windgen.org vawt discussion

Very cool, I was thinking the interface was USB, not Ethernet, but maybe the USB is only used to provide power.  I like Eithernet better because it means I can put the microcontroller in a box outside, and my computer can stay safely inside.  In fact my computer can be wireless anywhere.

The next chunk ...

dereksoftstuff wrote:

---

Bryan

Good links.

We can build our own anemometers (see next video)  and I'm not a mechanical engineer ! :-

a) for a lot less $

b) with a lot more accuracy

c) and be able to update or fix them when required

I've assumed a linear relationship between rpm and speed (using rotor diameter). If necessary a more sophisticated algorithm could be employed later.

---

True, but some of us would like to purchase one instead of build one. :-)

 

Bryan

Good links.

We can build our own anemometers (see next video)  and I'm not a mechanical engineer ! :-

a) for a lot less $

b) with a lot more accuracy

c) and be able to update or fix them when required

I've assumed a linear relationship between rpm and speed (using rotor diameter). If necessary a more sophisticated algorithm could be employed later.

Has anyone tried one of these:

http://www.davisnet.com/product_documents/weather/manuals/07395-032_IM_07911.pdf

You can buy the anemometer itself with a wind vane for about $120, but I haven't found any documentation about the electronics to hook it up to your own electronics.

Or you can purchase just an anemometer for about $55 here:

http://www.inspeed.com/anemometers/Vortex_Wind_Sensor.asp

It says that it has a reed switch that pulses once per revolution.  I haven't seen details about calibration, but it does appear that they assume a linear relationship between wind speed and RPM in their set-up guide below.

http://www.inspeed.com/anemometers/pdf/Setting%20the%20Velo8%20Wheel%20Size.pdf



-- Edited by Caleb on Friday 29th of January 2010 02:29:06 AM

Steve - just fooling with you on the video -> I thought you could make it yourself with all the info in this topic!

Probably post video etc next week. Would be good to get some testers for the system - some with real-live outdoor turbines.

Collaboration is always good -> it's the ideas that are difficult. I'll tell you about my vawt wings project one day ...

I used to use PIC 18's then discovered LMI cortex-m3 - no comparison - this is a real beast (well, a little one).

If anyone wants info (any) - just post here, and I'll put it in a video. I'm still learning about all this stuff myself ...

dereksoftstuff wrote:

---

Steve if you get a chance can you post the next video for me ...

---

Thanks for the info, I wanted to get that out before people started ordering boards then realized they needed some other parts to make the system work.  Now they can order it all together.

Sure I'll post the video, I don't see it on youtube yet.  Or you can do it yourself.  Right above where you type is a little icon that looks like a piece of film (second from the right.)  Click on it and just past in the URL of your video.  That's it!

I'm looking forward to part2.  Do you have a web page with some more information?

I'm thinking I'll get my RPMs from just one phase of my alt (I bring all 3 phases into my house.)  I can just clamp the voltage to protect the input.  As long as we can configure magnets on the shaft that should work.  (I'd have to enter half the number magnets I really have)

For someone named softstuff, you seem to have a good knowledge of hardware. :-)
I also do both, but mostly low level code, like what would run in a microcontroller.  GUIs aren't really my thing, except in Java.  Much of my experience has been with the Microchip PIC controllers, and I have done several USB devices that also create a virtual serial port.

Maybe we could collaborate on something someday.

Steve

Haven't got a battery, just using dummy loads (resistive, lamps ...) at moment.

The voltage sense and current sense circuits are entirely left to the builder, to suit their turbines. THe ADCs are 3V max - DO NOT EXCEED!

To use the microcontroller you must use appropriate voltage divider to get the max volts down to 3V for ADC (or fry it). My turbine is high volts low amps star, so I can get 70V +, hence the scale factor of about 29 in the user config, I think I've got 100k and 3.5k resistors for the divider.

For current sense, I'm using a high side system (you can use whatever you want), with a 10mOhm current sense resistor (solid metal band) and a MAX4080 IC with gain of 20. You must calculate the MAX current your turbine can produce (in any conditions), and using this as the full scale value - ensure that the output voltage you supply is 0 -> 3V for ADC. Obviously you want your calcs to be less than the 3V max, but close enough to use full range.

e.g. ADCs are 10bit, therefore approx 1000 values for 3V - resolution 3mV per LSB.

If you have a simple 12V battery system, say max volts = 15V, then need 15:3 = scale factor of 5 voltage divider. So voltage resolution is 5 * 3mv = 15mV etc


All ADCs are sampled at 10ms (i.e. 100 Hz),  the average is total/100, min and max for that whole second. Then start again for next second. Obviously scaled using User scale factors.

Speed - I use Hall switch (Allegro A3245) and very small (approx 5mm dia * 1mm) neo mags glued onto the rotating shaft (air separation approx 5mm). Each time magnet passes switch, a pulse is sent to a digital input on microcontroller. Microcontroller counts and timestamps pulses. Hence the need to know how many mags on shaft. More mags on shaft - more accurate rpm/speed (important for anemometer). Mags should be equally spaced! Best to use 1, 2, 4 ...

Uploading of program is via USB - no extra hardware required. Although you will need the Luminary Micro programmer. Free download from them.

Steve if you get a chance can you post the next video for me ...

Derek, you mentioned that the A/D inputs on the card are 0-3v.  I guess you have your current sense resistor on the ground side of the battery?

For the voltage input, do you take many samples a second and average or do we need to do some RC filtering?

How did you do the speed inputs?  Did you clamp the voltage of one phase of the alternator with a resistor and some diodes to the power supply and run that into a digital input?  Or does the speed come from counting ripples on the voltage input?

