Electric Motor Secrets

[gyula]

Hi all.

....So I thought that maybe the problem is not in the circuit, nor in the coils or timing, but in the changing inductance. I mean this time the ON time is the whole 60 degrees and in these 60 degrees the rotor gets in alignment with the stator, this also changes the 60 degree current waveform. If the rotor is not in alignment with the stator, more current is needed to get it aligned, but as soon as it gets aligned, less current is needed and this is clearly visible in the current waveform.
All we need to do is just chop this 60 degree ON time into many shorter pulses so we get most of it back. Might this be true? And if so, why didn't I see the same things with my V2.0 motor? The V2.0 motor had an air gap of 0.13mm, this one has an air gap of 0.08mm. Could this small air gap difference be the cause of such a current behavior?
Thank you,
Jetijs

Hi Jetijs,

I think you pretty much explained: it must be the bigger change in inductance that causes the difference in current in the first case when the air gap is bigger. In case of the smaller airgap with the V2.0, the inductance change must be much less, may it sound rather strange.
Maybe you could use a self-inductance meter to check the changes in the uH-mH values when you rotatate the rotor slowly by hand into alighment and then out of it?

The idea of chopping up the ON time into many short on-off pulses is a good one I think. This way you can reduce the negative effect of any changing self-inductance and create a more efficient recovery possibility of the flyback pulses. You may have heard of Doug Konzen's recovery circuit that could be used with your chop up circuit, see pages 3-4-5 in this PDF file:
http://www.panaceauniversity.org/BEM...0KoneheadX.pdf

He connects the AC input of a full wave diode bridge in parallel with the S1 switch (which switch would be the chopper switch in your case I think) and he uses another switch in series with the AC input of the bridge, slightly retarded in time wrt S1, see his Figure in page 3. I have not tested this but he has built several such recovery variants and described them in his yahoo mail group EVGRAY : zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz where some members built them.
The PDF file is collection on his hints.
Of course, you surely can use any other recovery circuits, the reason I mention this one is because I think it nicely fits to your intended chopping solution.

rgds, Gyula

[Jetijs]

Hi Gyula,
thanks for the reply
You got it other way around, the motor V2.0 had a bigger airgap - 0.13mm and it did not show this odd waveform even if one power pulse was the entire 60 degree period. The V3.0 motor has almost two times smaller air gap - 0.08mm and it does show this odd waveform.
Anyway, thank you for you reply and chopping idea

[gyula]

Hi Jetijs,

Well, somehow the two distances "twisted over" in my mind, sorry for that.

Surely the chopping process must greatly reduce the smaller gap's current distortion effect.

rgds, Gyula

[pranav2010]

about the turns of ev gray motors electromagnets & wire size

i want your help sir can someone help me please

can anyone tell me what is the number of turns in ev gray motors

electromagnets & what is the wire size please help me sir

[cody]

I had been pondering the same thing about the chopping method. As the rotor comes into alignment with the stator, the coil will have more inductance with each pulse. So the optimal frequency for the beginning of the of the power stroke will not be the optimal frequency by the end of the power stroke. I finally got around to building an attraction motor, but im going to try a completly different approach on the powering it.

[lighty]

@cody

You're correct in having to use chopping frequency which is adjustable depending on the position of the rotor (or rather depending on the inductance of the winding in any particular position). What most people don't realise is that there is a definite upper limit on the repetition rate of impulses and on their duty cycle. If the frequency is too high or duty cycle is too long the coils will act as choke. What will happen is that the current through coil will be virtually constant and one will get constant magnetisation. Which is OK for mechanical motion but it will reduce or completely disallow for inductive collapse recovery. There are two solutions- either to reduce inductance of the winding or to lower the frequency and/or duty cycle of the impulses.

I think that some months ago Jetijs measured exactly what I'm talking about (I don't have time going through past posts). As he raised chopping frequency more toward saturation point of windings recovered impulses energy was rising at first and then stuck at a certain point when he got near saturation plateau. So if you could compensate for changes in inductance of the windings at different positions of the rotor you would get maximum input/recovery ratio at all positions and all RPM's.

