Adams Axial Pulse Motor

[unimmortal]

More from the oracle

Thanks Chet.

Ok, while I'm waiting for rotors to be printed and a new multimeter - I thought I'd push the limits and see if my little monster could impersonate a vertical wind turbine 

There's still a little too much reluctance to be completely usable, but with a second monster mounted 30° offset will diminish this significantly. This matters as I'm working on mating the little terrors together. So instead of magnet to coil core attraction happening every 60°, this will become much smoother occuring every 30°.

Axial Adams VAWT in the sun.

[Chet K]

Here another bit of pulse motor info from TinselKoala
https://m.youtube.com/watch?v=z0sjqoshznU

Respectfully submitted

[unimmortal]

Here another bit of pulse motor info from TinselKoala
https://m.youtube.com/watch?v=z0sjqoshznU

Respectfully submitted

Thanks Chet.

Interesting, I had planned to do the same test, but with a 2 hall sensors and 2 coils, instead of a sensor coil - and flip flop building/collapsing fields. I think the reason why TK was seeing better results without the rotor, is because of the AC wave the magnet would introduce. Ironically, that's where I'm able to pull a little bit of current from my EM coils.

[unimmortal]

Ok, new multimeter day! ...and the results are very promising.

On a 9V / 0.109A input on the EM coils (yes, there is current draw from the PS), the output on the generator coils are 36V / 0.406A (rectified). There's also ~5-7VAC flotating around in the EM coil. The electrical setup is simply switched power directly to EM coils, no other components.

With the EM coils to be re-wound, I'm in two minds whether to do these as bifilar?

[gyula]

Ok, new multimeter day! ...and the results are very promising.

On a 9V / 0.109A input on the EM coils (yes, there is current draw from the PS), the output on the generator coils are 36V / 0.406A (rectified). There's also ~5-7VAC flotating around in the EM coil. The electrical setup is simply switched power directly to EM coils, no other components.

With the EM coils to be re-wound, I'm in two minds whether to do these as bifilar?

Thanks for the data.

Would you clarify whether you used any load across the rectified 36V output? 

OR you simply shorted the 36V output by the Ampermeter and it measured 0.406 Amper?

If you did the latter  i.e. the "load" was the Ampermeter only, then please use a normal resistor load across the rectified DC output. Resistor value is not so critical, any value between 82 Ohm and 100 Ohm would do.

This comes from my assumption that you used the Ampermeter only as the "load" to get the 0.403 A, so 36V divided by 0.4 = 90 Ohm, indicating the generator coils and the rectifiers represent about 90 Ohm internal impedance. Note that if you load the DC output by a 90 Ohm resistor (to get max power transfer), the output voltage would drop to around 18 VDC i.e. would go halves.

IF all these assumptions of mine are correct, then you would need to check whether the 109 mA input current changes when you load the output with the 90 Ohm (or close to) resistor.

A correct evaluation would be to measure
-- first the output DC voltage across your chosen load resistor (any value between say 82 and 110 Ohm)
-- then measure the DC current the load resistor draws (connect the Ampermeter in series with the load resistor)
-- then measure the input current when the load resistor at the output is connected

You can do these measurements using the new multimeter in the above order I suggest.

Then you can do some calculations on the input and output power levels. 

Gyula

[unimmortal]

You can do these measurements using the new multimeter in the above order I suggest.

Then you can do some calculations on the input and output power levels. 

Gyula

Thanks for the simple steps Guyla, I'll get this informattion to you shortly.

This new multimeter is a bit like a swiss army knife. Frequency (measuring against alow-side npn transistor) is a cool inclusion which helps put a timescale against things... these results were all occuring at ~50Hz.

I'm currently performing surgery on the LT's to make them into a single device. Broadly summarising the changes: the EM coils will be alternatively pulsing at 60deg at each end instead of 120deg at one end, there will be double the generating coils in the middle, and attraction will be occurring every 30deg instead of every 60deg - all of which should make for some interesting results.

