Friends, in news that will fail to surprise regular readers, I am conducting an experiment on my own home. I am altering the plumbing of my heat pump to remove a so-called Low-Loss Header. I plan to run the heat pump in a similar way to how it was run last year – and then in Spring 2024 I will assess whether the removal made any difference.
[June 2024: Skip to this article to read the results.]
My prediction is that the removal of the low-loss header will make almost no difference to the performance of the heat pump – I predict that in the coming heating season, my installation will have a Seasonal Coefficient of Performance (SCOP) of 3.5 ±0.1 , in line with performance over the last two winters.
So why am I spending money employing an elite heat pump ninja to modify my set up if I think it will make no difference? Because I want to understand how the Low-Loss Header is working in my home, and the only way to find out for sure is to do an experiment. Please indulge me: this is what experimental physicists do when they retire.
Allow me to explain…
Background
My 5 kW Vaillant Arotherm plus heat pump was installed in August 2021 and has heated my 162 square metre home in Teddington very well through the last two winters. I have a heat meter installed which I read along with an electrical meter once a week to assess COP through the year. I then sum the energy used and heat provided over a heating season (August to July) to assess annual SCOP. The results are shown below.
Click on Image for a larger version. Graph showing the performance of the heat pump through the heating seasons of 2021/22 and 2022/23. The red data (read against the left-hand axis) shows the weekly-averaged COP and the red dotted line shows the annually-averaged SCOP. The blue data (read against the right-hand axis) shows the average amount of heat pumped per day in kWh/day. Averaged weekly, 60 kWh/day corresponds to 2.5 kW continuous power.
The last two heating seasons differed in both heat pump settings and weather:
- In 2021/22 the internal temperature was 21 °C but in 2022/23 this was reduced to 20 °C. I would have thought this would increase SCOP somewhat.
- In 2021/22 7,380 kWh of heat were delivered but 2022/23 was colder resulting in 20% more heat (8,793 kWh) being delivered to the home. I would have thought this would decrease SCOP somewhat.
The measured SCOP has been 3.5 ± 0.05 in both years so these differences seem to have roughly compensated for each other.
On the coldest days last December 2022 I evaluated the performance of the heat pump and found that when the average external temperature was -3 °C the heat pump needed to deliver on average 3.5 kW of power to keep the house at 20 °C. And this was delivered with a daily-averaged COP of around 2.5.
Click on Image for a larger version. Heat pump data from December 2023. Left. Graph showing average daily heating power versus the average daily external temperature. Right. Graph showing averaged daily COP versus the average daily external temperature.
So the installation is performing well. But could it perform better? In 2020 before the installation I would have been very happy with a SCOP of 3.5, but the ‘state of the art’ in heat pumps has moved on. The Heat Pump Monitor dot org site shows live monitoring from many heat pumps with a SCOP greater than 4, including many of the same type of heat pump as mine. This has left me wondering if there is anything I could do to improve my installation.
Low-Loss Header?
When the system was installed I was told that a Low-Loss Header (LLH) was an important part of the installation, but at the time I struggled to understand what it did. I did figure it out eventually, although the words I use to explain its functioning tend to be different from those used by installers.
A Low-Loss Header is simply a large diameter pipe with smaller diameter inlets and outlets at each end. When a heat pump with a LLH is configured for heating radiators, the flow circuit looks like the figure below.
Click on image for a larger version. See text for details.
Instead of the water from the heat pump flowing directly through the radiators and back to the heat pump, the water flows through the LLH and back to the heat pump. This flow provides no heating to the house, but allows the heat pump to operate very efficiently because the flow impedance is very low: so far so pointless!
To actually heat the house, a second hydraulic pump draws water from the top end of the LLH, pushes it through the radiators, and returns it to the bottom of the LLH.
Click for a larger version. Detail from the figure above showing the flows and temperatures of the water flowing into and out of a ‘low loss header’ See text for further details
The LLH allows two hydraulic pumps to operate on the same heating system without competing with one another in an arrangement known as ‘hydraulic separation’. This is a smart idea when the system designer thinks that the flow impedance of the heating circuit might be too high for the hydraulic circulation pump in the heat pump to cope with.
