MMG - a high ohm motor with potential

[Slider2732]

The 'Dolphin' thread was beginning to incorporate many divergent ideas and I thought this motor deserved a thread of its own.

The MMG is a concept resulting from Lidmotors excellent 'Frog Hair' motor video, which can be found here:
https://www.youtube.com/watch?v=B0Nu4dKhUHg

MMG - Minnesota Mountain Goat, refers to the quip about the flatness of MN, when I moved to the USA from England.
The line being about tiny thin gauge wire "Thinner than a hair from a Minnesota Mountain Goat"...which of course don't exist.
Lidmotor's motor uses 2 coils of 2000ohms each and a simple switching transistor circuit, with an LED across the flyback from the pickup coil.

My idea, is to use DF coils (Dancing Flower solar dancers from Dollar Tree), to first of all emulate and then to carry the concept forward. Full idea, is to use principles of the Lasersaber EZ Spin and coil energy pickup, to see where we can go toward highly efficient running.

As such, the great news is, that the motor runs on 4x DF coils in series for the trigger coil and 4x DF coils as the pickup coils.
This first setup is very rough and scruffy, but works.
2x DF coil based motor coils - 1923 ohms and 2033ohms
Transistor - MPSA18
Resistor - none !
Power - 0.1F supercap charged up with a 3V coin cell, then the coin cell is removed once it starts up.
The motor ran for 25 minutes on its first test run.
From 1.683V to 1.523V during the run took ~5 minutes
The thing is, the LED is lit all the way through and offers capture potential through a diode.
Nothing at all is optimised yet, in fact the rotor started to wobble badly a few minutes into the run.
A very interesting phenomena, is that the motor needs a high spin velocity to start, but after which, it carries on even with much lower RPM's than it took to start it up !
Also, on connection of anything above 1V, the slightest of spin of the rotor will trigger the circuit into an 'on' condition, whereby the red LED lights continually. The rotor can then remain stationary or be removed and the LED stays lit.

Here is the video of the first run, apologies for the state of the build table:

[PhysicsProf]

 

Very good, Mark - congratulations on that first run, and no resistor!
Below are two circuits we've been discussing lately; are you using the one on the left?  (sans resistor)

[Slider2732]

Cheers Steve
Yep, it's the one on the left.
All magnets are North facing out, which indeed is the same for the Andrews circuit also shown.

The decision to remove the resistor was when trying out some disc magnets that arrived yesterday. At 3mmx1mm they were intended to be stacked 3 or 4 deep to optimize rotor balance and triggering and to clean up the look of the rotors. Turns out that there is a world of difference between N35 neo's and N50. These are too weak and don't induce much at all into the coil.
On changing to some cut pieces from a computer HDD magnet, the motor fired up.

Next is to put a variable pot onto the Base, then current measuring for differences and to reroute the LED output.
After that, another stack of 4 coils (as 1 collector coil) to be an energy harvester.
What I really like, is the way that the other systems have been erring toward around 1V as best efficiency...but this motor type spins a lot faster than they have for similar voltage and current draw. It means that more energy can be obtained from pickup methods.

[ion]

Just a few rhetorical questions about the EZspin project:

What fractional horsepower measurements have been made to ascertain that there is indeed a benefit over miniature and ultra miniature motors of conventional design in all of their various incarnations?

What practical use can one imagine for such a motor that has not already been provided by such above named motors of conventional design?

If a motor just spins for a long time, and can do no practical work by delivering fractional horsepower to a load, of what use is it?

Where do you envision the research will lead? What is the goal? Where are the goalposts?

[Slider2732]

Good questions ION.
Some of which I can answer, some of which are for everyone who wishes to develop efficient low power motor designs.
Rather than remain rhetorical, it would serve well to answer them here.

