2001-12-09 Rev. 2002-06-15, 2003-03-16, 03-30, 2004-01-25,
2007-01-21, -02-24, 2008-01-09, -01-14, 2009-01-11 (edits) 2010-03-16, 2011-02-22,
[Search on date pattern to find latest changes, more than one may be found.]
Return to Sitemap
|Progressive Examples below|
Whirligigs (whirly jigs, etc.) are follies that sit out in the wind and go 'round, doing no useful work. A weather vane at least points into the wind. When I moved into my house, the cyclone fence on the lot line was decorated with small whirly jigs made by the man who had lived next door since 1935. Most of these were made of metal, in several cases being beer cans with fins cut out of them. He is now gone and the neighbors have built a high board fence, removing his bit, so I guess I am keeping up the tradition.
|This post was installed in the backyard so I can fool around with things moving up in the air. The center 4x4 post is 12' tall and the 4x4 supports go 42" into the ground. All are pressure treated wood, no concrete was used, and the center mast swings down between the two supports like a flag pole. The rods through the posts are 1/2" all thread rod with washers and nuts on each end. They are installed with a hammer and removed with a wrench turning the nut and pulling the rod out. After it is out 4", it can usually be pulled by hand, especially if the mast is rocked to shift the lineup of the holes. In the right hand photo, it is rigged with an 8' 1x2 with a lead weight at one end and a thin aluminum curved sheet on the other. It moved very irregularly and drove the birds crazy.|
|Now (2000) the post has a
welded triangle frame about 6' tall that I hope to sag glass to
install in the frame, clear first for testing, then colored. As
shown below, for the time being I have sewn up a ripstop nylon
panel to provide some color and a windmill blade. Not especially
clear in the picture is that the two narrow point tips are not level.
They are almost over each other when this is laying on the
ground, so the panel is at an angle to the bearing and the whole
rotates in the wind.
|WHIRLY JIG -
Now the post has a welded triangle frame about 6' tall that I am
sagging glass to install in the frame, clear first for testing
framing, then colored. As shown, a ripstop nylon cloth is laced
on the bent end which acts like the glass to spin the frame. This
was also used to make a clay mold for slumping the glass. The
small diagonal rods are framing for the first pieces of glass.
The proportions of the small triangles and their distance from
center are such that the weight of the glass installed should
balance the glass in the twisted area. [It did when finally done,
30 months later.]
The larger tube from the base of the triangle has a lead slug in it to adjust balance. Bearings are a pair of wheel barrow replacement bearings, 1 3/8" x 1/2", which fit very nicely in the 1 1/2" steel tubing mount. Updates will be posted they occur. 1999-06-20
|Getting glass into this thing has been one frustrating project. It is nice having it swinging around in the yard, but attempts to sag glass (not a high priority project) have not gone well. I made a clay form and sagged some clear onto it, but the form shrank on drying and the glass was too small and both were eventually cracked before good lessons could be gained. I have tried about 3-4 times to make test pieces, but not pursued them. 2001-09-21|
|The blue cloth was tied in
place and ropes tightened it, like a canvas cover. The yellow
cloth was sewn to shape and held at the corners with carpet
thread, hand sewn around the metal frame. Notice the lump above
the yellow - this is a small c-clamp moved around to balance the
frame without glass so it would move - the lead could not be
moved far enough to balance. In the picture with the blue cloth,
the clamp is on the lower rail, about 1/4 of the way - the cloth/rope
was heavier, so the weight is closer to center.
Measuring and transferring the warped form in order to cut flat glass that will fit: Measure the 4 sides and one (or both) diagonal(s) and use the diagonals to swing arcs to construct the relationships of the sides.
|This is the frame with glass finally installed
in it. 2001-12-09 Note that with the glass installed there is no
c-clamp for balance. The point end glass was sagged on to a thin
clay form (below) which was made on a thick
clay ramp sculpted to match the shape of the steel, made somewhat
oversize so shrinkage would be less of a problem. The
installation was done with 100% silicon sealant, which tho clear,
shows up as a frosty clear. I want to try some E-6000. The metal
frame was wire brushed and then wiped down with mineral spirits
paint thinner before gluing. One of the smaller triangles was
cut a shade small from scrap so I put a bead of silicon sealant
in several locations around the frame and let it set, so there
was a place to rest the glass. Then I placed the glass and added
much more sealant to hold it. I will watch the effect of the
wind, etc. on the glass and sealant before I sag colored glass
for the tip.
