Hi,

I would like to know the best way to extract dc power from a dc motor that rotates in both directions.

I'm building an human energy harvester, which uses knee joint movement to rotate a shaft, connected to a dc motor via a gear train. I want to extract power from the flex/extension phases, in which the leg moves backwards/forwards around the knee joint, and the dc motor rotates clockwise/counter clockwise. I want to use a circuit which will output the maximum amount of dc power.

I was thinking of using a bridge rectifier, but I'm not sure if this would work/ if there would be too much voltage/power loss across the diodes. Are there any better circuits to use?

thanks!

 

It's true, a bridge rectifier will drop the voltage by 2 diode drops - about 1.4 to 1.6V for silicon diodes. For better efficiency at low voltages you could make your bridge using schotttky diodes. These have a voltage drop of about 0.3 to 0.5V each diode. They only go up to about 40V max reverse though.

It is possible to replace the diodes with power MOSFETS (that have a diode inside anyway) and drive them into the ON condition when they need to conduct (when reverse biased). Then, the voltage drop can be very low - just the voltage drop caused by the ON resistance of the device. But you need to design a circuit to control the MOSFETS - and this itself will use power.

It will be important, especially if your MOSFETS are in a bridge rectifier, that opposite devices do not conduct at the same time as this would short out the generator and possibly damage the MOSFET. So the driving circuit is not straightforward.

 

thanks JDT! The schotttky diodes sound like a good idea, and the 40V max should not be a problem if I use a 12V motor.

I've seen another knee based energy harvester design, which has a potentiometer connected to a shaft, to work out knee position/ control the power supply circuits, although it was using a brushless dc motor and would have different wiring.

If I was to use mosfets, I would have to use the pot in a voltage divider with the output connected to some kind of comparator circuit which would power the correct mosfets. I have no idea how to build an energy efficient comparator circuit? or what it would look like but thanks for the idea.

 

Perhaps a "motor" isn't the best possible way to extract energy from this type of motion.

Have you seen those flashlights that don't use a battery? The ones with the large magnet that you shake back and forth inside a tube?

If you had two small spring mechanisms and one of these power generating tubes with a magnet inside. The two springs could be used to rapidly turn the tube, with one end up and then the other. Something like the escapement mechanism on a clock. One side up, then the other side up. The knee movement would just trigger the latch which lets the spring move the tube. the faster you walked the faster the latch would 'tick' and the tube would just flip flop end for end stopping at each 180 degree point.

The magnet falls through a large coil of wire and generates a large voltage spike which is used to charge up some large capacity capacitors, then you simply use a power supply to harvest the output from the capacitors and achieve a steady voltage and power level. It would even continue to output for a short while after you stopped moving, until the charge on the caps falls to low to drive the power supply chip.

The trouble with a motor is getting a steady rate of rotation and keeping it. The rotation also has to be very fast to achieve usable power levels. This would mean gearing it up ALOT. That gearing would create some noticeable resistance to movement; especially from a stand still. Each change of direction of movement involves a stop and then a start in the opposite direction. That would mean a flywheel would be needed to maintain rotation and fancy expensive clutches that could keep the motor rotation going.

The idea of letting the motor stop and then turn the other direction would not work as well as you are thinking. there is inertia to consider for one thing. The motor rotor has mass and wants to keep on turning the same direction. Stopping it, and reversing its direction would waste lots of possible energy generating movement. the motor wouldn't generate useable power until it was moving at a pretty high RPM which takes time to achieve.

Probably should rethink the motor approach.

 

Have you considdered harvesting the power with some other medium, air perhaps.
I imagine that a double acting pump would be easy enough to construct around a joint, perhaps even both knee joints.
Single tube then to your generator, a micro turbine perhaps like those used in dental drills, not that much torque but speed by the bucket full, just what you need to spin a brushless motor fast enough to make some usefull voltage.

you can get dental turbines, cartridge type with the bearings included, on ebay for less than £20.

