There is a HEPA filter option that can be purchased separately. HEPA filters are the best for removing pollen and wildfire smoke during summer. I do not recommend the filters to remove ambient PM pollution during normal periods, a better idea is a good PC fan based filter appliance inside the home. This is because the filters cost a fair bit and reduce flow rate through the machine.
You must choose between the Automatic Storm Valve add-on and a filter, at any given time. The filters keep water out anyway, and also the pipes are slightly tilted so any water that gets in runs back out anyway.
Because of the new noise splitter, adding the filters is a bit awkward, when combined with the noise splitter, but it works ok. You also need an exterior plate and hood to use the filter, which increases price significantly. Beware it may not fit in some windows due to the way it interferes with (hits) the exterior window sill.
The filters are standard, widely available filters. You do not need to buy new filters from us at a high markup, and you do not even need exactly the right size of filter. You will always be able to get the filters you need for optimal functionality ,even into the distant future, like 20-100 years from now.
Any filter reduces the air flow rate. In Summer Mode (unidirectional supply) with the heat exchanger removed, maximum net provided fresh air flow rate with the HEPA filter is 60 CFM per module. The filters can handle bidirectional flow at these flow rates.
You can also get "standard" i.e. non-HEPA filters in this size. "Standard" car cabin filters are approximately equivalent to MERV 13: ("Performance Study of Five Different Cabin Air Filters in the Laboratory and On-Road Drive")
The "standard" filters impede airflow less, but I have not determined exactly what the figures are.
The car cabin filters I chose for the main filter kits are just the biggest car cabin filters that fit under the hood of the TW4. The Filters are Bosch 6075C HEPA filters. There are many of the same size.
I have also made an adapter for a standard 8 inch by 8 inch furnace filter, however I haven't tested it all out yet, esp in a storm/rain. I haven't measured the flow with these furnace filters yet.
The interface to the filters is a flat plate, so it's pretty easy to make adapters.
See the Learn more page for pictures of the filter, they cannot be embedded here unfortunately.
See the spreadsheet on the Learn More page for exact figures. In Ottawa, the calculations show a single pair of modules (one ERV unit) saves about 166 cubic meters of natural gas per year. That's about 150 kilograms of fuel! The monetary savings are shown in the spreadsheet.
The heat exchanger can include permanently bonded silica gel, which transfers water vapor between incoming and outgoing air with excellent efficiency, greater than 70%. This represents latent heat, because if the air coming in is too dry, you need to evaporate water to replace it, which absorbs heat. If it is too humid, the air conditioner has to condense the water, which requires primary energy.
You can save a copy of the spreadsheet, and then edit it, to enter the figures for your region for electricity and fossil fuel prices, etc. and see for yourself that the rate of return on a given quantity of capital/the rate of return on investment is far higher than photovoltaic panel installation and other HRV and ERV units.
There is a 5 year Warranty against defects in material or workmanship on all our products. Unlike with conventionally manufactured products, this is not an indication of design life.
In reality, all conventionally made products include what's called planned obsolescence - they are deliberately made to break soon after the warranty expires, and we have all come to expect this. When they do break, they cannot be fixed. The TW4 is not like that. There is no planned obsolescence.
Doing things in this way allows the customer to see the incentive and ability to employ planned obsolescence is gone, because it is not possible to demand high prices for replacement parts and labor when they can both be obtained from a large variety of sources, aside from the original manufacturer. First of all, there is no incentive to make things break as soon as possible after the warranty expires, and secondly even if they do you know you can get a fair price on replacement parts, and even into the distant future.
It is made to last as long as the fundamentals of technology allow. This improves the investment properties greatly. We believe doing things right is better than any warranty. We believe good engineering is better than a paper promise, especially for very long time periods, and a building element should last for 25, 50 or even 100 years.
Extremely long warranties don't really help very much in practice as companies go out of business or don't honour their promises, proof of purchase gets lost etc.
