Mitsubishi Ecodans ASHP issues...

[Gary]

Rise in dT would be the more direct method, but I'm not aware of anyone doing it that way.

[desA]

Rise in dT would be the more direct method, but I'm not aware of anyone doing it that way.

Based on a reduction in evap performance due to ice blockage?

[Gary]

Rise in dT = reduction in airflow

[Gary]

The only other way to substantially raise the dT is to speed up the compressor.

[desA]

For AWHP's, as Tw,out rises, Q'evap reduces, hence dT,evap will reduce automatically. This may present a challenge.

[Gary]

I'm not certain that the difference would be substantial, but in any case we would want the dT defrost initiation point to exceed any dT which could be caused by water temp.

More problematic would be defrost cycles initiated by a dirty coil or fan problems.

[MikeHolm]

Guys, in what way would the seemingly unlimited quantity of heat available at the condenser help in regaining the balance? Assuming that there was a proper sensing function at the evap of course. Boy do i risk sounding dumb.

[MikeHolm]

The question is more about the quantity of heat (or work, not the same thing I know) provided by the comp vs the coming from the condenser.

[desA]

Guys, in what way would the seemingly unlimited quantity of heat available at the condenser help in regaining the balance? Assuming that there was a proper sensing function at the evap of course. Boy do i risk sounding dumb.

Sounds just right for reverse-cycle (4-way valve) defrost.

[fixmyheatpump]

Well the installers have been intouch. They're going to have a good poke about the system with a Mitsubishi tech and see if they can find the problem.

I think it's bridging like other people have said. Nothing can be done about it, they're simply no good for my particular case, probably a very good system for somewhere that isn't as extreme.
Just no good for somewhere that drops below -15'c and then stays around -5'c for weeks at a time.

I honestly don't think they'll ever get it to work reliably, time to push for a GSHP system I think, at least that won't be affected by snow and ice

It won't be bridging I've never heard so much bull, and they work at -25C as stated in the manual. We have numerous systems installed and working fine at -15C and some of these are the W85 models which do not have the same low ambient technology as the W14. We've had constant -5C and drops to -15C all over the UK this and last winter so it's not just Scotland that gets these temperatures.

Everything points to the hydrolic water side of the system. This is obvious because you mention you have 4 x 16/60 Alpha pumps in series (2 x push - 2 x pull), this is flawed. By installing pumps in series you only increase the pressure they can overcome not the flow rate, this is done by installing them in parallel. The maximum flow you could possibly be getting out of these with a very low resistance is around 40 l/m, with the plates in the heat pumps and a radiator circuit you have a high resistance so must have low flow rate.

These units need between 20 & 40 l/m each constantly, but Mitsubishi recommend you are as near to the top end as possible, in your case it is between 40 & 80 l/m (nearer to 80 if poss) total balanced equaly between the two units, 15/60's in series are not capable of this no matter how many you have installed. If there is not enough flow there is not enough energy in the water circuit and the heat pumps will stop defrosting to prevent the water in the plate heat exchanger from freezing and cracking it, this would be the only reason they would partially defrost and stop which can be seen in one of your pictures where the top of the coil is defrosted but the bottom is not. Once you get a ON/Off demand from the stat or clock this cycle will start again. At the end of a defrost the fan spins at full speed which blows the moisture away to prevent "bridging" I've witnessed this hundreds of times and it works, so much steam comes out of the front of the units some customers have called thinking they're on fire.

A few questions - Have you got a low loss header installed? If not you need one to get the required flow rates when running two heat pumps on the same hydrolic circuit.

Have you got two flows and two returns coming from the header to outside, one for each heat pump with their own circulator i.e pumps in parallel? If not you need them.

Have you got a flow setter valve installed on either the flow or return to each unit, so two in total? If not you require them to prove and balance the flow rates.

As the whole system is massively oversized for the load you should have some control to switch off one units above a certain ambient to stop the units from cycling, this is simple and cheap with an external thermostat, have you got this installed?

If no to any of these then it is an application & installation error, fact.

This is an age old case of proven equipment getting the blame for poor application and installation, can't see the wood for the trees and blaming something we don't understand springs to mind.

If you want to waste your own money by ripping them out and installing a ground source system that's your prerogativebut don't be upset when you hear of a neighbour with an air source systems that is working fine after being applied and installed correctly.

[Gary]

If there is not enough flow there is not enough energy in the water circuit and the heat pumps will stop defrosting to prevent the water in the plate heat exchanger from freezing and cracking it...

Now, that makes sense.

[Bigfreeze]

The question is more about the quantity of heat (or work, not the same thing I know) provided by the comp vs the coming from the condenser.

The heat output from the compressor is only the equiqvalent (or close to) of the electrical input. All other heat output must be collected at the evap. Thats why you're inabilty to collect energy, due to cold weather or poor defrost, has the overriding effect it does on the COP.

[fixmyheatpump]

But... but... but... The defrost cycle terminates before the coil is fully defrosted.

Exactly Gary, as the system works with multiple sensors and very complicated software algorisms it is extremely intelegant and wouldn't stop defrosting until the required parameters were achieved, unless it was protecting itself for some reason.

There are sensors on the water flow pipe and in the plate heat exchanger which it would never let these drop below freezing as it would freeze the water and crack the heat exchanger. If the flow rates are too low then there is no energy in the circuit hence the loop drops to near zero and it stops to protect itself.