Alternatively, the system designer may not know what the flow impedance of the heating circuit is going to be, and so they install as LLH as an insurance policy. They can be sure that the heat pump will be happy – and if the auxiliary pump is installed and sized correctly, it should allow the heat pump to match itself to any flow impedance.
So for example, if the heat pump is supplying 20 litres per minute of water at 55 °C to the ‘low loss header’ and the return water is at 53 °C then it is delivering 2.8 kW of heat.
- The hydraulic pump draws perhaps 4 litres per minute of water at 55 °C leaving 16 litres per minute to flow straight through the header.
- The return water from the radiators is cooled to (say) 45 °C.
- From this one can calculate that the radiators have provided heating of 2.8 kW. It’s good to see that low loss headers obey the law of conservation of energy!
- So at the bottom of the ‘low loss header’ there is a mixture of:
- 16 litres per minute of water at 55 °C
- 4 litres per minute of water at 45°C
- When mixed together this makes 20 litres per minute of water at approximately 53 °C which is returned to the heat pump.
So using a LLH allows the heat pump to operate at high flow rates and heat water only over small temperature differences. And it simultaneously allows the radiators to operate with lower flow rates and bigger temperature drops. For those with experience of electronics, this is analogous to using a ‘transformer’ to alter the effective impedance of a circuit. [Edited on 11 October 2023}
Why remove an LLH?
So if LLH’s are so clever why remove them? Reflecting, and discussing the installation with heat pump ninja Syzmon Czaban, I have come to the conclusion that the LLH in my installation is probably unnecessary.
I think it was installed because at the time when the system was designed and specified, things were still a little pandemical and chaotic. So Szymon has removed the LLH and I will monitor how the system performs over the coming winter. Take a look at Symon’s video of the process: the poor quality of some of the video is because I was holding the camera!
Why I think it won’t make much difference
The modified system is now simpler. The photos below show the situation inside the heat pump cupboard before and after the LLH and auxiliary pump were removed.
Click on image for a larger version. Left: The installation I have been using for the last two years. Right: the modified version with the Low-Loss Header and pump removed. The alteration has carried out so the components can be easily reinstated if this turns out to be a bad idea!
Now instead of the auxiliary pump driving water around the circuit, the heat pump will have to do that work, so the saving of energy from removing the auxiliary pump will be compensated for by the extra work done by the heat pump.
So the difference in performance will be just due to the changed hydronic circuit. However to deliver roughly 8,000 kWh of heat with a COP of 4.0 rather than 3.5 would require the installation to use only 2,000 kWh of electricity instead of 2,285 kWh of electricity. In other words, removing the low loss header would have to save around 285 kWh of electricity per year or around 2 kWh/day during the heating season. I find it hard to believe that this simple change could possibly make that much difference.
What next?
So if the Low Loss Header isn’t keeping the COP below 4, then what is? Well I don’t know, but further issues – spotted immediately by Syzmon – are possible recirculation of cold air near the heat pump, and upgrading the radiators. But these are issues for another year: the Heat Pump Experiment of the winter of 23/24 has begun!
Tags: Heat Pump
Protons for Breakfast 
October 10, 2023 at 7:52 pm |
Hi Michael I have an aroTHERM 5kW that variously charges two SunAmp heat batteries – a 6kWh for a basement apartment and 12kWh for the main terrace residence – and then heats or cools the main residence via a duct water-air heat exchanger in the MVHR. It includes some evolving control systems and a 25l volumiser. The batteries charge at high temperatures 60-65° but still achieve COPs of 3-3.5. During the heat wave this year the cooling cycle worked very effectively at COPs of 3.8-4. All this according to MyVaillant app which you’ve previously noted may not be entirely accurate.
We’ve been operational for just over a year now so still learning, debugging, and adding refinements. I’ll be following your tinkering with great interest. My ASHP installer and mechanical engineer were both very good but your ninja sounds interesting as well.
I am particularly enjoying generating hot water at a COP>3 which seems like a magic trick!
October 11, 2023 at 12:46 pm |
Guy, Good Afternoon.