Click here for an animation .gif of the whirly jig moving. First attempt at automatic photographing with new camera and use of animation software. At 56K this will take about 1.5-2 minutes to load - to exit use back arrow, backspace or alt-left arrow.
2003-03-16 Over the last week, I sagged a blue replacement for the damaged glass end piece and installed it. Picture to follow.
|2006-12-13 A much belated picture of the
replacement mentioned above along with another whirly built some time ago
with a disk glued in place which is discussed below. These two are on
opposite sides of the post so they don't hit each other, but with enough
wind they clang their ends on the post occasionally. [Blue glass broke out
in heavy winds 2007-02-24, picture at right.]
The disk whirly uses a 3" solid iron cast ball which was drilled and tapped for 1/2" thread. The shaft is 5/8" 16 gauge square tubing which just takes 1/2" rod inside. The axle housing is 1.25" round tubing which was carefully drilled and filed to take the square shaft through its side and a 3/8" hole was drilled in the shaft at the balance point so the axle could pass through. The disk is a commercially bent 1/2" sq. tubing circle (as might be used on 'wrought iron' fences) with a 1/2" rod welded to one edge and a stained glass circle glued in with silicone adhesive. Balance was not exact although both the weight and the disk can be adjusted by changing the insertion of their rod into shaft; a small square of lead flashing wrapped around the shaft makes for a quick adjustment.
| When I found that King
Architectural Metals had solid iron balls and largish steel rings, I had to
get my hands on examples, so I bought a 3" solid ball and a 5.5" ring; The
iron ball was fun to play around with, felt like a very heavy baseball.
Soon my brain was perking a bit and I saw the iron ball on a tube a short
distance from a bearing and the ring at the other end of the same tube, a long
distance from the bearing, welded on at an angle with red colored glass in it.
Step 1 was to drill the ball and tap it for a bolt. Problem 1 - Cast iron is really tough on the surface and drilling was an adventure. Got that done finally. Problem 2 - Getting a tap started in the tough skin went nowhere. Answer - drill out first 0.25" with larger bit, start tap further inside.
Step 2 was the tubing. Had 3/4" square on hand. Problems - It looked too big, the 1/2" bolt was not a tight fit. Solution - 5/8" square tubing. Problem - nobody I called which were my usual sources of steel had 5/8". The nearest place someone suggested was Houston. Then I tried Metal Supermarkets and they had it. I padded up the rental car, roped the 20' piece to the mirror, door frame and trunk hinge and drove it to work and then home.
A couple of weeks pass. I balance the 20' piece and hang the iron ball and bolt on one side and the steel ring with a representative amount of glass on the other, moving them until they balance. Marking those points, I carefully measure distance. I subtract off half the diameter of the ball and half the diameter of the ring to get the length of tube to cut (so the center of gravity remains the same.) I cut the tube. I slide the bolt of the iron ball into the tube IT'S TOO D..N BIG!!!