They also need to run for 7 minutes before they defrost, the reason is to put enough energy into the loop before it tries to take it out again for defrost. If the system's cycling because they are oversized there is a potential this 7 minutes is not elapsing but again this is an application issue and can be overcome with some simple controls.

I've asked Mitsubishi for their recommendations on combining multiple Ecodans onto the same hydrolic loop just before we took on our first project of this type on. They told me they don't recommend or support it, the reason they gave is not the fact that it can't be done it's because it is more complicated to get right and they don't want people messing it up and giving their equipment a bad name.

In this application the installer has designed it themselves and made a pigs ear of it, this whole thread and equipment bashing is the reason Mitsubishi do not recommend or support this type of set-up. Very frustrating!

[Bigfreeze]

It won't be bridging I've never heard so much bull, and they work at -25C as stated in the manual. We have numerous systems installed and working fine at -15C and some of these are the W85 models which do not have the same low ambient technology as the W14. We've had constant -5C and drops to -15C all over the UK this and last winter so it's not just Scotland that gets these temperatures.

Everything points to the hydrolic water side of the system. This is obvious because you mention you have 4 x 16/60 Alpha pumps in series (2 x push - 2 x pull), this is flawed. By installing pumps in series you only increase the pressure they can overcome not the flow rate, this is done by installing them in parallel. The maximum flow you could possibly be getting out of these with a very low resistance is around 40 l/m, with the plates in the heat pumps and a radiator circuit you have a high resistance so must have low flow rate.

These units need between 20 & 40 l/m each constantly, but Mitsubishi recommend you are as near to the top end as possible, in your case it is between 40 & 80 l/m (nearer to 80 if poss) total balanced equaly between the two units, 15/60's in series are not capable of this no matter how many you have installed. If there is not enough flow there is not enough energy in the water circuit and the heat pumps will stop defrosting to prevent the water in the plate heat exchanger from freezing and cracking it, this would be the only reason they would partially defrost and stop which can be seen in one of your pictures where the top of the coil is defrosted but the bottom is not. Once you get a ON/Off demand from the stat or clock this cycle will start again. At the end of a defrost the fan spins at full speed which blows the moisture away to prevent "bridging" I've witnessed this hundreds of times and it works, so much steam comes out of the front of the units some customers have called thinking they're on fire.

A few questions - Have you got a low loss header installed? If not you need one to get the required flow rates when running two heat pumps on the same hydrolic circuit.

Have you got two flows and two returns coming from the header to outside, one for each heat pump with their own circulator i.e pumps in parallel? If not you need them.

Have you got a flow setter valve installed on either the flow or return to each unit, so two in total? If not you require them to prove and balance the flow rates.

As the whole system is massively oversized for the load you should have some control to switch off one units above a certain ambient to stop the units from cycling, this is simple and cheap with an external thermostat, have you got this installed?

If no to any of these then it is an application & installation error, fact.

This is an age old case of proven equipment getting the blame for poor application and installation, can't see the wood for the trees and blaming something we don't understand springs to mind.

If you want to waste your own money by ripping them out and installing a ground source system that's your prerogativebut don't be upset when you hear of a neighbour with an air source systems that is working fine after being applied and installed correctly.

The issue with these units is not their inability to do the job in hand. It is their ability to do it efficiently. Sure the system above may well be designed poorly on the hyraulic side and the defrost may not be fully completing because of water flow (which I alluded to in an earlier post). The plain fact of the matter is that the evaporator on these units is poorly designed for the job that is required of it.

You can argue all you want, but anyone who has an ounce of refrigeration knowledge will tell you that the closer your fin spacing the faster you will get ice formation on the coil. In a system that is designed to run for up to 24hrs a day in moist conditions, this is a fatal flaw. Defrost removes vital energy from a house and costs money for no net gain. The more you must defrost and the longer the defrost period, the more it costs.

Anyone with any knowledge of heat pump design knows that, when dealing with a GSHP you can only collect a certain amount of energy from a certain amount of area before you begin to freeze up the ground due to its inabilty to replenish at the rate you extract. That very same principal stands true with ASHP and evaporators. You undersize and you freeze. Freezing means a lower COP, more defrosts and high energy bills. Now you can poo poo other peoples observations on here all you want. The fact remains that they are valid points and just because you install them and have seen them work, doesn't make a case for their efficiency. And at the end of the day, heatpumps are about efficiency.

[fixmyheatpump]

The issue with these units is not their inability to do the job in hand. It is their ability to do it efficiently. Sure the system above may well be designed poorly on the hyraulic side and the defrost may not be fully completing because of water flow (which I alluded to in an earlier post). The plain fact of the matter is that the evaporator on these units is poorly designed for the job that is required of it.

You can argue all you want, but anyone who has an ounce of refrigeration knowledge will tell you that the closer your fin spacing the faster you will get ice formation on the coil. In a system that is designed to run for up to 24hrs a day in moist conditions, this is a fatal flaw. Defrost removes vital energy from a house and costs money for no net gain. The more you must defrost and the longer the defrost period, the more it costs.

Anyone with any knowledge of heat pump design knows that, when dealing with a GSHP you can only collect a certain amount of energy from a certain amount of area before you begin to freeze up the ground due to its inabilty to replenish at the rate you extract. That very same principal stands true with ASHP and evaporators. You undersize and you freeze. Freezing means a lower COP, more defrosts and high energy bills. Now you can poo poo other peoples observations on here all you want. The fact remains that they are valid points and just because you install them and have seen them work, doesn't make a case for their efficiency. And at the end of the day, heatpumps are about efficiency.