That sounds like a very complex installation, but I like very much the idea of ducted air heating – I imagine would make for very comfortable living areas. My installation was so troubled that I didn’t dare activate teh cooling option and instead went for a separate (small) air conditioning installation in the bedroom. In this hot weather, it has been transformational.
And I am impressed by your use of Sunamp units: I did think about them but in then it was – like teh cooling options – one complexity too many.
Best wishes: M
October 10, 2023 at 10:21 pm |
Hi Michael, this sounds like a well-designed experiment, I’ll be interested to read your findings.
And… have you considered adjusting the settings on your heatpump so that it provides hot water to the taps in your household more efficiently during the periods when it’s not also keeping its radiators warm?
I’m asking this question because I had formed the *impression* (when eyeballing your electrical-consumption charts) that your heatpumps kicks in (roughly) every two hours. If so, then I think this is a rather expensive (with respect to battery-cell life-expectancy in your PowerWall) as well as slightly inefficient (roundtrip maybe 90% through the battery backup?) way to time-shift the energy produced by your rooftop photovoltaics. The usual practice here in NZ is to let the HWC temperature droop to 45 degrees (or even 40 degrees) when the sun isn’t shining, opportunistically routing any “excess” PV energy to the
HWC unless it’s already hot (at 60 degrees), and to use grid (or battery-supplied) energy only when necessary to keep it above its minimum “normal operating” temperature *or* (and this is quite important) to heat it to 60 degrees at least once every few days to keep the dreaded Legionella at bay. But… for all I know, there’s so little storage in your hot-water cylinder, and its piping is so poorly insulated, that it would not do an adequate job of delivering hot water for showers and washing dishes if its circulation pump were not running every two hours with a water temperature in excess of 50 degrees as it exits the cylinder??? But even in that case, the heatloss might be substantially lessened, during months when it is not heating radiators, by tweaking the low-end of the heatpump’s hysteresis band downward by a degree or three? And … any significant lessening of the annual throughput on your PowerWall would significantly increase its life-expectancy, I should think, if I’m right to guess that a PowerWall that survives its burn-in period is much more likely to die (slowly) of cell cycling (at a well-controlled cell temperature, charging rate, and discharging rate) than of any fault in its power electronics, heatpump, or control circuitry.
October 11, 2023 at 10:18 am |
Good Morning,
The DHW tank is heated intermittently on a timed-cycle.
In summer, the tank is topped up in mid-afternoon after the battery is likely full, but while there is still excess solar. Heating in the afternoon helps because the air temperature is near its maximum which reduces the required rise in temperature (improving COP). This operates with a COP of around 3 which makes better use of excess solar generation than a simple immersion heater.
In winter, the tank is topped up overnight using cheap rate electricity. The outdoor temperature is generally approaching its daily minimum but that can’t be helped. COP is still between 2.5 and 3.
Target DHW temperature is 50 °C which generally involves an hour long cycle heating the flowing water to about 55 °C. A 200 litre cylinder is enough for my wife and I for a day, but when the children are home we need to run the DHW cycle more than once a day.
There is a weekly anti-legionella run raising the water temperature to 60 °C – without any additional immersion heating required.
We also have a blending valve on the outlet of the cylinder set for (IIRC) about 48 °C.
In winter the grid connection is sufficient to allow charging of the battery and DHW operation of the heat pump.
BTW, the LEAF is working well and I am getting the swing of estimating the real range. I have been monitoring the battery degradation using LEAF SPY and it is incessant!
All the best
Michael
October 10, 2023 at 10:28 pm |
Hi MIchael,
I saw his video yesterday. So that was your house? Yes, I had wondered about the low loss header. When I had my system put in I was asking if it needed a buffer tank and I was told they fit a low loss header which is “like a buffer tank on steroids”. I had to stop him because he was getting too technical!
It turns out the supplier always recommends a low loss header and fitting one is implied as part of the warranty condition as with an unknown radiator system hydraulic separation provides a more defined operating condition for the heat pump.
It also provides good air and dirt separation in the large tube because the velocity is low.
Measuring the temperature differentials can be revealing as it might show “distortion” across the low loss header when the mean temperature on one side differs from the other. Having your heat pump run hotter than needs be to get the energy flowing round the other side as required leads to inefficient running.
