Bang for Buck Heatpump to buy?

[2devnull]

Looking to replace a 3.5 ton R22 system. I see so many different tiers and choices in what appears to be advertising blogs more than informative sources. Since I have had good luck with advise from this forum, I wanted to see if I can narrow down my focus to a unit (or couple) that are know to be the best bang for the buck, reliable, efficient etc.

Thanks in advance.

 

[Dana]

Most existing equipment installed out there is suboptimally oversized for the actual loads, and even the best bang/buck most-reliable heat pump is going to underperform on comfort & efficiency if the oversized factor is too high. Replacement time is an opportunity moment to get it right. Have you or a disinterested third party (like an engineer or RESNET rater, not an HVAC contractor) run a Manual-J cooling & heating load on the place?

A company in the Atlanta GA area who does load calculations as part of their business took a few dozen of their Manual-Js and compiled this graphic plotting square feet of house per ton against house size. Most of these houses were in the Gulf states and were not cherry picked for this analysis:

The center of the cluster in the 2000-4000' house range is about a ton per 1400', implying that 3.5 tons would be good for about 4900 square feet of house. The one sample house closest to 4900' has a ratio of 1100'/ ton which would be about 4.5 tons, but the other houses in the 4000-6000' range came in between a ton per 1500' and 2000', which would be 3 tons or less.

The only way to really get a handle on this is to have a competent person run the numbers, but only the worst houses in the 2000-3000' range had 3.5 tons of load.

 

[Sponsor]

 

[2devnull]

Is this something one can run themselves?

 

[Dana]

Is this something one can run themselves?

There are a few online calculators that can get close if you are VERY aggressive on air tightness, shading factor, and air tightness assumptions on both duct leakage and outdoor air infiltration. Loadcalc seems to be the most consistent, some people prefer coolcalc, but I've seen/read of multiple instances on this forum and others where coolcalc came up with numbers 300% greater than what a professional tool delivers with proper inputs. Slantfin usually overshoots by 35%, but often by 50% or more, but can sometimes be pretty close.

Go ahead and run all three if you have the time, and be as aggressive as possible on the assumptions rather than consertative, taking every possible factor that could lower the load numbers into account (as stated explicitly in the instructions for ACCA Manual-J). If you don't oversize from the LOWEST result of the three you won't end up 3x oversized.

Sanity check the numbers against the graphic in my first response to this thread. If you have a 2500' house and the calculator is telling you your load is only 3/4 tons it's probably not right, despite the one outlier case in the plot of a ~3300' house with a ratio of a ton per 3300' (= 1-ton load). If it's telling you it's 5 tons that's also very unlikely, since that's worse than the worst house in the plot. The vast majority of all houses in the plot are between a ton per 1000' and ton per 2000', with only a handful of smaller houses falling significantly below the ton per 1000' mark.

The existing ducts might be oversized for the flow of a right-sized system and may violate Manual-D on duct velocity, but that's not as big a problem for comfort & efficiency as oversizing. As long as the ducts are reasonably tight and insulated the impacts of suboptimal duct oversizing are small.

 

[2devnull]

Thank you very much Dana for the explanations. This is for an upstairs. The downstairs unit is maintaining it's temperature correctly. Anything to note if this is for an upstairs only unit?

 

[Dana]

Run the cooling& heating load numbers for just the rooms covered by that upstairs heat pump, since that's the equipment being replaced. Assume zero gains/losses from the floor, since it is fully conditioned space under the floor.

A right-sized system won't allow you t0 leave if off all day and bring a stagnating 90F upstairs down to 75F in a short amount of time, but modest setbacks can still work. If you have become used to 3x oversized systems it might feel like it runs nearly constantly or is "struggling to keep up" during hot weather, but those longer run times are better for humidity control, comfort and efficiency.

 

[2devnull]

Here are the numbers for the upstairs.

BTW - I checked my current system and it is a 2-ton that is servicing the upstairs.

 

[Dana]

Knowing that this tool always overestimates by a bit, the load is within range of the 1.5 ton Fujitsu ARU/AOU18RLFCD, (20,100 BTU/hr maximum cooling). Unlike standard split systems this is a fully modulating system that runs nearly 100% of the time at nearly silent low speed/high efficiency. In it's normal cooling mode it tests at 19.7 SEER, which will be fine most of the time. When it's super humid out you may need to run it in DRY cooling mode for several hours at a time to bring the humidity down, but don't leave it in DRY mode, since it won't respond to the temperature in that mode and will over-cool the place. It's only a few hundred more expensive than a 2 ton single speed, but it's a lot quieter and more comfortable. Being a mini-split, the air handler/cassette runs at lower static pressures than most big air handler systems, but can still deal with longer-bigger duct runs than most of the mini-split competition. (There are 1200-1500 in California using just one 18RLFCD to heat cool & heat the whole house.)

The fact that it modulates all the way down to 3100 BTU/hr (about 1/4 ton!) in both cooling and heating mode makes it the best deal on comfort, and it's not outrageously expensive between $2-2.5K for the hardware at internet store pricing, which is only about $300-400 more than say, a pretty good 14 SEER 1-speed 2 ton like the Goodman GSZ160241 /AVPTC29B14 combination, and comparable to or cheaper than most 16 SEER single speed split systems.

One quirk about the Fujitsu is that in it's default mode it is set up to sense the indoor air temperature by the temp of the incoming air at the air handler, not the wall-remote/thermostat. It can (and should, in your case) be programmed to modulate it's output based on the temperature sensor in the wall unit.

 

[2devnull]

@Dana - so this is interesting. How would one go about introducing this into the system/duct I already have. I assume the suction pipe and refrigerant pipe are both reuseable? I set one part where the condenser is and somehow put the other part on the air-handler/evaporator?

 

[Dana]

The AOU18RLFC would replace the condenser, the ARU18RLF cassette would replace the air handler & evaporator. Functionally it is just a split system like any other, but with a funky little evaporator and an oddly shaped condenser. Technically while in heating mode the outdoor unit is the evaporator and the coil at the air handler is the condenser, and in cooling mod the roles reverse, but people seem to use the same terminology as if they were cooling only.

If the previous system used R410A odds are pretty good you can re-use the refrigerant lines.

It doesn't take the "Michelangelo of sheet metal" to marry a mini-split cassette to an existing duct system, but it would be good to have somebody who can do the math on the ducts to ensure the lower static pressure cassette can drive it adequately. Any contractor installing ducted mini-splits should be able to do that, but that's not to say they always do. Some rules of thumb guys working with conventional air handlers would likely be thrown by it an not know what to do, but it's not rocket science.

 

[2devnull]

Unfortunately, my lines are R22. They go through (from the upstairs Florida type attic) the inside of the house walls down into the slab under many rooms downstairs. There is no way to replace it without a lot of flooring damage. I don't even know how one would run new lines now to convert to an R410A system.

 

[Dana]

There are people who have successfully re-used R22 lines with R410A after some sort of cleaning process, but I'm not versed on just what that entails.