Entropy Wheel Heat Recovery Ventilator

Looks like the project has been getting some attention lately, more by others than by me... I've been busy with this insane FPGA contract project as well as job hunting/job finding/prepping to move to the new job 2300 miles away. I probably won't have the best environment in Arizona to test the condensation mitigation/monitoring I had planned.

It did get me thinking about a solar-powered "water from air" device to see how much water could be extracted from dry air on a 250W solar power budget. A couple of years ago I worked briefly on an experimental, industrial-sized WFA meant to solve logistics problems in Iraq. Dessicants and heating/cooling loop could produce several gallons of water per gallon of fuel, an advantage for locations where liquid water didn't exist and the fewer truck trips the better. Wouldn't ZERO gallons of fuel be perfect? I keep thinking of projects I don't have the time to complete.

Appears I might have the unevenness of the end faces dealt with, picked up a big honking belt sander over the weekend and that appears to do the trick. It does tend to melt the plastic a bit so I need a wire brush to reopen all the tubules. I hope it's that easy. I was alternatively thinking of using a power door planer to finish the ends, shouldn't have the melty bits that way.

I was thinking about where to get a brush assembly like they use on some 19" rack cable entry holes to block air but allow inserting cables with ease... but those were conceptually right but not quite the size needed. I had bought a $1 broom at the dollar store I was planning to cut off bristles and "glue" using 3D printer "ABS glue" acetone/plastic mix to bind the bristles, but I do imagine that could have become a mess, and the bristles would need to be frayed to make them less stiff.... then I notice the door where I work....

So now I have a source for long, fairly cheap brushes that should be great. At about $3.35 per foot, available as 3, 4, or 6ft lengths from Zoro.com, a couple of brands: "Value Brand" A626A-36 for 3ft, $10.01, Zoro P/N G0373642. Another option there is PEMCO 45061CNB36, similarly priced. I want to see some CAD drawings for this to be sure I know what I'm getting, but they look promising.

The brush fits in the midline of the heat exchanger, separating the top (outdoor air) from the bottom (indoor air exiting) and preventing most of the leakage from happening. The brush allows the cylinder to spin freely. The aluminum frame holding it should allow slightly bending also, as the ideal shape allows for the air to "purge" before the blower reverses the direction of the air as the Coroplast tubules cross over from hot to cold side.

Still looking for good ideas on brushless motors, inexpensive and 10000 hour bearing life. I'm going to prototype with bathroom vent fans, but those are constant speed. That leaves the cylinder turning motor to wrap up. I'm going to prototype that with a stepper motor, but I imagine that will be too loud for normal used. I've seen motors & drivers optimized for quiet mode using motors with less cogging at the expensive of max torque. The drum needs to spin in the 20-60 RPM range. Air velocity will interplay with that.

On another note, I'm thinking of creating another, similar project for clothes dryers. A mechanical/manufacturing engineer I used to work with talked about a venting scheme he had for his dryer (I'm thinking his was gas) that draws in outdoor air to the laundry to prevent negatively pressurizing your house and creating cold drafts. Ever consider the negative pressure issue? For electric dryers, you don't get to burn the incoming air so you need a heat exchanger to recover exhaust heat to make a real positive impact on the electric bill. Moisture condensation becomes the key to the design, so to solve this I'm thinking a cross-flow HRV pitched at 45 degrees will allow a gravity-flow condensation path, mounted above the washer drain for a pump-less solution. Auto turn-on based on temperature in the exhaust manifold. One ingress booster fan since the exhaust has a powerful blower already. Intake air could feast on preheated attic air (my laundry is upstairs, realizing most are not, YMMV) assuming that isn't a major building code problem (yes, filtration will be considered).

Of course, you COULD avoid the hassle by buying one of the new ventless, condensing dryers... essentially my goal here is to make a condensing add-on for existing dryers. Admittedly, the safety considerations here are considerable and static electricity will be present in copious amounts....

The clips are definitely the way to go here... I have finally completed the original glued wheel (just the wheel, not the whole prototype) and it's quite the chunk of plastic, 10" cylinder x 14" long, almost two full 4x8ft sheets. The glue is messy (3M High Strength #90) and slow owing to the tack time between layers. Took 3 cans and that means $40-45 just for adhesive. Not to mention the special cyanocrylate I bought that was even more expensive, and have yet to use. Disadvantages beyond cost is that it leaves "boogers" blocking some tubes owing to the squeeze-out, and layers vary in length a bit, which is bad for air sealing. I need to cut off a tenth of an inch or so from either end to clean the boogers & even it up. HOW to do that is going to be a trick as it needs to be neat and perpendicular and my bandsaw doesn't have a 10" capacity. The belt sander might work to even things up, but I know it will friction-melt and perhaps clog openings more. The there's the issue of mounting it in a balanced fashion so it spins easily with a small motor.