Removing the secondary pump along with the low loss header may actually increase pump loss.
I was going to suggest adding insulation to the pipes where the header was removed but if it’s in your heated space it probably doesn’t mater. Mine’s in the loft so it gets cold.
It will be interesting to see your results.
October 11, 2023 at 9:51 am |
John, Good Morning.
Yes – so many questions that seem unfathomable, but once the results are in I suspect the answers will seem obvious!
We’ll find out in the spring.
All the best
Michael
October 13, 2023 at 5:48 am |
I have also changed from having flow separation to a direct system. My 50 litre buffer tank was quite short and squat, with the pipes to and return from the radiators only 2-3 inches apart. The pipes from and return to the heat pump were similarly close together. This meant that vertical mixing was inevitable. I measured the flow temperatures of all 4 pipes as John suggests and found a 4-5 deg C difference between flow temperature from the ASHP and the flow temperature to the radiators. This would be expected to lead to a 12-15% reduction in COP.
If a LLH is tall and thin, then this vertical mixing will be less pronounced. I am hoping to see an improvement in COP from removing my squat buffer tank.
I also wonder what thermostatic control you are using, Michael. A 3rd party thermostat acting as an on off switch has been likened to driving a car with the accelerator pedal pushed to the floor and using the ignition key to control the speed. The use of well adjusted weather compensation and no thermostat is supposed to give the best results.
October 13, 2023 at 9:43 am
Mike H, Good Morning,
What an interesting experiment. And I had not factored in the geometry of the LLH being tall and thin, but now you mention it it makes complete sense.
I’m afraid it’s now too late for me to make those temperature measurements :-(but I will be keeping track of how things go.
You asked about the thermostatic control. Last winter (2022/23) I used WC with room temperature modulation – I am not sure what algorithm this actually used, but it did work pretty successfully. RT didn’t move from 20 °C by more than about 0.5 °C all winter.
The previous winter (2021/22) I am not sure what I did, but I couldn’t get the temperature below 21 °C! so I just left it there – my wife was happy.
Now, I am working systematically with Szymon – trying to use pure WC and see how good it could be! Stay tuned!
M
October 11, 2023 at 10:11 am |
I’m not sure that the analogy of an LLH to a shunt is very good as shunt (just a simple resistor) will have losses whereas the LLH shouldn’t (apart from a bit of heat leakage).
I think it’s acting more like a transformer, in this case a step up transformer. The input side has a high flow (current) and low pressure and temperature differences (voltage) whereas the output side has a low flow (current) and high pressure and temperature differences (voltage).
October 11, 2023 at 12:41 pm |
Ed, Good Afternoon,
I agree!
In the first draft – and in my head – I used a transformer analogy, but at the end as I looked at the diagram and switched it to ‘shunt. Based on your comment I have switched it back! Thanks for getting in touch.
Michael
October 11, 2023 at 11:55 am |
Great article Michael, I very much look forward to seeing the results of your experiment in the spring.
Regarding your SCOP calculations, one thing to consider is the accuracy of your system flow temperature and energy measurement. If you refer to the excellent miniseries by Glyn Hudson which I highly recommend (link below) he concluded his Samsung ASHP was under reporting the flow temperature by 20% and over reporting the energy usage by 5%. Both of which resulted in a SCOP calculation of 30% lower than actual. He knows this because he installed very accurate independent sensors compared to Samsungs inbuilt sensors. Food for thought!
Glyn Hudson YouTube Heat Pump Miniseries:
October 11, 2023 at 12:38 pm |
Daniel, Good Afternoon.
Regarding the measurements, I am very aware of possible errors – my previous career was making ultra -recision measurements at the UK’s National Physical Laboratory. I use a Sonntex heat meter and Billing Grade electrical meter on the installation. I don’t pay attention to the self-reported measurements from the Vaillant unit which I found – up to now – to be inaccurate. I wrote about that here.
But thank’s for links to Glyn’s inspirational videos.
Michael
October 20, 2023 at 8:19 am |
Hi there,
I have a question related to heat pumps (albeit unrelated) to this article but I value the opinions on here so I thought I’d go for it.