My own goals have included efficiency of run times, further than a previous record of 19 minutes for a 1F supercap some years ago. But, in doing so, the learning encompasses related common topics such as the SSG and small sized battery charging.
There's every reason to develop better rotor designs and circuitry, to do the same jobs.
If, and it's a big if of course, remote rotor pickup by magnetic coupling can return more energy than is being used by the motor, then that is of interest to everyone in our field of research.
EZ Spin design derivatives use less energy the slower they go. When a remote rotor is introduced, the speed of the motor is impacted through the magnetic cogging. However, that negative impact can be turned to an advantage. The remote rotors generate effectively and only the careful timing to magnetically lock with the driving rotor forms the limit to the amount of remote rotors that can be employed as generators. Moreso, remote rotors can have other remote rotors spinning from their own magnetic field interactions, to no detriment !

Horsepower measurements are of the wrong order of magnitude. They are also of little importance to, for example, a solar powered bathroom extractor fan, but would be for a pulley driven electronic curtains closer.
Practical uses are a good area to explore, else yes, some of the research could be deemed folly.
On viewing Lasersabers original videos of the EZ Spin concept, he demonstrates relatively high torque for his low power design...rather than a drinking straw being used to demonstrate through noise and friction, it would be able to rotate a fan blade.
 
Regular motors of miniature design may include popular modelling variants of 4mm, 6mm and 8mm diameter. Of coreless construction they generate ~20,000 RPM and high torque. However, their efficiency is within those parameters and can typically deplete a charged 3.7V LiPo battery in around 6 minutes.
These high ohm motors serve a very different sector and function.

Also, sometimes there need be no practical use. It is, after all, the concept of novelty desktop toys which brought the Andrews Patent of the mid 1970's. From which, even now, products are for sale and that Patent still supplies people with employment.

 

[ion]

Slider, thanks for addressing those questions.

[Slider2732]

A worthwhile exercise ION, because others likely wonder the reasons for such research

There have been some good progress developments with this motor.
The rattling and wobbling is now gone from the rotor, due to much care and attention being given to the balancing. As mentioned, the new magnets were found to be too weak (N35), but the plier cracked HDD magnets were too far apart in sizings to allow for a smooth run at most voltages and speeds.
All is now much smoother. The result being, the motor will run down to a measured 1.086V.
Lidmotor stated there were troubles with his setup at below 1.5V and, because his constructions are never bodged together, it's my opinion that the MPSA18 is a more competent transistor for this motor than his MPSA06. The current gain is far larger for one thing.
What that translates to, is faster rotation speed, more efficient power usage and less wear on the tile underneath.

A flyback capture method was tried, that unfortunately hasn't shown positive results.
Patrick (Cool Joule on YouTube) showed a video where he changed the LED for a diode and cap return method to the battery.
A diode runs from the Collector, into a 1uF HV cap, the other leg of the cap being connected to the EB diode of a transistor. That transistor then connects to the run battery/cap Positive terminal.
His theory, was that the internal transistor diode that connects to the battery would limit throughput to 14V (MPSA06). The rest of the coil spike being contained in the capacitor and presumably allowing 14V discharges into the run source.
However, when tested here, I was seeing voltages on the cap of approximately only 0.6V higher on the cap than the measured battery voltage. In other words, the transistor diode drop difference.
It's still basically the route to take though for flyback capture, Joule Thief circuits being able to capture in similar ways. I just think that the cap is soaking the spike up without being useful in storage to actually do anything positive for the run source.  
Whereas, say in a SSG, that spike is the charging method for a power source.

[Slider2732]

Current measurements:
Meter used - UNI-T 136B

I wanted to see how close this motor of mine runs compared to Lidmotors original. Hopefully in the same ball park for figures and a good base point from which to calculate changes in any case.
Various configurations have been tried this morning.

First of all, the 0.1F cap was charged up to 1.4V.
Current readings changed every half second or so as the meter sampled, from 1.4uA (?!) to around 77uA. The average being pretty close to Lidmotors 40uA and indeed works as confirmation of his results

Then the cap was charged to 1.94V, with a 3V coin cell. A germanium diode,  plus 0.1uF ('104' orange ceramic) cap were fitted in series from the transistor Collector to the Positive terminal of the 0.1F run cap. The idea, was to isolate the returning backspike...because a direct connection of the diode stopped the rotor. It's all similar to audio amplifier isolations and seems to work fine on motor outputs. Of course, if it did nothing then it wouldn't be useful !  