I think you're missing the point, the reason Tired Geek originally posted on here was because he wanted advice on why the units are not completing a defrost cycle and freezing up, not because he wanted to hear opinions on ground source being better than air source by ground source salespeople.

So, the reason for his problem is because the units do not have the correct flow rates and possibly because they are cycling, they have not been applied correctly and that is nothing to do with the equipment. Once this is corrected they will not freeze up, the system will not cost as much to run and the fin spacing will stay the same, so that is why those observations can be poo pooed. What you are suggesting is the same as saying "I have an unleaded car I've put diesil in the tank and the reason it doesn't work is because the wheels are too small".

I'm a fridge engineer by trade but that's irrelevant, what you are actually saying in the second paragraph is that Mitsubishi's doctorate design engineers in Japan who designed and patented the Zubadan flash injection circuit/compressor to guarentee operation down to -25C don't have an ounce of refrigeration knowledge. Anybody who has an ounce of common sense would probably disagree with this. I think you need to have a read up on the technology, try here .... h ttp://w w w.mitsubishi-electric-aircon.de/eng/zubadan.php ..... if you want a further explanation let me know and i'll run through it with you.

At the end of the day heat pumps are not just about efficiency they are about cost as well, the amount it costs to install them and the amount at which you will save money over an alternative heating system. If a system costs £20,000 to install have a season COP of 5 but only saves you £300 a year you would be daft to install it as a money saving decision.

The fact that we've installed this system for numerous people who have all seen a drops in their fuel bills and have worked fine through this winter and last does infact make a case for their efficiencies. The fact that I've been to many systems that were performing poorly and found installation issues (including in Scotland) also makes a massive case for what needed to be said here.

[desA]

If there is not enough flow there is not enough energy in the water circuit and the heat pumps will stop defrosting to prevent the water in the plate heat exchanger from freezing and cracking it, this would be the only reason they would partially defrost and stop which can be seen in one of your pictures where the top of the coil is defrosted but the bottom is not.

Would there be an advantage to have a hot-water storage/buffer tank in the circuit, which can allow the heat-pumps to continue full defrost, without robbing the house of too much heat?

[fixmyheatpump]

Would there be an advantage to have a hot-water storage/buffer tank in the circuit, which can allow the heat-pumps to continue full defrost, without robbing the house of too much heat?

Yes a low loss header is just a small buffer vessel so it's the same principle, if you draw off to both units individually with 2 x 15/60's on each circuit you should get the correct flow rates and around 5 degree delta T per unit. You can then draw off to the radiators at a lower flow rate and larger delta T if required and the energy required to defrost will be in the circuit. When the flow rates are low the flow temperature increase quickly and the unit achieves its target so thinks job done i've put the required energy in, however it hasn't so the unit is basically tricked.

As both units are running together and massively oversized a buffer would reduce cycling so this might be the best way to go here but I personally think it would be better from a running cost point of view to stop one of them and it's circulators when the ambient increases and the property load drops.

[mad fridgie]

It won't be bridging I've never heard so much bull, and they work at -25C as stated in the manual. We have numerous systems installed and working fine at -15C and some of these are the W85 models which do not have the same low ambient technology as the W14. We've had constant -5C and drops to -15C all over the UK this and last winter so it's not just Scotland that gets these temperatures.

Everything points to the hydrolic water side of the system. This is obvious because you mention you have 4 x 16/60 Alpha pumps in series (2 x push - 2 x pull), this is flawed. By installing pumps in series you only increase the pressure they can overcome not the flow rate, this is done by installing them in parallel. The maximum flow you could possibly be getting out of these with a very low resistance is around 40 l/m, with the plates in the heat pumps and a radiator circuit you have a high resistance so must have low flow rate.

These units need between 20 & 40 l/m each constantly, but Mitsubishi recommend you are as near to the top end as possible, in your case it is between 40 & 80 l/m (nearer to 80 if poss) total balanced equaly between the two units, 15/60's in series are not capable of this no matter how many you have installed. If there is not enough flow there is not enough energy in the water circuit and the heat pumps will stop defrosting to prevent the water in the plate heat exchanger from freezing and cracking it, this would be the only reason they would partially defrost and stop which can be seen in one of your pictures where the top of the coil is defrosted but the bottom is not. Once you get a ON/Off demand from the stat or clock this cycle will start again. At the end of a defrost the fan spins at full speed which blows the moisture away to prevent "bridging" I've witnessed this hundreds of times and it works, so much steam comes out of the front of the units some customers have called thinking they're on fire.

A few questions - Have you got a low loss header installed? If not you need one to get the required flow rates when running two heat pumps on the same hydrolic circuit.

Have you got two flows and two returns coming from the header to outside, one for each heat pump with their own circulator i.e pumps in parallel? If not you need them.

Have you got a flow setter valve installed on either the flow or return to each unit, so two in total? If not you require them to prove and balance the flow rates.

As the whole system is massively oversized for the load you should have some control to switch off one units above a certain ambient to stop the units from cycling, this is simple and cheap with an external thermostat, have you got this installed?

If no to any of these then it is an application & installation error, fact.