They have given me 3/4" square which costs about half what 5/8" costs. Angry phone call. They will cheerfully swap all the 3/4" for the proper 5/8". But only if I bring it in or pay $25 delivery. Damn. $25 is most of the cost of renting a car for weekend from Enterprise. So the experiment is on hold and the steel is inside the garage. 2004-01-24
Finished the piece finally, although it looks a bit long. Instead of welding in the end fittings, I cross drilled and cotter pinned them. Making the cross tube for the bearings was tedious with the square hole for the tubing. I messed up the plan to use a 4" hollow steel ball to cover the center pivot by not drilling the holes in antipode positions. The unit balanced really well right off because it was worked in order and remeasured for balance each step. The long arm is 62" and the short 23" from the center bearing to the edge of the mounted item. 2004-04-01 So here is a picture of the new and old. My overall feeling is that the glass area should be bigger, either a larger disk or a triangle around the circle. I kept the mis-drilled 4" ball and am thinking of filling it with lead and making another balance item. Strangely, these two which should turn the same rotation in the same wind often go opposite, etc. 2004-07-22
This whirly (below) pushes the limits of weight on this design. The shaft is again 5/8" sq. steel tube but here the shaft was cut in half at the balance point and inserted into supports of 3/4" 16ga.sq steel tube welded to the axle housing to both reinforce the point of greatest stress and avoid the drilling of the shaft. Alignment of the supports for welding was done by putting the shaft on a long 2x4 jig using its edges for straightness. [Could have been done better.] The ball is a hollow steel ball filled with lead with a 1/2" rod welded to it. The triangle is a 1/2" sq. tubing form with a 1/2" round tubing mount (hollow to reduce weight) and blue stained glass was sagged on a clay form to fit. The 1/2" mounts slide inside the 5/8" shaft and are held with cotter pins. The whole thing weighs about 35 pounds. The long section is 76" and the short 22" from bearing to end of tubing. Shown mounted on my sun shade frame as a temporary test mount.
|After construction without the guy wires, the unit, even though balanced statically, did not revolve and it quickly became apparent that the fit of the shaft in the supports and the strength of the shaft could not cope with the weight and leverage of the triangle and ball. The axle housing was drilled and tapped for 1/4" bolt thread and various bolts were tried, ending with the 6" slotted machine screws shown. Each bolt is locked with a stop nut at its base and the wire is run through the slot. Bracing wires were added, rather crudely, and finally turn buckles added when the wires could not be adjusted by twisting. After repeated adjustment, the unit balances dynamically. If this were do be done as an art project, I believe I would omit the turnbuckles by using pull-through crimps near the end, and of course the wires would be more neatly wrapped or crimped. 2006-12-13 This unit has been driving me buggy through out 2007, finally got problem solved today. Depending on the whims of ?? it would revolve or not and would stand still in stiff winds. I tried redoing wires and gluing heads of wire bolts. It persisted in sitting at a 45° angle, which basically told me it was bent, so the center of gravity was off to one side. I worked a lot with long upper part. I tried measuring offset from a taut line. Nothing. Finally, before finally giving up, I took it down and moved it to a mount on the workbench at head height. After dropping a plumb line it became obvious that the short length to the heavy ball was considerably off line - 1/2" or more in that short distance. Instead of trying to bend it or cutting it free and rewelding, I took a lead disk, tapped the center for 1/4"-20 and threaded it with holding nuts on the bolt opposite the tilt of the shaft - adjusting it to get balance. As soon as it was put up, it started moving very freely. 2008-01-09|
longer super heavy and the older heavy with the round disk, both with metal
fins for greater movement.in heavy wind.
Video from sequence of GIF images carefully trimmed to center.
Video from first time use of Nikon video mode on tripod
'RULES' FOR WHIRLYS -
|A two ended horizontal whirly has a dead point where it will not start revolving. Normally this is not a problem because the wind is vagrant enough, but if the whirly is doing some work or is out of balance, it may get hung until the wind shifts significantly. This is the major reason anemometers (wind speed meters) have three cups. Making a 3 part mount is harder than a 2 part.|
|A whirly must have enough weight to carry it around past the peak of drag. The sheet metal whirly below will stop broadside without the added weight. The stick above was very vagrant in the wind.|
|Balance becomes more important as the axis is angled. A unit balanced side to side on a horizontal axis may have the center of gravity of the whole piece off the plane of the axis, so when it is tilted it settles badly.|
|This whirly was built all of metal to test some
ideas of shape of glass and to see if braising the sheet metal to
the flat stock would result in something that could be hammered
to shape afterward - it can. Careful examination will show
that there are some links of heavy chain hanging just inside the
blades. When the weight is not there, the unit does not have
enough momentum to overcome air drag so it stops at a minimum
I have now learned some of the choices that reduce the size of an animated gif, so here is the glass 3 arm without saving loading because it is so big. Same post as item at right.
is another test piece fused in a half cylinder mold from
fragments of window glass. They were glued with E-6000 to L bent
flat steel stock. The glass pieces are by no means identical,
except in a selected profile, so the whole is adjusted to balance with a fold of
lead moved along the arm.
