Sunday, January 28, 2007
You can click on any picture here to enlarge it.
The DC gearmotor is in the green box to the right. The output shaft of the gearmotor is coupled to the galvanized steel drive screw shaft inside the box. Both ends of the drive screw are supported with pillow block bearings, a flange mounted ball bearing unit (Dayton 4X727) at the motor box and a bronze bearing pillow block (Dayton 2X567) at the far end. That is an ordinary pressure treated 2x6 that forms the support. You can see the yoke assembly in about the middle of the screw. The limit switches haven't been added yet.
I don't have an up-to-date picture of the inside of the motor box, but here is a picture of an early version to show how simple it is. I used a Dayton 2L008 Permanent Magnet DC gearmotor with a 560:1 reduction ratio. Dayton products are available nationwide in Canada and the USA through Grainger Industrial Supply (www.grainger.com). This motor only costs about $50 and is available from stock. It is rated for 12 VDC, and draws 1.2 amps at full load while turning 12 rpm. Only 15 watts to move the entire solar array! This little motor with the drive screw is plenty powerful enough for the job.
The 1/4" motor shaft is coupled to the 1/2" drive screw with a pair of spider couplers, like Woods L050 series which match the size of the two shafts and provide for a bit of mechanical misalignment if there is any.
I'm not that proud of the motor mounting which is just a pair of galvanized steel perforated strips, shaped to suit the motor distance back from the panel, but I was in a hurry to see it work and this seemed to do the trick. It's stayed that way for the last two years and works to hold the motor securely in place.
The yoke contains a captive 1/2 inch hex nut with the same thread pitch as the rod. The two outboard steel wheels just graze the 2x6 and keep the yoke from spinning around the drive rod.
The white goop on the thread and at the nut is white lithium grease. I connected the gearmotor to a DC supply and allowed the motor to turn steadily while I dabbed the white grease in it's path. To the left, you can see the grease spread out on the screw. I ran it back and forth over the whole distance I expected it to travel and kept dabbing grease on it until it was liberally coated smoothly and evenly. After two seasons, there is no rust and no apparent wear to the thread or the nut. I haven't re-applied the grease but probably will this spring.
Here are the two main components of the yoke prior to threading and gluing. You can see that I have brazed a short section of 3/16" steel rod into a drilled recess in the hex nut. This was actually my second attempt, since I am not a welder. Prior to drilling the recess for the rod, I centerpunched a guide dimple for the drill on one of the hex edges. I found that it was important not to drill the pilot hole completely through the side of the nut so that the brazing alloy didn't flow through and clog the thread. I suppose that if I had had a tap of the appropriate size, I could have run it through the nut to clean up the thread afterwards, but I didn't have one at the time.
Here is the yoke with the steel rod threaded 10-32 and the nut glued into the 1/2 inch band iron frame with industrial epoxy. The two holes in the legs of the frame are for the steel wheels. The threaded rod couples to the actuator rod.
If I was a better welder, I might have brazed this whole assembly. Because of the large area beween the nut and the frame for the glue to grab, I had much better confidence in this approach. I am using an epoxy by Locktite called Hysol 907. It is amazing stuff, but does require suitable curing time (24 hours is recommended for full strength), or a heat cure (150 degrees F for about 2 hours). The surfaces to be glued should be roughed with silicon carbide sandpaper and cleaned with de-greaser prior to gluing.
This is the pillow block which supports the far end of the drive screw. It is a bronze bushing which can pivot slightly in its housing to provide for some degree of mis-alignment. I used a bronze bushing pillow block out of concern for exposure to the elements. A sealed ball bearing unit was quite a bit more expensive and probably not necessary.
The oil cap is a nice touch and I even put a few drops of oil in it. The protruding oil cap did end up being a bit of a nuisance later as it interfered somewhat with the actuator rod which swings near it on it's way to the drive rod.
