I have a wind project that I have been working on for several months now. The intention was to use this as a trickle charger along with some solar panels, to a battery bank for a shed and backup power source for a well tank booster pump. The well has a submersible pump, that I would power with genset to fill a tank,and use a pressure booster pump to the pressure tank / filters / house supply.
But those plans have been on hold. For almost two years now. Meanwhile, I didn't want to put up a HAWT...I can hear inspectors now with some wild fabricated complaint that I have a 100-ft tower knocking eagles and turkeys out of the air. This will happen and I want to be able to laugh the first time and point to the slow-turning sav. I wanted to build something I could test here and take there, and more akin to 'yard art'. And learn something along the way, for when I get there and planted and can put up a tower to get those good winds I know are up there.
The property is on a hill, with the NE and N sides dropping into a valley, and the S and W sides rolling into open flat land before it drops into a valley. If I stand at the high point of the hill, I can see the treeline tops that start at the edge of the valley. There is good winds at the ground level, it blows paths into the tall grass around buildings.
Well, I've built 'something', it definitely qualifies as yard art, but will it actually put out a battery charge? I'll share what I did, and take all the congrats and admiration...and / OR the flames and flying tomatoes.
I'll stay in 'safer' territory with the first post and cover the servos, the second will cover the tower, and the third will cover the sav. And none of this will represent cost-effective power generation. (that's what the solar panels are for).
I've already posted some in Steadfast's thread about servo motors and what I have on hand. I have done some experimenting on these to see what they can do as a generator. For a VAWT, with a geared shaft speed anticipated between 90-180 RPM (for 50-100 RPM VAWT), I need to get the output voltage high.
I sat up a bench test to measure some outputs under a load and to get a feel of shaft load while generating. Motors I tested and chart below.
These are two of the motors I tested, a new-style 450 and new-style 850.
The test setup. Not scientific; RPM estimated using stopwatch and clicker; the load resistors were 50W max so higher measurements may be current-limited. I did heat them up quite a bit.
That chart helped me find a motor for SF's application, he needed a motor that lightly loaded the prop and would cut in around 240RPM, any sooner and the 'light wind' turbine would be stalled except in a windstorm (per my thinking). But the figures didn't help me find the motor I needed.
I had destroyed one of the motors for bearings and flanges, and did some experimenting on finding the coil junctions. All of the new servos use some plastic-epoxy mold pored after the junctions are made. I tried to remove this, using heat, a chisel, stuff wouldn't cut, I found the junctions but also damaged a couple of coils.
The disassembled 450 is identical to the motor I sent SF. 9 coils / two offset rows of 8 mags each on the rotor. Front shaft sealed bearing 32mm (1.25") shaft diameter / 3" outer diameter. I used the front bearing flange for a tower shaft and the rear for a chain sprocket shaft. Not sure if I can use the rotor or magnets yet, right now just keeping it from sticking to anything.
I found the star junction (on the older 850 with exposed stator) on top of the rear of the stator. I first tried testing it with IRP (parallel coils) but it didn't lower the cut-in significantly. After I tore down the bench test and shipped SF the motors, I thought about trying series coil / star connection, so I rewired the motor, set up the test again, and did some spins. Didn't chart it, but the voltages were triple, and cut-in around 60RPM. The shaft loading didn't seem excessive but it was progressive with the speed. So I selected the motor to use for the VAWT.
I am not sure what the difference is between the older and newer 850. I haven't pulled a rotor from either, but the size of the motor is double the newer one for the same ratings. But just the exposed stator makes the older one more useable. That may be a good point to finding an older yasakuwa [sic - intentional] or even the other, especially japanese, manufacturers. FYI the newer styles are manufactured after 1996. Another note on finding servos - they are used in all kinds of CNC machines, and its easier to find one without a brake in a process machine than in a robot arm.
One end of each coil was originally wired to the Amphenol, and the star junction was on top of the stator. I added extension wires to the terminal strip and labeled each of the 9 coils. I did NOT find the exact position of each coil, but a DC voltage on any coil set will spin the rotor to a quadrant point for one of the three phases. I would need to remove the rotor to find each coil location, but a battery can find each coil group.
I added a terminal strip in the encoder enclosure, relocated the motor Amphenol to the encoder box, and sealed the original power receptacle mount. It will be easy to get to the connections again and change the arrangement.
A note for spottrouble - you have a Fanuc servo that you say is wired delta? If the coils are series...it may be EASY to find all the coils. Check the leads to the Amphenol, there may be six coil wires in a sleeve for each pin. If not, I bet the junctions are on top of the stator with an extension wire crimped to it. You may need to ohm them out and then do a battery test to make sure you identify the phases and polarity of each coil set if it is not obvious.
I still want to find an easier way to pick out each coil position, I can find the 'groups". If I decide to use a stacked-coil cutin on a motor (meaning using only one coil per phase, and adding other coils), then the location is more important to prevent uneven shaft loading.
Next post will cover the design of the tower that will be used with the sav, with center shaft sleeve that will allow swapping the 'props' out if I try something different on the same tower in the future.