Safely undersizing heat pump
I’m selecting a heat pump system for my house. I’m caught between two sizes from different manufacturers and wanted to get some opinions on which way to go, and how much it is OK to undersize with respect to Manual J load.
House:
1250 sq ft single-floor, 3bd
Zone 5A (Central MA), 7F design temp
My Manual J: 28500 BTU/h at 7F
[after planned insulation/sealing work; used “semi-loose” for tightness]
My load calc from oil usage over three winters: 27000 BTU/h at 7F
[before insulation/sealing upgrades]
I’m going with a multi with three indoor units (2 heads + 1 compact ducted).
Options:
Bosch 27k BMS500-AAM027-1CSXHC (Max Performance system version)
(AHRI Cert #: 207760038)
Daikin 36k 4MXLH36WVJU
(AHRI Cert #: 212604432)
The Bosch fits OK, but is a bit short at the low end:
25000 BTU/h at 5F (vs 28.5k BTU/h load at 7F)
Just below 90% of design load then: 90%=25.6k.
Minimum at 47F is 4772 BTU/h (although this figure looks suspicious compared to other Max Performance system sizes; I think it’s actually around 8k).
The Daikin covers design temp comfortably, and has a minimum at 47F of 6550 BTU/h.
So, from using ASHP NEEP:
Bosch modulating range covers ~ 10F to 44F (or 51F if data is correct)
Daikin ~ -1F to 47F
Making the Bosch work has some benefits:
– cheaper (by ~$2000)
– installer has good contact with local Bosch rep, they have a headquarters in MA (but installer is also Daikin pro)
– Bosch heads have better design for cleaning blower wheel assembly (vs Daikin more fragile)
But the Daikin has a better range low down and looks on paper to have slightly better COP numbers through the range (whether that makes up the price difference in actual operation is another story…).
My main question is how low is it sensible to go in relation to Manual J load? I know Manual J does have some degree of padding, certainly seems to be the case from oil use, but I don’t know how far to push it.
Is the Daikin worth the extra cost or should the Bosch still work? Maybe it’s better to spend that extra money on solar, fixing more sealing/insulation (although a lot is planned).
And if there are any factors to consider related to either the Bosch or Daikin equipment, it would be great to hear.
It’s unfortunate the R-454B version of this Bosch doesn’t seem to be an option yet in MA, at least with my installer, as that has a 27k BTU/h output at 5F, which would likely be the best compromise!
Thanks for any help,
Alexis
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Are you asking if you'll damage the heat pump by undersizing it? If that's the question, then the answer is NO. What WILL happen is that the heat pump won't be able to keep up with the load of the space, which means (in the heating season), you'll "lose the battle", and the temperature inside will gradually drop once the outdoor temperature drops below a critical point. Basically the heat pump will hold your temperature setpoint until it gets cold enough outside that the thermal losses of the structure exceed the output of the heatpump, at which time the indoor temperature will "track" the outdoor temperature as it drops ratiometrically. The same thing will happen in reverse during the cooling season.
If you do end up undersized, you can always add another heat pump to add some additional capacity, making a sort of "two stage" system, with one heat pump carrying most of the load, and the second kicking in to help out when the first one's capacity is exceeded. You won't damage the units by undersizing them, you'll just extend their run cycles and have a more limited range of outdoor temperatures over which your system will be able to maintain your indoor temperature setpoint.
If you want to know EXACTLY how much you can undersize and still be OK, I'd say you're asking how much safety margin Manual J allows, and I don't think that's something you can predict very well without modeling the structure. As you get tighter and tigher in terms of removing excess capacity, everything gets more critical and there is more change something won't work as shown in a model so be careful.
Bill
Thanks for these considerations.
I wasn't worried about the heat pump itself, just the design strategy – 'safe' in the sense of a reasonable way to heat the space.
For adding the extra heat pump, were you thinking as a separate unit with its own indoor unit, or tied to the existing heat pump (a kind of cascade); I've seen the second option with air-to-water systems, not air-to-air.
