Heat pump water heater.

Icarus

We are working as energy consultants doing a beta test installation of 2 Rheem heat pump water heaters (HPWH) in series, the first as a preheat tank. The problem we have run into is that we cannot set the tstat lower than 110f on either tank so essentially we are asking both tanks to run full bore. The second problem is that the controls default with colder incoming water (~60f) to resistance mode for a long time, trying to keep up with a “phantom” demand. The whole idea is to reduce the load on the resistance to zero allowing each tank to operate efficiently AND produce enough hot water to satisfy demand. So the two questions are, 1, can I somehow lower the set temp, and 2 can i simply disconnect the resistance elements? I don’t think the incoming cold water is a big issue for the condenser/evaporator.

Thanks for all thoughts,

Icarus

Jamie Hall

Presumably, since this is a beta test, you won't mind fiddling... they have a heat pump only mode which should keep the resistance elements turned off. Selected from the Mode menu. There is also a Vacation mode which is alleged to hold the tank temperature at 82 F. I haven't tried it. It might serve for your first stage unit. Otherwise on temperature setting you're on your own modifying the thermostat (there went that warranty...).

As I'm sure you are aware, recovery will be painfully slow...

Icarus

The “heat pump only mode” still reverts to resistance with incoming cold water. My guess is this is prevent “complaints” about slow recovery. I did not know about the vacation mode (I am not on site regularly) so we will give that a try. I wonder if it will still kick on the resistance element. Also wondering, if. I simply disconnect the element will it fault shut down, or will the HP try to keep up.

Thanks for the input, keep those ideas coming,

Icarus

Zman

Is it possible to move the tank temp sensor?
You could certainly disconnect the elements, I doubt it would hurt anything but you might want to run that by the folks at Rheem.
As long as you are running tanks in series, how about putting the cold water in an uninsulated storage tank first. Since the Rheem is just taking heat out of the air in the space and putting it in the tank, why not move the first 30 degrees or so for free.

Icarus

An incoming uninsulated buffer tank adds a third tank to the mix, and we have a room issue. (We are also trying to come up with a plan for others to follow to very efficiently heat a lot of water strictly with HP. I have a note into Rheem, but I’m sure they would not be happy with a simple disconnect. My question would be will this cause a fault. I don’t yet know where the tank sensor is. Keep the ideas coming, thanks,

Icarus

Icarus

Doing a bit of reading...the manual suggests that “opening” the elements will result in a shut down fault. Vacation mode seems to be a fixed temp of ~68f, probably too low for a preheat tank. Gotta figure out how to either trick the tstat set point lower, or the vacation set point higher.

Icarus

ethicalpaul

I'm not sure exactly which model you have, but I can tell you that mine from 2019 (Gen 4) most definitely has a "heat pump only mode".

I'm not sure what your concern is about it running "full bore". Once you figure out the heat-pump-only mode...the first one is going to be heating the water that it can (set it to the minimum if you wish, but I'm not sure why you would), and what's the harm in that?

You want the water to get heated, right? The worst case is that the first one heats the water in heat-pump-only mode and the second one hardly has to heat anything.

Here's my usage from last month. That scale tops out at 2.4 kWh per day and I can assure you there's no resistive heating occurring there

ethicalpaul

Oh and one more thing. Look at this screen carefully. Note the bottom says "HEAT PUMP ONLY". But at the top there is a nag banner that says "This equipment should be set to ENERGY SAVING mode for maximum efficiency" with an Enable button

BUT DON'T BE FOOLED! The so-called "ENERGY SAVING" setting is the hybrid mode with resistive heating sometimes. They fooled me with that one once and I saw my electricity usage go up dramatically. I didn't need the faster recovery, and you with 2 heaters shouldn't either hopefully! So set it HEAT PUMP ONLY and not "ENERGY SAVING"

Larry Weingarten

Hi, I’m wondering about plumbing them in series. HPs don’t like warmed feed water. Could they plumbed in parallel? I’d be tempted to hunt down an engineer at Rheem and get some thoughts about what’s possible with their heaters.

Yours, Larry

ethicalpaul

@Larry Weingarten , how does the heat pump know what temperature the water was fed at? It just monitors the temperature at the top and bottom of the tank and runs the heat pump if necessary.