I'm not saying it 'WILL' work but i dont see why not right now the main advantage would be easy gearing, all you would need to do is vary the pump displacement ... one leaver!! The same strategy has been used in spain to harvest wave power with a segmented cylinder the uses hydrolics at each joint to achieve a unidirectional oil flow, from the reciprical motion, that then drives an alternator.

(Just musing )Al

 

Yeah!

and put those coil generators I mentioned above in the shoes and harvest that energy as well

 

thanks for your replies!

Kermit2, I'm interested in the mechanism used for flipping the tube. The spring mechanism used to flip the tube what does it look like, and how is it flipped by knee movement? If its like a clock would the spring have to be wound up first?

Also with regards to the motor, I never considered that it would be continuously rotating, rather that it would only generate power during the high velocity phases of knee flex/extension. Based on my calculations to get max power using an angular velocity of half the unloaded RPM of the motor, a gear ratio of 1:67 is ideal, which I don't think will put too much strain on the person's leg.

For a picture of the design, I've used as a guide,
http://www.gadgetreview.com/2008/02/out-of-juice-just-strap-on-this-knee-brace.html

But thinking on what you said, it does seem like alot of energy would be wasted in reversing the direction of the motor. In the design above they use a roller clutch to just use one phase(extension and not flexion) to ensure that the motor only rotates in one direction. Although from what I can see, there is no flywheel to make sure the motor keeps rotating continuously. But maybe adding a flywheel would be an improvement/ and then gearing it up slightly more, to get fast continuous motion. I'll have to test this.

and Dyslexicbloke, initially I wanted to use hydraulic intensifier circuit(that sounds like the wave harvesting circuit you mentioned) to apply force to a piezoelectricstack, rather than a generator, before realising that couldn't design anything good with hydraulics, and I really wanted to have something i could build and test easily/ plus i've always been more comfortable with electrical components than anything mechanical.

 

Yes the mechanism I pictured would be a clock mechanism and wouldhave to be wound up. And like a clock it would 'tick' for longer than you would care to walk strapped in the device. Further thought leads me to think that it wouldn't be as good an energy harvester as I thought, since the speed of the magnet would be determined by gravity alone and such a short fall would limit speed and therefor the voltage. The shoes would be better since you would get the inertia 'kick' and much more initial velocity on the magnet... springs on the tube ends could lend repetition to the travel. Pick up one of those flashlights that uses this mechanism and look it over next time you are in the stores shopping and/or looking at the ladies.

 

The reason I mentioned air is because you could potentally experiment with diferent turbine types.
A Turgo turbine for example rotates in the same direction with airflow in either direction so you wouldnt need any valves at all, just a bellows at the knee and a single tube.

I have no idea how efficiant you could make it, particulaly at the experimantal stage, but the losses arrising from reversing the rotation of some relativly heavy rotor will be huge !!
You sighted the design with a clutch/ratched, I suspect thats because you would waste more energy reversing a mechanisme that you loose by only harvesting half of what is potentially available.

Linear generation, the shaker tourch, might be a possibility if you were driving the magnet or coil directly but that brings you back to reversing the travle of a mass.

Also remember that the faster you can move the field relative to the coil the better things will get.

I am not pretending to know much about this stuff so I may be way off base, I just find the concept interesting.

Good luck

Al

 

Kermit2: Isn't a shaker torch being flipped round over and over again, kind of acting exactly like a motor generator? except a motors design would be optimised for rotation. But I probably will end up buying that flashlight, could be some good exercise for me.

Dyslexicbloke: yeah you might be right that it takes more energy to reverse motion then just using energy from half the phase. maybe (using cw/ccw rotation) extra energy will come from an added stress to the users leg, and that they may have to work harder to walk with the device, but ultimately this will lead to higher power production.

Or maybe it won't. I plan to build a lego NXT simulation of movement round the knee joint, test some variables like gearing/ taking power from diff phases to see what charges some batteries the quickest.

 

A Turgo turbine for example rotates in the same direction with airflow in either direction

Sorry talking complete rubbish ......
This should have said 'Wells turbine' a Turgo is not at all sutable.

Al