Without the diode and 0.1uF, current measured was an average of 130uA
With the diode and 0.1uF 95uA

So, how about at 1.4V ?
No higher than approx 55uA, with an average of around 25uA


Pic attached, not the best but hopefully shows the method.

[TinselKoala]

Hi Guys

I don't have any of the high-impedance frog-hair wire coils to test, although I'm trying to find some. All the "dollar stores" around here seem to be permanently sold out of the "dancing flower" things.

But... I know that the transistor switching circuit is itself inherently lossy. So I'm wondering why not try a simple reed switch with this motor. The high-Z coils, switched by the properly positioned reed switch, should give really long runtimes on the capacitor power supply, and you'll be able to harvest the backspike in the usual way, I think, without the complication of losses in the transistor.

I've found that the tiny reed switch inside an inexpensive DIP reed-relay chip can be used effectively, and the contacts can be protected against arcing and welding by putting a tiny ceramic cap across the contacts.

My little motor uses some pretty low resistance coils so I'm using a lot more current, but the supply voltage is much lower than you can achieve with a transistor switch. Also, my little "minipulse" motor uses a bipolar rotor, so I had to pre-bias the reed relay chip with an external magnet so it would only switch on one polarity of the rotor passage. The coil of the relay isn't used at all, just the reed contacts.
(warning: strobotac in use in the following videos, flashing lights warning...)

http://www.youtube.com/watch?v=zdNdR8Uz_dA

Here's an earlier video in the MiniPulse sequence where I explain the reed relay driver circuit, the location of the "sweet spot" for pre-bias magnet position, and show the backspike collection system firing a NE-2 neon (using quite a bit higher input voltage and current though).

http://www.youtube.com/watch?v=NFLBrRfXfJ0

I think with the really high impedance coils you should be able to get similar action with your much lower input power levels, and you should be able to harvest backspikes too.

I would guess that the reed switch system would allow even longer runtimes than you are presently getting, using the capacitor power source at low voltage levels.

I'm still looking for some coils like yours to try this idea, but it would be pretty easy for you to do it if you can get hold of a similar tiny reed relay in the DIP package.

Like these:
http://www.allelectronics.com/make-a-store/item/rly-495/5-vdc-dip-reed-relay/1.html

[Slider2732]

Thanks for the info and vid links TK  

In the dolphin thread and the EZ Spin replication thread, a reed was used.
This design of Lidmotors is, in effect, 4 'compass point' coils around the rotor, in 1 bigger coil.
So, it will have no problems at all with a reed.

Those reeds on the last link are very similar to some that I unfortunately threw out some time ago. Thought they were resistor modules lol and only later realised they were for the handset detection in some old cordless phone handsets.
But now, the preference is for small glass encased types. I cut the leads off right at the glass join (carefully) and then solder 30AWG wires to those points. It reduces the metals pull.
You have far more experience with this sort of motor than I, so am very appreciative of your input

I'll try a reed on this setup. A focus, in some ways, has been to see how efficient the electronic switching can be. I don't like the rotor drag of a reed....but have been very surprised at the robustness of them so far. 10's of hours of running and only 1 failure, probably because of trialing that setup at 9V.
Envisaged run voltages are approx 1.4V to 3V.
Your ceramic cap method of spike protection is a great tip - thanks !  


Edit to add:
Some tips for the dancers.
Boxes of 48 of the dancers can be ordered from Dollar Tree's website (dollartree.com).
An alternative way of doing things for small quantities, is to go in and ask the staff when their next shipment is due. Go in on that day and you'll get the things...but as told to me "we had 9 cases of these last year and they went in 1 day".
Look on the base for the design date. In mid 2014 they changed to 5mmx3mm neodymium magnets for the swinging bit and they work brilliantly on these sort of motors. It's a cost less to think about.
Dates such as 2013, as seen on the dancing Devils for Valentines at the moment have ceramic in them. While useful in their own right, they are too heavy and bulky. They are the equivalent strength of the neo's though and how China will have arrived at 5mmx3mm.
Later models have a better glue to affix the coil to the plastic base than earlier ones. All will likely have the better glue now.
To remove a coil from a base, the most reliable method i've found is to use an electric stove ring. Put the ring on the highest setting (orange glow). Bring the plastic base to within 1/2" of the element, making sure that the coil is also over it. Wait a few seconds for the plastic to start bending and folding with the heat and then remove. Quickly lift the coil from the edge where you were holding it and the coil will pull off easily. If not, heat again for 5 or 6 seconds. I've just removed the coils from the last 4 of my box of 48 and haven't broken any by using that method.  