This is an age old case of proven equipment getting the blame for poor application and installation, can't see the wood for the trees and blaming something we don't understand springs to mind.

If you want to waste your own money by ripping them out and installing a ground source system that's your prerogativebut don't be upset when you hear of a neighbour with an air source systems that is working fine after being applied and installed correctly.

Firstly I think you mis understand bridging, we are not talking about between the fins, but between the fins and the casing, and this is caused by external factors 'snow'. The coil could be completely clear but ice/snow is still on the case. You do have to be a bit unlucky for this to occur, low density snow, prolong periods of temps below 2C
Re water pumps. water pumps are pressure/flow related, increase the pressure drop reduce the flow,
So unless there is no pressure drop in your system, your statement that series pumps will not increase flow is incorrect. The pump selection is very likely to be incorrect, but without having detailed info its a blunt statement.
I am a believer of dedicated water pumps for the heat pumps and deadicted water pumps for the application (low loss header, or balancing pipe), except if there is no control on the heating circuits (underfloor system, constant flow)
Lets look at how much energy is required for defrost.
Face area of the coil looks to about 1M2, lets say ice is 5mm thick (enough to stop air flow), that will give 5litres of ice or just under 5kg of ice (we will call it 5Kg) 5*334= 1670Kj, lets round up to 2000Kj to allow for the coil.
Lets look at your water flow 30l/min (0.5kg/s)
.5*4.2*5 (10.5kj/s) defrost is 7 minutes (420second)
10.5 *420= 4410Kj
thats more than twice whats required.
how much water is required in the system to ensure defrost. Water temp start 35C (for rads i would say quite a bit higher), min water temp 10C to ensure freezing does not occur. 25C TD
4410 / 25 /4.2= 42 LITRES is required in the system to ensure correct defrost. Of course we still have the energy from the compressor to be added to the 4410.

[Bigfreeze]

I think you're missing the point, the reason Tired Geek originally posted on here was because he wanted advice on why the units are not completing a defrost cycle and freezing up, not because he wanted to hear opinions on ground source being better than air source by ground source salespeople.

So, the reason for his problem is because the units do not have the correct flow rates and possibly because they are cycling, they have not been applied correctly and that is nothing to do with the equipment. Once this is corrected they will not freeze up, the system will not cost as much to run and the fin spacing will stay the same, so that is why those observations can be poo pooed. What you are suggesting is the same as saying "I have an unleaded car I've put diesil in the tank and the reason it doesn't work is because the wheels are too small".

I'm a fridge engineer by trade but that's irrelevant, what you are actually saying in the second paragraph is that Mitsubishi's doctorate design engineers in Japan who designed and patented the Zubadan flash injection circuit/compressor to guarentee operation down to -25C don't have an ounce of refrigeration knowledge. Anybody who has an ounce of common sense would probably disagree with this. I think you need to have a read up on the technology, try here .... h ttp://w w w.mitsubishi-electric-aircon.de/eng/zubadan.php ..... if you want a further explanation let me know and i'll run through it with you.

At the end of the day heat pumps are not just about efficiency they are about cost as well, the amount it costs to install them and the amount at which you will save money over an alternative heating system. If a system costs £20,000 to install have a season COP of 5 but only saves you £300 a year you would be daft to install it as a money saving decision.

The fact that we've installed this system for numerous people who have all seen a drops in their fuel bills and have worked fine through this winter and last does infact make a case for their efficiencies. The fact that I've been to many systems that were performing poorly and found installation issues (including in Scotland) also makes a massive case for what needed to be said here.

I've already addressed the issues that tiredgeek had with flow etc on this and other threads and have moved on from that. Solving the defrost issue will not resolve the fact that whether they are derfosted or not these units do not suit the mans application. I've already given him my suggestion regarding dropping one of the units and backing up with oil when the temp drops. You don't even see an issue with rads heating a listed building connected to a heat pump and therefore I question your judgement and also your impartiality, as you, as a vendor, are the only one with vested interests regarding these units.

I am well aware of how these units work and I do not need to be schooled in heat pumps. You car analogy makes no sense at all.

But here is a car analogy for you in return regarding costs and return. If I was driving across the states I sure as hell wouldn't use a Kia, and when I buy a heatpump I want the bloody thing to last more than 8/9 years before I have to invest in a new one. So saving money up front is not what its all about. Do you realise how much money a COP of 5 verus a COP of 3 will save you over 20 years? Check it out and see if your figures stack up then.

Btw the introduction of these units was a reaction to a dropping AC market, an oversupply of manufactured units and taking advantages of economies of scale. They were not designed from the ground up, they were designed as AC units.

[TiredGeek]

Everything points to the hydrolic water side of the system. This is obvious because you mention you have 4 x 16/60 Alpha pumps in series (2 x push - 2 x pull), this is flawed.
These units need between 20 & 40 l/m each constantly, but Mitsubishi recommend you are as near to the top end as possible, in your case it is between 40 & 80 l/m (nearer to 80 if poss) total balanced equaly between the two units,

Each unit has it's own flow and return pumps. 1 x push, 1 x pull per ASHP.

A few questions - Have you got a low loss header installed? If not you need one to get the required flow rates when running two heat pumps on the same hydrolic circuit.

No. They flow and returns are both split off to each ASHP directly from the system. No buffer tank of any type in there.

Have you got two flows and two returns coming from the header to outside, one for each heat pump with their own circulator i.e pumps in parallel? If not you need them.

