Rotary heat exchangers in northern Swedish winters

By Erik Nyström

Our house was built in 1965 and originally relied on natural ventilation. When we decided to replace it, we chose a balanced mechanical ventilation system with heat recovery. In Sweden this is commonly called FTX, from från- och tilluft med värmeåtervinning. Our system uses a rotary heat exchanger.

Everything worked well until a period of genuinely cold weather in January, when the unit eventually stopped because of ice. I started searching for information and quickly found plenty of forum discussions from people with similar problems. I also talked to friends who had experienced the same thing. Some of them have problems with their ventilation systems more or less every winter when it gets cold enough.

I could not really accept that as normal operation. I have much the same attitude to ventilation systems as I do to computers: once properly installed and configured, they should do their job and stay out of the way. You should not have to babysit a ventilation unit, shut it down during cold weather or periodically thaw it with a heater.

So I started digging into how the system actually works. I read about the physics of condensation and frost in rotary heat exchangers, learned the control logic of our Acetec A100 in considerably more detail than I had originally intended, watched temperatures and airflow, changed settings and experimented during cold weather. That led to several changes. I became much more conservative about defrosting, concluded that we needed a preheater for the really cold periods and had an electric one installed. We also replaced the combined outdoor-air and exhaust hood that had caused us problems with two separate, generously sized hoods.

One claim kept appearing while I was reading about all this: rotary heat exchangers do not really need defrosting, and preheaters are unnecessary because the rotor handles frost by itself. After looking into the physics and seeing what happens in practice in our climate, I do not think that is good advice for northern Sweden.

TL;DR

If I were installing a rotary FTX system in northern Sweden, this is what I would do:

Those are my recommendations, not settings taken from a manufacturer. Swegon has published a very useful explanation of condensation and frosting in rotary heat exchangers, and I have used material like that to understand the underlying physics. The practical conclusions here are my own.

Rotary heat exchangers can freeze

Rotary heat exchangers cope with cold weather much better than many fixed heat exchangers. Part of the reason is simply how the rotor works. The matrix continually moves between the warm extract air and the cold outdoor air, allowing some moisture deposited on the extract side to evaporate again when that part of the rotor reaches the supply side. Frost therefore starts becoming a problem at considerably lower outdoor temperatures than it does in many fixed exchangers.

That does not mean frost cannot form. Warm indoor air contains moisture, and when it is cooled far enough inside the rotor, water will condense. If this happens below freezing, frost or ice can form on the matrix. Whether that frost continues to accumulate depends on several things, especially the outdoor temperature and the humidity of the extract air, but also the extract-air temperature and the efficiency of the rotor. In one example, Swegon shows that freezing can start below roughly −12 °C with 21 °C extract air and less than 30 percent relative humidity. Their calculations also show freezing beginning around −8 °C when there is already excess water in the rotor. There is therefore no single magic outdoor temperature where a rotary exchanger suddenly starts or stops being safe.

This is also where I think the phrase “rotary heat exchangers don’t freeze” causes confusion. Frost forming is not the same thing as the rotor immediately freezing solid. A properly controlled unit can deal with a certain amount of frost, and in many conditions the moisture will disappear again when conditions improve. The problem is prolonged cold weather where frost accumulates faster than the system can get rid of it. As ice fills parts of the rotor, pressure drop increases, airflow is restricted and heat recovery becomes worse. Swegon describes exactly this behaviour.

Why I start defrosting early

I want frost protection to become active at roughly −10 °C actual outdoor temperature. This is intentionally conservative. I have no interest in determining that my particular house, at a particular indoor humidity and airflow, starts accumulating frost at exactly −12.7 °C and then trying to operate as close to that limit as possible. Indoor humidity changes when people shower, cook or dry laundry, and the weather can change quickly.

I would rather have the unit defrost occasionally when it turns out not to have been necessary. Different units achieve this in different ways. They can alter the supply and extract airflow, change rotor speed or otherwise reduce heat recovery for a while so that the cold parts of the exchanger become warmer. Swegon describes reducing rotor speed as one method: efficiency falls, the exhaust-air temperature rises and the rotor can be brought out of the conditions where frost is accumulating.

Pressure drop across the rotor can also be measured and used to determine when defrosting is required. That is a legitimate control method, but it is not the approach I prefer for a house in this climate. An increased pressure drop means that frost has already restricted the rotor enough to become measurable. I would rather use outdoor temperature and time as the primary trigger and accept some unnecessary defrosting. When it is −25 °C outside, avoiding a few unnecessary cycles is fairly low on my list of priorities.

