Implementing Dual-Fuel house heating (Heat pump)


Looking to upgrade my old Air Conditioner, I decided to go for a dual-fuel heating solution. Heat-pump for cooling and moderate heating with my existing natural gas furnace for heating at low temperatures.

Does a Heat-pump work in a colder climate like Ottawa, Ontario, Canada? – YES! Even with a non-cold climate heat pump, about 70% of my heating (based on -9C before switching to gas). If you are heating with something other than natural gas, make the switch! You’ll probably have a return on investment of 2-3 years.

Update 14 Apr 2023 – some clarifications, updated running data & running costs. (70% heating on HP & ROI < 10 years)

Update 3 Nov 2023 – some clarifications and cooling season

Update 14 Nov 2023 – Background & FAQ + minor updates.

Update 10 Jan 2024 – Cost details


Heat Pumps have been used for years in mild climates and more recently have gained popularity in colder climates. They work by transferring heat from the outside to the inside in the same way an Air Conditionner works (transferring heat from the inside to the outside). A Heat pump can be operated in both direction providing heating in winter and cooling summer.

The biggest barrier in past years was the capacity of the unit, especially at colder temperatures. As the outside temperature drops, the heat pump can’t transfer as much heat (nor do it as efficiently) while the heating requirements to keep the house at temperature increase. Recent advancements have made heat pumps retain their capacity and efficiency at colder temperatures, some much below 0. However, most installations in colder climate require a back-up solution in case temperatures drop to a point where it cannot keep up. This backup can be through central or plank resistance heating (ie electricity) or, through a conventional furnace (natural gas, propane, oil etc.).

The setup:

I have an existing forced-air heating system with a semi-recent natural gas furnace and an old air conditioner. While the air conditioner worked decently, it was quite inefficient (SEER 10) and nearing its end of life. I live in Ottawa, Canada which has temperatures ranging anywhere between -30C (-22F) and +30C (86F) so both heating and cooling requirements are fairly high.

Equipment Options:

I have a few friends that recently replaced their older furnaces with cold climate heat pumps. Although they would be the most effective for reducing carbon footprint, they are quite expensive and still need a backup solution for low temperatures. Besides, I have a working furnace that has a lot of life left in it. (too much cost for my case – although these currently have federal rebates that can be over 5000$)

My real focus was to replace the air conditioner and see how much more a heat-pump would be. A heat-pump is essentially an air conditioner that is run in reverse, cooling the outside and heating the inside. In this setup, a heat exchanger needs to be installed in the duct work above the furnace (same thing for an Air Conditioner) and a unit installed outside. The heat-pump cannot provide all of the required heating in very low temperatures unless it is very large so the thermostat has to be configured so it works until a certain setpoint, after which the natural gas furnace provides the heating. (I already have a nest thermostat that can negotiate this setpoint)

Finally I could have just opted to go with an air conditioner only. (cheaper but no heating)

Heat Pump Sizing

With the approximate price of the Heat Pump being about 35% more than an Air Conditioner (2000$ more in my case), I figured it would likely at least break-even in terms of cost and reduce my carbon footprint quite a bit.

I had a home energy audit conducted as part of an eco-home renovation program so I had some estimates with regards to how much heating and cooling my house would require. I also had some data from tracking my Nest data via the API during the heating season so I could estimate what my actual requirements were.

With a system that is too small you run the risk of not being able to keep up with heating/cooling demand and having the system run continuously without being able to maintain the desired temperature.

With a system that is too big, you run the risk of short-cycling which puts additional wear on the system and can create some temperature imbalances or reduced humidity management.

As I was going for a cheaper single stage heat-pump, I had to balance my heating and cooling requirements. With a dual-stage or continual stage, the system can ramp up or down so if you have too much cooling in the summer, it can just run at a lower stage and avoid short cycling. In my climate, heating loads are 2-3x the cooling load, so there is no risk of sizing too small for heating.

As heat-pumps essentially transfer heat from one location to another, the more temperature difference, the less effective they are. In other words, as the temperature outside gets colder, the heat pump can’t pump as much heat inside. This is why there is a cross-over temperature in a dual-fuel system where the heat pump just can’t keep up anymore (output-based cross-over temperature).

