User Tools

Site Tools

Navigation Menu


  • StatusClosed
  • OP-ModePre-Launch
  • LocationN48 - E11
  • Localtime20:25
  • CountdownT-00D 00:00

Hot Projects


SEEDStack - Open 3D printable seed/sprouting systemDIY Food Hacking


UCSSPM - Unified Clear-Sky Solar Prediction ModelOpen Solar Power


picoReflow - DIY PID Reflow Oven Controller based on RaspberryPiDIY Reflow Soldering


PiGI - DIY Geiger Counter based on RaspberryPiRasPi Geiger Counter

DIY ARA-2000

Active Wideband Receiver Antenna for SDR - ARA-2000Wideband Antenna


DSPace - Map everythingMap everything!



The Tiny Vertical Axis Wind Turbine is a prototyping model for the Zephyr Wind-Park Construction Kit.


Gathering the material for the first prototypes. Do you1) have something available for this project? Please add yourself to this list and describe the parts, tools or experience you have to share.

TODO: post a list with all the parts needed for one TiVA.

NICE TO HAVE and still searching for this project: Someone with the ability to establish FEM simulations of different rotor type models and mechanics to analyze stress points in the mechanics and to optimize the rotors performance.

Alex Shure

Has access to an electronic lab and some parts at work, which could come in handy for a TiVA prototype. –shure 11:23, 3 April 2012 (UTC)

  • 16 high power RGB common anode LEDs2)
  • 6 AA NiMH cells, 2950 mAh
  • 1 battery holder for 4 AA cells
  • 4 MSP430 dev kits with debugging and hardware flash emulation.
  • 1 Arduino Duemilanove, Atmega168
  • 1 Arduino UNO, Atmega328, Atmega8U2
  • 11 NRF24L01+ 2.4 GHz wireless transceiver modules3)
  • 8 LM2596 DC/DC step-down buck converter modules4)
  • 19 IRF530 n-channel MOSFETs (no logic level types)
  • various SMD resistors, also some shunt-suitable values in 1206

Detlef Schmidt

Detlef offered to build at least one prototype for our wind turbine project.

TODO: write a list with resources, that we need for 1 TIVA.

shure: I would like to deploy 1-3 of these turbines at a nearby off-grid mountainbike downhill track, for which I am developing a MCU equipped timekeeping system. I hope to gain the interest for renewable energy / wind turbines of any passenger who rides or cheers there at a race.

Prototype BOM


Part Sum # PPU
magnets48 24 2
copper 0,5mm, 300g15 3 5
epoxy 5 1 5
discs 10 2 5
screws 2 1 2
Summe 80
Rotor	Achse	5	1	5
	Rund für Lagersitze	10	2	5
	Lager	6	2	3
	Schrauben	3	1	3
	disc mit Flügelaufnahme	10	1	10		Lochmaße
Summe	Rotor	34					34
Wings	Holz	36	3	12
Summe	Flügel	36					36
Electronic	Hühnerfutter	2	1	2
	MCU-board	30	1	30
	Schottkygleichrichter	3
	Synchrongleichrichter	10	1	1
	LED treiber	3	1	3
	hp RGB LED	4	1	4
Summe	Electronics	52					52
						Summe für 1 TiVA:	202


Power: (formula from the wiki entry)

  • h_1=0.32 m
  • d_1=0.32 m
  • A_1=0.1024 m^2
  • h_2=0.48 m
  • d_2=0.32 m
  • A_2=0.1536 m^2

Assuming a bad (20%) or decent (30%) turbine design and averaging with \rho_{turbine}=0.26

A rather bad permanent magnet alternator with \rho_{alternator}=0.75;

A normal synchronous rectifier with superb-by-design perfomance of \rho_{rect}=0.98

A buck-boost inverter with a good performance of \rho_{rect}=0.85;


Conclusion: A 0.32 x 0.32 drag-only VAWT generates about P_{mech} = 0.1...10 W and in average German wind conditions 5) about 0.5 - 1 W. If we have a good alternator (which will be easier at this size because of the high rpm) and a synchronous rectifier (rectifier not necessary if buck/boost power supply doesn't need DC, are there suitable packages for this mode?), most of the power will be available as an input for a buck/boost converter, which can operate reasonably well at these small power ratings. Chrono developed a PDU which contains high-efficiency buck-boost inverters - maybe a small scale version can be powered directly by this wind turbine, generating only 5 V and 3.3 V, omitting the 12 V.

