A Portable Windmill Experiment (Ongoing)

[Updated 02April09]

[Indian Territory is returning to temperate conditions for a few weeks, so I’ll be able to continue the windmill work: hooking up a genny to the 75 watt S-rotor, refurbing the break-down HAWT, etc.  Lots of stuff’s been going on lately, sorry].

What the… a portable windmill?” you may well wonder.
Well, my thinking is that I can’t make a very effective photovoltaic panel in my workroom, windpower has a lower cost per peak watt than solar panels, and I know I can make a windmill (see the Otherpower website).
The projected application is for emergency power (we had this ice-storm, see, and we were without power for 10 days, yeccch!) in a knock-down form that is easily stored and transported. I figured portability and cheapness would be ideal for disaster recovery and remote operations (I know this guy that does the ham radio emergency communications thang) and, hey, “infantry portable/field maintainable by indigenous personnel” has gotta be a plus for just about anything, right?

Besides, windmills are cool, especially if they are funkily home-brewed.

While perusing homebrew windpower efforts on the web, I came across an interesting approach by “Cowboywindmillbuilder” in his Instructable . Some of the key inspirations contained in his design were:

  • Cheapness and simplicity
  • Semi-rigid, fabric covered blades
  • Downwind design
  • Using a 90 degree drill adapter to bring the torque to ground level

Cheapness, simplicity and portability would be impossible in a high-tech, high-efficiency system, so in my homebrew system moderate efficiency and simple construction (and did I mention cheapness?) were my goals. [Cheapness and simplicity are, of course, requirements of anything (and anyone) that I do].

One strength of a downwind style windmill is eliminating the need for a tail, that part acting like a weather-vane to keep the prop turned to the wind. The drag of the prop, mounted on the “down wind” side of the pivot is enough to keep the blades facing the wind. Some efficiency is lost from turbulence around the mast, but that’s just pocket change in the bottom line. Lower weight and simplicity will make up for this.

I’m breaking this project down into several sub-tasks:

  1. Hub/Blade assembly
  2. Alternator
    • Coils
    • Wiring scheme

Making the Hub (Method 1)

I am fortunate to work at a company with a large variety of shapes and sizes of scrap aluminum and steel left over from the water-jet machines. Many of these pieces are mild steel and circular. I found a few 8- 3/16th circles of 3/16th steel that seemed perfect for my project. I marked the center (using basic geometry constructs) then used a protractor to mark the circle off in 12 even angular divisions, like a clock face.

To hold the 1/2″ plastic conduits for the leading edge supports I purchased a set of three 3/4″ set screw couplings for metallic conduit and attached them at 4 hour intervals around the upwind side of the hub. This scheme should make balancing the blade easier, just loosen two screws and fiddle around with the protruding length.

To hold the fiberglass rods for the trailing edge supports I purchased 3 steel 1 x 3/8″ pipe nipples. I used a flux-wire welder to tack these on the downwind side of the hub at intervals 1 hour behind the conduit couplings. I added a little bit of weld to the open inside ends of the nipples to act as a stop for the fiberglass rods.

The hub, all welded up

The hub, all welded up (by the worst welder in the world)

Making the Blades

For small wind turbines, there is a delicate balancing act between the lift and drag of the blades. For all practical purposes, 3 blades is the cannonical sweet spot, so I just went with that. The reason you see so many blades on the old farm windmills is for torque, since these were used to pump water, and that needs a lot of torque to just get started. To generate electricity (something I hoped to achieve), you need some speed, which the farm windmills don’t do so well, due to the drag of all them blades.

Since my windmill isn’t ever going be placed more than 10-20 feet up (I’m living just outside Downtown on a tiny plot of land), I decided that low mass was a requirement: if the spinning part didn’t weigh much, chances are it wouldn’t be able to deliver a head-ectomy if I stumbled into it.

For the original blade “frame”, I’ve elected to use 1/2″ ID PVC electrical conduit for the leading edge and fiberglass rods (originally driveway markers) for the trailing edge. The more flexible trailing edge supports allow for some self-regulation of speed without adding braking mechanisms: as the wind speed increases the trailing edges bend with the wind to reduce the prop speed passively. Since this is a downwind design, there’s no problem with the blades bending enough to contact the mast.

The plastic conduit is rated for above or below ground use, so it will be more resistant to UV aging than white PVC water pipe. It is also flexible, but larger diameter and less flexible than the trailing edges. The larger diameter of the leading edges vs. trailing edges will (I hope) improve the airfoil (once the frame is covered with a tight plastic skin).

Blade Frames

I cut 3-1/2 foot lengths of 1/2inch conduit and drilled a 5/16th inch hole about 1 inch (one hammer head width) in from the outside end. The trailing edges uses the cheap fiberglass rods ( found next to mailboxes at the local hardware store) that measure almost 4 feet long x about 5/16″ in diameter.

