What's the most efficient way to use a heat pump?

Last week I got my old gas furnace and central AC replaced with a brand new high efficiency variable speed heat pump. With the old furnace I would have the thermostat scheduled to turn the temperature down at night and then come on a half hour before I get up in the morning. But I am seeing some sources online that say with a heat pump it’s more efficient to maintain a constant temperature rather than turning it up and down at different times of day. Other sources (generally older ones, it seems), say you should turn it down when you’re asleep or away. So what is the Straight Dope on the most efficient thermostat settings for a heat pump?

I do kind of like having the temperature a little cooler a night. And I understand if you ask it for a big change in temperature the aux heat will come on, which will draw a lot of power. So currently I have the schedule programmed thusly: 60°F overnight. Then 62°F at 5:00 am. Then 64°F at 5:30 am. Then 65°F at 6:00 am, where it stays the rest of the day. Would that work? Or should I just leave it at 65° all night?

It’s going to be more efficient to turn the temp down at night, like you are doing. It’s colder at night, so the heat pump has to work harder to raise the temperature. Lowering the set point means it will cycle on less often.

If you have cheap electricity at night, it might not save much money, but if you like it cool, that’s great. Personally, I think 60° is pretty darn chilly, but that’s me…

That’s what I’d always heard for other types of heat, but some places are saying that heat pumps are different:

You may have lowered the temperature on your old heating system when you went to work or at night. A heat pump is different. It reaches peak efficiency by maintaining a set temperature. Find the right setting (see #1), then leave it alone and let it work! Exception: for absences of over 24 hours, go ahead and turn it down.

I’d ask them to “show their work."

Part of it would be that heat pumps get less efficient (and have less capacity) when the outside temperature get lower, so it could be more efficient to maintain the heat you have during the night, then lose it and try to regain heat when the outside temperature is lower in the early morning. Likewise during the heat of the mid-day your heat pump is going to be more efficient and have more capacity then trying to draw heat in the evening after sunset. I really don’t see much problem in having the heat pump drop to a lower temp after one leaves for the day in the early morning, but it’s better to warm back up earlier, perhaps starting at 1pm or so (whenever is the hottest part of the day), then wait till a hour before you get home. But with such a short setback time it’s questionable how much that effort is worth.

Another part of it is the harder you push it, the more likely you are to trigger a defrost cycle which is using energy to melt ice outside. It’s not uncommon for it not to need it at all or much for a low draw, but having the defrost cycle often when on high, trying to bring up the temperate rather then maintaining it as most modern heat pumps have multi speed compressors.

I think a lot of the efficiency depends on when your emergency heat is engaged.

For my thermostat, emergency heat is activated if the set point is more than two degrees F above the current temperature. So, if the room temperature is, say, 60 degrees and the thermostat is set for 65, it will turn on the emergency heat. If the thermostat is set to 62, only the compressor will run. I try to keep the temperature rise at two degrees per hour to keep the emergency heat from kicking in.

Check your thermostat’s instructions. Decent programmable thermostats that make you choose the type of system you have (gas furnace, heat pump, oil furnace, gas boiler, etc.) are usually smart enough not to engage the auxiliary heat when coming out of a setback, so you can avoid programming all those extra intermediate steps. The instructions need to say it does that though, usually called something like smart/intelligent recovery, otherwise all bets are off. That also assumes the thermostat controls the auxiliary heater at all. Normally heat pump thermostats are wired so the thermostat controls the aux heat, but it’s not guaranteed.

Hmm, the instructions didn’t specifically say that as far as I could find, but this thermostat is specifically for this heat pump. I was told that I had to use this thermostat to take advantage of the heat pump’s variable stages. It was mildly disappointing that I had to give up my Nest.

We have a nest for each of our heat pumps. Seems to work fine

The new heat pump is a Bosch. From Googling just now apparently Bosch tells their contractors that they don’t recommend using a Nest with it. Other people (including Google/Nest themselves) say it will work. My WAG is that the Nest will “work” in the sense that the heat pump will come on and run, but it won’t fully take advantage of Bosch’s features to get the most efficiency out of if.

The thermostat they gave me is a Bosch BCC100, which is actually a generic smart thermostat that you can buy from Home Depot (so it’s not specifically for that heat pump like I said before). But presumably it works better with it than the Nest. Or for all all I know maybe Bosch just says that so they can sell more of their thermostats.

At least it’s one they’ve vetted to work the way the heat pump is expected to work. But there’s always some idiosyncrasies that may not be well documented. For instance, at a previous office where I worked we replaced the building’s HVAC with new high-efficiency furnaces, high-SEER a/c, economizers, humidifiers, etc. It was a Carrier system with Carrier thermostats. The problem we ran into, which became apparent in cool weather when the economizer opened to bring in outdoor air for “free” cooling, was that the way the thermostat staged was very bad. Basically, if after a predetermined amount of time (something like 10 minutes), on stage 1 (economizer only), if the temperature had not reached the setpoint, then it would ramp up to stage 2 (economizer plus one of the two a/c units) and run that way until it was satisfied, at which point it shut down. Lather, rinse, repeat.