Two independent units is what I was thinking, in different areas (like opposite sides of a room). The "cascade" you're thinking of might be a "multisplit" type setup, where multiple indoor "heads" can run off of one, common, outdoor unit on a single refrigerant loop. Those systems work fine, but are a little more complex to size because the outdoor unit can be sized for the total demand of the indoor heads, or some fraction of the total demand.
Bill
Yes, the options I'm looking at are in fact multi-splits. Certainly a challenge to size with the more limited ranges.
The cascade setups I'd seen were specifically with air-to-water in the UK (Vaillant Aerotherm heat pumps).
If you go to the NEEP page for a heat pump and click on "Advanced Data-- System Sizing " it will ask you to put in your zip code and heating load and will tell you how many hours per year the heat pump won't be able to keep up. Usually what you'll do in those hours is rely on electric resistance heat, which is more expensive to operate.
There's no hard and fast rule as to what size is going to be most economical. Sometimes a smaller unit uses less electricity overall even though it uses more resistance heat because most of the year it's at a higher COP.
Thanks for these points – yes I've been using that Advanced Data tool on NEEP extensively!
With the options I am looking at, the Daikin has better COP through the range even though it is a bigger unit than the Bosch. But I'm not sure if it's really worth the extra install cost for the slightly more efficient performance.
It's tricky comparing operating costs as it's hard to project running costs, especially with the limited data points.
So, this is partly why I was trying to figure out how much to weigh covering design load, is scraping 90% of calculated load ill-advised.
The numbers on the Bosh unit actually look pretty good. If the turndown is as low as the NEEP datasheet, it should work well.
The NEEP sizing tool will show you how much energy in MMBtu your aux heat will need in a year. If the cost of this is minimal, than there is not reason to get a more expensive unit as the ROI on it will be very long.
One option to get a bit more capacity for about the same cost is to go for a two zone for the wall mounts and a dedicated outdoor unit for the ducted unit. Generally the BOM cost on this is about the same and you get better efficiency, a bit more output and some backup in case one unit goes down.
I think the NEEP data and Bosch submittal is actually incorrect unfortunately for the capacity range of the 27k I mentioned, the more I look into it. Bosch have a 'regular' and a 'max performance' version for the various compressor sizes (18k, 27k, 36k). Looking at the capacity ranges of those on the Bosch submittals, it looks like the minimum capacity of the 27k max performance is switched with the regular (I'll put data list below to show).
If I take the NEEP tool data at face value, I'd need 7.2 MMBtu of additional heating, around 2100 kWh, which would be over $600 per year given MA electricity rates. Now, not all of that load would need covering (a cold hour overnight here and there for example), but – at face value – there's some cost there. (Is it worth putting that money in solar or a better heat pump? Another question...)
Thanks for the alternative setup suggestion, I'll look into that. I feel adding another outdoor unit will bring up costs – more electrical (extra circuit breaker), extra outdoor unit, same number of indoor heads. Worth a look though!
That Bosch data:
18k regular: 4700-25100
18k Max: 7250-28850
27k regular: 8303 [!!]-29935
27k Max: 4772 [!!]-38550
36k regular: 7454-45628
36k Max: 12900-51650
At least on LG and Mitsubishi the cost of a two zone multi split is about the same as the cost of two single zone units. Or close enough.
If you have a basement, a single zone multi position air handler with ducts for the whole place will definitely be cheaper BOM cost than a 3 zone setup. The only issue is residential installers don't like doing ducts and pricing is sometimes silly.
Also your heating costs are not that much more. If you had a heat pump it would run at COP of around 2 in cold weather, so way better than resistance heat but not free.
Thanks for this.
Practically for my house, it would be 3-zone multi vs 1-to-1 + 2-zone multi. For these installers and being in MA, getting to three 1-to-1's would cost more from the extra electrical, outdoor units installation... Doesn't seem like the efficiency boost is worth it.
For ducts, only option is unconditioned attic (got feedback on this in this forum a while back). Going to have to have the compact ducted in the attic, but runs are short.
