First crack at wind turbine and weather station (pics)

[OhioDale]

Hey everyone i have been looking through the site for a year or so now, and some friends and i have started a project.  We are building our first wind turbine.  We want to add to it by creating a weather station that records wind speed and direction.  The genny is 3phase with 40 magnets (20 per rotor) and 15 coils.  The weather station is still being designed as we build it.  It is going to utilize a freescale 68hc12 microcontroller and a pc running a C# application.  The speed will be recorded using a spinning disk and Infrared LED's.  The disk will allow a set number of pulses to be transmitted to the microcontroller every repitition.  The direction will be determined by similar means.  A four bit grey code position encoder will be used.  the wheel will have four bit grey code that will translate the position into 4bit binary that the microcontroller can understand.  This information will be sent to a pc via serial port.  A C# application will listen for incoming values and record them.  The C# program will then create a real time graph and give the user the ability to graph archived values.  This will allow us to monitor performance over periods of time and compare data from different stators.  

As most people are more interested in the turbines themselves i will leave out all of the programming.  Anyone who is interested on how we will accomplish this feel free to ask.  We have been pretty busy with other things so our progress is slow but steady.  At this point we have a few coils wrapped, the alternator almost complete, charge controller is being assembled, weather station is being built, microcontroller program is partially functional, and the C# app is bare bones.

Here is the the alternator with on rotor attached.  The magnets were placed with a piece of plywood.  The bearing is from from a late 80's chevy (i think).  The steel was donated and machined from a local mill.

Open circuit voltage when hand turned and using #16 wire (first coil & hand wound) was around 3V, short circuit current test was around 3A.  I will post some better data once we get more the other test coils finished.

This is the first testing of the jacking screws.  They worked very well and allowed for a slow decent and easy removal of the second rotor. The same coil had an open circuit voltage of 9V, and s.c. current of 3A.  The current was still lower because it isnt mounted to anything and is difficult to spin.  

Wound some coils at work.  The first is #20 with 465 turns.  I also finished a #12 with 85 turns today.  I want to make three #16 coils.  two will have matching thickness but different diameter, and the third will have a matching diameter and a different diameter.  I will post the results once i have completed some testing.  I want to use either a drill press or lathe to get a steady rpm. and give my arm a rest.

src="http://img267.imageshack.us/img267/3883/clampsfp2.jpg">

The coils are being sealed with a basic enamel.  i coat the outer diameter and later pull one end of the plexiglass of and apply some more.  The wood clamps prevent the coil from expanding while the enamel sets.

I grabbed some scrap copper from work to make some L-brackets.  

The enclosure is made of smoked plexiglass.  The front pane will be transparent and have quick disconnects for the I/O.  The microncontroller will be mounted below.

[OhioDale]

The led's will be mounted in here.  The bottom will be the IR emitter.  The top is a detector (transistor) when IR is detected a base current is induced and our output terminal will toggle.  This will allow us to generate pulses for revolutions or determine position.  Ex.  no light passing equals 0000 correlate to 0 -> 22.5 degrees.  0001 can be 22.5 -> 45 degrees and so on (2^4 = 16 sections : 360deg/16 = 22.5deg).  the zeros are represented by a low (0V) and the ones by a hi (5V).  the grey code will allow us to eliminate most race conditions and reduce errors.  

 

Here we are testing the led's  they need to be shielded from outside IR sources to reduce interference.  Also the operating current is higher than regular led's therefore the series resistance is much lower.  This created a lot of problems for me since i was too lazy to look at the data sheet and went with a higher resistance at first.  Normally i would put 110 Ohms in series but these only need 18 Ohms(emitter). also we are using 3.3V and 5V. i will post a simple circuit diagram later.

here i am testing the led's inside the mount.  I alloted .5 inch between centerlines.  I should have checked the viewing angle and made sure they won't interfere with each other but once again laziness wins, and that would take the fun out of testing it.  

Here is a closer picture of the mount with the led's being tested.  I am using a piece of electrical tape to block the IR.  

[kurt]

your pictures are to big resize them to no more than 640 pixles wide

[OhioDale]

sorry my resolution is set at 1280 x 800

if this is better let me know and i will repost the rest

[s4w2099]

Hey man. Nice job really neat. I was trying to figure out, what is the voltage of your baterry bank? 48Volts? I hope it is because you said that you have 465 turns of #20 gage wire and 5 coild per phase.

I am not a big expert on the subject but I think it is a bit overkill. The resistance in each phase will be very large I am guessing if its single stranded and you will end up wasting a lot of the energy as heat in the stator.

I would like to know more details about your project just for curiosity. What size magnets? what grade? rotor diameter? coil size, thickness and inner diameter?

:-D

Take care man, and good job once again! I hope mine was was clean an yours

[OhioDale]

The coils are for testing possible outputs ( #20, #16, #12 and different sizes).  We currently dont have a battery bank.  The whole project is for school and we are going over how a wind turbine can be adapted to most regions of the world.  The magnets are 2x1x.5" N40.  The rotors are 15" dia.  and .5" thick.  The first couple of test coils are around 2.8 x 2.25" x .7 .  This is one of the things we are going to play around with.  I worked through a lot of equations last year to try and predict output, but with my lack of knowledge in constructing permanent magnet alternators I dont know how much flux density to expect and also how much losses from friction and such. It will be cool to see what the ideal output is compared to actual and show where these losses come from.