With the 1/2″ conduit and the fiberglass rods mounted in the hub, carefully bow the fiberglass rod and stick into the 5/16th hole in the 1/2 inch conduit. This will give you a nice curved frame for the blades with a low angle of attack (pitch).

Blade Covers

Once you have all your blade frames bowed up, it’s time to cover them. Continuing my cheapness theme, I made my blade covers using black Visqueen plastic (LDPE sheeting, I’m guessing 4-6 mil) that I had laying around. Clear would have been cheerier, but I figured the black would add more UV resistance. It is thick, strong and bonds well using a “Eurosealer”, no glue and no tape required. The Eurosealer was picked up at the local drug-store’s “As Seen on TV” aisle a few years back. The gadget was marketed for resealing plastic food bags, but it’s seen more use making hot-air blimps and oddball projects.

The Infamous Eurosealer

The Infamous Eurosealer

Note: there are battery only models out there, but mine also came with an AC adapter. I’m not sure how long a set of batteries would last making long seams like this project requires.

I put the hub/frame assembly flat on the ground, then draped the Visqueen over each blade frame in turn. Folding the plastic over the 1/2″ conduit, I sealed it around the curved trailing edge using the Eurosealer. I Placed a few “tacks” near the narrow ends and around the center, pulling the plastic as tight as I could easily, then went back and did as continuous a line as I could around the curved trailing edge.

Once the blades were covered (took me a half-hour or so), I tightened up the covers by loosening the set-screws on the couplers holding the 1/2″ conduit to the hub and pushed the conduit into the center of hub just a little further. This bowed out the trailing-edge fiberglass rod to hold the covers taut. (Yes, having an extra pair of hands would be helpful at this stage, since the bowed rods can be kinda strong, but this is not required).
Photos of the process are coming just as soon as I get around to repairing the one blade covering that the dog managed to gouge while the thing was sitting on my porch.

Making the Hub (Method 2)

Cutting PVC lengthwise can produce very usable blades as well. I’ve recently come across some 4 and 6 inch diameter pipes salvaged from a construction site (legitimate scrounged), ann will look at a multiple rotor design along the lines of the “Sky Serpent” method (which I find really impressive).

Measuring Performance of Hub/Blades

Before making your alternator it is helpful to have some data about how the spinning-around-thing performs.
You can spend as much as you want to on a wind data logging unit… or you can spend $10-20 on a bike computer that can tell you a lot that you need to know. For instance, your windmill’s revolutions per minute.

I picked up the cheapest bike computer I could find (a Schwinn). It consisted of a largish digital watch looking thing to mount on the handlebars, a cable/magnetic sensor assembly, and a magnet to clamp to a spoke of the wheel. Instead of a bicycle wheel, attach the magnet (or a magnet, anyway) to the hub of the windmill, and attach the sensor to a part of the windmill mount close enough to pick up the passage of the magnet (it should swivel with the hub/blade assembly). Get the bike computer into set-up mode (I think I had to hold down both buttons for three seconds) and get to where you set your wheel circumference. If you set the wheel to be 1.666 meters in diameter, the KPH readout will equate to RPMs if you discount the decimal point ($10-12 test equipment is GOOD).

The Alternator

For this, my first attempt at homebrew power generation, I took a cue from the gang at the Otherpower website and the very informative How-To file and went with a three-phase dual rotor axial flux design. Unable to decide on which wiring scheme is optimal (star vs. delta) I decided to dress all the coil leads out so that I could switch between them while experimenting. The initial wirir=ng was in ⭐ configuration.

Since my generator is smaller than any I’ve read about (about 4 inches in diameter, (perhaps too small viz the relationship between disk diameter, RPMs, angular velocity and power output) with only six coils and nine sets of magnets mounted to the rotor disks (a very minimalist design, setting me back $10 in 3/4 x 1/8inch circular rare-earth magnets from the hobby store), results would be uncertain. This is especially true since I made the rotor disks first, and then started doing the math on the coils sizes required, so, anyway, here’s the warts and all situation report:

I’d salvaged two steel disks from the recycle barrel at work (call them 3.9 inch disks). These were used to make the the rotor (magnet-bearing) disks.

I gave them a good de-greasing and washing (and a mild pickling with dilute muriatic acid), smoothed the burrs, and drilled holes in their center points, then marked them up like a clock-faces, with radial marks (center to edge) at 30 degree increments.

I positioned nine 3/4″x1/8″ Neo magnets (from the craft store) about 1/4″ in (one butane lighter’s width) from the edge on each of the two steel disks, let some super-glue ‘wick’ under them, then made a masking tape wall around the edge of the disks, just deep enough to cover the faces of the magnets. I filled the depth of this wall (to the rim) with polyester resin. This encapsulated the magnets, for better or for worse.
That much worked fine.