[unimmortal]

Ok, here is the latest iteration of the Adams Axial. It has now evolved enough that I'll detail the various aspects for anyone playing along at home.

Physical - To contain the pull force of the magnets I've had to move to a more user friendly design that allowed easy testing of various coil housings, and a way to have them isolated from the bearings. The pine board is ~300mm dia x 18mm and the threaded rods and hardware are M10.

Magnetic - Each of the 6 holes of the rotor has 2 20x20mm N40 (10gm) Neo's (15kg pulling force each), with a 5x10mm N40 between them set in a 5mm shelf to hold it all together. The 6 coil cores on each side of the rotor are 12mm in diameter and are offset 30 degrees. The PCD of the rotor magnets is 80mm. Total rotor weight is around 350gm. Airgap between the coil cores and magnets is approx 6-8mm.

Electric -  6 electromagnets, each made with 28m of 1mm wire, wound on a 40x12mm core, wired in series, 4.2 ohms. 6 generator coils, each made with 70m of 0.5mm wire, wound on a 40x12mm core, wired in series, 50.2 ohms.

My switch leaks current due to the hall sensor being at ~2.4v, which is opening the NMOS just enough to flow (as it grounds a voltage divider) but not enough to hit gate threshold on the PFET - nonetheless, there is no increase above the .109A I see when I turn the switch on without the coils connected.

So with an input of 9V, I'm getting back ~16-20VAC on the EM coils, 18VAC on the Gen coils @ 4600rpm. I'm over the moon with these results. The magnetic reluctance is now so low that there is more resistance in a doorknob.

A couple of anamolies though, as the rotor increases in speed I see VAC at the EM coil rise to 5 almost 6, then the multimeter drops back to 0 before jumping to 16-20. I would have put up a photo of my LED train from previous posts, but as it jumps up to 16-20 it took out 8 LEDs... burnt out, dead, on the spot.

The second anamoly - with high speed I'm generating a great deal of EMF and hammering the PFET, so I can only run it for short amounts of time. I've added a lead to each join between the coils to measure what is happening between each of the coils. So with one lead of the multimeter connected to the join between the first and second coils, the multimeter returned around 20V. Now when I touched the other lead of the multimeter, this increases to 57V - not sure how to interpret that number, did I just become an earth reference for EMF or a path for EMF?

I've had this running briefly at 18V and it is scary how fast it wants to spin and how quiet it becomes.

[gyula]

Hello, Happy New Year!

You have achieved nice progress on this newer setup. On the anomalies:

1)  It is possible that as the rotor increases in speed there is a certain speed where the multimeter needs time to average out the AC peak voltage spikes that are created by the collapsing EM fields. The peak amplitude of the spikes gradually increases as the rotor advances in RPM.
[ A waveform check by an oscilloscope across the EM coils would help explore what is going on. A HV probe for the scope would be useful to protect its input, I mention this if you have a chance to borrow or have access to an oscilloscope. ]
The LEDs received the spikes and their voltage and current specs were surely exceeded (consider their reverse voltage rating too if applicable). 

2)  Yes, your body becomes part of the EM field the pulsing electromagnets and the rotating magnets create. The multimeter is an indicator here and serves as a rudimentary field strength meter with forever changing display as you move the leads, your hands etc. Your body conducts some nanoAmper displacement (capacitive) current via the 10 MegaOhm internal meter resistance.

Perhaps, when you feel like doing the measurement steps on this newer setup I outlined above in December, that would be great.   
Now you have about 50 Ohm internal generator resistance, so a 50 Ohm load resistance would be a good match for maximum power transfer. You can use a full wave diode bridge and say a 470 uF or even 1000 uF puffer cap.  Say you measure 10 VDC across the 50 Ohm resistor, the dissipation in it would be around 2 W.   

Gyula

[unimmortal]

Thank you Gyula, and a happy new year to you.