The difference between 25,000 and 28,500 BTU/hour likely falls within the Manual J fudge factor.
The worst-case scenario is that you might be using backup resistance heat more often during the coldest hours of the day. But most of those coldest hours will be late at night or very early in the morning when your family might still be sound asleep and not even aware that the temperature has dropped a few degrees.
If you complete the envelope upgrades, the temperature drop will of course be less and could easily be zero. It won't take much to reduce the design load from 27,000 to 25,000.
Thanks for the input, I was hoping the 'padding margin' extended down close to that.
And yes, some of those hours will be a cold hour here and there overnight, so that kind of dip isn't really a problem. Hard to know how to estimate what deficit will remain: taking the NEEP tool data at face value, I'd need 2100 kWh of supplemental heat, which would be expensive in MA. But if actual load is more like 90% of Manual J, it's closer to 1200kWh deficit. Still, suppose I have to cover half of that, it's still over $200 a year in extra heating.
The Manual J load is calculated with envelope upgrades included in the inputs. (Load from oil use is before of course.) It's a state insulation program, and one area they won't be improving is the walls. So, there is potential (though not a renovation we'd undertake too soon) to improve walls.
The amount of padding depends on the experience and biases of the person doing it. But if you are confident in your fuel-based calculations, I'd use 27k as the pre-upgrade design load, and just treat the manual J result as a sanity check on that.
Is it possible to get them to run a pre-upgrade manual J as well, keeping all the other assumptions identical? Knowing roughly how much load reduction you can expect would remove a lot of the uncertainty.
Thanks for this.
I put as much detailed info as I could into my Manual J using CoolCalc (so it's my Manual J, not the company's) – I know values for pretty much all aspects of the house (and after improvements), and followed the 'aggressive' instruction. Infiltration is probably the biggest unknown, and I'm not sure I'll have time to get a blower door test done after sealing but before HVAC equipment purchase. (And even then, there's the conversion from ACH50 to ACH natural.)
Fuel-load calc followed the article here on GBA. It's fuzzy from a few angles: boiler is 30 years old (so probably not the 84% efficiency I used), it also served direct hot water, and we kept an unused back room at 55F. But I did use a few different base temps to get a range (and suit what we set thermostat to), and even taking 61F base for all years, it's only just above 28k BTU/h.
Are you suggesting it's worth running a Manual J for pre-upgrade to compare to oil use numbers?
I think it would be worth knowing roughly how much impact the envelope improvements will have in terms of cutting the heating load at design temp. Suppose it shows that you'll get a reduction of 4,000 BTU/hour. Then a 25,000 BTU/hr heat pump should be fine as long as the current design load is no more than about 29,000 BTU/hr.
You'd then have to decide how likely it is that the current load is above or below that value. It sounds like you have already done some sensitivity analysis, running through the calculations several times with different assumptions, but you may want to take a closer look at bad-to-worst-case scenarios.
[replying to your comment #28 here - can't seem to reply directly to it for some reason]
Interesting idea. It leads to some odd numbers, but maybe some clarifications.
Upgrades will be (and current levels are (1950s house)):
- main attic loose cellulose to ~R44
(currently very old fiberglass batts + ~3" old cellulose = ~R15??)
- back room attic to ~R50
(currently pink fiberglass batts = ~R19?)
- bedrooms over garage: garage ceiling dense pack insulated to ~R21
(currently only plaster ceiling below floor joists)
- back room over closed crawlspace, floor insulated to ~R30
(currently old pink fiberglass under plywood = ~R11?)
- sealing throughout attic and basement rim joist
(currently nothing)
Here's what CoolCalc gives me:
If I revert all improvements in construction details to estimates of current levels, and bring air sealing down one setting (from semi-loose to loose), the heat loss goes from 28.5k BTU/h to 39.5k BTU/h (!).
Nearly 40k BTU/h for current heat loss seems quite off, so something is wrong.