<note tip>lx: I would go for a 0.32 x 0.48 m^2 VAWT and I guess the overall efficiency will be about 0.08 - 0.1.</note>

Assuming a worst-case average electrical power of 1 W after rectifying and regulating, one can still charge a cheap 4-pack of NiCd/NiMH (4 x 1.3 V = 5.2 V) which provides power for the system and for high power demands, e.g. activating the LED pattern at night. <note>Charging all of the cells with 1 W from 0 % to 100 % takes (4 * 4 Wh) / 1 W = 16 h. At a wind speed of 8 m/s = 28,8 km/h and P_{el} = 48 W * 0.16 = 7.68 W, the batteries will be fully charged in just (4 * 4 Wh) / 7.68 W = 2 h 5 min.</note>

One AA cell contains 1.3 V x 2500 mAh = 3.25 Wh of stored energy. We don't fully discharge the batteries, thus only 3 Wh will be used. However, taking charging and internal resistance losses and a safety margin into account, we need about 4 Wh of energy to store and retrieve about 3 Wh of energy.

4 AA cells equal 4 x 3 Wh = 12 Wh of energy. Without simultaneous recharging, this is enough to provide:

- five hours of one hp-LED shining at full brightness in white color or - ten days of one hp-LED flashing at full brightness with one color at a duty cycle of 10%, e.g. on for one second and off for nine seconds. - one MSP430G2231IPN14 16bit micro controller working for ages, at as low as 2V, it may consume 1 mW = 1/1000 W. Typical no-load best-case values from the MSP430 datasheet: 0.1 μA RAM retention 0.4 μA Standby mode (VLO) 0.7 μA real-time clock mode 220 μA / MIPS active

Excellent values! An 8-bit Arduino looks pretty old school against these numbers. ;-)

# lx: I have 6 MSP430 in a DIP form factor in my lab and 3 spare ti MSP430 Launchpad proto boards with onboard hardware emulator and debugger. Chrono, do you want one? ## hmm, I was thinking, I have the tools to develop avr but nothing to develop msp, if you have one dev-kit to spare and linux tools are freely available I'd like to consder going for msp as well, although I'd really recommend staying on avr for bigger projects since many people can do arduino now, so they won't have to much trouble with pure avr. Another arch always reduces the amount of people who can deal with it yet :( ### I have a spare dev kit in my lab. there's a msp-linux community … I don't have the tools for AVR, except an Arduino. So no debugging, HV-programming or hardware emulation. The full dev kit for an MSP430 is dirt cheap at $4.30, including two MSP430s in DIPs, a hardware emulator, spy-by-wire, debugging etc. I agree with the Arudino-publicity argument, and I would always try to incorporate an Arduino, as it is the most simple and comprehensing development tool there is for beginners. However, the ti.MSP430s are relatively new. A downside is their not-so-easy dev environment. Eclipse or IAR or propietary, free software from ti can be used. I have not yet experimented with it, but I have Arduino experience. It would be new for the both of us.

pro MSP430, con Arduino:

- the price! can be bought with a programmer for $4.30 vs Arduino $25 or a third-party Arduino for maybe $18. This is a serious difference. - even the single MCUs are cheaper, also, the AtMegas for an Arduino bootloader are hard to get. - less external parts for operation at high speeds, Arduino/atmega168 and 328 need an external oscillator to operate at full speed (16 Mhz) - runs stable over a wide range of input voltage down to 1.8V - an excellent sleep mode with RAM retention at only 0.1µA and great power efficiency. 220µA in full operation mode is an excellent figure for off-grid low energy applications. Almost no load to the turbine. Can also be powered by a “Joule Thief” and a single old AA battery, or just two old AA cells in series (3V). That should last for ages, at a constant current of 0.25 mA and an old battery of 1000 mAh, the unit will still run for 180 days, and the MSP430 can be operated with a supply voltage as low as 1.8V.

con MSP430: - less memory, but this depends on the package, (there are top-end msp430 processors which cost less than $1 vs an ever-expensive-avr) - less libraries available, smaller community

At a small production run of 10 TIVAs and the demand for USB ISP, Arduino vs MSP430 would equal 10*$25 = $250.00 vs 10*$4.30 = 43.00 (!)

A nice solution: ⇒ Write clean C-code and let it be compatible with MSP430 and AVR compilers. Some Arduino projects were easily ported to the MSP430.

In realtime without battery backup, the hp-LED may be pulsed at full power and 10% duty cycle at quite low wind speeds and 100% at >6.25 m/s.

<50cm long parts can be cut out at almost every small CNC machine.