Although not very far off the ground, it moves gently in the
wind, more easily than I expected; which is why one makes a test
A six foot piece of 1/2" square steel tubing is the mount
bar. The L's are temporarily held in place with small C-clamps
which are also used the the center to hold the mount bar to the
bearing. [Later bar tapped for 10-24 machine screws.]
The bearing is about 4" of 1.5" steel tube with 1/2" ID 1-3/8" OD ball bearings pressed in each end - these are sold as replacement bearings for wheelbarrow wheels and cost about $2-3 each. 1/2" steel L's are bolted to the side of the tube and extend enough above the bearing to allow a security shaft collar to be applied when needed. Below the bearing a shaft collar holds the bearing on the 1/2" rough steel rod pushed into the ground. [This image shows the end of the rod threaded for a security nut with neither the nut or collar installed. Security in this case assumes it is mounted on a shaft attached to the ground where people might lift off the top to take it or destroy it. The small piece of metal between the L and cross bar is a shim to adjust for slight bending of L in fabrication to make it level. The rusty condition is typical of my test projects which often don't get painted. 2009-10-18]
|I wanted to make a 3 arm whirly, while taking steps to balance it as much as possible in the making. The arms were cut even and heated and hammer formed around a piece of steel pipe filled with lead held in a vice. After forming the arms to the central cylinder and after the glass was fused, the other ends (2 shown here) were formed to make them match the glass as much as possible, for good gluing and visual effect. The considerable angle of the bottom one is due to the distortion in the mold.|
|This is the glass first used. It was cut from three
pieces of stained glass in my stash, cut from a paper
pattern with not too much care for rough edges (obviously)
The modified arms were glued with E-6000 and assembled on the
axle cylinder (no picture before modification.)
The action of the spinner was slow and limited and considering the good action from a casually made hemi-cylinder, rather than fooling around with bending the arms, I decided to make a new mold and re-slump.
|After viewing the first mold, made in 3" metal
conduit section, the clay was resoftened and molded into
a setup built around a 4" thin wall PVC drainpipe 1/2
To re-slump the glass, I would have to get it off the arms. My first attempt at burning out the E-6000 by going to 600F failed and left the glue hanging on. A run to 850 burned it out and left the glass free. E-6000 turns out to be a contact adhesive and for NON-porous materials it is supposed to be exposed for 10 minutes after application and before clamping. When I followed the directions - applied a coat to the metal arm, pushed the pieces together, set them apart for about 5 minutes - I found that the behavior was much closer to a contact adhesive - i.e. harder to reposition. So I can choose the adjustability.
|The ends of the arms were reformed to match the glass - a line on each arm places the glass the same distance from the root. These pieces are much more alike than after the first slump. All have a rise over the rim and a dip. The glass was supported with an angle iron to align for gluing as described above and clothes pins clamped the glass.|
|This is the finished spinner. The central hub is 1-1/2" 16 gauge steel tubing with 1/2" ID wheel barrow bearings pressed in place and 3 holes tapped 10-32 around the side. The support is 3/4" square tubing mounted at angle (about 45°) to a workbench to be moved to a post in the front yard. The block at the top of the support is two steel plates braised to 3/4" square tubing which is tapped 1/2" - 16 for the axle bolt. The plates are radiused to 2" so a 1/4" pivot bolt 1/2" from the end of the support and 1/4" lock bolt aligned on the 2" radius can hold the adjustment for making the axle vertical. 2002-06-03 One color broken free of most glue in strong wind 2004-12-20|
|2002-06-15 After moving this spinner to the front yard, swapping it with the sheet metal one above, I got very little action. Close investigation showed that it was out of balance on the somewhat tilted axle, so I brought it back for testing. The thing was wildly out of balance, aggravated by the arms' axes not passing though the axle centerline. The dark blue arm is much heavier than the other two. After using clothes pins to clamp scrap glass and then cutting short lengths of steel I could slide up and down the arms, I was able to balance it with 4" pieces of 1/2x1/8" steel (plus the weight of the wooden clothes pins) overlapping the lighter blue glass.||The steel arms were carefully cut to
matching length (therefore weight) and were carefully
marked to align the glass on them. The glass was cut to
the same pattern, but casually and by measurement, the
dark piece is perhaps 1/4" taller that the other two.