Saturday, January 27, 2007
The stud on the yoke couples it to a steel actuator arm (the rusty rod running up at a slight angle) which drives the end of a 1/4 inch steel drive rod that runs down the entire length of the array, under the collectors. Each collector is coupled to the drive rod through an aluminum arm and an eye bolt.
As the threaded rod turns, the yoke travels along the shaft and pushes or pulls the collectors into position via the drive rod.
Two limit switches (seen along the bottom edge of the wooden beam) sense a small magnet on the lower wheel of the yoke to tell the electronics to stop motion at the east or west limits.
Here is a close-up (click to enlarge). The arrangement is such that the collectors will rotate through about plus or minus 45 degrees, thus catching the sun's hottest rays at mid-day.
The overall gear ratio is about 6000:1. You can see the shaft turning about 12 revolutions per minute when the motor runs, but you can barely see the motion of the collectors.
The motor drive wooden beam is positioned at about the mid point of the collectors and drives them through the aluminum arms attached to the middle rib on each.
At this stage, I had not put a concrete footing underneath to support the motor drive so I had suspended it between the ends of the array on a wooden cross piece. It flexes a bit, but is solid enough such that it worked well for the first season.
At each collector, two small coil springs and two collets on the drive rod provide a bit of decoupling to keep the wind from buffeting the motor. The eye bolt can rotate because it is fastened with an aircraft nut at the back which is not completely tighted on the eye bolt's thread. If you can picture the drive rod moving away from you into the picture, the front collet forces the spring against the eyebolt which pushes the collector aluminum arm away from you. This action will rotate the open face of the collector towards you in this illustration.
The collets can be loosened and the position of each collector adjusted by sliding the collets and springs along the drive rod and then re-tightening. In this manner, each of the collectors can be adjusted at installation so that they all point in the same direction.
You can see that the drive rod has already started to rust. I intend to replace it with one made from stainless steel. Otherwise it will soon be impossible to adjust the position of the collets on the shaft. Interestingly, the collets which are supposedly not stainless, are holding up rather well with the exeption of the set screws.
Tuesday, January 23, 2007
The collector tubes span an eight and a half foot gap between the support rails. It is important that the collector tubes do not sag. If they do sag, they will not be in the focus of the parabolic reflector and will not receive concentrated sunlight.
I'm often asked if it is possible to use black PVC pipe for the collector tubes. PVC is a lot cheaper than metal and besides, it is already black, a great light absorbing color. The metal tubes need a coating of flat black paint or black heat shrink to make them effective absorbers.
In this picture (click to enlarge) you can see a comparison of black PVC pipe to stainless steel, with one of the solar heaters in the background. The sag in the PVC pipe is readily visible. The tubes in the picture are empty. Normally the 1-1/2 inch pipe used in this installation contains about 10 pounds of water over the length of the span, which makes the sag greater! ABS pipe has a similar problem, although not as severe.
I did a sag test last year using copper pipe (the DWV type, with the thinnest wall) suspended over the span, filled with water, and I found that the sag was less than a millimeter.
It is important to keep the pipe in the focus of the parabollic reflector and it is not possible to do this with the PVC unless it is somehow supported mid-span.
The metal tubes while more expensive are significantly stiffer than the PVC and they span the length of the collector and remain accurately positioned at the focus with no problems.
The second problem with the PVC is what happens to it if the water drains from the system and the PVC receives the focussed heat from the reflector without the cooling effects of the water. In dry tests, the collector tube reaches over 300 degrees F. I did this test with a black painted copper pipe, not PVC. The PVC would likely sag more, or burn. The metal tubes are unaffected by this heat. ABS platic pipe is better at higher temperatures, but the material is white and would require a coating to improve thermal absorption.
Finally, the PVC pipe does not have as good thermal conductivity as metal. Not only is plastic not as good a conductor of heat, but the wall thickness is significantly more than the metal for a given pipe size, further impeding solar heat transfer to the water.
I prefer metal pipe over plastic for the collector tubes for all the above reasons.