50% of home improvement product type companies go out of business within only 3 years. 70% are gone within 10 years.
The design and production of the storm valve module is done, but the functionality to use it is not yet implemented.
If you live in an area where storms occur with enough severity to force air or water through the unit, and the unit is not installed in a sufficiently sheltered area, you should use the Automatic Storm Valve add-on module. This will close the intake when the system detects extreme wind or when it is turned off, preventing the intrusion of air during a storm. The unit/pipe is installed at a very slight angle so that in the unlikely event any water gets in, it runs back outdoors again.
It's generally still advisable to install it on a wall that gets relatively little wind hitting it, or less head-on, because high winds will overwhelm the fan and preclude normal operation. The valve will close, but continued normal operation is better still. Also at some point as the wind pressure gets stronger, efficiency will be impacted due to asymmetrical flow through the device, but not until about 30 kph head-on. This depends on the flow level commanded. If the wind is 30 kph but at an angle, that implies a lower pressure, so you might be good to, say, 40 kph. The winds in an area tend to not come from random directions, but are influenced by geography and so on, so choosing the right side of the house helps a lot with this.
Open source has major implications for the long term return on investment, and real-world reliability of the system. You also know what your are getting before you buy a lot better.
Any of this nature appliance must have some custom parts, but by making them 3d-printable and making the source code available, you can know that you will always be able to obtain replacement parts.
When a system is made from common, widely available parts, and you have the blueprints, you will never be at the mercy of any one company, which will keep you waiting for long periods or charge too much for replacement parts or service. All conventional companies engage in what's called Planned Obsolescence. They design things to break and in such a way that you have to come back to them for parts and service - and then they eventually stop supporting a product or go out of business, leaving the product an expensive piece of trash. Which is built into your wall.
If you want to amortize the cost of a product over a long time period, it has to last a long time. 25 or even 50 years is sensible for a building element.
Open source also provides a certain inertia and staying power, and is more respectful of the user in many ways.
The design can be seen and criticized by anyone, helping prevent the existence of problems at an early stage.
For a rare few of us, the ability to modify, reprogram and enhance the unit, or shimmy it into a different role, is valuable.
Environmental friendliness also shakes out of this designed-to-last approach. It's way less stuff in the trash, and less stuff produced means less pollution.
We support the Right To Repair movement. This is especially important for farm equipment and other basic infrastructure, such as housing and building technology, which we all depend on and form a major part of the cost of living.
Consider: fans never last more than a few years of continuous operation. If you purchase a conventionally made unit, will you be able to find a replacement fan when the time comes, several years from now? Will the company even sell you one at all? Better ask them before buying. What if the company is out of business? What is the plan?
What happens if something gets damaged, due to a lighting strike/power surge, fire, severe storm, or just careless mechanical trauma? Kids or pets? How much will it cost to rectify? These are part of life for a building element. With the OpenERV, these things can all be solved relatively quickly and cheaply.
All the other ductless ERV and HRV units on the market are prone to getting an excess of air blowing through them, which greatly reduces real world efficiency. This is partly due to the limitations on fan technology. The efficiency they advertise is merely under laboratory conditions with no wind etc, low flow rates, and only part of the cycle, and is not realistic. The OpenERV automatically monitors the wind speed and compensates. It has a built in solid state air flow sensor on the indoor side.
This improves comfort by helping prevent drafts, and improves the return on financial investment. It also allows the fan to reduce it's speed during calm conditions, reducing noise levels.
There is a standard part of which is used to measure the impact of wind, in some ways, on these type of units, DIN 13141-8. There is S1, S2 and S3, as ratings. S1 is the best. The TW4 actually exceeds S1 requirement by a considerable margin.