Using the 'loose' infiltration may be too much (and so maybe I should be using something closer to 'average' for the after-improvements calc). And maybe my assumption for wall insulation is too pessimistic: walls are not being upgraded, but currently there are v old fiberglass batts in 2x4 framing, and I'd used R3 for these. (Back room is a ~1970s addition, the pink fiberglass in those walls is in better condition but still old, I'd used R6 there.)
It's possible CoolCalc is overestimating some other losses with existing setup:
- two bedrooms over a garage that currently has no insulation in ceiling (there is a plaster ceiling though);
- back room over crawlspace (the addition) with minimal insulation currently.
I've also used BetterBuiltNW to get a rough number (not as many options to set different construction details though like some crawlspace, some rooms over garage...). I had 29k BTU/h load there using R44 for the attic and an improved ACH nat of 0.5 (otherwise fairly default). There, if I change the attic insulation down to R15 and put the infiltration in winter at default 0.7 (ACH nat), I get a load of 34.5k BTU/h.
Going back to CoolCalc, if I keep values at my estimates for current levels, but use 'semi-loose' and bump walls to R6, I get 35k BTU/h as a load.
For various reasons (solar gain, back room at 55F), my oil load estimate of 27k BTU/h is likely a little low (although (light, 2-person) hot water use wasn't factored and boiler is 30yrs old and not really 84% efficient). Actual load derived from oil might be more around 30k BTU/h (just considering bringing that back room up to normal temp). If we expect some padding from Manual J, then maybe something like 35k BTU/h is a reasonable number for current load and tallies somewhat with oil load if that's seen as an under-estimate.
So, all this would suggest improvements would be very roughly reducing load by ~7000 BTU/h. A lot of vagueness in this though... and it feels a bit circular to judge improvements in this way.
It is a single-floor house, so attic area is the full 1250 sq ft of floor space. Improving that from a poor ~R15 or so to R44 should make a reasonable difference, and then the sealing should also count for something.
To me, some useful takeaways: it does suggest maybe the infiltration setting should be closer to 'average' for the Manual J after sealing improvements, and maybe the walls should be considered to be better than R3.
Remember your oil use calculation is an average number, not a max number. I would personally not undersize for heat, central mass can routinely get in the negative single digits, so having a unit that puts out heat in that scenario is worthwhile, IMHO
Good point – but I thought the calculation gave a fair design load figure? It's not an average of fuel used exactly, it takes the average oil use and then, using degree day info, extrapolates BTU/h needed for design temp. Manual J isn't a max number either. Both aim to give load at 99% design temp (7F in my case).
But you make a good point. With design temp of 7F, if the system is already a little undersized close to 10F, there's a good chunk of heating needed that's maybe not covered by the heat pump. My other thought related to this is how robust/efficient/quiet is the defrost cycle on the Bosch(Midea) vs the Daikin...
If the house has significant solar gain, the fuel use calculation will underestimate the heating load. Otherwise it should give a max number, similar to Manual J.
Interesting point on solar gain and calculation from fuel use. Our house is SW facing with good sun exposure. Hard to know if insulation upgrades would equate to solar gain discrepancy.
Sounds like I was misusing the term 'max number' – load at 99% design temp isn't the total/max heating load of the house, but I suppose it is the practical max value when planning a heating system.
Unless you change the windows to lower gain, the solar gain will not change.
I am not aware of any way that seasonal fuel use can capture peak use.
My boilers computer has a time run counter. I have kept track of that in the past, but it is fairly granular[1 hour] and I don't have the time to do it every day or every hour, but I have used it combined with fuel deliveries to make estimates.
Even that does not really tell me peak usage
"Unless you change the windows to lower gain, the solar gain will not change."
The fuel use method is trying to predict what the heating load will be on the design day, based upon extrapolating from what the average heating load and average temperature are over the course of the heating season.
If the average heating load is lowered by solar gain, that extrapolation won't be valid, because the design conditions almost certainly won't happen when the sun is shining.
Thanks for this discussion and clarification – the explanation of the solar gain problem helps.