I recently had a company come round and perform a heat loss survey and quote for a heat pump installation, replacing my gas boiler.
The outcome is that I’d need a 14 KW heat pump and the total cost of installation would be £10k, making it £2.5k after the government grant is applied. This includes decommissioning the gas boiler, and fitting a hot water cylinder. For context, I have underfloor heating throughout the downstairs and I’m not replacing any radiators (which are all upstairs).
Does £2.5k sound too good to be true? Every other quote I’ve received is (at least) £6k after the grant has been applied.
The two alarm bells that ring out for me are that the company (Megawave Energy) were incorporated 12 months ago and they import heat pumps from China and rebrand them with their logo. In itself not necessarily a problem but will I be stuck with a heat pump that is unserviceable except by this company?
In other news, my 7.5kw solar install is happening in 3 weeks time, along with the 13.5kw battery.
October 20, 2023 at 2:53 pm |
Karl, Good Afternoon.
First of all, the easy bit. Congratulations on your new solar & battery installation. I feel sure you will be very happy together. I expect it will change your life more than you expect.
Regarding the heat pump, I don’t have any knowledge of your home or the heat loss, so it is hard for me to comment. However I note that 14 kW is a very big heat pump. Whenever I talk to people about heat pumps I encourage them to think quite hard about it and act slowly. So in your case, if you feel uneasy, then I would advise you to wait until next summer for the installation. Things are not likely to get significantly more expensive or cheaper in that time. You can then use this winter to measure the actual heating power required in your home.
To do this, read your gas meter weekly (I recommend a Saturday or Sunday morning) and then work out how much gas you use week by week. Find out how much gas you use (in kWh) in the coldest week of the year. Divide the total by 168 hours to get an estimate of the average heating power of your gas boiler. You can then see if it is close to 14 kW. You can do this day-by-day if we have one or two very cold days. 14 kW corresponds to 336 kWh per day which is a very large amount of heat.
Alternative, (or in addition) spend £300 or so and have an independent MCS heat loss survey. This will likely over-estimate your heat loss.
If you do this by next spring you should feel more confident about whether the heat pump is the right size.
Regarding installers, if you can find one, see if you can get a quote from a Heat Geek trained installer.
Do get back to me if you have any specific questions.
Michael
October 20, 2023 at 6:25 pm
Hi Michael,
Thanks for taking the time to reply. I did think a 14 kW heat pump sounded a bit much. Although I’m guessing that 14 kW is the max power output and it should rarely run at 14 kW (or at least not for very long).
I’m getting a smart meter installed next week by Octopus so hopefully measuring kWh used per week will become easier. I have last year’s figure for the coldest month and we used 1400 kWh that month. That only works out at around 2 kWh per hour (we still had the heating turned right down last winter).
Measuring it on the coldest day will be useful as I guess you want to size the pump to be able to cope with heating demand on the coldest days (not just the average winter day).
I have looked at Heat Geek but am admittedly put off a bit by their pricing structure. They charge £2200 to “…assure that the performance set out in our proposal is met”.
Surely, if they set out a performance target in their proposal then its incumbent on them to meet that target. Not charge extra to meet it.
Anyway, I’m getting ahead of myself. I’ll take your advice and bide my time this winter.
Thanks again,
Karl
October 20, 2023 at 9:34 pm
Karl, Good Evening,
If you used 1400 kWh in the coldest month then your home was adequately heated with an average of (1400/30) which is about 50 kWh/day which is equivalent to 2 kW on continuously. Even if you doubled that or tripled that, it seems unlikely that a 14 kW heat pump would be appropriate. Heat pumps can reduce their power but below about 40% of their nominal power they switch on and off which is thought – possibly incorrectly – to be bad for them. This was teh case with early heat pumps but I doubt it is true now.
Regarding Heat Geek. There are Heat Geek Trained installers such as Syzmon Czaban (https://www.urbanplumbers.co.uk) who runs a great You Tube channel. And there are installations by the heat geek company.