It also means that heating gets less efficient as the outside temperature drops. So it takes more electricity (and costs more) to keep the temperature. So at a certain point, it becomes less expensive to heat using the regular furnace on natural gas than the heat pump (economic-based cross-over temperature). As natural gas is fairly cheap compared to propane or oil, I expect my economic cross-over temperature to be fairly high. If you have propane or oil, your temperature would likely be below the output-based cross-over temperature.

The lower the cross-over temperature is, the more cost savings and more carbon reduction you can get. So sizing correctly is important to make sure it is both economical and eco-friendly.

Essentially I had 3 scenarios to consider, 1 estimated loads pre-eco retrofit, 2 estimated loads post-eco retrofit and 3 existing sizing/nest data. On top of that, do I use the economic cross-over to just run when it is cheaper or, do I try and use the output cross-over to save more CO2 emissions?

Sizing procedure and calculations

I calculated what each scenario would give and essentially figured slightly oversizing for cooling loads would give the best cross-over temperatures. I created an excel sheet with the various options and my calculated loads to see where my cross-over points would be and corresponding cooling requirements/output. (Horizontal axis is in F, vertical is in BTU/h) (Please let me know if you are interested in this spreadsheet, it is possible to enter your own variables or pick some options from a list and see how it would look like)

To do this, you basically have to plot a line between where you assume there will be no heating required up to where you need the most heating (design temperature – 99th percentile of how cold it can get). In this graph you can see my worst case, where the heating load is 80kBTU, the optimal post-retrofit load at 40kBTU and in the middle, what my estimates from existing equipment/data gives around 50kBTU. This line basically maps the increase in heating you need as temperature drops while assuming it is a linear increase. For my region we assume no heating at 60F (15.6C) and the 99th design temp at -7.24F (-21.8C). I looked up the design temperature based on historical data for my location.

The same kind of procedure for the cooling where I used a range between 80% and 125% of the estimated cooling load. The cooling data typically only has one data point so the target is to size to fall in between that range. On this side however, I had a bit of an issue with the estimated loads. My existing AC was already beyond these ranges and on hot days can run 12hrs in a day or more. So while I wouldn’t say it is undersized, it certainly is not *that* oversized to fall out of the range, at least in my mind.

The next step is to plot the output curves of the actual heat pump options. In this case, I used what data was available on the spec sheets. There is some better data available from the manufacturer but they don’t provide it unless you create an account as a dealer/supplier etc. so this is the best I could do. Specs usually have a cooling capacity, and two heating capacities one at 47F (8.3C) and the other at 17F (-8.3C). As noted before, typical heat pumps have less capacity at lower temperatures and also are less efficient. (However some cold climate heat pumps have more output at lower temperatures so using only these 2 data points may not be a good estimate). I added several different heat pump models to this spreadsheet so I could easily compare models of different size and efficiency.

Output Cross-over temperature

Now that we plotted the heating load and the outputs of the various sizes, we can calculate the output-based cross-over temperatures. Essentially that is the temperature where the heating load line crosses the output line. At this point (if the temperature was stable for hours), your heat pump would run 100% of the time and not be able to provide enough heat so the house would slowly cool down. In the table below is the comparison of 3 options:

Heat pumpOutput cross-over Temp% of yearly heating load
2.5 Ton4.4C (40F) to -2.6C (27F)11-44%
3 Ton1.5C (34.9F) to -6.5C (20F)24-60%
3 Ton (more efficient)1.9C (35.4F) to -6.3C (20.6F)23-59%

Here the difference between the more efficient 3 Ton heat pump was a bit more efficiency at higher temperatures when heating and more efficiency at cooling but a bit less output. Essentially, it would use a bit less electricity to run, but not be able to be on for as much of the time plus, more efficient units are more expensive.

Nevertheless, there is a huge range of cross-over temperatures and heating load diversion between the scenarios ( worst case, optimal, realistic). With a variety of house sealing and window replacement I had done and with the data I had, I was fairly certain it would be closer to the higher % in these ranges (lower cross-over temperature).

Economic Cross-over temperature calculation

I also calculated my economic-cross over temperature based on the tiered electricity pricing (and lower cost on weekends), as well as the cost of natural gas. (More details on this calculation here) Essentially I blended the rate for electricity as a rough estimate and looked at the coefficient of performance (efficiency that varies depending on temperature). The temperature when it would start being more expensive on average ie, the economic cross-over temperature was something like 3F, much below the output cross-over temperature. (for 6 hours a day on weekdays its actually more expensive to run the heat pump than gas, but the average over a week is quite a bit lower – you could fine-tune it for cost savings, but I rather save the planet than a few cents and the trouble).