[Slider2732]

The motor has been converted to reed switch operation.
At 1.4V, speed has been markedly slower than with the transistor circuit, with current spikes of ~250uA.
Otherwise, the current is zero between reed closings..presumably average current will be (RPM / 4) / 250.
There is only 1 coil needed though, instead of 2 and a reed would allow much lower running voltages.
The rotating speed is what i've been looking at for energy harvesting, all else being broadly equal.
With an envisaged operating voltage of ~1.4V, i'll probably go back to the transistor set up, with its ~25uA usage for now.

Also, a 3rd set of 4 coils has been made, which has a slightly lower reading of 1923 ohms.
That will be trialed as a pickup coil.

[PhysicsProf]

  Good information for future builders, Mark - thank you!

  Can you tell which approach might be better for harvesting fly-back energy,  reed-switches or transistors?

[TinselKoala]

@Slider:
You may now be seeing duty cycle issues with the reed switch. To get the optimum pulse width you may need to play with the radial position, or even pre-bias the switch with a little magnet on the side of the switch away from the rotor. Optimum timing by circumferential position adjustment. My intuition tells me that you should be able to get the same RPM with the same number of coils, but at lower power levels, than with the transistor, once the reed is adjusted just right. This might be more trouble than it's worth though, since the selftriggering transistor system "automagically" takes care of these issues.

I also think that the reed system will ultimately deliver greater collapse spike than a transistor could, for a given input power,  because the reed is potentially able to switch with faster rise/fall time than the transistor will.  The tiny reed switch inside the DIP chip might be better for this motor than the larger discrete reedswitch that you are using, maybe. Do you have a link for the exact part you are using?

[Slider2732]

Thanks guys.
My problem is that reeds do always win out in own builds, removing the need for >0.6V and minimum current.
But it's been confusing to carefully set the reed and see a lower RPM, with also seemingly very similar current draw.
The fastest collapse is a goal and that should entail a MOSFET (small package type) or another way with no parasitics...which would be the reed.

To confuse everything, the motor was redone as transistor fired and now behaves in a remarkably odd way that I can't fathom.
Changing out the transistor removed the thoughts of the original starting to fail etc.
It only needs the firing coil, no pickup coil and will fire for a few seconds with a hand spin. The LED flashes strongly as it goes, with nothing connected to the Base of the transistor !
RPMs pick up, so it's not just the action of induction causing the flashing. The flashing subsides to a flickering glow and then the motor stops some while later.
Will get a video of it.
No matter what I do though, that flash of the LED can't be translated to a pulse to the Base.
Originally thought to be something SSG type related, it doesn't seem to be. I tried a 1000uH axial inductor and small ceramic caps in parallel, variable resistor and other ideas, in the hope that a delay could be introduced. The idea being to emulate a reed firing a few degrees different to TDC (to use a car term). It seems it should work, somehow, to the eye.

Here are the reeds:
http://www.ebay.com/itm/Snail-Mail-10pcs-10x-Reed-Switch-Glass-N-O-Low-Voltage-Current-2x14mm-/121149315417

[Slider2732]

Here's the video.
Of note, the rotor spins very well and then loses speed quite rapidly after a point. That point was when talking of the other tile that i'll swap over to. When I go back to the rotor, it's slowed down quite a lot compared to how it was running.
Everything is probably clearer in person.
The pickup in rotor speed makes it of interest, rather than mere induction spikes causing the LED to light. The LED is across the coil, so that was the first thought.
Also, the LED flashes at a much slower rate than 4 speeding magnets past the coil would seem to suggest would be correct. The flash rate you see in the short video is the flash rate in person.

Any help in understanding the effect would be appreciated.