Between these two options I like the Daikin best. The amount of supplemental heating needed in the Bosch case seems like it could be both costly and inconvenient, especially since you don't have an obvious nice source of aux heat (existing gas furnace, pellet stove, etc.) and would need to use electric space heaters.
And while you may have some conservatism in your Manual J, there can be errors both ways, particularly if infiltration is one of the causes of the load, and you haven't had a blower door test.
I undersized my mitsubishi system (2 MXZ-SM42NAMHZ outdoor units w/ 3 and 2 indoor units respectively) on my 100 year old house, based on an aggressive manual J, and largely due to infiltration, I undersized a bit too much at first and our first winter was a bit cold - there were a few weeks where the house dropped to 60F which caused some familial frustration. As we improved air sealing and insulation, and I repaired the direct vent gas insert for a little supplemental, this winter is off to a much better start.
In your case the Bosch example is even more aggressively undersized than my case, and I think you'd be setting yourself up for frustration unless you have a very clear plan for how you are going to drop the load through air sealing and insulation, and have modeled that case. I'd say the Bosch case is starting to look pretty good at 22,000 BTU/h of design load, but that is a 30% improvement over your model, which I wouldn't chalk up to simple conservatism.
Thanks, really appreciate this perspective.
(2x SM42's and you were still undersized!)
Our experience with the oil and my rough calculations from oil usage before any envelope upgrades suggest that the Manual J figure is unlikely to be an underestimate. However, it could certainly be what we actually need in Jan/Feb.
And you make a good point on when the Bosch looks good from an output standpoint: I'm looking at the max output capacity (typically this is OK when matching 100% at design conditions), and that's still slightly below 90% of the Manual J load.
One other aspect in here: the Bosch is a version of the Midea units. They are fine, but finding information (specs, manuals) is very hard. Indeed, there seem to be clear errors in the Bosch submittals and on NEEP. It was, however, relatively easy to find the full engineering data for the Daikin, find capacities for different combinations of indoor unit, etc. This is more encouraging certainly.
Ha, yes, it is a 2700 sq. ft. house, 102 years old, historic district, structural masonry walls, field stone foundation, and 53 single-pane double-hung windows. Original heat load is probably around 90k btu/h at 0F. Through interlocking weather stripping of all the windows, air sealing, knee wall foam and sealing, door weatherstripping, etc., I think I'm closer to 75-80k right now.
Part of the undersizing issue is that one of my outdoor units is underloaded right now because I'm reserving capacity for the basement when it gets redone (and don't forget the 5-10% altitude derate in capacity for being in CO). After the next round of attic insulation, finishing the knee wall/vault bay transition sealing, and quality exterior storms my hope is to get down to 70k, and eliminate the last of the resistive heat. That will make the heatpump system "just right"!
Old houses are beautiful, but improving their energy performance is definitely a long process!
Wow, 53 single-pane windows! Sounds like a nice house though. Lot of work to land it in the capacity pocket.
"since you don't have an obvious nice source of aux heat (existing gas furnace, pellet stove, etc.) and would need to use electric space heaters."
Air handlers will normally have electric resistance strips in them to provide supplemental heat.
I think part of the issue here is that he won't have an air handler - just two wall units and a compact ducted (aka slim ducted?) unit, which wouldn't have provisions for aux strips. And even with aux strips, the high electric prices in MA would make them very unfavorable to use.
Correct – compact ducted can't have resistance strips. Only ducted option is in unconditioned attic, so I want to keep anything up there to a minimum.
I think if I were to risk undersizing[and divorce] I would install some of these where they may be needed.
https://www.homedepot.com/p/Cadet-120-volt-1-000-watt-Com-Pak-In-wall-Fan-forced-Electric-Heater-in-White-with-Thermostat-CSC101TW/100078620
Cheapety cheap cheap and cost nothing to run when not needed.
had one in the otherwise unheated bathroom in my last house and loved it. Nothing better than being able to crank up a little hot air on the knees on a cold morning.
Might also solve an inspectors demand for heat in every room.
Ha! I'll bear those in mind...