It is probably also thicker, though I don't have a good
measure of that.
LESSONS: besides cutting the glass to the same size, weigh the glass and trim before sagging. Perhaps balance each pair of arms before gluing. Build in an adjustment for balance as with the triangle above.
Now I need to pull the sheet metal spinner and check it for balance, hummmm.
One of units damaged 2012-06-13 in major hail storm.
|I was standing in the backyard today
[2004-01-24] and realized that while I had several
whirlies working out there, I had apparently never taken pictures of
some of them.
This one is a vertical (obviously) with slumped window glass 'vanes' glued in place. The vanes are undersized for the thing which is about 2 feet across, but were made with available width cheap stuff.
This is an exercise in three activities -
drilling, welding and forging a good circle for the rim;
drilling the hub and adjusting the three spokes for balance
slumping and fitting the glass.
The mold for the glass was made by scooping out damp sand from my foundry sand box and laying a rolled out sheet of clay on it. The pattern for cutting the glass and the clay was shaped of paper to get the corner angle and glass size. That was copied onto aluminum flashing for a more permanent record.
The slumped pieces (3 runs because of only one mold) were held with clothes pins while the silicone glue set.
Lead flashing was folded over the rim for balance (lower left and center right)
|This is a rebuild of the ring above with sagged blades more like windmill blades, glued between a 23" outer ring and the inner with the glass being 7" long tapered 4" to 3" at each end. The 3 pieces were cut to size and sagged on a clay form. It is intended that colored glass blades will be inserted between the clear blades. Turns much more than the one above. 2008-01-14|
|Same frame, with three colored glass vanes added, and a mount to hang it beside house to it can be seen out window. 2008-03-19|
|And this is what remains of the whirly above, taken during the hail storm 6/13/2012 that destroyed it and the two flat units above as well as doing $200+ million damage in narrow swath down through Dallas county. We don't get storms like this often but when we do people lose roofs and car windshields. And glass. 2012-06-18|
|This image is provided to show an actual working wind mill (click to link to source site) that pumps air to aerate a pond. The tapered, twisted curve of the blades can be seen as well as the outer ring support. Be aware that the decorative windmills sold in garden centers usually just spin on a fixed shaft, which is simpler to make, like my whirlies above, but do no work and are hard to adapt to drive a small generators. 2012-01-28|
|This is an old bicycle wheel with bent aluminum bar stock as the mounts and supports for the mounts. Clear window glass sagged was on a clay form, perhaps the same pieces as below. An aluminum bar stock bent as a hook at the top and to an offset flat drilled for the bike axle. With good lube, the wheel moves freely in the breeze. 2008-01-14 [One glass unit fell off as glue aged and I just mounted a lead disk to balance and it still turns. 2009-07-30]|
|This is a test piece in a couple of ways. Originally the glass pieces each had two flanges, but it was very sluggish, so I cut off one of the flanges, breaking one of the glass vanes. After slumping a replacement, I mistakenly glued it on reversed (2 down, 1 up). After it had run for a while, it kept slipping, so I re-bent the aluminum bar mounts so instead of wrapping around the hub, they were bolted up it. Works good.|
||On the front peak of the house, I custom welded and installed a bracket for mounting whirlies. (Click on pictures to enlarge) The mount includes wire to send a pulse signal down to the house when these rotate. There are actually three points on the bracket to mount 1/2" threaded rod - on top of the top bar as this one is, on the sides of the top bar, and on the front of the vertical bar. This whirly is made of sheet metal braised to flat bar stock and was originally a test to see if bending could be done after braising - yup. The flat runs on both sided of the hub, held with a 10-24 bolt. The chain dangling from the sheet metal gives this two vane whirly momentum to get through the dead spot two vane units have.|