I've done a whole lot more testing over the past few days, and have almost arrived at a stable baseline - tuning out vibration is a painstaking process. Input consumption is around 9V@40-70mA w/ 10mm air gap.

Attached is the obligatory 2 x 12v globes running on the em coils' rectified emf output. The meters are showing current and volt outputs for the gen coils.

[bistander]

Thank you Gyula, and a happy new year to you.

I've done a whole lot more testing over the past few days, and have almost arrived at a stable baseline - tuning out vibration is a painstaking process. Input consumption is around 9V@40-70mA w/ 10mm air gap.

Attached is the obligatory 2 x 12v globes running on the em coils' rectified emf output. The meters are showing current and volt outputs for the gen coils.

Hi unimmortal,
You appear to have the Astro A1 meter connected incorrectly. Read the manual.
bi
https://www.astroai.com/user-manual/up/100019

[forest]

Yes I think for amperage measurements you need to put red probe to 10A socket

[gyula]

Hi unimmortal,

Thanks for doing the tests.
To err is human of course.   Please check again both the input and output currents. 

Would like to ask:  the 40 - 70 mA input current range means that 40 mA is taken when the 2 globes i.e. bulbs are unconnected and the 70 mA is taken with the bulbs connected? 

I cannot see a puffer capacitor across the DC output of the diode bridge, maybe the viewing angle of you camera hides it.

I mention this because when you have unfiltered DC output the meters may or may not not show correct voltage or current values even though you have a true-RMS meter. It is better to filter the rectified AC.

The 17.8 V output voltage may be close to the actual loaded output voltage because of the moderate brightness of the two 12V bulbs connected in series.
   
When measuring the output current, indeed use the 10 A input of the meter (with the range switch set in A as is shown in your picture)
and in case the bulbs' current is less than 400 mA, then you could plug the red meter probe to the mA input instead of the A input and set the range switch to mA too.
This way you can get more accurate measurement on the output current.
 
Gyula

Thank you Gyula, and a happy new year to you.

I've done a whole lot more testing over the past few days, and have almost arrived at a stable baseline - tuning out vibration is a painstaking process. Input consumption is around 9V@40-70mA w/ 10mm air gap.

Attached is the obligatory 2 x 12v globes running on the em coils' rectified emf output. The meters are showing current and volt outputs for the gen coils.

[unimmortal]

Thanks Gyula - what size capacitor would you recommend for filtering rectified AC?

I've moved to an external timing wheel as I was getting too much interference at the hall sensor from the coils. Using 10mm dia magnets on a similar PCD to the rotor, the timing and pulse is much sharper now. Testing this last night got me back to where I wanted to be, with inputs as low as 9V@6mA (meter was placed between the power supply and switch). Output is quite low at ~12V@50mA on the gen coils, but that is to be expected until I start switching them. I also haven't measured any output on the EM coils and look forward to verifying that against input current draw.

As a separate and previous test, I placed shottky diodes facing into the EM coil junctions, thinking maybe they'll break down from the EMF when the switch opens and flow power to the connected lights. As it turns out I could moderately illuminate 3 x12V globes  connected to the DC negative of the power supply. Again, once I get rid of a few more gremlins that keep popping up, I can retest and validate. Exciting times.

[gyula]

For puffer capacitor use at least 470 uF and higher like 1000 uF electrolytic.

[unimmortal]

For puffer capacitor use at least 470 uF and higher like 1000 uF electrolytic.

Thanks Gyula.

Ok, finally some consumption figures. 18V@460ma gets me around 23VAC, with a rectifier 20VDC, with a  puffer capacitor (+ to -) I stopped it at 30V (25V, 470uF).

Running at 18V is giving me more emf to work with, and with the reverse biased shottky's hanging of each coil join (6 including ground) and aggregating back to ground, the transistor is staying very cool - largely because the timing window is around half the size reducing current draw, but mainly because each coil has its own exit path reducing the emf evac time before the next pulse.