There are three of us in the group one traditional and two digital.  I am digital but i have a decent knowledge of the analog side as well.  

Thanks for the input and i hope i answered your questions.  

[RP]

That looks to be 800X600, still big.

I can't tell from the pictures but it looks like the long axis of the ID of your coils is much shorter than the magnet length.  If so, you are giving up some voltage.  Any part of the winding at right angles to the path of the magnets produces nothing.  Try to get the coil legs at least close to the length of the magnet.

[willib]

you seem to be going about the the hard way.

i realize your mags are rectangles , but to give you something to think about, i'll describe my machines

the mags i use are .875" dia. 7/8" ..

i make the OD of the coil a little less than twice the the magnet dia.

and i make the hole smaller than the magnet dia. this way i get more turns per coil , and the turns in the middle dont really add that much resistance

this gives me a coil "leg" of over .6 inches , on a 1.75 dia coil.

using 0.045" dia wire i was able to get 80 turns out of a coil just 0.270 thick by a tad less than 1.75"dia..

with a resistance of somewhere around .135 ohms per coil , dont remember exactly i could check my notes but i'm too tired.

so wind a test coil , single strand , of about 1/4" thick by say 1.5 inches long inside length , inside width of 1/2 inch , this will save a lot on the total resistance per phase.and still get you the number of turns , withe these inside dimentions i imagine that your coil legs will be over .6" wide ,

remember coil shape relative to magnet size is very important , and the more turns you can squeeze into the X-Y plane , the less you need in the Z.

what are the LEDs for?

three hall sensor can give you the position of the rotors, if that is what you need ?,and and the rpm, at the same time.and are uneffected by infrared .

pc powersupply fans all use hall switches ,old CD rom drives use hall sensors.

[RP]

I believe the LEDs are for a wind direction sensor, not the rotor.

[Flux]

To give you a rough guide you can expect about 600mT in a 3/4" air gap with that magnet set up.

If you don't have any particular voltage in mind I don't see what you gain by messing about with different size wires. You can deduce all you want from one test coil. Experimenting with coil geometry may gain you some knowledge.

You are going to end up with a very high voltage machine. Any tests you do into resistive loads will give very different results from rectified tests into batteries,

If you want to settle things for the two types of use you will come up with quite different options.

Don't make the hole in the coil significantly shorter than the magnet length, you may find some advantage in reducing the width of the hole below magnet width.

With your magnet spacing fixed you don't have much latitude to change much.

Nice work you are doing there.

Flux

[kurt]

you keep posting your pictures at 800 x 600 pixels that is to big for people who run there monitor resolution at 800 x 600. the file size is fine it is the physical size of the pictures that is to big. please reduce them to 640 pixels wide or less.

[OhioDale]

when i hosted the one previous i selected 640x480.

i tried 320x240 this time.

is there any way i can edit my post so i can resize the images?

[willib]

any progress today?

the new photo is fine for me

[thefinis]

Not sure but I think you can just change the file saved at that file name in your files where you uploaded it. Should work as long as you replace the existing file at that file name with one with the right size jpg in it. It is a pretty big deal to many of the folks who are in the boonies and/or off grid and have limited power and web connect resources.

Great looking stuff

Finis

[Experimental]

  I had no problem at all with your pics and liked the larger ones, as I could make out all details --( perhaps my computer !!???)

   Any way, I enjoyed your post and you are doing nice work -- Kind of curious where you are going with all the different coils (perhaps learning experiments) and look foreward to further posts.........

   Great work, happy building, Bill H..

[OhioDale]

Hey guys thanks for all the feedback.

Experimental, yeah we are making the coils to practice and also to physically see the different outputs.  I have never wound any coils until these ones.  Speaking of coils has anyone tried using flat wire?  I think it would be much easier to make.  The smallest we have here is .015"x.375"  This wire isnt insulated but we also have some stuff called vartex (spelling?) which is .005"thick and can be wound with the copper.  

willib, not too much progress over the last couple of days.  Snow storm closed a lot of stuff down.  I did order a new copy of C# 2005 since i lost my product key to 2003.  Once i get that program going i will put up some screenshots of the interface.  Other than that i have been doing some reading and looking into winding a flat wire coil.  I might stop to pick up the microcontroller and stuff so i can work on that program.  Its been a while since i have looked at it and i dont want to forget what i had in mind.

Also i think i will post pictures at smaller 320x240 and post link to larger ones.    

[Experimental]

Hi Dale,

   Yes, our benefactor Dan B has made coils with flat wire (check the archives) as well as a couple of others -- but as I recall, there were a few problems with them--   Dan,s were on a diesel powered generator they built !!

As yours are not coated with an insulator, I doubt the benefits of using it and just the fact that most gens and motors use normal magnet wire, would lead me in that direction !!  BUT, don,t let what the "crowd" does, stop experimenting, as that,s how new invention comes into being !!!