No header tank, but that's how the pumps are.

Have you got a flow setter valve installed on either the flow or return to each unit, so two in total? If not you require them to prove and balance the flow rates.

Yes. Flow setter on return, 25L flow indicated on each ASHP.

As the whole system is massively oversized for the load you should have some control to switch off one units above a certain ambient to stop the units from cycling, this is simple and cheap with an external thermostat, have you got this installed?

No, but we have tried isolating one unit by tripping the CB and the "plug in electric usage thingy" from the electric company didn't show any significant change in usage. One or two running, it's about £12 / day.
They don't cycle very much, just ramp up and down...

These are good points and it's a shame that when the Mitsubishi technician came out last year he failed to notice any of them.
He did say the flow rate at 18L was too low and he wanted to see at least 20L, which we did by modding the plumbing a bit with more 28mm pipe.
He never mentioned a buffer tank, turning one off etc.

Another issue that hasn't really been touched on yet: the water from a defrost, this just drains through the bottom of the case. When it's really cold it just freezes in there and is building up to threaten the fans. I know a heater would cure it but that's more electric used and I don't want to have to afford it. Any other ideas?

They want to have another look, Mitsubishi and the installers, I'd love them to get it working properly as it fits our needs so much better than GSHP (low starting current, no digging up the garden etc).
Maybe a buffer tank is the answer. With regards to that, isn't bigger better? I have room for at least 300L tank, would that be better than a 100L?

[fixmyheatpump]

These are good points and it's a shame that when the Mitsubishi technician came out last year he failed to notice any of them.
He did say the flow rate at 18L was too low and he wanted to see at least 20L, which we did by modding the plumbing a bit with more 28mm pipe.
He never mentioned a buffer tank, turning one off etc.

Another issue that hasn't really been touched on yet: the water from a defrost, this just drains through the bottom of the case. When it's really cold it just freezes in there and is building up to threaten the fans. I know a heater would cure it but that's more electric used and I don't want to have to afford it. Any other ideas?

They want to have another look, Mitsubishi and the installers, I'd love them to get it working properly as it fits our needs so much better than GSHP (low starting current, no digging up the garden etc).
Maybe a buffer tank is the answer. With regards to that, isn't bigger better? I have room for at least 300L tank, would that be better than a 100L?

It will work once it's right, regardless of it being a listed building or radiators, if the heat losses are less than the output of the units and the radiators are sized correctly there is no reason why it will not.

If you have two pumps per unit but no header or buffer how are they pumping around the heating system, I don't understand exactly how they are installed? Do you have another pump drawing to the heating system?

[desA]

Maybe a buffer tank is the answer. With regards to that, isn't bigger better? I have room for at least 300L tank, would that be better than a 100L?

Have your installers estimate the probable draw off the system to perform defrost under cold conditions ie. not depleting piping energy. This 'spare' energy should be held in the buffer tank. If you have the space, then larger is better - in my view.

I'd be interested in seeing other opinions on this.

[fixmyheatpump]

There should be no need for a buffer cylinder just a header as the units are inverter controlled. If you did want to install one as a header it needs to be small otherwise you will get poor control of the flow temperature, I would say max. 50L.

A rule of thumb for fixed speed systems is 9L per KW but this would be too big for this system as they are inverter driven and increase/reduce their output dependant on flow temperature and ambient temperature.

[TiredGeek]

I must admit I thought of a buffer tank long ago, but for a different reason:
Some of the radiators are a bit noisy due to the high flow rate around the system needed to keep water going thru the Mitsis at an acceptable rate. I was thinking that using two 210L tank (one per ASHP) with the Mitsis only heating them, and running the central heating through the tank coils (so isolating the circuits from each other) would allow high flow to the Mitsis but low to the rads.
The Mitsis keep the tanks hot, the central heating water picks up the heat as it goes through the tank coil and warms the rads.
I thought about just using one water pump per ASHP into the "buffer" on high setting and then the other two pushing / pulling the water to the rads on a lower setting. (maybe use a "solar" tank and utilise both coils for the C heating water?)

It would also allow me to reduce the amount of anti-freeze in the central heating side as this makes the water harder to pump around to some of the more distant rads. At the moment the water is a bit thick (compared to pure water) as the anti-freeze is needed to protect down to -25'c for the ASHP.

I can aslo leave the ASHP keeping the tanks warm 24/7 and just run the heating water pumps morning and night for the heating, or when I want, would this maximise efficiency?

Does any of this make sense?

If it does work in principal, does anyone know of a cylinder that has 28mm inlet and outlets?

[fixmyheatpump]

I must admit I thought of a buffer tank long ago, but for a different reason:
Some of the radiators are a bit noisy due to the high flow rate around the system needed to keep water going thru the Mitsis at an acceptable rate. I was thinking that using two 210L tank (one per ASHP) with the Mitsis only heating them, and running the central heating through the tank coils (so isolating the circuits from each other) would allow high flow to the Mitsis but low to the rads.
The Mitsis keep the tanks hot, the central heating water picks up the heat as it goes through the tank coil and warms the rads.
I thought about just using one water pump per ASHP into the "buffer" on high setting and then the other two pushing / pulling the water to the rads on a lower setting. (maybe use a "solar" tank and utilise both coils for the C heating water?)

