs like a higher flow rate and, whilst 26lpm seems typical, the minimum is 16lpm. I wouldn't want to go down that far, really just to get closer to the secondary side (at least based on my crude figures). 

Intuitively, moving to PWM control does seem sensible, given the interactions of 3 processes; primary pumping, hydraulics of the buffer and secondary pumping. I guess I'm gravitating toward a step-wise solution - sort the tandem running, increase the secondary side pump capacity then look at full PWM.

Thanks for your help. 

Simon

@dunlorn no worries you mean you have a life outside of heating system fettling 😉 ?

I don't have that pump but the manual refers to an "SC" version which yours has printed on it - that version is controlled by the buttons on the front, not by PWM. But you still have the 1-2-3 to play with.

if you reduce the flow rate through the HP, the DT that the HP sees will go up. that doesn't necessarily reduce output, so long as its within the flow rate bounds that the HP likes, and within the DT bounds that the HP likes.  I have a G6 16kw , mine will run down to just about 12l/min.  I think the 12kw is essentially the same box - if you look at the customer visible hardware layer specs  like size, weight, refrigerant charge , they're identical. perhaps  some kind of different internal tuning (about which I know nothing)? on that note, have you ever asked Samsung if they can "re-tune" your 12 to a 16? then you could run with 1 and be bufferless, keep the other as a spare?

with the buffer and 2 HP's, your goal is to  balance flow rates each side , so that is be a good reason to play with the pump speeds. Probably

• if only running 1 heat pump, keep it on 3
•  if running both heat pumps, drop it down to 2 (or maybe even 1.. you will have to see)

the mass flow rate / DT maths method you used to estimate flow rate on the secondary side is workable but highly prone to any errors in the temperature sensors causing a larger error in computing the flow, as the DT is small. Investing in a proper heat metering setup on the secondary side , or at least a flow meter, or a pump with a reliable accurate flow rate gauge, could be worthwhile.

Thanks @iancalderbank. Yep, the calc is, er, somewhat sensitive as you say but as an approximation it's OK for now. I'll look into the metering you describe as well as trialling a few pump speed options. Some fun for the Christmas period 😂

I do think, from a heat loss point of view, a single 16kW unit would be fine. I even think 12kW would do probably do the job, were it not for defrosts and DHW. If the two operated effectively in tandem then that issue goes away. Yes, more installed capacity than I need but at least it should all be purring away quite happily. 

I'm just dropping a note to my installer on the issues of tandem running and pump capacity/control so I'll see where that takes me.

The more I get into this the more I realise quite what a complex, interacting system it all is!

Best regards 

S