It's generally still advisable to install it on a wall that gets relatively little wind hitting it, or less head-on wind, because high winds will overwhelm the fan and preclude normal operation. This applies to all such units, but the TW4 is better than any other I have been able to find on this subject. At some point as the wind pressure gets stronger, efficiency will be impacted due to asymmetrical flow through the device as the fans are already compensating as much as they can, but not until about 30 kph head-on, which is pretty windy. If the wind is 30 kph but at an angle, that implies a lower pressure against the house, so you might be good to say 40 kph. The winds in an area tend to not come from random directions, but are influenced by geography and so on, so choosing the right side of the house helps a lot with this.
Yes, but this has not been added to the firmware yet. You will be able to command the unit over MQTT to operate in supply, exhaust, or energy recovery mode. You can remove the heat exchanger, which only takes a couple minutes, and can be done from the indoor side, to increase the airflow rate during summer (no sense impeding airflow with the heat exchanger if you don't need to).
We should ask this question of any appliance.
The TW4 requires almost no maintenance. If the unit gets dirty due to unusual conditions, you can remove the components from the pipe and clean them, but that should not normally be needed or useful. The HEPA filter may need to be cleaned or replaced, it should be good for at least one season, depending on pollution levels.
After about 7-12 years of elapsed time (more elapsed time if not run at full power etc.), the fan motor bearings will need to be replaced. They cost 25 cents each and you need 4 of them. It can be done by hand with basic tools in about 20 minutes. I have made a basic youtube video to show the process. This process of wear has some randomness to it and depends on usage levels.
If it is damaged due to extreme weather, mechanical trauma, a power surge or lightning, the relevant parts can always be sourced and replaced, either with identical 3d printed parts, or commodity, standard parts. Even the electronics can be replaced with standard modules. All screws are stainless steel, so they will never rust.
Yes, it is a generic "MQTT" device. This protocol is one of the most generic, broadly compatible and easy to use home automation/internet of things protocols. This means with some configuration, it can be controlled using Google Assistant, HomeKit, Alexa and other major home automation systems. You can also configure it to be controlled more directly by smart switches and knobs, or using smart switches/knobs that communicate with e.g. Alexa, which then commands the ventilator.
We plan to make it Matter compatible when suitable software modules are released for that system. The configuration is not completely trivial, but we will release a guide on how to do it.
The electronics actually have a Bluetooth module in there, but it's not being used and there are no plans to use it.
Ductless ERV manufacturer's frequently indicate a laundry list of supposed standards the units are compatible with. Some even claim inapplicable standards, such as those made for fans or ducted ERV or HRV units.
Fundamentally, the efficiency, influence of wind and other air pressure on the device, flow rate and electrical efficiency of the fan in terms of watts per CFM is what determines compatibility with these standards.
The OpenERV TW4 is RoHS and WEEE compliant, meaning it does not contain any of a list of toxic substances, such a lead.
By typical standards, a pair of OpenERV TW4 modules is suitable for about 1200 square feet of interior space. 60 CFM gives an ACH of 0.35 for 1,142 square feet of floor area with 9 foot high ceilings, which is "good" ventilation. If you just need some supplemental ventilation, you can get away with a larger floor area per pair. You should realize that ventilation isn't about floor area, though, it depends what and who is actually in the dwelling unit. If there is a lot of stuff and people, sources of humidity or smells etc, more ventilation is warranted.
ASHRAE 62.2 is the main standard for setting goals for indoor air ventilation rates in America. It says 0.35 ACH is a good goal, you can see more details here.
Some manufacturer's claim compliance with DIN 1946-6. This is merely a standard for the amount of ventilation air for a building. A ventilator unit cannot itself be compliant, only a building with adequate ventilation can be compliant. The TW4 can be used to make a compliant building.
DIN 308 is a legitimate standard for measuring the efficiency of Energy Recovery ventilators and Heat Recovery Ventilators. But it does not apply to ductless units, only conventional ducted units. The testing protocol is specific to ducted units. It also does not factor in the impact of wind pressure on efficiency. With ducted systems, this is reasonable as the fans are high pressure. In ductless units, this does not make sense as wind has a very large impact on units that do not have compensation systems (the TW4 does have a good wind compensation system).