I don't see a way of doing that by mechanical approach or simple logic circuits aside from using some MCU or other programmable device which would then have implemented lookup table for any particular motor (which is not so good if you change any parameter of your motor) or it could measure input/recovery ratio at all times and self-adjust (which can then be implemented on any version of the motor). It would represent a kind of PID control which requires some skill to implement programmatically but it's definitely doable. If I had some spare time I would even do it myself and perhaps I will if the problem is still not solved by the time I finish my commercial projects.

[cody]

lighty,

As usual, your electronics knowledge is very welcome here I figured what you were describing could be done, its just beyond my current skill level. Ill leave that circuit for you to figure out Im going back to using a single pulse, but will be using the collapse a little differently. Too early to say much about it, but it may offer another option.

[lighty]

I think chopping impulses is very good approach recovery-wise. No matter how you recover impulses if you use single pulse you will have rather poor input/recovery ratio. Look at it this way - you will spend some amount of current on a single pulse and you will get only a small fraction of that energy recovered depending of the length of input impulse. Also, once you got inductor saturated any further current input will be wasted. It's much easier having chopped impulses and letting the inductor produce what is essentially virtually constant current. It's done pretty much the same way in BLDC motors although there are a number of different reasons for doing so.

Implementing PID should be fairly easy if you have experience and time. I don't have time at the moment and I have limited experience in implementing PID algorithms. I do have a friend who is experienced in that area of MCU control so I could cooperate with him when I have some spare time. I doubt I could even start before first quarter of 2010.

Since Jetijs is willing to share his blueprints in SolidWorks, for me it would be easiest to just replicate his mechanical setup which is already proven and to develop self-adjusting PID system from there on.

BTW- if lower inductance of winding is needed one could always make inductor out of several smaller diameter wires in parallel- it will induce net inductance. That's why I suggested using Litz wire in the first place and it also have much better response for shorter impulses with steep rise time curve.

[Eric]

ok after several 100 hours of work off and on over many months i have finally finnished my 3 stage mechanical cam switch for my version 3 motor. only time spent working on a variety of electric circuit possiblities will tell if the time spent in front of a little lathe mounted on my big mill was worth it. eather way i learned a lot and it was fun!!

here are some shots

this is a shot of the harbor frieght pressure washer motor i modified. this is a sizable motor i am guessing with the original windings it was a little under 1 HP after i packed on alot more copper as a bifilar winding who knows what the power rating is. this part of the job was easy. i just hacked out the copper in the rotor and used my mill to cut out just enough of the rotor lam core to leave me a cross style rotor so the rotor will get a pull at every 90 degrees. also if you look closely, befor i rewound the 2 stator poles, i used my mill to shave away some of the width of the stator lams so that the stator width matches the width of the rotor sections. each winding is 18gauge at about .3ohms of course the original rotor will have a larger air gap. at the moment this is not a concern for me the focus here is in exploring
a variety of control circuit possibities of which i will apply a final/best possibility to my version 4 motor which will be the switched reluctance motor i purchased via the link found earlier in this forum.


here is a side view of all 3 stages of switches installed on the motor. the idea here is rather than have 1 switch with a 4 pattern cam i will have 4 switches with a 1 pattern/lobe cam. the reason is i wanted the flexiblity to rotate power sources up to every 90degrees. the reason for 3 stages of switches is 1 set for each winding and the 3rd set to control when a shaft driven generator will apply a charge pulse to my power sources. starting with 12 volts and one of the windings on one of the poles using just one cam switch on just one stage of switches i will start with the simplest circuit and work up from there on a variety of ideas i have in mind. i could even make a 2 or 4 lobe cam shaft if i wanted to.


here is a end shot inside view of the first stage of switches notice how the switch housing seats right onto the bearing houseing of the motor. this will allow me to fine tune the timing while the motor is running.


here is a close up of the switch components notice how the allen set screws allow me to vertically adjust both how long the cam piston will contact with the cam shaft and the other side of the contact can be adjusted to further control how long the contacts stay connected. this should allow me to tune the pulse width of the circuit untill the rotor is sucked in and the contact releases.