|| For the wind tunnel,
etc., I needed an anemometer of some kind. This one was made from a
film canister pierced vertically by a brass tube with three brass wires
bent to fit through six holes and around the tube (tricky). The tube
telescopes through a spacer and into another tube to form a bearing
The cups were made by having one half dome plastic cup (not shown) that just fit in a small mushroom can, so plaster was put in the can, the dome was pushed in, and the inside of the dome was also filled with plaster. When set, the dome was removed, leaving two plaster forms with a gap between. Squares of thin (single weight glazing) Plexiglas were heated in an oven to about 240F until limp. Using gloves, they were removed, centered on the plaster dome and the can was pushed down to force the plex into the dome shape and flatten the corners on the work surface. As each cooled it was removed and examined - reheating if a redo was needed. The extra plastic was sawn and sanded off, holes drilled and silicone adhesive used to glue them on the rods.
Indoor Whirlies - Twice I have
installed whirly blades above the roof and led a thin rod down to turn sheer
fabric hanging panels inside the house. The first time it was
successful, the second time it leaked and was plugged up. Right now I have
similar panels hanging on ball bearing fishing swivels from the HVAC vent but
it isn't the same as the wind blown one hop and dance with gusts.
The design is in 3 parts: blades, connecting mount and rod, and hanger inside. My hanger was brass rod in a + shape soldered with a single upright in center. I could have hooked this to the down rod, but soldered it with a thin brass tube for tighter wind reaction. The blades on the top were variously half tin cans or curved brass sheet soldered to arms leading to a square brass mounting tube. In the roof and in the ceiling 1/4" diameter thin wall brass tubing was fitted in aligned drilled holes. The lower tube merely guided the down rod. The upper one acted as a bearing and support for the mount.
The top of the rod had several telescoping brass tubes to center the 1/16" rod in the 7/32" sleeve that fits in the 1/4" roof tube. On the sleeve are also soldered a short 1/4" round tube as a thrust bearing. The next tube inside (3/16") extended up to take a 7/32" square to take the 1/4" square tubing on the blades. Design sizes were driven by available telescoping square tubing sizes. Tubing Source 2011-02-22
|Tailed Whirl - In this case, I wanted to add a tail to a fan blade wind whirly (down below) by forging a 1/4" rod out the back.. I tried to draw some ideas (right), but found it only gave me hints. I was sitting there next to the copper wire and began playing with it. I was undecided whether I wanted something that wrapped back with a 3D curve or a spiral. I had to be able to balance it while mounted in the back of the fan shaft. I also had to be able to forge it. The picture was taken with the wire laying on the drawing outdoors. Both the tracing of the wire and the actual shadow make clear that my drawings miss the sharpness of the dip as the spiral turns the corner. After doing the whirlies, I made a couple of people, which are shown right, with the round nose pliers I bought a couple of weeks ago. 2004-12-10|
|This is the spiral shaft after forging. The 1/4" rod is threaded into a tapped hole in the end of the 1/2" shaft of the fan blades. Forging was about as much of a challenge than I expected, getting the twist in place and working it smooth, and getting the balance was a challenge. Even cold, I had to jigger for several minutes of bending to get it centered.|
|NOT MY WHIRLY - This image and videos were taken at the Ft. Worth Main Street Arts fest in 2008 and show very large (over 6 feet) metal and glass whirlies, some of which are geared so one part drives other parts Andrew Carson.|
Contact Mike Firth