Can you fill me in on what the issue would be? I'd like to learn what it would be because I can't think of a reason. The feed comes in at the bottom and will mix with whatever temperature is in that tank right?

Larry Weingarten

Hi, Not sure I’m actually answering your question, but HPs like a nice delta between incoming water temp and water exiting the unit. If it’s preheated, efficiency drops in the downstream unit. Also, splitting the flow should double the water’s residency time in the tank, letting the HP catch up better. That make sense?

Yours, Larry

ethicalpaul

Efficiency drops as the temp increases I can get that, but any heat pump has to deal with that every day so I see no difference there.

I see the same double residency in series plus no concern about which way the water chooses to travel. I’m not sure how you’d ensure equal pulling from both in parallel. But thanks for the explanation I’ll keep thinking!

Jamie Hall

I think, perhaps, that we need to know just what @Icarus is trying to accomplish. Not in terms of the nuts and bolts, but why is he considering putting two water heaters in series?

I can't see any advantage to it...

On specific hardware, the temperature sensors on the Rheem are thermistors, and they are located at the resistance heating elements. They feed a variable resistance (variable with temperature) to the control computer.

Robert_25

Jamie Hall said:

I think, perhaps, that we need to know just what @Icarus is trying to accomplish. Not in terms of the nuts and bolts, but why is he considering putting two water heaters in series?

I can't see any advantage to it...

On specific hardware, the temperature sensors on the Rheem are thermistors, and they are located at the resistance heating elements. They feed a variable resistance (variable with temperature) to the control computer.

He is trying to heat the water using only the heat pumps. The recovery is low enough that one heat pump water heater won't do it.

I think a resistor could be added to the temperature sensor circuit to make it appear that the tank is warmer than it really is, and avoid turning on the elements.

psb75

HP water heaters are programmed to a "fare-thee-well." I'm sure Rheem engineers would not be happy with any modifications. I've had issues operating the units NORMALLY and then had motherboards go bad. Rheem has been responsive in those cases with their warranty policy. I also have had similar experience with Brad White. When the screen goes blank--with power to the unit (220v at the jct. box and 120v at the top resistance element) then the board is "cooked." Replace board.
I'm sure ANY modifications will void warranty. Probably even operating units in series would be "questionable" at the Rheem offices, if not in their "fine print".

Solid_Fuel_Man

I would pipe the two in a reverse return parallel. Then figure out a way to make it heat pump only mode. 

You get the capacity and the efficiency. I'm not familiar with the Rheem units, only Bradford White, and GE. 

I've found the B-W units to ba able to produce (recover) about 10-15 gallons per hour. 

If you drain the tank completely, fill with fresh cold water, and time the recovery to full temp. You will get an idea of recovery with your particular ambient temp/humidity and incoming water temp. 

I have an old Nyletherm, predecessor to the Geyser, attached to my 120gallon indirect for summer use. It will run about 9 hours to recover the entire 120 gallons to 130F. 

Jellis

Install your storage tank first inline and use a re-circ loop to maintain the temp above the resistance heating cut in.



This would at-least temper the water coming into the heaters, i believe this would reduce the amount of time spent in full electric mode

Jamie Hall

Funny thing about heating water. It takes a certain number of BTUs to heat a certain amount of water from one temperature to another. If we suppose incoming water at 50, and outgoing at 120 (140 is better; fewer bugs, but let's say 120) it's going to take 70 BTUs to heat each pound of water that much. Or just under 600 BTU per gallon.

It doesn't matter how you add those BTUs. Heat pump, electric resistance, oil burner, gas burner, solar, whatever. Nor does it matter whether the heating elements are in series or in parallel (except as it may affect the efficiency of the units). Consider the series vs. parallel concept. If you have two heating elements in series, each only has to raise your X gallons of water half as far -- so you would be correct that you could raise 2X gallons with the units in series as you could with either one alone. But... in parallel, you can raise twice as much water the same distance, so you can raise 2X gallons with the units in parallel as you could with either alone.

So two water heaters will recover twice as fast as one. No matter how you hook them up. There is a practical consideration -- you do have to arrange your plumbing so that the flow through both of them is close to the same, but that's trivial.

