Electric Motor Secrets

[Peter Lindemann]

Recycle Circuit

Jetijs,

I've had some time to think about this situation and have a few more suggestions. It sounds like your air-gap is pretty small right now, so that is good. One other possibility is that your return energy readings are low partly because of the sampling rate of your meters. Metering the energy quotient from the inductive collapse has always proved difficult. Some people believe that there is no significant energy there, but the recent development of high efficiency DC-to-DC converters shows there is, if the switching and core materials are properly configured.

In Bedini's SG motors, the "metered" energy return always looks lower than the effect it produces in recharging the second battery. When I worked for John's company, we rarely bothered metering the output directly for these reasons.

What I would like you to do is take the output pulses from the isolated windings and apply them back to the front of the circuit to a capacitor placed between the front battery and the motor. The capacitor must be isolated from the battery with a diode on the negative line (cathode toward the battery) as I discussed before.

Then you can look at the input current between the battery and the capacitor and as well as the input current between the capacitor and the motor. The difference is the amount of energy you are recovering back to the capacitor. I got a 65% current reduction on the battery drain using this technique on my first try, using one of my little Bedini SG type motors, but I have not tried to maximize this effect in either an SG or an attraction motor yet. The transistor switching, coil/core response, and inductive coupling between the windings are the keys to high return.

As you have seen, your twisted wire pairs are working well, with only a small difference in recovered energy between the isolated windings and the power windings.

Try this experiment, and see if you get an indication that your recovered energy is higher than the 24% you are currently metering.

Peter

[Peter Lindemann]

Incomplete De-magnetization.....

Peter,
The air gap should not be the problem. Before I changed the magnetic bolts that were holding the rotor together, I noticed, that the gaps were slightly more wide on upper surface of the rotor and narrower on the bottom side. That's why I changed my rotor upside down when I changed the bolts. This time it was a lot harder to get the rotor in place so that it would not touch the startor. The gap is now so wide that I can get only one sheet of paper between startor and rotor, but not two sheets of paper. That is why I think that the gap is not the problem. Also I noticed, that changing the bolts to non magnetic steel ones, did not do any good, because the rotor still tends to align itself to the startor when no power is aplied. That means that the material of the rotor/startor plates retains some magnetism but just a little. That could also be the problem. Maybe this type of steel just can't demagnetize so fast? When I made the tests, it seemed that I simply can't get enough power to the coils if the impulse is choped to too many smaller impulses. I am sure, that the circuit and the 555 timer is right just in the schematic abowe. I did not solder the 555 timer circuit to a base plate instead I used a non soldering break out board, this way I can easy and fast adjust the capacitor value and thus the frequency just by adding more capacitors, or removing some. Also, this pulse chopping approach makes the transistors work cooler, they get only slightly warm. Since I am now using only 3 strands of wire and the non isolated output, maybe I should add another 3 transistors to the circuit and use the remaining strands for powering?
These tests were not very precise, because I was powering the motor with batteries and they slowly went down as you can see in the chart. But still, the results should be better. I will make some more tests tomorrow and see what the scope shows on the output section.
Another interesting thing I noticed, the output current seems to drop as the motor gets up to speed now. This is interesting, because when I ran my motor with only the comutator timing and no 555 circuit, the opposite was happening - as the speed increased, so did the output current also the input current dropped.

Edit: Just a thught. Lets suppose that the problem is the core material that can not demagnetize fast enogh. Wouldn't a slower rise time help? Maybe I should connect the three unused wires in series with the other three coils so that the inductance increases, this way also the rise time should increase, right? This way there should also be more time left for the material to demagnetize because we could use a smaller frequency on the 555 timer.

Jetijs

Jetijs,

The electro-magnetic effect that produces the output is the collapse of the magnetic field. This falling field produces a "rate of change" of flux in the output windings that produces the VOLTAGE of the output pulse. The current associated with this VOLTAGE is determined by Lenz Law, so the higher the voltage, the higher the current.

If your core material retains some of its magnetic field, that characteristic is going to slow down the magnetic field collapse and reduce the amount of energy the return pulse can produce.

This is the most probable reason you cannot get your recovery energy above 24%. Until this issue is resolved, your project motor cannot hope to produce a COP>1. Slowing down the rise-time of the inductor does not overcome the slowness and energy loss of the hysteresis curve of the core material.

Where we go from here is up to you.

