Posted by: @mike-h

↑

Beyond the upstairs radiator, a 15 mm pipe supplies 4 radiators and another 15 mm pipe supplies another 4 radiators. 

If the buffer tank and flow separation were removed and the heat pump was running at a flow rate of 30 L per minute, then the flow rates in the 22 mm pipework would be 30 L per minute and in the 15 mm pipework the flow rate would be around 11 L per minute. From my reading of HeatGeek articles and others, these flows are too large for the size of pipework. 
Replacing the pipework that supplies my

I’ve been evaluating the similar option for my 8.5 heat pump. Although your model is different to mine  Several things in the info you have given suggest certain limitations…

1. You have 4 rads off one 15mm pipe which has a limited energy carrying capacity. (2.75kw at 0.9mps and a DT of 5c) 

2. a flow rate of 30 ltrs per minute seems quite high and doesn’t allow or will struggle to absorb enough energy through the HP since they are typically optimised at about 10 to 20 ltrs per minute and at that flow rate only delivering a DT of about 5C to 8c. 

3. The data on your HP model should specify the minimum primary flow pipe diameter and I would have expected a HP without hydraulic separation with a 12kw output to require 28mm primary pipe diameters continuing to the first split of the central heating pipework where probably 2 x 22 mm pipes could take over. 

I have tuned my DT to about 6 or 7c with a flow rate of approx 12 LPM. We also have a maximum of 2 rads off a 15mm pipe with typical 1.2kw load each. 

we are still learning but feel we are headding in the right direction. 

Posted by: @sunandair

↑

a flow rate of 30 ltrs per minute seems quite high ..... since they are typically optimised at about 10 to 20 ltrs per minute

The min and max flow rate for the 12kW Samsung Gen6 is 12 & 58 lpm respectively, so I am interested in how you avoid 911 errors running below the minimum.  [I am not sure if Mike has a Gen6 - but can't think it would be that different if it was an earlier model]

My advice based on the experience of a DIY install of my own 16kW Samsung Gen6 is to "do the maths".  I did not know all the maths when I started and I have undersized my pipework a bit for optimal perofrmance.  I installed 28mm, but probably should have put in 35mm [due to having 44 meters of pipe run due to placement of the heat pump]

Posted by: @mike-h

↑

Am I right in thinking that this would be the only way to achieve an efficient system?

You can wear more clothes or move your house to a warmer location 🤣

Replacing the pipework that supplies my radiators sounds like a major project.

Yup, I did that already, and should have put in 35mm not 28mm (at least to the first manifold) - the heat pump is a long way from our house.

I wonder if it explains in part the frequent use of 4 pipe buffer tanks.

Probably, the installers do not want to come back because your heat pump keeps switching off.  Buffer tanks are for reliability, not efficiency and enable an install without very detailed calculations and some risk taking.

This will be a longish post, bear with me.

The "carnot equation" only has 2 parameters in it.  Flow temperature and Ambient temperature.  You can only [realistically] control flow temperature in an ASHP (though over a 10 year period you should be able to get a good idea, for your location, what ambient temperature will do through the year)

Efficiency:- The lower you can get your flow temperature, the more efficient your system will be. 

Unfortunately, while the optimisation process starts out with a very simple single factor to control, flow temperature, this parameter is coupled to many other parts of the system, and there is more maths involved there.  I have done a DiY install of a 16kW Samsung Gen6 heat pump, I got some stuff wrong [missed the maths in places] but here are my thoughts.

Setting up an efficient system needs you to

1. Understand [and do] the maths
2. Understand the limitations of your specific circumstances and work within those

The ambient [outside] temperature, size and insulation levels dictate your heat load.  Each room has a specific heat load, and your emitters [panel radiators/UFH] can meet that heat load at a specific average flow temperature and room temperature.  You will need a specific flow rate and delta T to deliver the required heat.  Your pipe sizes will limit your max flow rate.  Best practice suggests not going over 0.9 m/s flow speed (for noise and other reasons).  [I have been forced to go a bit over this for some days in the year, I am ok with that]  While this information is probably not directly helpful, it is important to demonstrate that the coupling between the carnot equation turns something simple into something that requires a lot of maths and engineering knowledge and experience.

I am still trying to fix the stuff I got wrong through lack of knowledge (my final efforts are on reducing the phenomenal 10m head my system has)

I would take the following approach.

1. using the heat load [at the design temperature you chose] work out the flow rate & dt needed.  This is easily done in a spreadsheet with dt and flow on the two axes (you will have many combinations of dt and flow that could, theoretically meet your need).  Samsung documentation says min/max for heating dT is 5-15 (though mine accepted a setting of 3 and delivers 3 - I think recommended is 5)
2. then work out the average radiator temperature needed to emit required heat at design room temperature for that heat load
3. check your radiators will actually deliver the required dT at that temperature (if not loop back to your spreadsheet)
4. calculate the resistance of your pipe work and all components on it (this is important to ensure your pump can achieve your desired flow rate on a single circuit)-(ignore vertical head in a closed loop system - so only pipe and components should be counted)
5. check the flow speeds for your diameter pipes at the flow you have ended up with are within tolerances
6. check that the pump you have (or will buy) can meet that flow rate you have chosen at the pump head you calculated

Repeat this process until you get to the numbers you need.  MCS design temperature is to cover 99% of the days in the year.  My approach was to accept less than that, and wear more clothes on those days, in the interests of efficiency.  (it costs me nothing to put a comfortable jacket on for a couple extra days of the year).