So... don't mess with the controls of the heaters, except to put them in heat pump only mode. Spend a little thought on arranging the plumbing so that the flow resistance is as close to exactly equal to the two units as possible, and stop right there.

If you need more hot water for a short term demand, use a storage tank and recirculation or simply a bigger tank for the water heater. You won't help the average demand over a day or so, but you will help the short term demand peak.

Robert_25

The other thing to consider is just how much heat can you pull out of the space these heat pump water heaters are placed in. Running in heat pump only mode only to increase the standby losses of a nearby boiler may not yield the results you expect.

ethicalpaul

In my case, my somewhat poorly insulated basement is pulling away the excess boiler heat anyway. I haven't noticed a difference in basement temperature in summer or winter from the water heater, although I'm sure there's a little (maybe a degree F?)

Icarus

Sorry for the lack of response..I have been a bit under the weather of late. I need to read all these posts. Somewhere along the line, one post had the concept dialed. We want to use 100% HP, and with the two tanks have enough capacity to recover overnight that way. The intuitive idea is to heat the first tank ~80-100f, and have the second take that incoming water to 120. I understand the ability of a room to deliver heat, Delta efficiencies and so for. What I really need is to figure out how to lower the set point and remain in HP mode throughout. We know that we can recover daily use with the two tanks this way...if we can configure them properly. (1 tank delivers daily use, but needs to rely on resistance).

Thanks all, I’ll stay in touch. Icarus

Solid_Fuel_Man

Parallel tanks. 

Pipe them reverse return for equal flow. 

Jamie Hall

Solid_Fuel_Man said:

Parallel tanks. 

Pipe them reverse return for equal flow. 

This. Don't try to do it in series. Getting the controls to play nice with each other will entail modifying them -- probably fairly extensively -- and then you are the test pilot. In parallel? Plug and play, same result -- as I pointed out above.

ethicalpaul

I don’t understand what could possibly go wrong with a series connection with them both set for heat pump mode at whatever temperature, either the same, or increasing.

they will just heat whatever they have in them at the time.

care to enlighten?

hot_rod

parallel with a couple Caleffi Quick Setter Plus, so they exactly share the load.

motoguy128

Efficiency and therefore total output for the same input electricity will be highest in parallel.  

By reverse return, I assume they mean when using a recirc loop.  

Depending on total peak demand, you might still need another electric storage tank.  

Canucker

Reverse return is for the tanks to balance input and output flow. If you don't pipe it that way, you could get one tank supplying most of the load, even though the fittings are the same due to minor variations.

Jamie Hall

"By reverse return, I assume they mean when using a recirc loop. "
No, not exactly. For simplicity in words, suppose that you have the two tanks lined up next to each other. The water feed comes from the right hand side as you face the tanks. As you go by tank 1, there is a T which goes to the inlet of tank 1. When you get to tank 2, there is an elbow which goes to tank 1.

Now starting at tank 1, the outlet goes through a heat trap then to an elbow and heads over to tank 2. At tank 2, the outlet goes through an identical heat trap and up to the pipe from tank 1, where it joins through a T and then on out to the left somewhere to feed the house.

If you now follow the pipes, no matter which tank the water is flowing through there is the identical set of pipe lengths and fittings, although in different order, and thus the identical head loss -- and thus the flow will split evenly.

hot_rod

Reverse return, vs balance valves

Jamie Hall

And the advantage of reverse return is that it will stay balanced, regardless of flow rate.

Solid_Fuel_Man

Yes ^^^^^^^^^^^^^

Reverse return is self balancing. Cheaper than balancing valves in pretty much every case. 

Icarus

You guys are making this too complicated. The idea...is to heat a daily load of water using HPWH in heat pump only mode, using overnight to allow for recharge. We DON”T want to heat two tanks to ~120f. 1 tank, heated to 120f is plenty to satisfy demand, but in order for it to recover it always clicks into resistance mode thinking “WE” want it to cover the demand.

The idea is to preheat the first tank to something under ~100, and the second tank to 120F. The idea is that the first tank can operate 24/7 in heat pump only mode as needed to maintain the set temp, while the second can take the ~100f water and bump it to 120f over night to recharge, then hold it for the larger evening demand.