To understand the impact of wind without compensation systems, the typical fan pressure is 3 mm H2O for a fan at maximal power in this type of unit (the TW4 can do 4.4 mm H2O). The pressure against a building can be calculated from the wind speed blowing directly against a wall (9.8 pascals is about 1 mm H2O). 1 mm H2O is only about 4 meters per second, which is 14.4 kilometers per hour. The average where I live is about 11 km/h. Flow is proportional to pressure in this instance. So at 14.4 kph you are going to get 33% more air flowing in one direction and 33% less in the other, which means 0.66/1.33 = 0.49, only 49% of the air flowing one way is matched by air flowing back the other way. You can't recover any energy from that 0.51 volumes of air that didn't even pass through the heat exchanger. The efficiency with which heat is extracted from the 49% actually goes up a bit, but if it was already 85% it doesn't have far to go, you cannot possibly exceed 100%.
And that's at full blast. Without wind compensation, if the fan is turned down, the whole thing becomes even more susceptible to wind, and can easily be totally overwhelmed with normal winds.
The TW4 has an excellent quality wind compensation system.
A useful standard for these ductless machines is DIN13141-8, which is about measuring thermal efficiency *in the absence of wind or static pressure*, and also some other things such as the degree of unintentional ingress/egress caused by wind or other air pressure across the wall.
I prefer the PassivHaus testing methodology for completeness and relevance of the specification and actual protocol, but they haven't published any test data with it for ductless units, because the prior generation devices were not adequate for passive house use.
At very low settings, it is truly inaudible in a normal environment. At higher setting, it is not perfectly inaudible if you are in the same room and there is zero other background noise, but it is still quieter than any other such unit. On medium, it will be drowned out by a typical computer fan, refrigerator, an open window with the typical background roar of the city, a forced air furnace or central ventilation fan of any kind, and many other normal noise sources.
Talking about and conveying an accurate idea of what exactly something sounds like is prone to error because individual perception, the type of noise, background noise, the acoustic environment (including the tendency to reflect echos, etc.) and other factors all play a role. But we can do some quantitative measurement, which all else being equal gives you a reasonable idea.
I prefer dBa because the meters are easier to get. In fact, I've never seen a meter that reads in Sone. The TW4 should be about 38 dBa on medium and is 41 dBa on max power, as far as my equipment indicates although I usually say it's slightly higher, to be sure. My house with absolutely nothing on in the house (so just street noise) is 36 dBa. When the furnace comes on, it's 43 dBa, 1 meter from the vent. So the TW4 is remarkably quiet, but it's not silent. (I am very careful about my home environmental quality and hate it when people say their stuff is "silent" when it's not.)
The fan is located on the outdoor side, which helps reduce the radiation of sound energy into the home, compared to putting it on the indoor side, or fully inside the room. This provides a profoundly quiet unit, at higher flow rates and efficiency levels than other units on the market.
I have chosen to include the acoustic cover by default, because although most people will consider it overkill, I take pride in doing really good work. It makes no sense to cut corners when you are going to be using something for 50-100 years.
One thing to note about these decentralized/ductless units is that previous generations had a minor problem with the quality of the noise, not just the dBa rating. They had reversible fans, which have to of course reach zero speed and therefore zero noise level every time the airflow reverses direction. This pulsation has been identified as annoying. The TW4 solves this from two directions: one, the noise is so low it is inaudible entirely if you are in a different area or the unit is turned down. Secondly, more critically, the reality is there are two fans in each module (half of a pair), and I have been able to arrange things such that one fan turns on slightly before the other one turns off. This actually increases average flow slightly because there is less "dead time" while airflow reverses, but more critically, it makes the noise apparently continuous because the dip is mostly gone. A future firmware update with a more sophisticated system is expected to practically eliminate the variation in noise level/character, making the device appear continuous in sound, which we can tune out quite effectively, even while falling asleep.
