Dear Ewebie,

Yes, I understand what you are trying to say. The idea is to keep the inductance low so the rise-time of the current is extremely fast. The magnetic field can then be produced by a very short burst of current. Then just shut off the input current and short-circuit the output. The magnetic field wants to collapse instantly, but generates a current that opposes the change in flux, which in this case sustains the magnetic field longer for no additional power input. During this "free field sustaining" period, the magnetic field can still be attracting the rotor piece to the stator core, producing "free" mechanical energy.

This "little trick" is part of what Bob Teal was doing in his Magnipulsion Engine. He used two coils in parallel controlled by a single switch. In this arrangement, the two coils charge up in parallel, but then try to discharge into each other in series, thereby sustaining the field for much longer than the time it took to create it.

There are many such "tricks" that can be employed in these motors to get "free" mechanical energy production. Thanks for bringing this up.

Peter

Peter,
thanks for the suggestions. The reed witches did work after I froze them and heated up several times, but I still could not get the perfect timing. I could either get two pulses perfect, one a tinny bit off and the third pulse wont fire, or one pulse perfect with other three pulses way off. I tried to center the timing wheel but with no luck, if I move a reed switch near enough the magnets I can get all the pulses, but they are not equal in pulse width. When I move the reed switch just a half of millimeter away, I can get only three pulses. Yesterday I spent all day adjusting and readjusting the timing, I tried three different timing wheels - no luck. The problem is in my method of attaching the timing wheel to the shaft:


As you can see it is only bolted to a thread in the shaft end through three
6mm thick aluminum spacers. No bolt is gonna be precise enough. It's a shame I did not think more about the timing when making the shaft, I would made it better I know the reed switch would work well for this purpose, but due to my bad job on the shaft, I have to consider different switching methods. I really want to get the motor working and I have already spent too much time with reed switches and no luck. The first try was good, but also there the timing wasn't exactly very accurate. So I went to local electronic shop today to buy some more 2N3055 transistors and also bought some optointerrupters, they had only two pieces of ITR8102. They were very cheap, so I bought both of them. Here is the data sheet:
ITR8102 pdf, ITR8102 description, ITR8102 datasheets, ITR8102 view ::: ALLDATASHEET :::
The gap fo the optointerrupter is 3mm wide, so I made a timing wheel out of 2mm thick aluminum. The wheel is designed so that there will be 4 pulses at 90 degree angles, each one 15 degree in duration (exactly half of the rotor leg width). If this wont work well, I can make new timing wheels easy in the matter of minutes. Also the gap of the optotrigger is deep enough to allow that little off center I have.



I figured the circuit should be something like this:


Of course I am not sure if this optotrigger will handle more than one transistor this way. Maybe I need this second stage amplifier. In that case I have some small 2N3584 transistors, just like in the John Bedini's cap pulser setup, where these transistors are used to trigger the SCR. Maybe I can use that? Here's a data sheet:
2N3584 pdf, 2N3584 description, 2N3584 data sheets, 2N3584 view ::: ALLDATASHEET :::
I think It shouldn't be a problem for me to solder it all together, I have some experience on that (have built some cap pulser setups). I just need to verify that the circuit I drew is correct and if someone has any comments about this. I intend to power the motor from 24V with two batteries in series.

Thank you all
Jetijs

with my little electronics knowledge i can't see any mistakes in the circuit.
I do wonder though, if those little triggers can switch the transistor directly. I would risk only one of the triggers first, see what happens.
I know a lot about frying parts, because im a master in doing that

can't wait to see it running, especially since i ran out of parts, so cant do anymore testing myself.

@Jetijs

It would be better to use opto-reflective switch since then you will only need to have one disk with reflective surface and you could simply change your timing with a bit of a non-reflective paint. No need for any slot or additional machining. On the other side the electronics would be more complex because of the need to control the levels of the photo-transistor in cases of the changing reflexivity. So, in my opinion you should stick with photo-interrupter switch for the moment.

Also, it would be easier to drive MOSFET- they are basically driven by voltage rather than by current so they're simpler to work with in logic circuits or in this case in switching circuit. IRFP450/460 comes to mind since they're rather common and not expensive. Also, when driving MOSFETs with voltage it would be harder to burn out your opto-interrupter switch (not impossible though).

I usually use much more complex circuit topology since for various reasons I need to have my signal conditioned, cleaned, pulse response advanced and to ensure extremely fast switching with advanced control I use MOSFET driver ICs. You don't need all that so the schematic I suggest is the following.

Please note that the first schematic has an inverted logic - that means that when the photo transistor is active (when the slot allows for the passing of light) the MOSFET will act as an open switch.

If you want MOSFET to conduct current when the slot is passing the light then simply use the second schematic called "non-inverted" (it has one bipolar switch added in order to invert the signal logic and also to ensure better control over MOSFET- in my experience for this simple application a small transistor like 2N2222 or BC635 or something like that that you can get your hands on).