48cm wings can be made out of:

- styrofoam, Styrodur etc with a hot wire CNC cutter - the famous 2-by-4s with a planer - like an R/C plane wing with wooden rips and a foiled surface - sheet metal, aluminium sheeting bent over cores [rips] - wooden sheet material - plastic pipes

fixed main shaft: Do = 8 mm, 608ZZ radial single race bearings rotating turbine assembly, rotor shaft where the bearings seat: Di = 22 mm

At this size, a single I-beam design should be suitable, not a dual-bridge-H-rotor assembly.

A V rotor looks promising, too. Resource demand is further reduced with this type of rotor. two bladed or three bladed? apparently, two bladed designs have severe problems with low wind conditions and self-starting issues? ⇒ three bladed. ack, three seemed most promising in most scenarios.

V-rotor advantages: * least amount of material for a given lift-type wing surface * best wing volume vs static structural volume ratio * only one wing-fixture-point * no bridges, less moving parts * less connections, less machining operations, less screws or welds * dissassembly is easier * uses a higher surface at a larger height, less turbulences at the ground * (tbd) less prone to oscillations? * snow can't set onto most of the rotor * can be adapted to also use up-winds in urban environments, especially interesting at the top of buildings.

__      __    <-test if winglets make a difference                   
 \      /                                                            
  \----/    <- a rope to cope with centripetal forces at high rpm    
   \  /    <- wings in V-form                                        
    \/    <- plate with wing-fixtures and seats for the two bearings 
    ||    <- shaft/rotor coupling with two bearings                  
   _/\_    <- any type of stand or clamp, generator, electronics     

I was thinking that we can't have a reliable _absolute_ measuring device, so if all devices are built the same way then we can have a _relative_ measuring device… That is right, because there is not wing-tip-speed ratio at drag devices. no-load drag wings have a wing-tip-speed ratio of one, thus being as fast as the wind ;) we would have to have half-cups as wings to form an actual absolute wind speed measure device, like those things you can buy..

build a lovely grid and show, that wind turbines can be fun. it would be easy to deploy and portable, system voltage of 5V would provide charging power for mobile phones etc. USB power output would be cool, fed by a 5 V buck-boost converter and 4 AA cells. measurement is difficult anyways, because devices would have to be calibrated in a (diy) wind tunnel or something, to get reliable results…

*can serve as a measure+log device for wind speeds *has on-board electronics: switching power supply, 3.3V or 5V system voltage for MCU and electronics+LEDs, goldcap ?, mcu recommendation: either a low power ti MSP 16bit on a launchpad or the ordinary Atmel Atmega 328(pu) with an Arduino bootloader (or derivative) → both would be diy-friendly and cheap. logging shield with shunts and opamps, goldcap, hprgb shield with logic level mosfets, software pwm. *can be deployed on a field, in an urban environment etc and may be connected (for example cheap NRF24L01 node-based-network) *one high power RGB LED acts as a universal signal: can be an indicator for wind speed, keep-alive.. or a 3 x 8 bit digital pixel. If deployed in an array on a field or in an urban environment: in low natural light conditions, at sundawn or in the night, the wind pattern can be determined by the flash+color pattern of the small wind turbines. MCU logging onto e.g. micro-sd card or via wireless link is an optional step. It would be a very cool art piece at night if all turbines would be connected to a master (which would be easy outdoors on a field) or connected in a grid. A pattern could be generated and all turbines could flash in sync. single flashes could be emitted with full power even if the wind conditions are bad, just the off-periods may be pretty low then.

small VAWTs could be attached to the top of a tree, especially to free-standing ones. No pole required and higher wind speeds gained: win-win

For a rough estimation, if the VAWT is working and how the wind condition is: Stick a thick wool or thin polythene (bin liner) tell-tales onto the top of the blades. This gives an indication of the relative speed of the blades and it is quite simple to see if the turbine is just being blown around by the drag on the downwind rotor or 'actually' running.

-A tell-tale in the centre of the rotor between the blades; so one can see the airflow through the rotor. As the rotor starts this will still blow out sideways, but when the rotor is running (without any load), it will hang limp indicating very little air flowing through the turbine. A gust or putting load on the turbine/generator will cause it to blow out sideways due to the wind which gets through.

1) yes, I mean you, my dear reader!
2) attached to aluminium star shaped heatsinks, each with three 350 mA rgb emitters, 3 Wcontinuos, 4,6 Wpeak. best light vs current value may be at 180-260 mA, still visible from long distances.
3) (3V3) populated on a small SMD board, PCB antenna
4) IN: 4-40(memo:check cap ratings!), OUT: 3,2-26, populated on a small SMD board. <200khz. iirc 70-90%, could be tuned with better coils.
5) needs further research, 3 - 4 m/s??