This low noise and continuous nature of the noise makes it so it can even be used reasonably in a bedroom.
No, it is designed and tested to work all year round, including during Canadian weather at -30 degrees C. This is a serious leg up over other units of various kinds.
Frost up and condensation issues don't generally occur with the TW4, even without sorbent, because the water that gets deposited by air on the way out evaporates or even sublimates again into air on the way in. It has been tested at -25. At such temperatures the pressure bias is important and may need slight adjustment, to avoid frost getting on the tips of the blades and causing a slight ticking noise (jamming does not occur). The factory default should be fine, but this is one thing that is still being sorted during beta.
The best defence against frost up is very high efficiency. The TW4 is extremely efficient, and gets more efficient at lower flow rates.
The heat exchagner will be available with Silica gel or without. The silica gel coated one has an even larger margin against frost getting on the tips of the fan blades, because it has a higher latent efficiency.
In the future, yes, but this is not part of the firmware right now. The electronics can be waterproofed during manufacture, to resist the water spray encountered in bathroom air. Waterproofing is accomplished mainly with dielectric grease made for marine use (this is a special type of material used for waterproofing electronics on ships).
Additive manufacturing has been proven to be capable of producing quality, reliable components.
All units are extensively tested before being released from Quality Control. The sub-modules we use are standard, and used in many other quality products. The power supplies, for instance, are made by a reputable manufacturer and are CE marked and UL listed. We get quality components from respectable manufacturers, assemble them with custom mechanical and acoustic components, then test and quality assure the result. The system runs on low voltage DC, is low power and inherently safe.
I believe that quality engineering and manufacturing is essential.
The warranty helps, our warranty is actually much longer than the industry standard for this type of machine.
CSA and CE mark are still in the works. Unfortunately a lot of stuff from overseas (i.e. aliexpress) does not have these things and people rarely seem to notice or care, but it would be good to get that done before the Beta phase is complete.
We present a part of a new paradigm for getting good things done, in the near and long term, locally and globally, and invite you to be a part of it.
The OpenERV is an Energy Recovery Ventilator, which means it recovers at least some of the water vapor. This is better than an HRV, which does not recover water vapor, but is otherwise the same.
This recovery of water vapor is valuable because if water vapor escapes from your home unduly, it leads to an uncomfortably low humidity in winter. Secondly, when the humidity in a home is reduced, the evaporation of water is accelerated in the home, absorbing energy wherever that happens.
During summer, the ingress of water vapor is also impeded, keeping humidity in the home lower than an HRV.
Sometimes you don't want to recover water vapor, but would prefer to flush it out. For this reason, the ERV is available with and without the material which increases water vapor transfer, but under the right conditions, both types will transfer water vapor, the uncoated one to a lesser degree than the coated one. Vapor condenses on the way out and then evaporates again, with the uncoated type.
There is a bug mesh that clamps on under the storm valve add on module. You can clamp any sheet of bug mesh under the clamp. However you can filter out no-see-ums, blackflies, or whatever you wish. Mosquitos can't get past the fan and regenerator anyway. Most people won't need the bug mesh, but it's an option if you do. The HEPA filter of course also prevents bugs from getting in. You can also just cut a piece of bug filter mesh and just tape it on. The bug mesh kit requires the outdoor plate and hood.
The machine is designed to be able to handle this in general, but I have started using a polymer called MPLA for many components, to focus on cold weather roll out first. Thus I cannot recommend that it be employed in very hot areas like the Nevada desert right now. The main concern is that the combination of direct sunlight and very high temperatures could lead to damage or malfunction. It just needs testing before I can promise it will be a trouble free experience.
The maintenance cost per year plus the yearly payments on the capital cost amortized over the life of the unit is far superior to a conventional unit. See spreadsheet on the learn more page for details. Unfortunately, the figures for efficiency and noise are generally not reliable in the marketing materials for these kinds of units, making comparisons difficult.