Arranging the coils inside a coffee-can lid mold didn’t go as well; I think it was the process of placing the weighted lid over the top that disturbed their alignment that threw the coils off-center. Note that the pink stuff around the bolt in the center of the (left) stator disk in a latex masking agent used to keep the plastic coatings from fouling the threads. You can just make out the dental floss I used to tie the coils together in the stator disk before embedding them in polyester resin.

The disk on the left is one rotor, with magnets embedded in polyester resin. Another identical disk is under the stator disk on the right. All those green enameled wired coming from the right-hand disk are from the coils of the stator (connected in “star” configuration, another diagram coming soon).

The scarring on the top rotor disk (left) are from magnetic fragments attracted to the magnets (the silver circles embedded in polyester resin). The stator disk (with the coils) sports a fresh coat of urethane conformal coating followed by a thin coating of silicone conformal coating, to protect the coil wires that had been abraded in early testing. I also increased the air-gap about another .060 inch with an extra washer.

At 1000 RPMs I saw about 12 volts peak-to-peak per phase, which should result in 15 volts or so after rectification with Schottky diode bridges and an output cap on each bridge section. Not bad for a 4 inch diameter brushless alternator weighing less than five pounds built with craft-store magnets. I used no bearings in this early design, as it was a prototype, soon to be reconfigured for extended used.

Since I work with BFMs (big freaking magnets) 8-5 weekdays, I was lucky enough to have a gauss meter/teslometer at my disposal, whereas most home ‘mill builders don’t. If I read the crazy dingus right, with my rotor disks about 1/4 inch apart I was reading 3-4 thousand Gauss. So, that’s the number I’ve plugged into a spreadsheet (of unknown lineage), for calculating the number of coils to use in a given alternator configuration (as well as this one, which could be of use to those ‘rolling your own’). I am assuming these to be simplistic and rough approximations… but they give a starting point, use at your own risk (and I’m new at Google docs, gimme a clue if you got one to share).

[UPDATE 2 FEB2009:  I’ve been updating my spreadsheet lately.  Bear in mind that the strength of the magnets are of paramount importance, followed by the air gap, then thickness of magnets is my perceived order of precedence in making an effective alternator.  Get the strongest magnets you can, the thinnest coils you can (like single-layer) and spin them as fast as you can…. all else is commentary, except to say if you can’t get a full T (Tesla) in the middle of your air-gap (where your coil is) you will have a wimpy thang instead of an alternator, okay?].

16Nov08 Note: the spreadsheet has evolved (not posted yet) but I’ve added an S-rotor calc page, and linked a lot of the variables together from different pages to make it easier to optimize a design for wire AWG, magnet configuration and strength, RPM/windspeed relationships (theoretically tied in with tip-speed ratios and rotor coefficients). It isn’t ready for publication yet, but it should be soon, and I hope it is some help to others in the homebrew community. Chasing all these variables around on separate sheets was like chasing rabbits and starting to make my brain hurt…).

Rethinking the Alternator

Update 16Nov08: Well, the little alternator isn’t performing as well as it needs to for my arcane purposes. I’m only getting 10 volts from it at 1000 RPM, far under expectations. Along with doing some powerful cypherin’ for a new stator/rotor combination (remember that PCB coil thing I was yakking about in another post?), I’m going to remove the excess thickness of plastic ‘potting’ over the rotors and try to tighten up the air-gap, at least until I test to destruction (and can then get on with doing it the right way the second time).

My .75x.125 inch craft-store magnets appeared (after calculation) to be at least N42 NeoFeB disks, which ain’t so bad, but the flux goes to crap at much over .1 inch separation, and I guessing that getting them just a tiny bit closer together could see useful output.

Just noticed that K&J Magnetics has a special on their DX06, a nice little N42, 1 inch round by .375 thick. I’m trying to order a 50-pack for a project at work, and see how they’ll do (Bog willing and the PO goes through).

Some “Real World” Blade/Hub Testing

In my latest test attempt to gather RPM and output data to my windmill (more photos, prose coming on this), I managed to hoist the knock-down bladed windmill up on the 10-foot 1.25 inch length of conduit, only to have the windmill head fall off and land on my foot. Yea, verily and for sooth it did hurt savagely, falling as it did with the rigid conduit straight down. (Note to self, find a way to use a light weight aluminum bade-hub in future designs).

“Oh, fudge” I said. This little piggy got an ice-pack.

Understandably, I’ve backed off just a bit since then, sorry… but more information is coming, as soon as I beef up the mounting method used for the blade/hub assembly (I got in a hurry last time, and it hurt lots).