I also looked at the average cost by integrated the coefficient of performance and temperature (basically how much efficiency the system would have from when I need heating down to the cross-over point and weigh the number based on how much annual heating there would be at each step). My averaged economic cross-over was something much below whatever the output-cross over would be. In fact, since most of the heating would happen at efficient ranges, it would never be more expensive to run the heat pump on average.

Sizing choice

With the economic and output based cross-overs in mind. It was fairly clear to me that the 3 Ton size was correct, although it was a bit oversized for cooling loads, the huge difference between the 2.5 and 3 in the heating season seemed worth it (and the cost was only around 5% more for upsizing). Going to 3.5 or 4 ton was not really an option as it would risk being way too oversized for cooling. With a 2-stage or variable stage, this would have been something to consider, since the heat-pump can run in a lower output stage during cooling, you can oversize without having to worry about it too much. However, my original justification was getting a new Air Conditioning + adding heating as bonus, so I’m going the cheap single-stage route.

Cost savings in using dual-fuel?

I calculated a range of savings (running costs) I could expect based on the estimated heating load and costs. Depending on the load scenario and using a blended electricity rate, comparing that to the cost of the furnace running on gas etc. I calculated running cost savings in the 10-20% range.

That’s great for running costs but in terms of capital return? Adding in heating functionality to the AC would pay itself off in 15-30 years. That’s not great seeing as that could be more than the life of the system, but the price of gas is forecast to keep climbing with increases in carbon taxation and extraction costs. On top of that, this unit also reduces my cooling running costs by 40% (my current AC is super inefficient) so it will pay itself off faster than that grim forecast. I decided to go for it anyways, divert some CO2 and save the planet!

Dual-fuel heat-pump system in use

I bought the system and got it installed, no issues getting it up, they just did a drop-in where my old AC was, piped in the lines, installed a new coil on top of the furnace, wired it all up with my Nest and good to go!

Defrost mode

The system was installed and configured so the gas furnace would turn on when the heat pump needed to defrost. This happens when some frost is detected on the outside unit (since it gets colder to “pump” heat inside, the temperature can go below the dew point and lead to frost which reduces the efficiency/effectiveness of the unit.). To get rid of the frost, the unit runs as an Air conditioner to heat the outside coils and remove the frost, this does cool the coil in the air ducts so heating them up with the gas furnace reduces the defrost time, and avoid “cooling” the house during this temporary defrost. Yes this does create some extra CO2, but it also extends the range/time where the unit can operate so overall likely a net benefit.

Cross-over temperature & findings

With the nest thermostat dual-fuel setup, I set the cross-over temperature below my calculations as an initial trial to see how things would go.

The heating is certainly much slower than when the natural gas furnace runs. It’s a slow and steady kind of heating vs a fast rise. It’s something to note and something to keep in mind (especially if you have a large temperature rise like if the house was in eco-mode while on a trip). But I would say its different rather than bad, the temperatures in certain areas of my house are more consistent with this type of heating. Adjusting to it might be difficult for some that are expecting a huge heat input, I could see some people thinking it doesn’t work if they aren’t expecting it to be different.

My cross-over temperature is also quite a few degrees lower than my calculations (which is a good thing!). I was in the range of -6.5C and at -9C, it seems to at least maintain temperature (perhaps slow rise). I haven’t had temperatures much below that for a long time since it was installed so difficult to say if I could push it lower.

EDIT Jan 2024– Lower temperatures have occurred since then, -9C is pretty much my current minimum temperature but at that point really depends on wind and sun conditions. Lots of wind means the HP is running pretty much 100%, lots of sun means it runs quite a bit less and one could push to lower temperatures.

Essentially with this lower temperature, more of my heating is diverted to the heat pump, less carbon emissions and more cost savings. This might not always be the case but I did forecast my cross-over temperature to be on the better side of my estimates so it’s fairly aligned (but better) than what I was expecting.

My calculated range of 24%-60% of heating being taken over by the heat pump is actually closer to 70%! That’s quite a bit better than I was expecting & should translate to better running costs too!

EDIT Jan 2024 – With the huge drop in natural gas prices and increase in electricity moving the economic cross-over to -7C for this season and -9C being marginal in some situations, I upped the cross-over to -8C (63% of heating). I assume with some ongoing updates to house insulation/sealing and the natural gas prices going to rise from this 2-year low, I may change it next season.