The unit is designed to last much longer than any other on the market. A conventionally made unit can't reasonably be expected to last much longer than it's warranty period. If it does, it's a fluke and, arguably a mistake of the producing company, because they forfeit that ability to charge for labor, parts, or a new unit.
When the fan will wear out is an important question, and how to get another one. No axial fan lasts forever in constant operation, that is beyond current cost-effective technology for this application. Many of us think of the fans in our forced air heating systems or similar, which do last for a very long time before the bearings wear out. Such fans rotate at relatively low speed and the bearings exist at low temperatures. The type of fan used in a ductless energy recovery ventilator is fundamentally different, they operate at higher RPM and motor temperatures and thus don't last as long. Bearing life in this context is primarily determined by grease life, which is affected greatly by temperature.
Mean Time Before Failure ratings for fans are statistical measures that do not represent average usable life span, for several reasons such as the exclusion of some causes of failure, and unrealistic criteria for failure. Failure due to normal wear is actually excluded as a "valid" cause by most companies. Secondly, the test is only done for a limited time period. Analogously, if the same is applied to a human, the MTBF of a 25 year old human comes out as about 800 years.
In engineering practice, we need to actually know how long things will last, not marketing nonsense, and we use measures like L10 for fan life. This indicates how long 90% of the fans you buy will last. The fans I'm currently using in the TW4 have an L10 of 40,000 hours. That's 4.56 years, at full power,100% constant operation, indoors (so higher temperature), and 90% are still functional. So I generally think of 7-12 years of elapsed time in our actual context as a fair figure.
With an open source unit made from standard parts, the incentives are inherently aligned differently. Because you don't depend on us for parts or labor, there is no incentive to try to rig the design of the unit to drive you back to us, before or after the warranty period expires.
Taken together, the incentives are better aligned this way, and that paves the way for a more promising financial proposition - when the cost of a unit can be amortized over 40 years instead of 5, that's a factor of 8 more service life - for the same capital cost, and for lower running costs, because maintenance is less costly.
Open source is no mere philosophical proposition, it has very tangible benefits.
The unit is actually designed so the pipe can just stick out the exterior wall. This is a unique design feature among ductless ERV units as far as I know. The seal around the wall and pipe is accomplished with caulk and foam and/or a rubber airtight grommet. The exterior components actually clamp to the pipe with a radial clamp, tightened by a screw. This also makes (rarely required) types of maintenance and repair much more practical.
Ventilator units of any type, whether they have heat recovery or not, can be tested to ascertain the amount of outdoor sound that they allow through/block. This matters in a city because there are traffic noises and things that can be annoying.
Clearly, most units do not provide the same level of sound attenuation as a solid wall. But they do provide more than an open window.
It is measured in Decibels, A-weighted, usually (dBa). The test is that they install the unit in a wall in a laboratory, and then produce sound on one side of the wall, then measure the reduction in sound volume that arises after the sound waves pass through the unit. So there is one microphone on the outside, one on the inside, and you measure the difference between the sound levels that they detect. This process is repeated at a range of frequencies, and then an algorithm is used to summarize the data points into a single number.
The higher the number, the better.
It's helpful to think in terms of earplugs. Those yellow or orange foam earplugs give a sound reduction of about 30 decibels, if you use them right. The so called NRR, the Noise Reduction Rating, is included with the earplugs and should be stated on the package. So 30 decibels is a pretty large reduction in sound level.
A typical ventilator of this type gives about 30 decibels of sound reduction, in reality. They come with various accessories sometimes to improve this figure. Usually the accessories are in the form of indoor or outdoor hoods with a small amount of sound absorbing foam or fiberglass. They don't actually help that much.
The TW4 already includes a fiberglass panel build into the interior cover, by default, no extra charge. This reduces the sound from the fan, and also helps significantly with traffic noise. It's possible you could put some sound absorbing felt, foam or fiberglass (felt is the best for sound, but it could get wet and go moldy, the fiberglass is the next best, foam is not so hot unless it's memory foam/viscoelastic foam like earplugs are) under the outer hood, but there isn't really space if you also use the storm valve.