19Nov08 Update: Have laid in some new materials to correct prior shortcomings in the hub mounting. More pics and prose as they develop.

Some “Real World” S-Rotor Testing

It has to be said, I have a one story house, between a pair of two-story houses (with big trees all around); the worst case scenario for wind-power testing close to ground level. The “good” news is that tonight there are severe thunderstorms moving through the area, with enough wind to really spin the two 5-gallon bucket S-rotor at the top of a 10-foot section of metal conduit I’ve got ‘posted’ in the back yard. I’ve go the bike-computer strapped to it, and will be publishing wind-speed and RPM data as soon as possible tomorrow.

The bottom plywood disk has an 11 inch zinc pulley screwed to it around the central 1/2 inch shaft, and I’ve obtained some angle-iron to mount an axial-flux generator (currently being scaled up) to this with a pulley ration of 11:1 3/4. For the belt I’ll be using the tubular neoprene seals from the bucket lids (it’s only garbage if you don’t use it).

Update 06Nov08: The little Savonius that I built using two split 5-gallon buckets (“Homer” buckets, from Home Depot [cough-sponsorship-cough]), and some 2-foot diameter plywood disks survived the “severe” weather last night. The whole assembly was mounted on a 10 foot vertical metal conduit on a flange bearing, with data collected by the Schwinn bicycle computer. With the lightning flashing last night, I was up on the pole trying to secure a magnet and pickup for data collection, St. Ben (Franklin) was telling me, “Dude, I got fried like this once. Get DOWN!”, but I hung with it. Here’s the data I got:

  1. Max RPM:247
  2. Average RPM:47

WeatherUnderground shows gusts up to 32MPH for that 24-hour period (I currently lack on-site wind-speed instrumentation. Plugging that data in to my calculations shows that one of two things happened:

  1. Either it gusted much higher than 32MPH, or
  2. My TSR is greater than 1.2, which is highly doubtful

My current working hypothesis is that I let St. Ben get me panicked, and I forgot to clear memory, leaving previously collected data stored in the bike computer to screw up my results.

Even so, dropping the two-bucket stack (twice) from 10 feet didn’t seem to phase it, so , inefficient or not, the Savonius is a rugged design. That all aside, the rotor did throw the speed sensing magnet sometime during the night. I’m applying more permanent mounting for the magnet (better than electrical tape, anyway, St. Ben was really trying to rush me at the time).

Update 16Nov08: After rearranging the magnet-sensor (and remembering to clear the memory) I re-hung the S-rotor for 23/75 hours starting 2PM 15Nov08. Today the bike computer showed:

  • Max RPM: 120
  • Average RPM: 40
  • 3.75 hours of travel (over 3KPH)

Weather Underground showed a gust of 33 MPH just as I was getting set up, so I’ll call that the 120 RPM maximum figure. These sound like more reasonable numbers for a TSR equivalent of between 0.5 and 0.6 (and 3.75 hours * 4.0 KPH = 15 KM correlates well with the odometer readings), so I’m calling the numbers at least not totally bogus.

19Nov08 Update: Figures from a 3-day trial are in line with last reading:

  • Max RPMs: 120
  • Average RPMs: 37

If not totally bogus, at least the figures are consistent with my preconceptions. Hopefully upgraded instrumentation will verify these figures.

21Nov08 Update: After a day of strong northerly winds with gusts to 34 MPH:

  • Max RPMs: 222
  • Avg RPM: 57

Have corrected the dimensions used in calculations. 1.6 ft wide, 1.84 feet high, 2 ft end-disks (1/2 inch plywood). Looks very close to a TSR of 00.52!

08Dec08 Update: It’s been a good and windy couple of days!

  • Max RPM: 242
  • Avg RPM: 49

(Sorry for the slow progress, pressure of business is keeping me in over-time mode).

Experimental Contagion

The bug is catching… my comrade that runs the water-jet machine is starting to work on his own design, so there should be some interesting synergy. He’s getting interested in the Lenz-type turbine as well. I’m still sticking with the S-rotor (just because of the low level of machining involved and building with mostly scrap materials/found item art aspect).



Tags: , , , , , ,

2 Responses to “A Portable Windmill Experiment (Ongoing)”

  1. kaysommers Says:

    hello!! thanks for the useful tip you left me! 🙂

    i typed some questions for you there as a reply, i’m not sure if you got it or not. (still figuring out WP being new)

    so if you’d kindly answer, i’d be so grateful. do you have any picture instruction on your tip? i don’t think i get it completely but i do have some ideas. wish i payed more attention on this subject in school back then :\

    thanks again!

  2. offlogic Says:

    Kay, I followed up in email.
    Good luck on your jewelry!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s