It would also allow me to reduce the amount of anti-freeze in the central heating side as this makes the water harder to pump around to some of the more distant rads. At the moment the water is a bit thick (compared to pure water) as the anti-freeze is needed to protect down to -25'c for the ASHP.

I can aslo leave the ASHP keeping the tanks warm 24/7 and just run the heating water pumps morning and night for the heating, or when I want, would this maximise efficiency?

Does any of this make sense?

If it does work in principal, does anyone know of a cylinder that has 28mm inlet and outlets?

How is it piped? You can't have 2 pumps per unit and send it around a radiator circuit without a header to seperate the flow, that is your issue.

[fixmyheatpump]

I have a pipe schematic I sent to a customer in Kilmarnock with a similar issue which is now working perfectly.

If you want me to send it over let me know.

[TiredGeek]

If you have two pumps per unit but no header or buffer how are they pumping around the heating system, I don't understand exactly how they are installed? Do you have another pump drawing to the heating system?

I'll try to explain how it's plumbed, it's actually very easy to understand if you could see it.....

The system is as follows: All 31 rads are plumbed onto a 22mm pipe-run thoughout the house, as upstairs and downstairs circuits. These are linked together in two places so there is no actual "upstairs / downstairs" seperate circuits. The early stages of both these runs are in 28mm pipe.
The water pumps are 28mm, the central heating valves are 28mm (the DHW valves and pipes are 22mm 'cos thats the size of the coil through the tank).

So, each ASHP has a push and a pull pump.
One ASHP feeds (and returns) upstairs with 28mm pipe, the other downstairs with 28mm pipe, the flows and returns are linked near the water pumps and further on in the circuit.
Where the 28mm pipes end, they are linked together with 22mm running from upstairs to downstairs, and also branch off to different areas of radiators in 22mm pipe (so in effect, either ASHP can supply water to any rad).

The system seems to work well as each ASHP is getting 25L flow rate, and all the rads get warm equally. The only downside is some rads at the beginning of the downstairs circuit (in the kitchen) are fed directly from the 28mm pipe and are a bit noisy due to the high flow rate.
Even when the upstairs rads are turned off manually the flow rate remains unchanged through the ASHPs.

For DHW it's like this: flow from the ASHP goes through the water pumps (both water pumps are linked here on the output side), they then hit a "T" which diverts the water to the central heating valve or the DHW valve.
If it's doing room heating it goes off to the rads via a 28mm valve.
If it's doing DHW it goes through the 22mm DHW valve, thru the cylinder coil and then hits a "T" with the heating return connected to the other side. Each cylinder return and the heating return goes to the return pump and then the ASHP (In effect each ASHP has it's own cylinder but both are linked at the pumps outputs so either ASHP can heat both cylinders).

Wish I could draw it, much easier to explain

[TiredGeek]

I know I'm going to regret this due to spam, but as I can't do PM I'll post an email addy in a modified form

my-forum-name-exactly-as-it-is-at-hotmail-dot-com

Hope that defeats the spam bots......

Just put your e-mail details in your personal profile and the members can pick it up from there.... Frank

[sinewave]

Very intresting this thread!

It does sound like your pipework setup is causing control problems possibly?

My own setup albeit using an Altherma uses a low loss header.

I have 28mm flow & return from the indoor unit going to a Vaillant low loss header.

The 2 outputs in 28mm drop down 600mm or so and then T off to 3 22mm heating loops round the house (3 storey house)

Each loop has it's own Grundfoss Alpha2 15-60 pump serving each floor via a programmable room stat and 2 port valves.

The pump from the Altherma does nothing more than pump from the indoor unit to the header and back.

Below pix may explain better!

[desA]

There should be no need for a buffer cylinder just a header as the units are inverter controlled. If you did want to install one as a header it needs to be small otherwise you will get poor control of the flow temperature, I would say max. 50L.

A rule of thumb for fixed speed systems is 9L per KW but this would be too big for this system as they are inverter driven and increase/reduce their output dependant on flow temperature and ambient temperature.

Your system response will depend on the thermal inertia (lag) of the flooring. Practically, the buffer tank volume may not be a huge issue in terms of thermal lag. You'll need to check your loads, I'd imagine.

For radiators, I'd expect that the buffer tank would have some influence on room thermal lags. Would be useful to estimate this.

[TiredGeek]

Neatly done

It's making sense to me that a lack of header(s) is causing problems.

The Mitsi doesn't have an inbuilt water pump, so If this is what is needed then I need to re-plumb to a degree.
I'm going to need to really think this through to get it right as the Mitsi also does the DHW, so then where does the header tank go in the circuit, where do the pumps go etc.....

[fixmyheatpump]

Very intresting this thread!

It does sound like your pipework setup is causing control problems possibly?

My own setup albeit using an Altherma uses a low loss header.

I have 28mm flow & return from the indoor unit going to a Vaillant low loss header.

The 2 outputs in 28mm drop down 600mm or so and then T off to 3 22mm heating loops round the house (3 storey house)

Each loop has it's own Grundfoss Alpha2 15-60 pump serving each floor via a programmable room stat and 2 port valves.

The pump from the Altherma does nothing more than pump from the indoor unit to the header and back.

Below pix may explain better!

Spot on! This is how it needs to be configured

[fixmyheatpump]

Neatly done

It's making sense to me that a lack of header(s) is causing problems.