   Keep at it and happy building, Bill H...

   

[OhioDale]

Here is a picture of the circuit for the led's.  The diodes give off IR and the transistors detect it.  The IR generates a base current in the transistor.  This is why the physical part only has two leads and the schematic does not have anything hooked to the base.  The 4 bits allow for 16 positions each one with a radius of 22.5 deg.  The reason we are making the weather station with led's, microcontroller, and C# is because it allows us to get experience in each of these areas and fulfills criteria for our project.  

larger image http://img137.imageshack.us/img137/8827/4bituv5.png

the top circuit contains the emitters.  Each diode turns on at 1.2V and 100mA.  The circuit supplies 125mA and 5V this is in the operating range.

The transistors only give the maximum values.  My laziness led to some reverse engineering.  We looked online to find a similar design and found that the resistance was too small for our setup.  The 100k Ohm allows for a significant enough voltage drop once the transistor is "on".  The labeled nodes "bit4,3,2,1" are our outputs and are high when no IR is present and drop once IR is detected.

 larger picture http://img131.imageshack.us/img131/2264/dsc01941at4.jpg

larger image http://img265.imageshack.us/img265/5303/dsc01942qc7.jpg

here is the transistor circuit wired with temporary breadboards to allow for quick modifications.

[MechEngShawn]

Hi Dale,

I've done a few things with IR before and had a suggestion if you're having trouble with current setup.  You might consider hacking an old printer for the optical switches used in an angular position sensor (ie position encoder), a limit switch, or one of the switches used to tell when you have access doors on the printer open.  They're nice because the contain the diode & photo transistor already set and focused at the right distance.  If you haven't seen one, they're packaged in plastic and [ -shaped with about an 1/8th in. gap between emitter/detector, so it would allow you to make your own encoder disk (thin sheet metal disc with a few holes) to spin in the gap.

If this doesn't give you enough flexibility for your design, you may want to look into the IR emitters & detectors used in a VCR.  They send the signals over a ~36-40kHz carrier frequency so it doens't mess with the ambient light.  You can get the detectors pre-made, but I think you may have to make a 555 circuit to pulse the IR LED at 40kHz.  

Someone else made a comment about a tachometer on the turbine itself.  I don't think that was your intention, but had you considered putting one on?  With the weather station alread up and your data logging software already running, it may be useful (and not too much additional labour) to match your windspeed and other stats with RPM. Dare I also say voltage & current monitoring, which would require some nifty A-D conversions?

Which C# utility are you using?  I got my Visual C++ for free from Microsoft, the Express Edition of their Developer suite.  Works well. C# & others available too.

I like the work you've done!  Hope to read more as the project progresses.  I'm actually in Mech Eng north of the border in Hamilton at McMaster Univ.  Its nice to see more people with hands-on rather than textbook-only experince!

Shawn

[OhioDale]

Shawn,   Thanks for the ideas.  The description you gave for the IR setup sounds like it would work very well.  We are using 1/8 plexiglass for our encoder wheel.  Our current setup seems to work well but i will defintely keep you idea in mind.  

We are currently planning on taking readings from the turbine (I,V, RPM)  this would be used with the wind speed speed to compare performance.  We have not gone too far into that design.  I do know the microcontroller we are using has built in A/D capability.  If anyone has any ideas on how to get that information to a format Analog or digital that is in a microcontrollers operating range (not sure what the range is yet) i am open to suggestions.

I was going to use the Visual Studio 2003 version, but i lost my product key and my old laptop died. I ordered the 2005 C# book with what i thought had the software, but i now realize i didnt buy the deluxe edition.  Hence the super cheap price......Now i might have to go with the express version, or get the student version.

[TomW]

Dale;

but i now realize i didnt buy the deluxe edition.  Hence the super cheap price......Now i might have to go with the express version, or get the student version.

Have you ever considered the Linux Operating System and it's extremely complete Gcc and compiler tools? Literally millions of lines of open source code is out there, too, with many code examples and full ready to compile program code that are not hidden under the mask of proprietary source code. You can even run it from a live CD in several flavors such as Knoppix so your hard drive stays unchanged. Just a download away.

Absolutely free right down the line. I don't code at all so not sure what that other stuff is other than "it takes money to use".

Just tossing that on the fire.

Several of us over on IRC are dyed in the wool Linux users should you want to try it we would be more than happy to help drag you thru it ;-].

Cheers.

TomW

[Ungrounded Lightning Rod]

You might consider hacking an old printer for the optical switches used in an angular position sensor ...

Or tear down a roller-ball mouse.  Two sensors like that on a PC board with star wheels, microswitches, and a little logic board that has it all set to talk to a PC.

Get a serial mouse and hook it up to an RS-232 port.  (Put a cheap modem next to it if the distance is too long for a serial line and run phone wire to another modem near the computer.)

The protocols are published and drivers are available.  Look at the open-source mouse drivers in Linux for sample code, then write your own stuff to watch the mouse and clock and interpret the result as speed rather than pointer movement.