Can this microcontroller board be programmed through the USB, or do we need a separate programmer?

As soon as I can figure out what I need, I'll be reducing Digikey's stock by one.


-- Edited by sjh7132 on Thursday 28th of January 2010 02:02:36 AM

keithturtle wrote:

---

Digikey has 16 in stock

http://search.digikey.com/scripts/DkSearch/dksus.dll?lang=en&site=US&WT.z_homepage_link=hp_go_button&KeyWords=EKK-LM3S6965+&x=14&y=22

It'll be 15 in a few hours

Turtle, jumping in

---

Before you order you might want to figure out what current sense resistor / voltage divider resistors you need and put them on the same order.

 

Digikey has 16 in stock

http://search.digikey.com/scripts/DkSearch/dksus.dll?lang=en&site=US&WT.z_homepage_link=hp_go_button&KeyWords=EKK-LM3S6965+&x=14&y=22

It'll be 15 in a few hours

Turtle, jumping in

Keith

Not used Arduino - people say lots of good things about them. I think you have to buy into the Arduino way of doing things to use their kit.

The ARM cortex-m3 chip used by LMI and others is very powerful and feature rich, and you can use standard C/C++ and free development environments to develop.

Also the 6965 kit is just what I wanted for my projects, plug in ethernet to PC - easy.

I think LMI would do well to consider piggy-backing other boards and making a more modular hardware system like Arduino. Particularly wifi, and zigbee type radio comms.

In researching this project, did you consider Arduino?

I have them, but have yet to learn them

Turtle, slow

The clue's in the name ...

Jerry

Yes about $70, SanDisk micro SD card 2GB about $8.


Steve

SD memory is a constant size (and small kb) - you can run the microcontroller until it dies.

History data is cyclic - store the last 60 minutes, last 24 hours, last 7 days, last 53 weeks then start again ...

The microcontroller and PC are independent, only need connect PC to view data.

The power to the microcontroller is ONLY through USB, no power -> no data collected.

No calendar time onboard microcontroller - gets it from PC when connected.

If microcontroller powered down, then when re-powered and synced with PC - will update time and reset data for power down period.

e.g turn off for 2 hours -> last 60 mins = all zero, last 2 hours = all zero ...

If you're going to power down for long period, and have data you want --> use backup facility. Can also CLEAN all data for a fresh start.

The 4 ADC's on lm3s6965 (hence the limit of 2 power ports -> 2 Volts + 2 amps) are ranged 0 -> 3V (STRICTLY) . Hence the need for, a simple step down circuit from Vturbine to 3V max is required -> see the user config section of video for ScaleFactors for ADCs. A couple of resisters basically for volts.

I'm using a 10mOhm current sense resister designed for the purpose, it does not look like it's going to fry!

Sorry, I'm full of questions about this system.

Can the averaging period be changed?  With the 1 second averaging what's the memory usage on the SD card per day? (I'm wondering if you had a few GB SD card if it would be good for a month or so?)

If it had another power supply will the software handle being disconnected from the USB and just storing data?

Is the data time tagged with real time (is there an on board clock), or is it relative.  (I'm wondering if I powered it by  the mill itself, and the wind stopped, how it would handle the gap while it was off.)

There must be some limits to voltage and current.  An A/D that's meant to work on with 12v would either hit it's limit with 250v or would have limited resolution at 12v.
And with current, if you put in too much the Shunt resistor will fry.

DerekSoftStuff, are you an EE or mostly a software person?

Project wrote:

---

The controller is only $69.00 at Mouser but they have none in stock.

http://www.mouser.com/ProductDetail/Texas-Instruments/EKK-LM3S6965/?qs=vnwGVgFuQiaEThtYIQjC0g%3d%3d

Thats cheap unless I read it wrong.

---

The price is right for a microcontroller demo board.  How did you figure out which board it was?

 

The controller is only $69.00 at Mouser but they have none in stock.

http://www.mouser.com/ProductDetail/Texas-Instruments/EKK-LM3S6965/?qs=vnwGVgFuQiaEThtYIQjC0g%3d%3d

Thats cheap unless I read it wrong.

Jerry

Cooooooooooooooool!

dereksoftstuff wrote:

---

Steve

Have you got an audio alarm on any new posts on the forum, or perhaps just a bad sleeper ...

---

No, just email alerts.  It's morning here so I was awake. :-)

 

Steve

Have you got an audio alarm on any new posts on the forum, or perhaps just a bad sleeper ...

1) free - both microcontroller and PC programs - will post on my share site with next video

2)  http://www.luminarymicro.com/product/lm3s6965_ethernet_evaluation_kit.html

3) none

4) No

5) It's up to the user to setup the loads - this is just a tool to measure performance.

6) Anemometer is DIY - will include build details in next video. Will have to calibrate yourself.

I like it!  I was thinking of doing something like this.  The video brings up more questions than it answer through, maybe they will be answered in part 2.  But they are (so far.)

1) Is this software for sale or free ware?  If for sale, where, how much?


2)The hardware looks like a mix of a commercial board and custom electronics.  Is there a plan to produce the custom electronics (or the whole thing) on a board?  Kit form?  

3) What are the limits for the voltage and current inputs?

4) Is there any practical limit on the speed input?

5) Can the circuit load the turbine to make bring it to different operating points / TSRs to find the peak power point?   Or do you do that manually by changing the resistive / voltage load?

6) Is there a reasonably priced anemometer for sale that interfaces nicely to this system? (and is already calibrated)

Very nice!  I want one.

Just reposting with a more friendly link.

Time to get the performance ball rolling ...


http://www.youtube.com/watch?v=djoMJORGAog

-- Edited by dereksoftstuff on Tuesday 26th of January 2010 03:10:01 PM