here is a shot showing how i will start with just the first stage of switches.

the pink nail polish with the lines scraped in mark when the rotor begins to lineup with stator and when it finnishes. (on time/off time)


now the fun begins!
Eric

[Eric]

forgot to post my first test.
lol it was fun and messy all i tried was the simplest of mechanical circuits
pos of source to 1 winding neg through one of the switches and a diode on the neg side of the coil hooked to a cap and light bulb to collect and burn off the bemf. this is the absolute simplest and weakest condition. at 12v it did spin slowly there was still sparking and a lot of heat on the contacts. i am guessing to much sustained amps combined with the sparking. if i am going to continue to explore driving the full amperage mechaniclly through the contacts i am going to need to learn about a good spark snubber circuit. can anyone help with that?
the other option i tried was to drive an NPN mjl21194 using the contact to control the P base after the 100ohm resistor bridge. this is literally as close to the simple circuit as i have ever got to dupiicating the exact circuit idea shown near the end of the DVD outside of useing a reed switch. interestingly enough this was much worse. there was no power in the coil to even budge the rotor and the transistor instantly got very hot. i did go back to look at a scope of the first circuit without the tranny. the shot did not look like a clean vertical rise time at all. so i am geussing the contact is not cleanly opening the tranny up all the way effectively turning the tranny into a poor resistor. this is very troubleing as most of my more complicated circuit ideas will probably need discreet componets like transistors. if i cant solve that problem then i might have to write off the whole mechanical idea and switch to optics instead. any ideas on how to solve this as well? i was curious if something like a smidt trigger might help to use the dirty mechanical signal and turn it into a clean digital? signal for the tranny base.

cheers to anyone who can help!
Eric

[cody]

Eric,
Well i was waiting for someone else to help you but no one has said anything, so ill give it a shot. But first off let me say that that is a nice and interesting switch idea. Ok, now to try to help. The first thing that is really sticking out to me is the geometry of your rotor and stator. Your stator poles are very large compared to your rotor poles. If you go back some pages you will see the optimal sizes they should be. The way you currently have it, its probably attracting the wrong rotor pole and actually slowing the motor down. This in turn means that your motor is going very slow and the power going to the coils is on for too long so its drawing too many amps and burning up your transistors. Im afraid my best advice would be to consider a different rotor to stator geometry. Sorry, i know you were looking for an easier solution. Also you should not be getting sparking on your contacts if you are collecting the collapsing field correctly, you may want to double check that. But its probably related to the slow speed issue. Maybe someone else has something else to say.

[Eric]

Eric,
Well i was waiting for someone else to help you but no one has said anything, so ill give it a shot. But first off let me say that that is a nice and interesting switch idea. Ok, now to try to help. The first thing that is really sticking out to me is the geometry of your rotor and stator. Your stator poles are very large compared to your rotor poles. If you go back some pages you will see the optimal sizes they should be. The way you currently have it, its probably attracting the wrong rotor pole and actually slowing the motor down. This in turn means that your motor is going very slow and the power going to the coils is on for too long so its drawing too many amps and burning up your transistors. Im afraid my best advice would be to consider a different rotor to stator geometry. Sorry, i know you were looking for an easier solution. Also you should not be getting sparking on your contacts if you are collecting the collapsing field correctly, you may want to double check that. But its probably related to the slow speed issue. Maybe someone else has something else to say.

hi cody!

i have been running some tests and i think i have figured out the problem. first, to address the stator/rotor idea. its very difficult in the picture to see but i have actually cut out part of the iron lams in the stator so the width of the stator matches the width of the rotor. and there is plenty of space between rotor widths that i am pretty confident there is no reverse pull on the rotor after it completes its allignment and freewheels to the next pulse.