Your design day should be zero degrees f… no? Just a thought
I would go with the Daikin and supplement as needed when it is not enough…
Worcester Airport is 6.8F per ASHRAE 2021 (https://ashrae-meteo.info/v2.0/). For some reason, in some tables, they give odd places for Worcester MA – I have seen 0F listed from towns miles away for the Worcester temp. (But I can see the airport from my house!)
The Daikin would hopefully not really need any supplement (its balance point, if I take my Manual J at face value, is -1F).
According to NEEP the design temp for Worcester Airport is 7F. However, on average the annual low is -12F and 24 hours a year below zero. So there is kind of a long tail.
Ah, thanks for pointing out this nuance.
Where are you seeing this data? Hours at different temp bins was something I never really found a good source for. Still, 99% temp means 99% of the time it's at or above 7F – so the tail is in the sense of how low does it dip in that 1% because that can be +/- BTUs to supplement (and could potentially be a cold snap longer than a day below design).
Looking at https://ashrae-meteo.info/v2.0/ :
99.6% temp is 1.7F
Extreme annual mean minimum is -4F (standard deviation 4.8)
Extreme minimum: -7.5F [5 yr window], -10.4F [10 yr], -13F [20 yr], -16.6F [50 yr] [!]
Various ways to draw conclusions (other than wow look at those minimums rise). The extreme lows are going to be overnight for a limited number of hours in Jan/Feb. Having the house dip by some degrees there is acceptable (vs being oversized most of the time).
However, going light on capacity at 7F design temp means there's a not insignificant chunk of heat loss that would have to considered in terms of impact and coverage. Browsing weather tables, over the past 10 years, there have been occasional two-three day stretches at a time where the temp didn't rise much above ~1F (and dipped down to the -10s overnight). Does the cost of supplementing those days with resistance outweigh the more expensive installation of a bigger heat pump?
But it's a good point – the tail of how low it does dip below design temp.
I don’t have an answer for you but I have to give you kudos for the effort you are putting into this. You are doing the correct leg work and asking the right questions. The crazy thing to me is that this is the sort of scenario that an HVAC company should be able to effectively advise on but based on my experience most companies only have interest in selling the equipment and little interest in getting it right. The HVAC moto in most cases seems to be is “go big and go home”.
Thanks! I have successfully irritated/alienated a few installers over the past year, and also been told a few times that I know more than most HVAC installers (not a good sign!).
I'm amazed how often companies have told me I'm going 'way beyond' any other customer when I simply bring up capacity of equipment at 5F vs the nameplate number (not even the 47F rated value!). It's a long slog from there to "how much if at all is capacity de-rated with this combination of heads on a multi".
I could write essays on how frustrating different aspects of this are.... You'd imagine a $22-35k project would involve expert analysis and design, and good to-hand knowledge of equipment limitations.
It's true there are the companies who just give you a quote for one large system to make sure there's no problem on the coldest temps, don't go into any detail or discussion, and then only use sales-talk. But even with the installers who are more genuine, there seems to be a lack of knowledge and thoroughness on the design side of these projects. Not having any certified standards or regulation is probably a factor (although here in MA, you have to be a MassSave approved installer for the rebates, so maybe there are some kind of standards...).
I had originally wanted to do air-to-water, and read/watched a few UK sources on heat pumps for that. Although there are also plenty of poor installers there, there are enough expert companies who do a better job of assessing houses, analyzing systems, and giving a realistic projection of how the equipment will operate to the homeowner. They also have some standards agencies (for example HIES), so if you pick an installer who is part of a particular standards organisation and you have a problem (oversized, bad install), you have somewhere to go to get it sorted out.
Of course, will all this homework pay off? At this stage it doesn't feel like it!