I'm still trying to work out the best type of motor. Stepper-sized BLDC's I have seen run very expensive, like $80. I might initially use a stepper but I expect that to be much too loud. Not sure a hobby BLDC would be much better as they "cog" also and typically use non-sinusoidal coil excitation, unlike brushless servo motors that run with sinewave drive instead. If you have sourcing suggestions, please message me!

See my Thing on Thingverse with the glue-less clip-together assembly parts. http://www.thingiverse.com/thing:485101

And also a cutting tool for rapid, safe cuts of Coroplast! http://www.thingiverse.com/thing:482141

Still need  a high-reliability BLDC drive motor-- have a little brush motor for now, needs a frame to stand on and also to suppose the other parts. Oh yeah, and software/drive electronics. This is going to take a while, but you'll have about 72 hours of 3D printing to do to make all the parts anyway!

My first attempt at using OpenSCAD to design a "comb" to hold the coroplast layers together. Looks very promising. First prototype is a bit tight, accumulating a bit of tension over the 7 layers I tried. I still have to devise how to tie these together with a printable frame and tie to a central bearing - might leave that to SketchUp.

This project was something I began working on before the Hackaday Prize was announced, and sensors and connectivity were always part of the plan. At the very least, temperature and relative humidity need to be measured as condensation can be a problem without a drain to send it to. In certain circumstances, shutting down fresh air input to function as exhaust only would allow it to function as a ventilator but drawing fresh air into the home through the normal leakage points. For R&D purposes, the temperature sensors allow me to estimate efficiency and understand how things are working. Pressure sensors can detect other situations: filter blockage and fan speed compensation (if I find a BLDC fan solution) and also wind/venturi effect compensation -- or weather prediction via barometric pressure.

Other thoughts would be passive IR motion detection/presence sensing and perhaps LED lighting to go with it. Options of nightlight level or full 10W LED lamp at about 800 lumens.

Connectivity to the unit itself will be via low-cost, long-range, low bit rate Nordic Semiconductor 2.4GHz NRF24L01+ modules. The Cypress PSOC 4 I an eyeing as the embedded processor has just 32KB of program FLASH and 4KB RAM, so Wifi would be hard pressed to fit and overkill for the application. Since I have plans for many nodes constrained by minimum cost & minimum power, collecting sensor data I think would be best served by a bridge device with Ethernet or Wifi connectivity, be it Arduino, Raspberry Pi, BeagleBone Black, etc. Optimally, integration into a whole-house automation & alarm system would be the final goal.

With the sensor package and wide range of potential usage preferences, the short packet exchanged would carry sensor readings from the heat exchanger and also carry configuration information to the embedded controller. Example configuration would specify things like motion activation activities (fan on, light/nightlight on), timed auto turn-off, scheduled activation, filter alarms (timed or back pressure-based), and whatever else I can think of. Spooky noises mode Easter egg?

The Panasonic FV-04VE1 is the benchmark I am competing with. These are commercially available via Amazon and other distributors. The Panasonic design is based on a "capillary core" heat exchanger. A 40CFM unit is just under $350 and consumes 23W. I believe the Panasonic makes use of BLDC motor drives for quiet, reliable operation - something I'd also like to incorporate, for filter blockage compensation/pressure changes. Accessories for the Panasonic, such as Y-fitting ducts to retrofit existing exhaust ports, could be used in my project. They combine incoming and outgoing air in a single duct in such an application - side benefit of being a counter-flow heat exchanger by coincidence. The Panasonic units are made for bath fan retrofit/new installs, as opposed to whole-house ERV's. Given the price, it might take a while to recoup costs, if ever... not that my R&D unit would be any cheaper!

Also on the drawing board is a 3D-printable razor guide to easily and SAFELY perform all the cuts needed. A guide insert to guide the blade dead-center of the coroplast "tubules" and maybe a lower piece to cover the exposed blade. This way you can insert the guide in right tubule and make a perfect slice in seconds. Thinking of devising a "guide comb" to help identify the tubule positions to begin cutting, and a guide to assist in slicing the huge 4x8 ft panels down to 16" strips. One must pay attention to the orientation of the tubules when cutting.... 

Some lessons learned with coroplast. This stuff prefers mechanical binding methods, as very few glues will stick to this "low surface energy" material. 3M Hi-Strength 90 is ok, but you will need at least 2 cans @ $13 each to make the wheel, possibly more. I also have a specialty cyanocrylate type adhesive, Tech Bond BP Blue. This stuff is expensive, but should be much less messy and cause less thickness buildup, as the 3M #90 goes on rather like wood panel glue or caulk. 

So a better choice might be a mechanical clip with barbs to insert in the ends. This can also serve to bind the pieces to the center hub and support ball bearings. Working on a design, trying to learn Open SCAD.