the problem i think is the ohms are too low in the 18 gauge windings. i connected all 4 of the windings in series to maximize the ohms. in the pure mechanical circuit there was much less sparking and a lot less heat. in the simple transistor circuit (using the mech switch to trigger the base of an NPN) was the strangest of all. i tried 2 windings in series still lots of instant heat and a very weak magnetic field, then 3 windings in series the rotor did turn weakly and a little less heat from the tranny. it wasnt untill i had all 4 windings in series that the transistor circuit ran at the same magnetic strength as the machanical circuit. and the tranny was back to performing normally as in my other 2 motors with little to no heat.

when i tried to run an ohm test again with my multimeter on one of the strands i kept getting different readings from .0000 to ,0002 to .0004, my guess is the ohms are so low the meter is having a hard time giving me an accurate reading. also in the equation v=IR if we have to divide the V with a low decimal R the I gets big really fast. i am guessing this has something to do with why the tranny was failling to operate normally with 1 strand of the winding. so i am looking to rewind the motor with eather 22gauge or 26 gauge to increase the ohms in each of the 4 strands to closer to 1 ohm i have plenty of both gauges.

i went ahead and ran a test using all 4 strands in series again, this time i dialed in all 4 switches tied together to activate 1 transistor 4 times per revolution (90 degress apart) the motor really flew!! not bad for starters.

more improvments to come.

Eric

[captainpecan]

Motor secrets video

Forgive me if I am asking the same question that has been gone over already, as it is going to take me weeks to get all the posts read here.

In the Motor secrets video, at the end, Peter specifically says that you can create a generator with no bemf also. Could anyone point me in the right direction as to where he may go into more detail about this design? I'm very curious to learn more about it, because with my understanding, it is the fact that there is no magnets that makes it so much easier to eliminate that bemf in the attraction motors. How is Peter suggesting to do this when a generator is used. I am of course assuming the generator will have magnets in it. It could I guess work like an alternator and have no magnets but the same principle applies either way because the magnetic field that builds up is what is generating the energy.

Could anyone be so kind as to point me towards the light?!?!

[LowTechIsCool]

attraction generator

I would be grateful if someone would take some time and explain any mistakes I am about to make in my reasoning.

CaptainPecan, I have thought about the generator myself and think you might appreciate the following:

One of the reasons that we have trouble getting above 70% recovery on each pulse is that as the rotor aligns with the stator, the inductance of the coils increases and this chokes off some of the recoverable energy? This is seen from Jetijs' pictures taken of the current on his most recent coils?

What I would like to try is to take a motor and attach it to the shaft of our current motor, which will be acting as a generator for this thought experiment. Pulses will be applied to the coils of our 'generator' starting when the rotor is fully aligned with the stator. The motor will force the rotor to an unaligned position, if it is exerting more torque than our 'generator' is to become realigned.

Here is where the generator comes in, the amount of electrical power from the coils should be greater on the recovery than on the charge, because the forcing of the rotor to an unaligned position has resulted in a reduced inductance of the coils.

It does require some initial electrical energy, but you should be able to reuse a portion of the recovered energy for the next pulse and store or use the excess to run some device.

-Chris Corkum

[captainpecan]

I think I follow what you are trying to say. Not exactly the direction I was going with the question, but thought provoking. In the scenario you talk about, if I am understanding it correctly, it will take more brute force from the motor the turn it. Adding pulses into the motor/generator seems it would create even more BEMF than if it was just turned normally with an attraction motor. Yes I agree, pulsing it could allow you to recover maybe 70% of the energy introduced into it, but it seems you could save 100% of that energy if you don't pulse it all? Thanks for the reply though, maybe I am looking at it wrong?
Maybe what you are referring to is a little closer to the workings of a normal car alternator that uses incoming current, to generator more output current from the help of the engine turning it.

Problem is, I still see quite a bit of Lenz Law drag in those scenario's. The more energy generated, the more drag created. Maybe leaving open coils until generating current is at peak and capture only the peak, but still for that split second, there is maximum drag. Peters video does very well in explaining how to terminate BEMF entirely or close to it when motoring. But only hints at being able to do it while generating... Does anyone follow my question? Maybe there isn't an answer yet, It just seemed like Peter was eluding to the possibility that he was going to tell how at the end of that video.