The next step is to switch the generator coils in order to generate an emf kick that should see the current draw of the EM coils drop drastically.

[gyula]

Yes, the rectified peak voltage in the puffer cap can go as high as the peak AC voltage feeding the rectifier if there is no or there is only an electrically little load across the puffer capacitor.

Regarding your next step(s), do as you wish of course, and focus on avoiding self-deception when doing measurements.  I still suggest you consider the evaluation steps I wrote in my Reply #54 above 
             https://www.overunityresearch.com/index.php?topic=4539.msg109356#msg109356   Choose the load resistor according to the generator coils overall internal resistance, maximum power transfer happens when your load resistor halves the unloaded  output voltage.

Thanks Gyula.

Ok, finally some consumption figures. 18V@460ma gets me around 23VAC, with a rectifier 20VDC, with a  puffer capacitor (+ to -) I stopped it at 30V (25V, 470uF).

Running at 18V is giving me more emf to work with, and with the reverse biased shottky's hanging of each coil join (6 including ground) and aggregating back to ground, the transistor is staying very cool - largely because the timing window is around half the size reducing current draw, but mainly because each coil has its own exit path reducing the emf evac time before the next pulse.

The next step is to switch the generator coils in order to generate an emf kick that should see the current draw of the EM coils drop drastically.

[unimmortal]

It's been a long while between posts.... so here's a good news update.

First pic show an input of 10W, a load of 15W 240v led globe. Open circuit max = 147V. When tuned right, I can turn the globe on and off with no change in current draw and little if no change in rpm.

Second pic shows how it looks now (sans wiring). Bottom coils are pulse, working together counterwound top / bottom and switching alternatively into N up facing rotors. Top coils are generate, and are also counterwound (bottom CW, top CCW). Induced voltage flows via a diode from CW > CCW. Each of the generator coils are AWG24, 150grams, 90 metres.

Learnt techniques:
Magnets are in part mechanical. The airgap for the pulse coils (~12mm) is double that of the generator coil (~6mm), however they are so balanced a baby could move the rotor even though they are 20mm cores only mm away from N40 Neo's with a holding force of 30Kg and there are 6 in each rotor. Basic leverage applied to magents - two magnet faces for the pulse, one magnet face for the generator. As such the tyranny of cogging due to proximity is now all but gone.

Alternating cores is also part of this trick, the direction the magnetic field is being attracted from constantly oscillates, and in the case of the generator cores 'distracts' the rotor from the pulse cores allowing a good degree of tuning. Without the generator coils, the motor runs with a higher current draw and slower rpm.


So bucking coils on a pulse motor work. Who woulda thought?

[Verpies]

First pic show an input of 10W,

How are you measuring that power?

...a load of 15W 240v led globe.

LEDs are not linear. Do not use non-linear loads.
How are you determining the power dissipated by that load ?

[unimmortal]

Hey Verpes,

Power is measured via a meter between PS and mosfet (blue screen) right hand side of pic.

As for load and measurements, I'm just happy playing around with this electromagnetic meccano set looking for as much voltage as I can get, and right now I can drive globes and small pedestal fans. A re-arrangement of coils/mags will see mains power voltage with a bit of effort.

The main aspect at the moment is validating bucking coils, but the techniques I've come across, I think, are much more important.

[Verpies]

Power is measured via a meter between PS and mosfet (blue screen) right hand side of pic.

Does this meter measure input power or some other quantity ?

[unimmortal]

Does this meter measure input power or some other quantity ?

Yes, it's an inline DC meter (Volts, Amps, Watts - or Power) - typically for monitoring power draw. Ignore the 'Energy' figure, that's over time.

The motor will run on as little as 5V@400ma, but by that point the screen turns off. Tuned I've had it still turning under 1W.

[Verpies]

Yes, it's an inline DC meter (Volts, Amps, Watts - or Power) - typically for monitoring power draw.

Volts and Amps are not units of power.
Please decide which ones it measures and which ones it calculates.