[Slider2732]

Update -
Have swapped everything over to a better white bathroom tile.
The cause of non running was found to be a broken coil.
No idea about the perplexing run characteristics of the single coil running above, but in one way it was only that the other coil broke that allowed those findings.

Here it is running properly and from the table lamp as power source, using a solar garden light panel.
Panel is putting out 2.1V at 173uA.

Now to tidy up again and test out some harvesting ideas.

[Allcanadian]

@TK

You may now be seeing duty cycle issues with the reed switch. To get the optimum pulse width you may need to play with the radial position, or even pre-bias the switch with a little magnet on the side of the switch away from the rotor. Optimum timing by circumferential position adjustment. My intuition tells me that you should be able to get the same RPM with the same number of coils, but at lower power levels, than with the transistor, once the reed is adjusted just right. This might be more trouble than it's worth though, since the selftriggering transistor system "automagically" takes care of these issues.

I would agree, I found a solution to this problem by using two reed switches in series a little over one magnet diameter apart. The fist switch closes but does not close the circuit until the second one closes just as the first is opening because the magnet has passed it. As such the pulse width can be very very small as it is not determined by one reed switch but the difference between the two.

I have not tried more than two reed switches however I would suspect the pulse width could have extremely low values possibly ns however a consistent pulse width may become an issue.

AC

[muDped]

I would agree, I found a solution to this problem by using two reed switches in series a little over one magnet diameter apart. The fist switch closes but does not close the circuit until the second one closes just as the first is opening because the magnet has passed it. As such the pulse width can be very very small as it is not determined by one reed switch but the difference between the two.

AC,

That  is a truly ingenious solution to the problem
of finding a simple non-electronic way to adjust
pulse width!  A Magnetic And Gate with period
(Clock) overlap to achieve a desired pulse width.

Knowledge is a beautiful thing!

[Slider2732]

That really is a great idea and certainly worth trying out !
Will probably do that tomorrow as we have a forcasted 67F for the temp.

A personal goal has been achieved just now, something small but previously out of reach.
The idea, was to step up the game from the solar panel powering and move to light power with no solar panel
The system now runs from the CFL table lamp, using the same sort of slightly modified harvesting circuit as YouTube user 'Inventor3' demonstrated. My version has no connection between the first diodes and the join of the 2 output capacitors.
It's the same thing that runs the digital clock seen in some videos.
I use a CD or HDD platter as collection metal, an old radio antenna as extra pickup, that then is isolated with tape as it is held in place by a PC heatsink. The heatsink forms a Ground, which the circuit connects to.
Anyway, the result is a constant 0.98V, under the load of the running motor.

Here's a quick demo video:

[TinselKoala]

@Slider: that harvester circuit is really neat, you are powering the motor from "electrosmog" as I like to call it, and it's all around us as wasted power from our conventional systems. Finding ways to harvest and use it is one of my own areas of interest.
Thanks for the link to the reed switches, I'm ordering some this afternoon (but from China probably, at about 15 cents each...)

@Allcanadian: Very clever idea with a lot of flexibility, that use of two reeds in series. Thanks for posting it!

@Slider again: I'm wondering if those little reeds will close from the field of a coil wrapped around the glass envelope...   

--TK

[Slider2732]

@TK - Electrosmog is a good name for it and a term that average folks can understand. Smog also has that 'air' about it of us wishing to remove it and utilise it.
I somewhat dropped my guard with the present setup last night and it's one of those bizarre things that may not be realised from such a low output power source. The problem happens when a load is disconnected, the voltage builds up. I had simply removed the HDD platter from the top of the antenna last night, to let the motor spin down, because the input isn't high enough without the platter. Got up today and put the platter back on while making a coffee. Went to start experiments and turned the meter on...14.3V haha
With 16V capacitors things can get hairy for safety, because left unchecked i've had 35V showing on such circuits !
Must say though, that the thought of a cap blowing from the emissions of a table lamp does appeal  

Hmmm, to run at 60Hz you mean ? that would be very cool.
If you have a 3V or so wall adapter you'll often see the mains hum carrying on through to its output. A simple pulse motor rotor and a coil directly across the adapter output wires will allow that motor to run at multiples of 60RPM. No circuit required. It takes some deft flicks of the rotor to get it synced up, but is neat to then see it running.