I also took a look at my estimates for electricity costs. I used a blended rate as a first pass. We have tiered pricing; every weekday the cost varies depending on the time of day: low 12 hours, medium 6 hours, high 6 hours. On weekends and holidays the cost stays low. My blended rate just assumed heating would happen equally at all cost tiers when in actual fact a bigger proportion happens in low, off-hours. (cost calculation here )It ends up being 8% less than I had originally calculated this along with the higher % of heating diverted translates to running cost savings being 40% better and the return dropping to about 10 years. With the forecast increase in costs of gas & this not being the limit of the system, the heat-pump will pay itself off even sooner than that!

Funny thing to note is that Enbridge, our gas provider charges for gas based on manual meter readings every 2nd month. One month they estimate your consumption, the next they send someone to do a reading and adjust. Well for several months of the heating season after the heat pump was installed, I kept getting negative balances. The estimate month was high, and the actual reading one month later was lower than they had estimated the previous month (leading to them owing me money). I’m not sure if Enbridge will fix their estimates but you would think they would have some sort of smart meter system instead of sending people out to read the meters every 2 months.

After a cooling season, I have to say it does run quite a bit less than my previous AC. I used to have some 12+ hours running times on really hot days, which is now closer to 8 hours. The output, efficiency are both higher and it is newer so probably working better than the previous AC.

Heating season is starting up again and I’ll see how it performs after a full season (with colder temperatures too).


There is a lot of talk lately about heat pumps and I was surprised at what some experts were saying. It didn’t really seem to make sense physically or only seem to apply to certain cases.

Does it cost more to run a heat pump?

No, for me, it’s actually 10-20% less expensive than using natural gas! If you have propane or oil it’s probably a lot cheaper to use a heat pump. However, electricity and fuel rates vary so it might not be the same for everyone. Tiered electricity rates are lowest at night, when it is coldest (ie you need most heating) and while the heat pump is less efficient. The difference in electricity cost, far outweighs the efficiency drop so it is actually cheapest to run the heat pump at night when you need it.

Can I use a heat pump on my house?

Yes you can. I heard an expert say only if your house is well insulated and newer. In my mind, that’s BS, if your current furnace can heat the house, a heat pump that can provide the same output at that temperature can do it too. Heat output is heat output regardless of the source, if you buy something that is too small yes you will have problems heat-pump or not. What the actual dialogue should be is that getting a bigger sized heat pump for a leaky not-well insulated house might be quite expensive (and you’re wasting $ with your furnace too). But you aren’t limited to going just heat pump or just furnace, you can blend it with a dual-fuel solution like I did.

Does a heat pump work in cold climate?

Yes, it certainly does but depending how cold, you might need a backup system (electric, or existing gas/propane/oil furnace) for the few times where the temperature is extremely low. Keep in mind, that in Ottawa, where we do get some -30s C sometimes, the actual temperature is higher than that most of the time. With a cross-over temperature of -9 C I can get about 70% of my yearly heating on the heat pump. (The National Research Council has some fancy charts you can look at to see what outside temperature as cross-over gives you in terms of yearly heating)

Do I need to buy a cold climate heat pump or will a regular one work?

No, mine isn’t considered “cold climate” its output drops faster with decreasing temperature than a cold climate one would but I have run it at -9 C without a problem. To cover lower temperatures, I would not only need a cold climate heat pump, but one with greater capacity as well. Since this covers me for about 70% of my heating, it was not worth the cost at this time even thought I wasn’t able to get any gvt grants for it. However, depending on your location you might have different results.







7 Responses to “Implementing Dual-Fuel house heating (Heat pump)”

  1. Don Sproule Avatar
    Don Sproule

    I found your article very informative. I too am from Ottawa and have just had installed a ccASHP to complement my natural gas furnace. In addition to daily readings, I actually can get on-line hourly readings from my Lennox S30 thermostat of: gas load, ASPH load, external temperature and internal temperature What I don’t understand is that even during the very warm weather we’re having this week 30F to 34F) , I’m still get hourly periods where I’m getting significant (and unexpected) gas demand. I’m suspicious the settings have not been done properly and so I’ll be reaching out to my installation company. In the meantime, could you send me the spread sheet that you offered? I’d like to go back to first principles and determine what the expected cross-over points should be. Thanks. If you’re willing I’d like direct email contact as well.