At the end of this you should have the following "numbers"

Flow rate, delta T, flow speed, total head in the circuit, pump flow rate capacity at calculated head

It should be reasonably safe to proceed with changes at this point if the flow rate is sufficient for your heat pump.

** NOTE **, if you proceed without being certain that you can

1. consistently achieve a flow rate sufficiently above the minimum your heat pump requires
     and
2. the radiators can deliver the dT at the flow temperature you chose

you will end up with a heat pump that will keep switching itself off as I have experienced - see first picture (result of flow temp not compatible with emitters)

@william1066 many thanks for your detailed reply. The current flow rate for my heat pump is around 22 litres per minute (it is a non PCM pump), but I have no way of knowing the flow rate to my radiators. I too am happy to wear a bit extra when the ambient temperature drops really low. It looks as if I will need to live with my buffer tank or have new pipe work fitted.

@mike-h With your Internal Diameter of 20.6 and flow of 22 you have a speed of 1.1 m/s.  That will give you 6.27 kW of heat at dT 5.  I will, at some point, find or build a google spreadsheet with the calculations mentioned in my post, then you can take a look.  I think that taking your time is the right thing to do. If you did the maths on the efficiency delta between having and not having a buffer thank, it is probably not great in the short term.  Over 20 years, it likely adds up, but I think taking your time to make the right changes is important.

Posted by: @mike-h

↑

I have a 12 kW Samsung air source heat pump with a 4 pipe buffer tank and flow separation. As @heacol points out, this is an inefficient system. I would love to get rid of my buffer tank, but my problem is the size of my existing pipework. I have 22 mm piping that runs from the buffer tank and supplies two downstairs radiators and a third upstairs radiator. Beyond the upstairs radiator, a 15 mm pipe supplies 4 radiators and another 15 mm pipe supplies another 4 radiators. 

If the buffer tank and flow separation were removed and the heat pump was running at a flow rate of 30 L per minute, then the flow rates in the 22 mm pipework would be 30 L per minute and in the 15 mm pipework the flow rate would be around 11 L per minute. From my reading of HeatGeek articles and others, these flows are too large for the size of pipework. 
Replacing the pipework that supplies my radiators sounds like a major project. Am I right in thinking that this would be the only way to achieve an efficient system? I doubt if I am the only one with this problem, so I wonder if it explains in part the frequent use of 4 pipe buffer tanks. 

However, if it is working now and heating your house, there can be a significant improvement by removing the buffer tank and installing a bypass at the furthest point (after the first 3 radiators), removing the TRV's on those radiators and all other radiators in common areas, replacing with high flow lock shields (go to the merchant, ask to see all the lock shields they have and look how big the hole is that allows the water through, choose the biggest one, usually the cheapest), allowing full flow and maximum heat output, use the Samsung controller to control the internal temperature, not third party thermostats, retain TRV's in rooms that you want cooler. I would install at least 3, 22 mm auto bypass valves, in parallel to ensure good flow. Remove all low flow boiler valves (the ones with a silver box on top of the valve) and install a high flow valve that will not sap the pump pressure, remove the flow meter and the magnetic filter if it is installed. Install a 1/12" Y strainer and if you can, glue a small magnet in the lid. The idea is to stop unnecessary parasitic loads on the pump, so it can do the job that it is there to do, push water to the radiator. Once this is done, close the TRV valves on the radiators, run the heat pump and open the auto bypass valves until it stops faulting on low flow. If you want to go a bit further, install the biggest radiators in the primary 22 mm circuit that you can.

This is in no way the ideal solution, but you will get significantly grater performance than you currently have. As I said, we do not live in a perfect world and I realize that it is not always possible to achieve a perfect scenario, but the best you can get with your current installation, will still be cheaper to run than a gas or oil boiler, and significantly cheaper than it is currently.

@heacol I recently ordered a 28mm fill/flush/flow setter valve and was shocked at how small bore it was internally. Don’t know if I’ll bother fitting it on my system…

Posted by: @sunandair

↑

a flow rate of 30 ltrs per minute seems quite high ..... since they are typically optimised at about 10 to 20 ltrs per minute

 

The min and max flow rate for the 12kW Samsung Gen6 is 12 & 58 lpm respectively, so I am interested in how you avoid 911 errors running below the minimum.  [I am not sure if Mike has a Gen6 - but can't think it would be that different if it was an earlier model]

@william1066 thanks for your feedback. Totally agree about doing the maths. 

But I must admit I had no idea Samsung flow ranges were so high, though I thought I’d read somewhere heat pumps generally operated in the same way with similar parameters. Clearly not.

I started out planning to give Mike some general principles since nobody had answered at that point and his post looked a bit lonely. 

I have an Ecodan 8.5 which, now I know has a much lower flow range than your Samsung. (Circa 10.8 to 24 L)  Thanks for the extra information. The data sheet also states 22mm primary is acceptable but we opted for 28mm primaries.

On our installation I have recently made a massive leap in COP from the original installer settings so was a bit excited to hint at the changes that worked. We dropped our pump speed within the allowed range until our DT hovered just a little above 5c. This simple act changed our COP from 1.49 to a more promising 3.17.

In answer to your question, the ecodan flow error on our settings doesn’t cut in until below 5 ltrs per minute. I’m not sure but I think this may be adjustable since I’ve read that some writers have stated flow rates even lower….

@heacol system filtration?

I’ll stick the flow setter on eBay.

@hughf 😀  A standard Y strainer is adequate if your system is clean and well treated. We ise 1 1/2 " for flow.