Peter

[Jetijs]

Peter,
The air gap should not be the problem. Before I changed the magnetic bolts that were holding the rotor together, I noticed, that the gaps were slightly more wide on upper surface of the rotor and narrower on the bottom side. That's why I changed my rotor upside down when I changed the bolts. This time it was a lot harder to get the rotor in place so that it would not touch the startor. The gap is now so wide that I can get only one sheet of paper between startor and rotor, but not two sheets of paper. That is why I think that the gap is not the problem. Also I noticed, that changing the bolts to non magnetic steel ones, did not do any good, because the rotor still tends to align itself to the startor when no power is aplied. That means that the material of the rotor/startor plates retains some magnetism but just a little. That could also be the problem. Maybe this type of steel just can't demagnetize so fast? When I made the tests, it seemed that I simply can't get enough power to the coils if the impulse is choped to too many smaller impulses. I am sure, that the circuit and the 555 timer is right just in the schematic abowe. I did not solder the 555 timer circuit to a base plate instead I used a non soldering break out board, this way I can easy and fast adjust the capacitor value and thus the frequency just by adding more capacitors, or removing some. Also, this pulse chopping approach makes the transistors work cooler, they get only slightly warm. Since I am now using only 3 strands of wire and the non isolated output, maybe I should add another 3 transistors to the circuit and use the remaining strands for powering?
These tests were not very precise, because I was powering the motor with batteries and they slowly went down as you can see in the chart. But still, the results should be better. I will make some more tests tomorrow and see what the scope shows on the output section.
Another interesting thing I noticed, the output current seems to drop as the motor gets up to speed now. This is interesting, because when I ran my motor with only the comutator timing and no 555 circuit, the opposite was happening - as the speed increased, so did the output current also the input current dropped.

Edit: Just a thught. Lets suppose that the problem is the core material that can not demagnetize fast enogh. Wouldn't a slower rise time help? Maybe I should connect the three unused wires in series with the other three coils so that the inductance increases, this way also the rise time should increase, right? This way there should also be more time left for the material to demagnetize because we could use a smaller frequency on the 555 timer.

Jetijs

[Peter Lindemann]

Excellent Work

Jetijs,

Well, it looks like you have the circuit working right and it is functioning properly. So now, the commutator is switching ON when in the right position, and the Stator Coil is being switched On and Off at its inductive rise-time.

Why you are only getting back 24%, at best, is not known to me. It looks like there may be some problems in your magnetic circuit. Possibly the large air-gap is part of the problem, as well, although this seems less likely.

It could something else. Let me think about it for a while.

Lighty, do you have any ideas?

Peter

[Jetijs]

Peter,
I made some tests with the 555 timer pulsing the LED. The results are not too good. Here they are:


I used a 1 KOhms for the first resistor and 85K for the second one on the 555 circuit, the greated difference between these values, the closer we get to 50:50 ratio of ON and OFF time. I changed the frequency of the pulses by adding or removing some 1,5nF capacitors on that timer. The more capacitors in parallel, the longer the pulses. The resistor between the output of the 555 timer and the IR diode is 470 Ohms. The comutator wheel has four gaps, 25 degree of an arc each. In the chart you can see, that by reducing the pulse width, I get more spikes per 25 degree window and also getmore back on the output. But I can't get more than 24% back. Also the more spikes, the slower speed.

Here are some scope shots:





Notice how they get shorter in height when there are more spikes per 25 degree window. That means that I can get the pulses even shorter than the rise time, but nevertheless I can not get more than 24% back. Also the higher the frequency of the spikes, the higher pitch noise the motor gives out.
Any suggestions which way to dig next?

[Peter Lindemann]

Steven, are you still here?

Steven,

After showing us those excellent photos of your motor a few weeks ago, we haven't heard from you.

Can you give us a progress report?

Thanks,

Peter

[amigo]

Jetijs,

If you need a good square wave oscillator, please check LTC1799, here's a link to one circuit: Electronics -- Small Footprint Square Wave Oscillator

[Jetijs]

Ok, thanks Peter
I have not made a new comutator with a 25 degree gap yet. I will first solder the 555 circuit and test it with a scope.
I think I have all the info now to proceed

Jetijs

[Peter Lindemann]

Pretty Good

Thank you Peter!
The idea of using only one 12V battery to power the 555 is so simple that it is almost embarrasing that I did not think about that Anyway, here is the picture of the whole circuit that I am indending to use:

I think that it is all right here, but just to be sure I need someone to verify this
Thanks,
Jetijs

Jetijs,

Looks close. Pin 4 of the 555 needs to be tied to the +12. I think if the capacitor is about .01uf, the resistor between Pins 7 & 8 is a 10k resistor, and the resistor between Pins 7 & 6 is a 1M pot, you should be able to play with the pulse width to find the inductive rise-time of the coil and a good percentage energy return. Play around with it, as I may still have this wrong. I don't use 555 timers in the standard mode very much, and I don't have this stuff memorized.

You may also have to lower the value of the resistor on the LED to get the same brightness as the 1k across 24 volts. The output at Pin 3 is not the full +12 volts.

Other than those minor things, it looks good to try.

Have you opened up the window in your commutator wheel to 25 degrees?