In the summer it would similarly run stage 1 with one a/c for 10 minutes, then bump up to stage 2 and bring the second a/c online until satisfied. The problem in both cases is that it led to blasts of cold air followed by the system shutting down.

We then had the thermostat replaced with a White Rogers model that used a temperature differential to manage the stages, and it would cut out stage 2 and roll back to stage 1. So instead of the system going 1-2-off, 1-2-off, 1-2-off, it would go 1-2-1-off, 1-off, or 1-2-1-2-1 as many times as needed leading to much more consistent temperatures. It also greatly reduced the maximum power demand charge, eliminating it entirely in the winter time because it could just hold the economizer open and never ramp up the a/c unit.

Good luck getting information like that from user manuals. Also, I had to pour through the furnace manual, because it was a twinned system (two furnaces operating together), I had to pull a jumper off one of the control boards to allow it to defer staging to the thermostat.

Now that made me think of something. With the way I have the have the thermostat programmed right now, even if the auxiliary heat isn’t coming on, when it the temperature ramps up in the morning it seems like it runs at stage 2 for a long time to bring the temperature back up, then 1, then off. Maybe if I just left the set point constant it wouldn’t need to run so long on stage 2, but would just go 1-off a few times during the night to maintain the temperature, or maybe 1-2-1-off.

Basically a heat pump is a pump. (duh) It’s like your air conditioner mounted backwards. Instead of pumping calories from a cold house to a hot outdoors, it’s pumping calories uphill from a cold outdoors to a warm house.

Because of this, by it’s nature, a heat pump is most efficient when “uphill” is less. So lowering the thermostat obviously means a smaller number of calories to pump because the “uphill” is less. Lowering the thermostat when outdoors is cooler is also more efficient. The question as mentioned above - which only an analysis of the performance and local temperature stats can answer is - what happens in the morning when it’s still cold and you are trying to make the house hotter?

And the other question regarding auxiliary heat would be - at what outside temperature would raising the house from, say, 69°F to 70°F be more cost efficient with auxiliary heat than waiting for the heat pump? (My gut says never). So really, the question is how fast your pump can raise the house temperature by one degree and whether that speed is acceptable in the morning. But all that is dependent on your heat sink’s characteristics and the house, also.

Maybe you set the start point for raising temperature to earlier? But it seems to me a “smart” thermostat would learn the heat characteristics of the house and adjust accordingly.

Maybe, but since residential electric service doesn’t have demand charges, it doesn’t matter if the heat pump runs for 60 minutes at stage 1 versus 30 minutes at stage 2.

This is untrue.
Our rate has quite an onerous demand charge: Time-of-Use 4pm-7pm Weekdays with Demand Charge

Back when I had a recording weather station I noticed in the fall and winter the coldest outside temperature was right around sunrise to an hour or so after, 7 to 8am when most folks were getting up to go to work or school.

My hypothesis was the ground was slowly cooling and heating the air above it less all night, then after sunrise, it took a while for the sun to heat back up again. This was most noticeable when it was clear all night – cloud cover would keep the drop from happening as much.

You’re right I was being overzealous in my generalization. Smart metering with variable rates are becoming more common as the grid gets more stressed, though at least in the US I think it’s still mostly opt-in on the residential side. Demand metering has been de rigueur for industrial customers, and commercial customers that have large air conditioners, pumps/motors, or heating equipment (think restaurants, retail stores with a lot of lighting, gas stations, etc.)

The kind of demand metering I’m familiar with is a simple maximum power draw in kW over the course of the month that’s used as a multiplier in your bill. In that case it doesn’t matter if you used the peak draw for one hour or 25 days, but of course the more kWh you use the higher the bill is going to be anyway. The electric meter simply has a separate dial for the current power draw, kind of like the speedometer in a car, and another secondary dial on the same spindle that it can push up, but which won’t fall back down, indicating the highest reading. The meter reader would reset the secondary dial every month with a little twist lever. I figure that’s mostly antiquated now.

A further complication is that the coefficient of performance is often related to how hard a unit is working as well as the temperature differential of ambient air across the pump.

Half an hour at full power will likely not shift as much energy as one hour at half power.

Everything is eventually bounded by Carnot. But those boundaries are defined by the temperature of the heat exchangers, not the air. More surface area and more air past the exchangers helps. But the performance can degrade dramatically from the baseline specification.

When things are really bad the COP gets close to unity and you may as well save wear and tear and just use a resistive heater.

A freezing outdoor unit wrecks everything as well. That puts a floor under what you can achieve. Especially as it is most likely first thing in the morning.

It seems reasonable that keeping a relatively constant temperature is a good first cut answer.

All heat pumps have a defrost cycle.
This is annoying, but allows operation down to 20°F or below.

They do, but it places a floor under the efficiency they will operate with. Whilst they defrost they aren’t heating - or if they do the COP is below unity. What the overall COP might be will depend upon propensity for unit freezing, which depends on temperature and humidity.

Eventually the only reason to use a heat pump to heat is to get better than unity COP. Whilst a unit is able to defrost itself, whether it remains a cost effective heating proposition is worth worrying about.