The Mitsi doesn't have an inbuilt water pump, so If this is what is needed then I need to re-plumb to a degree.
I'm going to need to really think this through to get it right as the Mitsi also does the DHW, so then where does the header tank go in the circuit, where do the pumps go etc.....

Will send you the schematic shortly

[TiredGeek]

http://www.arca53.dsl.pipex.com/
(damn, can't post a link..... if you put the usual in front of it, it should work....)

So I need to configure my system to use a low loss header as per the drawing there?

FixMyHeatPump, can you try that email please so we can discuss.......

[TiredGeek]

LOL, crossed postings

[back2space]

Some excellent responses guys.

Tiredgeek when they attend show them "fixmyheatpumps" detailed post above.

I feel you will resolve this. Regarding the water draining away you could ensure there is something under the units to stop the water freezing on contact with the ground eg salt under the drain hole.

When u get the defrost working properly I think it will drain out of the unit ok anyway as it will be warm water draining away so majority should drain. Presently your defrosts are not long enough so it doesn't drain away properly.

[frank]

Another issue that hasn't really been touched on yet: the water from a defrost, this just drains through the bottom of the case. When it's really cold it just freezes in there and is building up to threaten the fans. I know a heater would cure it but that's more electric used and I don't want to have to afford it. Any other ideas?

Of all the units I've seen with a similar problem, where the defrost water hits the base plate and then freezes, I've thought that a small 1/4" hot gas pipe spead across the base pan fitted with a small sol valve would do the trick.

Following defrost, the small sol valve would activate, allowing a small proportion of the hot gas discharge to be sent around the base plate, possibly for 10 minutes, but open to suggestions or trials on this.

This would ensure that the base plate remains hot so ice build up is prevented, but.... more importantly, the indoor heating is not inhibited, just runs at a slightly lower capacity during this time.

Another option with this would be to start the hot gas base plate heater 5 minutes before any defrost cycle, therefore making the base plate hot before defrost.

Obviously, this option would need nicorporating into the control logic and not just be an 'add on' fix

I know that the electric base plate heater is available, but from an energy point of view, this remains in circuit based on ambient temps, therefore running for all the time the unit is not in defrost cycle

[TiredGeek]

If there is not enough flow there is not enough energy in the water circuit and the heat pumps will stop defrosting to prevent the water in the plate heat exchanger from freezing and cracking it, this would be the only reason they would partially defrost and stop which can be seen in one of your pictures where the top of the coil is defrosted but the bottom is not. Once you get a ON/Off demand from the stat or clock this cycle will start again. At the end of a defrost the fan spins at full speed which blows the moisture away to prevent "bridging" I've witnessed this hundreds of times and it works, so much steam comes out of the front of the units some customers have called thinking they're on fire.

This makes so much sense I really can't understand why the install company (who are not a small unknown firm) and Mitsubishi's own technicians haven't figured it out ! I have witnessed the first bit where it tries to defrost and gives up, I have passed all the codes the units flashed up when this happened on to both Mitsi tech and the installers, they didn't figure it out.
I agree with the amount of water vapour that blows out, it's like fog when both of them do it. I must say I have worries about this building up on the vent covers in below freezing ambients, but until it's working properly we won't know

I'm now tending to believe that if you say these things will work at the house, if we get them plumbed in correctly, then they will
I sure hope it's just a simple replumb job and we're good

[TiredGeek]

OK, more questions

1/
I've been looking at low loss headers, and they're pretty simple looking affairs, with a simple purpose.
Is it possible, on a temporary basis just to prove the theory (or if it works, permanently), to knock one up out of pipe fittings?

I'm thinking of a section of 28mm copper, an elbow at the top as the hot feed in from the ASHP, below that a 28mm T for the water out to the radiators, bit more pipe, another 28mm T for the cold return from the rads, and finally another 28mm elbow to return to the ASHP. Basically a low volume 28mm diameter header.

Obviously I'd have to reconfigure some pumps, but do you think this would work? I see no point in spending lots of money to prove a point when I have plenty of fittings lying about.....

2/
For two ASHP, would it be the best to fit a header to each, and then tap off both of these and mix the water further in the system, or use one header with two inputs and two returns for both ASHP to share, and take the water into the system directly from this?

Thanks for all the help guys, we feel much more positive about the prospects of this sytem actually working in the long term now

[sinewave]

TG, my low loss header is manufactured by Vaillant and only cost £150+ VAT.

For two units you really need one header with two inputs & returns to the ASHP's though I'm asuming it would work with only one input/return but T'd on the connections.

Vaillant do 3 sizes I believe so you'd probably need the biggest one but at the end of the day it needs specing by a proper comercial heating designer to be sure.

[MikeHolm]

I make all my own low loss headers and the thing to remember is that the ends should be caps, not elbows and that, for the flow rates you are using I would use 50mm pipe with 28mm branches (assuming you can get this type of tee, we are still using imperial measure here). You could also take the opportunity to put a sediment faucet at the bottom which will help with keeping any small bits from the pumps.