The problem I have is 1: I can’t set the first tank low enough, 2: the first tank reverts to resistance when it gets it’s first dash of cold, negating the whole concept. I could do this quite simply preheating the water with a room temperature uninsulated tank, or a nat gas tank (which we already have but are trying to abandon) and set it to ~90f and then feed the second tank, or feed it through a demand gas. All these options defeat the concept that we are trying to achieve.

So...question is...I can I adjust the first tank to lower than 110f, and how can I prevent it from trying to go to resistance mode? That is the issue.

Once again, I like all the thought that has gone into this, but I think we are over thinking it. Bottom line, we don’t need two tanks of hot water, simply looking for a way to heat one (using two) most efficiently. Thanks

Icarus

Jamie Hall

Have you tried the heat pump only mode? That should keep the resistance elements off, and give you maximum efficiency in electrical use -- whether it's one tank or two.

If you have, and it still runs the resistance elements, it's not operating properly; contact your technician or installer.

OK Now we have it so it will only run in heat pump mode.

The next question thing is: when it is running in heat pump only mode, what you have for hot water is what's in the tank, plus painfully slow recovery. As I have said before, the recovery rate, in gallons per minute from cold temperature input to desired output, is determined entirely by how many heat pumps you have, and it makes absolutely no difference at all whether they are connected in parallel or in series. None. Whatsoever. If you could somehow magically split the rise between two in series, that would get twice the flow in the series circuit. If you could split the flow in parallel, which is not magic, you would get twice the flow in a parallel circuit.

Now there is a difference between series and parallel: in the series system, which you seem wedded to at this point, you only have half the storage -- just what is in the high temperature tank.

Now. You say "Bottom line, we don’t need two tanks of hot water, simply looking for a way to heat one (using two) most efficiently." If you don't need two tanks of hot water, you don't need two tanks of hot water. Period. If one tank of hot water with recovery spread out over the day is satisfactory, do that. If one tank of hot water with recovery spread out is not satisfactory, you need two tanks. In parallel. Period.

I just don't know how to make it clearer.

Solid_Fuel_Man

Exactly. There is no free lunch. Tanks in series dont work well. The first ends up doing a bunch of work the second is just along for the ride and still has standby losses. 

You either need the HP to heat the water, or the storage to keep it. 

Icarus

Jamie...

I”m not too obtuse or thick here, I understand two things....1: that the recovery rate will be painfully slow, and within certain limitations we don’t care. @: I understand that water heat is simply a case of BTU introduced into the water pure and simple, T incoming, BTU input/efficiency= T outgoing. The Delta is only relevant here as to the HP sweet spot.

Let’s agree on the following assumptions. 1 tank, heating to 120f cannot keep up with (specific timed) demand over a 24 hour period, operating strictly on HP. We get that!

Second my thought is, that if the heat pump tank were delivered higher temp water (from what ever source) it’s chances to deliver the demand increase. (Essentially a smaller Delta). Do we agree on that?

3: if assumption #2 is correct then the question is, how best to deliver water to the “final tank” at a higher temp? Are idea is quite simple, heat the preheat tank from ambient incoming to something short of 120, say 90f or 100f, using HP mode only, then deliver that preheated water to the second tank to top up the temp to 120f.

The problems we encounter are two. The first is that we cannot figure out how to lower the stat temp of the first tank below 110...too high for our experiment. The second is that when cold water (~55f) is introduced to the first tank, the tank defaults to resistance mode. My question is, does it do that to protect the HP (though I would think the HP could deal with that delta) or as I suspect, does it default because it “thinks” we like most people want speed over efficiency?

So, as for this beta, we simply want to drop the tstat temp of the first tank, and “ban” the use of resistance in both tanks.

Bottom line, from an efficiency stand point, having the 1st tank only heat partially keeps that tank in the “sweet spot” of HP ef most of the time AND with a smaller Delta between the room and the water, there is smaller standby losses. Second, by sending the second tank preheated water, it reduces the recovery time for the demand.