Peter

[Jetijs]

Thank you Peter!
The idea of using only one 12V battery to power the 555 is so simple that it is almost embarrasing that I did not think about that Anyway, here is the picture of the whole circuit that I am indending to use:

I think that it is all right here, but just to be sure I need someone to verify this
Thanks,
Jetijs

[Peter Lindemann]

A Few Ideas

Hello everyone.
Today I played around with 555 circuits. I could get the frequency easily, but the thing is that a simple 555 timer can't get you a perfect 50% duty cycle. You can use a flip flop to solve this. I did this a while ago when I atempted to buid a Tesla switch. Unfortunately I lost all my circuit examples when the hard disk array on my computer crashed. But I guess I don't need a perfect 50:50 duty cycle, 50.001:49:999 will do just fine. I just need some positive voltage regulators, because my system runs at 24V and the 555 timer can handle max 16V. I think the LM7812 will do just fine. Here is the data sheet:
LM7812 datasheet pdf datenblatt - Fairchild Semiconductor - 3-TERMINAL 1A POSITIVE VOLTAGE REGULATORS ::: ALLDATASHEET :::
Also I have a question. When I make the efficiency tests, the output battery gets charged with each test and there is only a little pause between tests. The battery gets to a higher voltage with each next test thus giving me better results, but this is not acurate. Also if a test is longer, the battery gets charged higher at the end of the test. Maybe someone has a different idea how to calculate the output energy more precize?
Thanks,
Jetijs

Jetijs,

I love the way you just jump on this stuff. Anyway, here are some simple solutions to the problems you are having.

1) Don't worry about the perfect 50% duty cycle with this set-up, it just doesn't matter. 51% - 49% is plenty good for this.

2) Run the 555 timer circuit and the LED portion of the optical commutator from just ONE of your 12 volt batteries. This portion of the system isn't connected to anything else, so it doesn't have to be run on the 24 volt system.

3) For efficiency tests, go back to running the little light bulbs.

Good luck,

Peter

[Jetijs]

Hello everyone.
Today I played around with 555 circuits. I could get the frequency easily, but the thing is that a simple 555 timer can't get you a perfect 50% duty cycle. You can use a flip flop to solve this. I did this a while ago when I atempted to buid a Tesla switch. Unfortunately I lost all my circuit examples when the hard disk array on my computer crashed. But I guess I don't need a perfect 50:50 duty cycle, 50.001:49:999 will do just fine. I just need some positive voltage regulators, because my system runs at 24V and the 555 timer can handle max 16V. I think the LM7812 will do just fine. Here is the data sheet:
LM7812 datasheet pdf datenblatt - Fairchild Semiconductor - 3-TERMINAL 1A POSITIVE VOLTAGE REGULATORS ::: ALLDATASHEET :::
Also I have a question. When I make the efficiency tests, the output battery gets charged with each test and there is only a little pause between tests. The battery gets to a higher voltage with each next test thus giving me better results, but this is not acurate. Also if a test is longer, the battery gets charged higher at the end of the test. Maybe someone has a different idea how to calculate the output energy more precize?
Thanks,
Jetijs

[MHZathras]

The first waveform shows the peroid of the pulse. The curving rise time replesents the slew rate of the charging cycle of the on switch. It has a low pass filter tendency (slow rise time). The second wave form is from a longer period of repetition or lower rpm, same approx slew rate. The third wave form, assuming the controls for the scope are not changed, is an even slower rpm (ergo longer period of cycle) It has reached a saturation of some sort(most likely the power supply maximum voltage). Spectrally you could deduce their are more high frequency harmonics in the wave form. Mostly odd harmonics... But with the non linearity of the circuit with it high constituents of inductive reactance. I would guess that there are some even harmonics as well (sum and difference componects of the sum and diference componets, that are odd harmonics (3f-5f=2f etc.)

Wow, sorry guys I was trying to reply to a post on 11-22-2007, 01:55 PM and it landed here. I guess that is why I am called a junior member. I have been a linear and digital electronics engineer for a very log time. But I still have problems with BBs, Forums, and postings.

Sorry, maybe someone can put this post where I meant it to be. Just trying to contribute.

[Peter Lindemann]

Make it Variable

Thanks Peter
The idea about the 555 timer flashing the LED is great, because I can't really squeeze more than three pulses in this 30 degree window on my timing wheel due to the need of very small endmill bit diameter. I should be able to do this LED flash trick easily. Flashing the LED with 50% duty cycle will also make it consume less power. I will try this and post the results

Thanks,
Jetijs

Jetijs,

If you make the 555 timer circuit with a variable pot, you will be able to dial up the perfect chopping frequency to maximize the energy recovery. Once you know the right frequency, you can hardwire the circuit for those values.

Once you have the new rotor with the micro-clearance air-gap, we can go through each system one more time to balance the motor at a higher power level.

Great work!!

Peter

[Jetijs]

Thanks Peter
The idea about the 555 timer flashing the LED is great, because I can't really squeeze more than three pulses in this 30 degree window on my timing wheel due to the need of very small endmill bit diameter. I should be able to do this LED flash trick easily. Flashing the LED with 50% duty cycle will also make it consume less power. I will try this and post the results

Thanks,
Jetijs