The storm valve blocks a good bit of noise when it is closed, too. The storm valve closes automatically when the unit is turned off.
Low frequency noises are particularly hard to block. The reality is that any ventilator will allow a small amount of noise in. However it is far less than a window that is open a even small crack. Even a closed window doesn't usually block more than about 25 dBa. That allows you to see some idea of what it will be like.
It helps a lot to install it in a wall that does not face the street.
The heavy duty attenuator blocks even more noise than the sound splitter.
I have not tested things in a laboratory yet, that's not going to happen any time soon, but the heavy duty attenuator appears to add about 22 dBa of noise blocking so it should be about 50 total, which is pretty good, that's comparable to a wall.
This is an important question. They are expected to last about 7-12 years before the bearings need to be replaced, however the bearings can be replaced without replacing the rest of the fan. Bearing life in this kind of context is mainly a function of temperature and grease life, not number of rotations, because there is almost no load on the bearings.
The general plan is to include a set of replacement bearings and a spare circlip and instructions on how to replace the bearings, which is even easier than getting a new fan when the time comes. They are also standard bearings that are really easy to get ahold of. There is more info in the manual on sourcing them, you just search the right keywords etc and they are easy to find.
It's important to understand that all conventionally made appliances have a serious problem: when the parts break, best case scenario you can get a replacement from the original manufacturer. Our culture has come to accept this without really understanding it.
I've noticed people looking at the TW4 will ask this question, but somehow believe that a conventionally made device will last forever. This is not correct. The technology is the same in both cases. The units like the Blauberg Vento (which is my favorite conventional unit) do not use ceramic bearings that last forever or something. Magnetic, hydraulic bearings etc. do not last longer than good ball bearings in this context. People think because the parts don't touch they will last forever, but that is not correct, because dust etc always gets in, the grease eventually is no good due to escape of volatile components, etc. Also the parts in a ball bearing are also constantly separated by a fluid film, during actual operation.
We want to go beyond that and use standard parts that are available from a large number of manufacturers, so you don't get trapped or extorted or unable to get parts, even in the distant future.
In fact, when I inquired as a homeowner, one manufacturer of ductless ERV units told me they absolutely do not sell replacement fans. At all. To anyone. Only complete units. So when your fan fails, you can't get a new one. And I looked for other fans that could replace it, and there appear to be none available to consumers.
Replacing the bearings in the TW4 is not too hard, I have made a video on the subject. Getting the circlip off is the only hard part, you don't even need a screwdriver otherwise. The circlip is the only thing holding the fan together. The bearings just fall right out once you take it off, basically. There is a method with two torx screwdrivers that works well, and the circlips are also common standard parts so we can include extras, and you can get more if you break it or lose it while taking it off or whatever.
Because the system is open source, if you can't get an exact replacement fan you can do a fairly respectable job of putting a different one in. Thus, any standard waterproof 120 mm PC cooling fan of good quality can work, you can trim the corners of the cases off with a saw and they fit in the pipe. It's even possible to waterproof a fan that is not waterproof.
I believe that with waterproofing the electronics using dielectric grease, a method I developed for use when employing it as a bathroom fan, that this could work ok, actually. The fans are IP67 so actually they are completely potted in epoxy, except the bearings. The bearings might rust. They have metal seals but I don't think they are stainless steel, you could replace them with stainless or ceramic bearings. Even 304 stainless steal can corrode over time when exposed to NaCl, I am not well acquainted with how things go near an ocean but I have to say it sounds like a pretty harsh environment. I cannot warranty the units for such a scenario without testing, which is not going to happen any time soon. I can only hope that people who live in such areas have ways of making regular stuff work, because it must be pretty hard to find stuff that is made for that scenario, and are aware of what materials etc are ok and which are not.