I get that I can plumb these either in series or parallel and understand the results of each, but I don’t need 120 gallons of 120f water giving back heat to the room to supply the demand. The irksom problem is that I can’t override the programming of the Rheems and Rheem Tech support has been of no help.

This is after all an experiment. We do energy consulting, one of us an academic, my self a tinkerer. Together we are trying to come up with strategies that can compete with natural gas for carbon efficiency. HPWH does very well as long as it doesn’t drop into resistance mode. Yes, would could have started with a larger HPWH, but these are truly cheap (two of these were cheaper than on larger unit)

So, once again, I think I understand the issues clearly, and hope I have expressed our interest clearly. I appreciate all the thought that many of you have put into this, and willingly read and try to absorb them all, but right now, I am back where I started,

Icarus

Jamie Hall

I don't really think there's much confusion here, @Icarus . In principle, the two stage operation which you envision should work as well as the parallel operation I've been talking about. Whether it would be more or less efficient... that depends on the specific characteristics of the two heat pumps involved in the two stages -- if they are the same, and if their efficiency differs at different temperatures, one or the other will not be operating at its "sweet spot".

Then there is the question of control. If you can't get the Rheem to operate in heat pump only mode (their literature says you can, but it wouldn't be the first time that literature and reality conflict with computerized controls), then you are faced with the necessity of modifying the controls of one or both of them (probably both; many hot water heaters, and this may be one of them, simply will not turn on if the incoming water is too warm) in such a way that they manage to split the difference in temperature rise. This is not a trivial problem, since the temperature rise required will be changing constantly if water is being drawn off, and it's a delta T problem, not an absolute temperature problem (for instance, at the beginning of a water draw, the low tank may be at, let's say, 80 and the high tank at 120. Incoming water is at 40. As water is drawn, both tank temperatures will drop, and so the target temperature for the split will drop -- it won't stay at 80). Or to put it another way, your controls are going to have to be linear, rather than on/off.

I think, so far as control goes, you will be in a situation where you will be designing your own sensor and control systems to interface with the mechanical operation of the heat pumps.

Your second assumption is correct, of course -- the heat pump can heat more flow with a lower delta T; all else being equal, with half the delta T, you should get twice the flow (I might note here that all else isn't equal -- the capability of the heat pump is also affected by the difference in temperature between the heat source (probably air in the basement!) and the heat sink (the water you are trying to heat) -- and this must be factored into the control algorithm mentioned above to keep the load shared equally.

Again, with serial operation it is clear that you will need to engineer and build your own control systems for both heat pumps to get best efficiency (while you can't get the low temp one to cut out below 110, you may also have trouble getting the high temp one to cut in above 80 or 90 or so... and that target temperature for the low temp one isn't fixed anyway).

Not that it can't be done. It can, and it would be an interesting exercise in sensor and control integration. If I were doing it, I'd run it through a full scale computer, and allow the operator to vary the set points and curves at will, so as to optimise the operation (which, I might point out, the factory boys do with the equipment you buy!).

The storage vs. recovery question is somewhat different. There you really need to determine what your demand characteristics are. Oddly, this is a fairly common problem in water supply! The two extremes are one, the demand never exceeds the recovery rate of the equipment -- instantaneous (so called!) water heaters operate on this premise. The other is that the recovery rate is enough to supply the average demand over 24 hours, but the actual peak demand is concentrated over some much shorter period -- at which point you need to determine the storage volume to meet that peak demand (the obvious and completely impractical extreme is that all the hot water is needed in, say, 2 minutes!). At that extreme, clearly, you need the total volume for the day stored, to be replenished over time. Reality is somewhere in between -- but it's probably not unreasonable to operate on the premise that all the hot water will be needed within half an hour; the storage volume required then would be the total demand less the recovery volume provided in half an hour.

I don't know what that volume would be for your application. Clearly, if you don't need 120 gallons of water to meet that peak demand, you could get away with a smaller high temperature tank -- or tanks. The smallest Rheem hybrid I can find is a 40 gallon tank unit. That might suffice? In terms of losses, remember that the heat loss from a storage tank -- for the same interior temperature, exterior temperature, and insulation -- varies as the 2/3 power of the tank volume -- a 120 gallon tank will only have twice the loss that a 40 will, for instance, or to look at in another way, two 40s will lose as much heat as one 120.