We went through a similar process a few years ago. Our actual heat loss calculated from gas usage was 57k btuh. We used the ashrae formula of 1.4x which brought the size to 80k btuh. We then penciled in all of our air sealing and insulation upgrades which brought it down to an estimated 62k btuh. We installed 4 Fujitsu 1:1 with roughly that capacity(according to the NEEPs max field) and no backup heat. Figured if they couldn't keep up, we'd do more air sealing, install cellular insulated shades, run resistance heat or just live with it. Since then, we had the extreme lows of -15f (design temp is 9f)a few winters ago. We lost a couple of degrees off the setpoint but not a big deal. I believe that's the coldest it's been in 25 years here.
Thanks for the info – useful to hear.
Covering 25-year lows would, in my view, be a nice but not necessary design condition. Depending on equipment options for the house, it can be hard to balance capacity at extremes with the more typical operating conditions.
I didn't think the ASHRAE 1.4x design load for equipment capacity was a good method for heat pump sizing, more for gas (although I suppose it's for 'heating equipment' in general so valid, at least for the colder temps; might have to look at details for heat pumps when it comes to efficiency and comfort over the full heating season). I've generally seen 90-130% as the rough factor for sizing heat pumps to design load (maybe I'm mixing up tidbits I'd heard from Manual S... or the MassSave rebate limits for sizing...).
I see you used gas usage load and then reduced that according to planned upgrades, rather than taking a Manual J figure and multiplying by 1.4. Maybe I can try and do something similar from my oil usage, although it starts to get quite fuzzy.
With a 1:1, the 1.4x could work OK still. They give you a lot more wriggle room. If oversized, modulation down is often so good to still be OK in shoulder seasons. And their max is often better than what a multi of a similar size can ramp up to.
If I sized a multi to 1.4x design for my house (using Manual J figure), that would be an output of just under 40k BTU/h at 7F. This would generally mean a rated output at 47F of at least 40k BTU/h. With a ~40% turndown (optimistic for some manufactures, although others do better), that's a 16k BTU/h minimum at 47F. Actual house load is more like 9k BTU/h there, so almost 200% oversized already at 47. Even if you avoid comfort issues with properly sized heads/IDUs for spaces, you take a hit on efficiency and running costs.
(I'm really not clear on what an acceptable point is for being oversized in shoulder seasons and by how much – 150% at 47F? I'd been trying to get at worst to minimum output capacity = 100% load at 47F, but it's tricky with multis. Probably too many variables to have a solid guideline.)
I'll definitely take a look though on this larger side to see actual numbers. It's a useful point, and certainly another reason for me to be a bit more careful with potential undersizing.
I did a manual j as well to size the ductwork for each room and to calculate the size of the unit serving those rooms. I think the 1.4x is to account for defrost cycle. At any rate it turned out just fine. If you can swing it, I'd go with the 1:1. Also, these are all Fujitsu cold climate except for one. The total rated capacity is 4 tons but they put out more heat than that.
Thanks for the point about factoring defrost through the 1.4 sizing factor, and good to know it's working successfully as a system.
Certainly 1:1's have benefits. I think part of it is what equipment and extra electrical/labour is affordable for the budget. The mid-tier 1:1's aren't that much better than the Daikin multi for modulation range and efficiency. Going top-tier 1:1's really pushes up the budget.
I was looking at maybe splitting it is as a 1:1 and then a two-unit multi. More capacity at cold temps, if planned properly could end up just using one unit for some of the season.
Another unknow for fuel based sizing is actual boiler efficiency. With roughly 25k load, the smallest oil burner is about 4x oversized so burn times are pretty short. I doubt you are getting nameplate efficiency out of the unit. Quick google doesn't turn up the study but I remember efficiency dropping to well under 70% for short cycled units.
Edit: here it is
https://us.v-cdn.net/5021738/uploads/editor/eq/b07ko78bw0iz.png
Thanks for this, very interesting, appreciate the link.
It's a 30-yr old Weil-McClain WTGO-3, from memory something like 80k BTU output. Think it has a 0.85 nozzle on it too, and we have 117ft of baseboard!
So yes – run times are pretty short, 15 mins or so when it's fairly cold out, usually under 15mins. In the recent cold spell we had where it was in the 15-20F range for 24hrs, it would be on for just under 15mins, and come on roughly every hour or so.