    1. sunshine Avatar

      Thanks for the comment and happy new year!
      I can certainly send you a copy of the spreadsheet, may take a few days to dust it off and update some of the values in there/make it easier to use. I’m not sure how well it works with a ccashp where the curves for output and efficiency may not be as linear (or increasing), some extra interpretation may be required.

      I can think of three possibilities for seeing gas load in your system 1. The system may fire up to speed up defrost mode (when the outside coils drop in temperature below the dew point, frost starts to accumulate on them and it needs to be purged. The heat pump runs in reverse for a few minutes to do this and you can configure the furnace in a dual fuel system to turn on to speed up the process).
      2. The system may be configured in a sort of low/high mode as a multi stage furnace would be. If the heating load is high (big temp rise required), the temp does not rise fast enough or the heat pump has been on for a certain time, it might switch to the gas furnace to try and reach temperature more quickly.
      3. The cross-over temperature may be set too high (or the heat pump can’t keep up at lower temperatures).

      Hopefully that is helpful, I’ll take a look at dusting off the spreadsheet.
      Thanks again for the comment!

  2. David Avatar

    An interesting read. I recently had a Lennox SL25XPV-024-230A heat pump installed with a Lennox Furnace SLP99UH070XV36B as a dual fuel setup. Based on my utitily bills the spreadsheet I use currently calculates the cost decision point of switching to natural gas as -1.61 C. I’m also in the Ottawa area served by Enbridge and Hydro Ottawa. Natural gas is currently ridiculously cheap, so I expect as NG prices rise I will be adjusting my low balance point.

    1. sunshine Avatar

      Thanks for the comment and happy new year!
      Yes it’s totally possible depending on efficiency of unit and current prices. My original numbers used electricity rates before the rise we just had and gas rates were quite a bit higher. However I also drilled down the actual cost charged for both (there’s extra fees/rebates applied per kwh/m3 that aren’t obvious unless you look them up – they are on the bills as seperate line items) and time of use for electricity also makes a good difference.

      Note that I blended the electricity rate ie. Weekdays -low rate 50% of the time, medium 25% high 25% , weekends low rate 100% – so in total 64% of the time it’s at low rate, 18% medium, 18% high. Most of the heating happens at low rate and holidays are low rate too so this blend should still over estimate the electricity costs a bit.

      I can take a look at my spreadsheet while i update it and let you know what things look like now.
      Thanks again for the comment!

    2. sunshine Avatar

      I just re-ran some numbers here fairly quickly looking at all the rates/rebates with latest available (some kicked in Jan 2024). I took out the fixed prices as those won’t change regardless if you use the HP or the nat gas so these are all the variable rates (from all the different fees they charge except HST which wouldn’t impact this calc as its the same on both – I have a dirty spreadsheet with this if you want it, I might bundle it in a tab with my main one too).
      Nat Gas -> 0.41848 $/m3
      Electricity low -> 0.094767 $/kWh
      Electricity medium -> 0.124342 $/kWh
      Electricity high -> 0.175042 $/kWh
      my blended (average) electricity rate is 0.1144 $/kWh

      So for my situation the break-even COP is 2.7 which is just about 19F (-7.2C) at the time of writing – about 63% of heating load would be on the heat pump based on historical temperatures (and probably more with the milder climate recently).
      Let me know what you think and if you’re interested in seeing more details.

  3. donald carswell Avatar
    donald carswell

    Great article. I was interested to hear that almost 70% of gas use can be met by a single stage heat pump. Would you be able to share your ballpark installation cost? We are trying to convince others in our area (Peterborough, ON) that a dual-fuel system would be a cost-effective idea but there is nothing like an example to help.

    To commenter Don I would suggest that the staging delays and/or differentials have not been set the way he would wish. When I installed my Ecobee thermostat I had to be less aggressive than the defaults or it would kick on additional heat stages which I did not wish for.



    1. sunshine Avatar

      Hi Donald and sorry for the late reply.
      Yes about 70% of gas use is met with a heat pump (in my application) – this will depend on housing size/insulation, regional climate and the size of the heat pump.
      My cost was just under 7200$ (incl HST) for a 3-ton unit which included removal of old AC/coil and installation of the new unit/coil (they also ran a new electrical wire and new breaker as existing one for AC was too small).
      Prices may change depending on region, layout of equipment and time of year as well so grain of salt.
      Hope it helps!

Leave a Reply

Your email address will not be published. Required fields are marked *

Exit mobile version