It's an interesting problem, and I certainly applaud your quest for absolute maximum efficiency, whatever the difficulty. I'm afraid that I personally tend to go with what I know will work!

And I've got to admit that I'm still gently mystified why Rheem would claim it has a heat pump only mode, but you have found, evidently, that it doesn't (the discussion of mode is on page 19 of the manual I have available).

ethicalpaul

I don't know if you saw my posts with the phone screenshots, but my Rheem is most definitely, without doubt, without trickery, without question, running on heat pump only.

You have a Rheem also, right? What model?

Icarus

Jaimie...You have it in a nutshell, probably better spoke than I in fact. I understand that you can’t have both HPWH running most of the time at the “sweet spot”. That makes perfect sense. The idea would be to use set points that satisfy demands both short and long term with as much of the total btu input coming from the most efficiency source.

It would be a very simple exercise in this beta format if we could simply control each HPWH both for max T as well as HP only. Right now we may be able to force it to run in HP only but we cannot control the lower T set point at present.

As for Ethicalpaul’s comment. I know you can set it to HP only, but there are times when it will revert to resistance even in that configuration. Those times are usually when the first (preheat tank) receives a bunch of cold water.

Finally, I am at a bit of a handicap on this project right now due to Covid. The two of us because of each of our families medical concerns are not allowing ourselves to be in the room together, so I have to rely on information from my “partner” in this. As I said, he is an academic, I am the wrench turner.

Interestingly enough, he has been able to drop his daily kwh consumption by nearly 1/2 most days, since we did the initial install. He does this by setting the primary temp at 110 (vacation mode produces no heat and is not adjustable as far a he can determine.) and setting the second tank to 120, but has it bump to 130 for several hours before peak demand. That simple step dropped the kwh consumption nearly in half. We also understand that we are in the least efficient part of the year. Summer, with higher incoming water temps, and warmer air temps should drop the total kwh consumption to ~3kwh/day. Our theory is if we could do as we are trying to do, we should be able to drop it more, as well as satisfying demand.

Thanks...keep the conversation going,

Icarus

Solid_Fuel_Man

Do you have a meter directly on the water heaters recording Kwh? 

I'm very curious as to how this configuration would reduce overall energy used. It seems the two major factors in any heat pump configuration are incoming temp and standby loss. 

I'm gonna go out on a limb and say the standby loss between a set point of 90 and one of 120 makes quite a small difference. The incoming water temp would have the biggest difference. Both units receive the same air to cool, so that wouldn't be a consideration for this exercise. 

I'd be very interested in a dedicated water meter on the cold side, and a meter on the electrical side. 

What is your peak demand duration and gallons?

Jamie Hall

As an engineering exercise this really does fascinate me -- and I can see a number of applications where the general concept could be really useful (low and ultra low refrigeration? high temperature water or even steam? Hmm...)(how about thinking about a way to retrofit steam heat -- after all, you are using a phase change heat transfer system for steam heat and it would be a matter of finding a refrigerant with appropriate vapour pressure/temperature relationships... evaporate at low ambients, compress, condense in the radiators... hmm... might even be able to still use water, with appropriate pressure reducing gear on the radiators... we're light years beyond the folks talking vacuum steam operation here!)

The real problem, for you, as I see it at this point, is that to really optimise this arrangement one would need to really control each heat pump from a single control algorithm which would determine the final target temperature and the input temperature and the air temperature, and then using the characteristics of the heat pumps determine what the load on each should be. Ideally one would want to be able to modulate each heat pump, thus being able to choose the range of each so that you did the most lifting out of the one which was operating in its best efficiency range.

I rather suspect the two heat pumps would -- or rather, should be! -- optimised for a specific range, and certainly be able to modulate at least the evaporator pressure if not also the refrigerant pumping rate.

Way out in left field here! And not, I think, what you can do with off the shelf units -- and certainly not consumer grade stuff at this time!

Icarus

Solidfuelman, We use an Emporia energy CT monitor on each circuit, so we can see consumption in real time as well as cumulative over times. I’ll have to check with my partner to see what the estimated daily demand is.