[Allcanadian]

@TK
Hello TK,
It gets better when we run with this thought. I imagine we all have had to address the same issues at one time or another and one of mine was very high energy impulses with extremely sharp rise and fall times in the low ns range. They just didn't make an affordable single mosfet that could handle large current at high voltage because the gate capacitance would always rise in proportion to the current rating. That damn gate capacitance meant a large mosfet driver was required to even come close to the spec switching times. Then it occurred to me...why not two?, why not use a 556 dual timer with two separate frequencies to drive two cheap mosfets in series as one switch. If the real ON time is the difference between the two input frequencies to the two mosfets then two slow switching mosfets can have any ON time into the low ns range so long as the second mosfet turns on just as the first mosfet turns off.

It opens up quite a few possibilities simply because we could have switch pairs scattered about a circuit routing power here or there based on a single threshold or the difference between two or more and it solved many problems I was having which seemed to have no solution. I guess it is true that necessity is the mother of invention because most of the things I want to do in my circuits seem to have no easy solution if any at all so I have to keep making up stuff as I go along.

Edit: I should mention my nemisis has always been common ground paths and isolation with mosfet switching so if anyone has any suggestions I would be happy to hear them.

AC

[Slider2732]

@AC - that's some pretty good stuff you've made up then !
Makes complete sense, split the difference.
Plus, opens up non electronic means, such as the reeds.
A question though. The timing difference for a pulse motor would be based on the speed of the rotor ? As such, would the rotor have to be speed governed or do you think there is a range of fast switching..as in, above 1000RPM the system will work and any faster would be extra performance.
That may be a little link to the 60Hz thoughts. To use a known frequency as speed governor. AC rides on top of a DC voltage readily and need not draw power from a wall outlet as such, but use electrosmog for the function.


Here's a quick vid. Nothing too exciting, I just wanted to explore the properties a bit of micro input power and load handling.
The table light was switched off and the 470uF cap allowed to discharge down to around transistor switch on voltage. Then, the light was switched on and I wondered how things would stabilise out, indeed would it do that or would the rotor stop.
(the blurry imagery clears up after the first few seconds)

[Allcanadian]

@slider2732

A question though. The timing difference for a pulse motor would be based on the speed of the rotor ? As such, would the rotor have to be speed governed or do you think there is a range of fast switching..as in, above 1000RPM the system will work and any faster would be extra performance.
That may be a little link to the 60Hz thoughts. To use a known frequency as speed governor. AC rides on top of a DC voltage readily and need not draw power from a wall outlet as such, but use electrosmog for the function.

I discovered the 2 switch reed switch while building Ossie Callanans motor way back when which is still one of my favorite circuits -- http://www.syscoil.org/medias/pdf/documents/a_working_radiant_free_energy_system.pdf

It solves many of the issues concerning reed switch arcing which is an issue and the rotor switching at speed does become an issue as you suggested. One option is to use a secondary coil instead of a PM to bias the reed switch such as Ossie was doing in which case the changing induced current in the coil biases the reed switch to produce more consistent results. Basically we want to optimize the switching at the sweet spot of the motor where the input and output/drag forces balance.

I'm not even going to tell you how many damn reed switches I have arced together, it always starts with a small click and ends with a big puff of smoke, lol.

AC

[Slider2732]

Thanks for the further info.
TK's cap trick seems to work fine across a reed, or an LED/snubber diode. Very gentle things though and I hate it when removing a rotor say and a magnet sticks directly to the switch !
Have heard a lot about the Ossie motor, without knowing much about it. Thanks too for the PDF.


Here's another goal met, albeit again only a sideline small thing.
To run a motor from dirt outside in the back garden. It's been a bit of a quest for some time. One thing that would have to happen, no watering or salt added ! it hasn't rained here for over a week, making it a good day for such a test.
It also allowed another test of micro power, from a different source.
A line of thinking being, that if some energies are around us all the time, then a few uA of a problem area preventing self running needn't matter. A device would run anyway.

Am going to build a crystal radio