TO assure the performance of a heat pump is not simple. The key things you want to know is how much heat was delivered (kWh_th) and how much electricity was used (kWh_e) and COP, the ratio of the two. Heat pumps typically report COP but their estimates are not trustworthy. To assure performance, one needs to install a separate monitoring system. This includes a small web-based microcontroller, a heat meter, an electricity meter, and a couple of thermometers. The results are available to you and the company and can independently verify how well the installation is working. The hardware costs quite a bit and the installation is beyond my capabilities.
You can see the results of people who have elected to publish their data on this web site:https://heatpumpmonitor.org you can sort by installer and see how they perform.
If you are going to measure over the winter, then make sure the house is at the temperature you want so that the measurements are meaningful. In other words: don’t be cold!
Best wishes
Michael
October 26, 2023 at 12:06 pm
Hi Michael,
Thanks for the response and the referral of Syzmon Czaban. I’ve managed to book an MCS-accredited heat loss survey with him in early January.
Regards,
Karl
October 26, 2023 at 12:18 pm |
Karl: Great: I am sure Syzmon will get things sorted. M
December 14, 2023 at 8:24 pm |
Hi Michael, I like your articles in general and have recently challenged the Tritium rarity for the success of Fusion Energy. Tokamak say they have Tritium Breeding Blankets to help generate more.
December 14, 2023 at 11:22 pm |
Peter,
Good Evening. Thank you for letting me know you like the articles: I appreciate your letting me know.
The “Tritium” problem is just one of the problems facing a practical fusion reactor. Let’s assume that a fusion reactor can be built and made to operate at 90% up-time for 25 years. I don’t think that will ever happen but let’s start there.
How is then energy extracted from the reactor? The plan calls for neutrons carrying perhaps 100 MW of energy (equivalent heating power to 100,000 1-bar electric fires) are absorbed in a ‘blanket’ of heated liquid lithium compounds surrounding the reactor. In order to be sustainable, every neutron must generate act least one atom of tritium. The plan calls for the tritium to be generated by a nuclear reaction with the lithium. This will release tritium – a gas – which will bubble through the molten lithium blanket to be captured and purified.
Obviously not all the tritium will be captured, and some neutrons will travel straight through the blanket. To make up for this loss of tritium, the neutrons will also hit a ‘neutron multiplier’ which will generate typically 2 or 3 neutrons each neutron by which it is struck.
I agree that this process is technically possible: but it is extraordinarily difficult and none of the steps (to the best of my knowledge) have been validated at anything close to the scale required.
On current timelines, companies with address these problems in perhaps the 2040’s or 2050’s – too late to be relevant to our current climate crisis.
Best wishes
Michael
December 15, 2023 at 12:10 pm
On another track, found your Blog as a result of the IET lecture event at The Adelaide. Thought the practical Greener energy focus would interest the Chessington District Residents’ Association in RBK especially as you now highlight the shortcomings of neighbouring Richmond BK. The Secretary has forwarded Blog just to the Executive Committee members and bearing in mind that they can get to Surbiton of Kingston talks by either car or TfL buses. I think the Kingston LibDems could note the findings of Richmond’s records since their RBK Council also boast ‘wonderful’ green policies as they actively progress with over-development of the borough by allowing the building of high-rise apartment blocks wherever property developers see fit for making maximum profit.
January 14, 2024 at 5:07 pm |
[…] place for heat pump installation. As shown in different situations, like the experiments explained in Protons for Breakfast, having a heat pump in the cellar comes with benefits such as natural insulation and effective heat […]
January 14, 2024 at 5:45 pm |
If you are considering this: please check where the heat is being collected from!
June 8, 2024 at 6:00 am |
Hello Michael,
If there’s a “results” post, you should link to it and the end of the article 🙂
June 8, 2024 at 8:56 am |
Dan, the conclusion is yet to be written… I am behind with my blogging! But it’s a good idea to put a link in the original article. Thanks: M
January 6, 2025 at 1:09 pm |
Hello, it would be great if you could update us on the performance without the buffer tank.
thank-you
January 8, 2025 at 12:36 am |
Jack,
Good Evening. Please forgive my delay in replying: it’s been a difficult few days.
I’m afraid I don’t understand your comment.
You are commenting on this article which describes the experiment, and which contains a clear link to this article which describes the result of the experiment.
What is it that you require?
Best wishes
Michael