[Bigfreeze]

I don't agree that a low loss header is suited to this application, you will be short cycling some warmer water back on top of the return water, which will drag down your efficiency. A buffer is a far better solution. Both heatpumps could feed into the tank and the second machine could be set to only kick in once below a certain temp and if the tank hasn't reached temp in an hour. Low loss headers are suited to oil/gas boilers as the return temp is not as influential as in a heatpump installation

[mad fridgie]

I don't agree that a low loss header is suited to this application, you will be short cycling some warmer water back on top of the return water, which will drag down your efficiency. A buffer is a far better solution. Both heatpumps could feed into the tank and the second machine could be set to only kick in once below a certain temp and if the tank hasn't reached temp in an hour. Low loss headers are suited to oil/gas boilers as the return temp is not as influential as in a heatpump installation

I have to agree and disagree on this one, I do not like the pre manufactured units as it looks as if the water streams will mix, but, I also use a similar system to Mike, I use a vertical pipe in excess of a meter long (50-100mm diameter). Hot flows at the top, cold flows at the bottom.
With a buffer tank, unless you have quite high extremes in temperatures stratification is vey limited especially with the number of flows being injected into the tank. I do like tanks with phase change materials but that another thread.

[MikeHolm]

Speaking of phase change materials, time for a new thread. Stay tuned or beat me to the punch.

[Bigfreeze]

I have to agree and disagree on this one, I do not like the pre manufactured units as it looks as if the water streams will mix, but, I also use a similar system to Mike, I use a vertical pipe in excess of a meter long (50-100mm diameter). Hot flows at the top, cold flows at the bottom.
With a buffer tank, unless you have quite high extremes in temperatures stratification is vey limited especially with the number of flows being injected into the tank. I do like tanks with phase change materials but that another thread.

We use buffer vessels (either 300L or 500L most commonly but depending on heatpump size) they are 1.8m tall, our flow from the HP would enter at about 1.4m and return leave at .2m. We find there is usually a 5k difference between the top and bottom of the tank. That is your circuit for the HP and all the heatpump cares about satisfying. It has constant flow rates and allows the lowest return temp possible.

After that you can take the water to the system anyway you wish i.e take an 1 1/4" pipe off the tank at high level and header it off to 2/3/4 pumps on 1" or 3/4" and do similar on the return where you connect back into the tank at a similar height to the HP return

[MikeHolm]

These tanks and the BFs method are standard fare in the solar world. I have seen solar tanks 10m tall and 2m in diameter to create as much stratification as possible and I have made SS tanks 600mm by 3000mm because, as with HPs we want the coldest and un-mixed temps going back to the collector as possible and in the winter, it is sometimes difficult to get 50C needed at the top to prevent triggering of the back up heating system. I will dig through my archives, find a photo and post it.

But, I see Mads point on the design of the header. I would think that the flows will blend quite well before become laminar in the central part with hopefully minimal temp change at the bottom. True, this has been in boiler/solar applications and not HPs but especially with the tighter dT, I don't see the issue.

[TiredGeek]

I have had an idea floating around my head for a while that was aimed at trying to reduce noise in some radiators, but now I'm wondering if it would work and solve both issues.

I'm thinking a 200L DHW tank per ASHP, the flow and return from the HP go into the cylinder via the cold water feed and hot water outlet, not the coil as would be usual. This would give a large amount of water with minimal flow restriction for the HP to draw on for defrost, and the HP is only trying to keep the cylinder at temp. Essentially it's a 200L low loss header, with a coil running through it.

Now, my central heating circuit is run through the coil, picking up the heat from the cylinder and circulating it around the rads.

As the two circuits are now not sharing the same water I can reduce the amount of anti-freeze in the radiators, allowing me to reduce the flow rate through them while still getting heat to all the rads. Slowing the flow will stop the noise I have in some of my radiators.

To me this makes sense, the principal seems sound, but you guys have more knowledge so please correct me if I'm wrong, again

[Bigfreeze]

I have had an idea floating around my head for a while that was aimed at trying to reduce noise in some radiators, but now I'm wondering if it would work and solve both issues.

I'm thinking a 200L DHW tank per ASHP, the flow and return from the HP go into the cylinder via the cold water feed and hot water outlet, not the coil as would be usual. This would give a large amount of water with minimal flow restriction for the HP to draw on for defrost, and the HP is only trying to keep the cylinder at temp. Essentially it's a 200L low loss header, with a coil running through it.

Now, my central heating circuit is run through the coil, picking up the heat from the cylinder and circulating it around the rads.

As the two circuits are now not sharing the same water I can reduce the amount of anti-freeze in the radiators, allowing me to reduce the flow rate through them while still getting heat to all the rads. Slowing the flow will stop the noise I have in some of my radiators.

To me this makes sense, the principal seems sound, but you guys have more knowledge so please correct me if I'm wrong, again

Have a look at the buffer set up I proposed above. The tank should be direct feed, no coils and you do not need any antifreeze in the system. You also can control the pump speeds to what ever speed suits as you do not have to worry about flow speeds through the heatpump. Each heatpump would have its only dedicated pump feeding into this tank, guaranteeing you sufficient and constant flow at any particular time

[Bigfreeze]

Have a look at the buffer set up I proposed above. The tank should be direct feed, no coils and you do not need any antifreeze in the system. You also can control the pump speeds to what ever speed suits as you do not have to worry about flow speeds through the heatpump. Each heatpump would have its only dedicated pump feeding into this tank, guaranteeing you sufficient and constant flow at any particular time

You can see the set up on the attached word document

[sinewave]

This is all well & good in theory, but what about in practice?


How many domestic homes will have space for all this paraphanalia?

This is where the low loss header will rule as it needs zero space in comparisson to buffer tanks etc.

You don't get many homes with room for a dedicated ASHP Plant room do you?