Why I also use a preheater

Defrosting only gets you so far. During prolonged periods of very cold weather, I want another layer of protection, which is why we had an electric preheater installed in the outdoor-air duct. Mine is set to start at around −22 °C actual outdoor temperature.

The purpose of the preheater is not to make the incoming air warm and certainly not to heat the house. It only has to raise the outdoor-air temperature enough to improve the conditions inside the heat exchanger. Swegon describes preheating as particularly useful when low outdoor temperature is combined with enough moisture in the extract air that the rotor becomes overwhelmed with excess water. They also point out that the incoming air does not need to be heated very much to avoid the conditions where excess water and frost accumulate.

For me this makes an electric preheater an obvious addition in Luleå. It spends most of its life doing nothing and only starts operating during the colder part of the winter. There is of course an energy cost while it is running, but that has to be compared with the alternative. I would rather use some electricity during the coldest hours than have a ventilation system that gradually fills with ice or has to be shut down when ventilation is still very much needed. Swegon also notes that because preheating may only be needed for a relatively small number of hours, an electric heater can make sense despite electricity being an expensive source of heat.

I therefore do not see defrosting and preheating as alternatives. Defrosting handles the conditions where frost begins to accumulate. The preheater gives the system more margin when the weather becomes cold enough that defrosting alone is no longer something I want to depend on.

When the rotor finally stops

A rotor that has frozen solid is usually the end of the process rather than the beginning. Frost can build gradually while the unit continues to operate, and repeated defrost cycles may remove some of it without necessarily clearing everything. Depending on the design of the unit, water from partially melted frost can also reach colder areas and freeze again.

Eventually the remaining ice can become substantial enough to obstruct airflow severely or physically interfere with the rotor. At that point the stopped rotor is simply the most obvious symptom. This is another reason I prefer to intervene early rather than use the rotor itself as an experiment to discover exactly how much frost it can tolerate.

The exhaust hood matters more than you might think

Something I underestimated when we installed our system was the importance of what happens outside the ventilation unit. We originally used a combined outdoor-air and exhaust hood, and it became part of the freezing problem. The exhaust air leaving the house contains moisture, and at −20, −25 or −30 °C it can produce a surprising amount of frost around the outlet. A fine grille or some other narrow restriction can lose a large part of its free area before the amount of ice looks particularly dramatic.

If the exhaust side starts icing up while the outdoor-air intake remains relatively clear, resistance on the exhaust side rises and extract airflow falls. The supply side may meanwhile continue bringing roughly the same amount of air into the house, so the ventilation becomes unbalanced and the house moves towards positive pressure. There is also a second effect that is particularly bad for the heat exchanger: with less warm extract air passing through the rotor, the rotor itself becomes colder. That makes further frost accumulation easier. More frost in the rotor adds still more resistance, which can reduce extract airflow further. The system can therefore enter a downward spiral where icing at the exhaust hood contributes to poorer airflow through the rotor, the colder rotor collects more frost, and the airflow becomes worse again.

We eventually replaced our combined hood with separate outdoor-air and exhaust hoods. After that experience I would do the same from the beginning in a climate like this. I would size both generously and leave plenty of margin for airflow and pressure drop. A low pressure drop when everything is perfectly clean is nice, but what matters in January is whether there is still enough free area after frost has started building.

I would pay particular attention to the exhaust hood. It should be open and generously sized, with as little fine mesh or restrictive grille as practical. Ideally the grille should be accessible from outside and removable if necessary. If something is going to collect ice, I at least want to be able to reach it without dismantling half the ventilation system.

Why I still prefer a rotary heat exchanger

None of this has made me regret choosing a rotary heat exchanger. Quite the opposite. It recovers heat efficiently, handles low outdoor temperatures very well and can return some moisture to the supply air, which is useful during our dry winters. My issue is with the much stronger claim that because rotary exchangers handle cold weather well, frost protection can simply be ignored.

That does not match the physics, and it does not match what led me into this subject in the first place. When our own unit stopped, I found many other homeowners describing similar problems. Friends here in the north told me about systems freezing during cold spells, and some deal with it repeatedly during cold winters. Forum discussions contain the same kinds of stories: ice accumulating, rotors stopping, ventilation being shut down or people manually thawing equipment when the temperature falls far enough. One example from nearby Boden describes an FTX system that is normally shut down below roughly −25 °C because of recurring freezing problems. It is only an anecdote, but it is a very familiar kind of anecdote in this part of the country. Byggahus: discussion including the Boden example

I still think a rotary heat exchanger is one of the better choices for a house here. I just think it should be installed and configured on the assumption that a northern Swedish winter is actually going to happen.

Further reading