As for the values of the resistors it can easily be calculated when you specify the exact voltage you'll be using. Also, if the inductive collapse voltage should exceed semiconductor switch maximum rating you can easily add some fast transil diode in parallel in order to protect the device- the drawback is that if the protection diode conducts the most of the recovered energy will be grounded (but at least it will protect your semiconductor until you come up with some solution for the problem). For example the good old 2N3055 is rated about 70V CE voltage and anything higher than 80-90V would probably kill it. 2N3584 would be much better choice because IIRC it has about 250-300V CE voltage rating but I'm not sure about it's current capability (did you measure the power consumption of the coil) since I didn't look at the datasheet yet.

Anyway- I suggest you use MOSFET since it's much easier to work with (although not without it's quirks when working in more severe electrical conditions) but Peter might suggest otherwise. I guess you should listen to his advice- he's the chief of this parade anyway.
All you need to do is to replace MOSFET with BJT power transistor and add another resistor between base and emitter to ensure proper shut-down.


P.S.
Why do you need to machine a new disk every time- why don't you simply mask out the slot (if you need narrower slot) with some cardboard and adhesive tape?

P.P.S.
Since you've asked about whether your circuit would work the answer is most likely but it depends a lot on the power ratings of the opto-interrupter transistor. I will look at it more closely tomorrow, it's almost 4AM and my brain is shutting down.

Jetijs,

Apparently, you would like to abandon the magnetic reeds and move on to the opto-interrupters. Without a thorough study of what happened, you will never know. I believe the problem is with your magnets, the aluminum wheel they are mounted in, and the steel bolts nearby causing field distortions. In my estimation, you have not actual learned WHY the magnetic reeds are misbehaving at this stage of testing.

OK. On to the optical approach. As I said in an earlier post, the difficulty with this method is in the intermediate stage. The Data Sheet on the ITR8102 states that the photo-transistor output stage can only handle 50ma of current and a maximum C-to-E voltage of 30volts. This current level is insufficient to drive three output transistors directly.

As you can see, a number of different suggestions have already been made. Any of them can be made to work to one degree or another. Each solution will have its own issues. My suggestion is that you stay with the Bi-polar Transistors for now and go with one stage of signal amplification in between the photo-transistor and the power transistor.

The system I use is quite simple. Replace the magnetic reed switch with a medium power PNP transistor. So, the first stage of control will remain similar to what you had with the reed switch. The emitter of the PNP is connected to the Positive Supply and the collector is connected to the first resistor in the voltage divider. The base of the output transistor is still triggered between the two resistors of the voltage divider.

Next, we are going to create a second voltage divider operated from the NPN photo-transistor. So, the emitter of the NPN photo-transistor is connected to the circuit ground (negative of the supply) and the collector is connected to the first resistor of this new voltage divider. The second resistor connects to the Positive Supply and the mid-point between the resistors connects to the base of the PNP. So now you have positive control of your POWER NPN through one inverted stage of signal amplification. The resistors in the voltage divider operated by the photo-transistor can be 2k ohms each, or higher and still be sufficient to turn the PNP on quickly and completely. Just keep the total current in this section below half of the rated current maximum of the photo-transistor, or about 25ma.

Set this system up for one power transistor and one coil. I suggested in an earlier post that you run some tests with the values of your primary voltage divider to see how much higher the resistor values can be raised above your currently used 200 ohms and still get clean switching of your power transistor. See if the transistors switch cleanly with 680 ohm resistors. If they do, then just make two more voltage dividers to run two more power transistors to run the other two coils. Parallel the voltage dividers and run them all from the collector of the PNP.

Make sure you have an output diode on each of the coils when you test the system.

Good luck,

Peter

Ok, tkank's Peter
For what I understood from your post, the circuit should now look something like this:

Is that correct?
Thank you

Edit: Can I use the MJ2501 PNP transistor for triggering? It is a high power transistor, but I found some of them laying around. The max base current of this transistor is 0.2A, but that is the absolute maximum rating. Here is the data sheet:
MJ2501 pdf, MJ2501 description, MJ2501 datasheets, MJ2501 view ::: ALLDATASHEET :::
If I can use this, then I need no aditional parts and can start soldering.

Jetijs

Why bipolar transistors? I am not asking this because I doubt your judgment but because I really want to know. Is there any difference at this stage? I mean from first hand knowledge I know there is a significant difference when dealing with dielectricity but at this stage it's simply a switch. Or not? I never experienced any adverse effects when I compared MOSFET and BJT as switch (although the MOSFET proved to be faster and had lower turn-on resistance). Is there any difference?

Jetijs,

Yes, that is the idea. Try the PNP you have on hand. Also, see if the 200 ohm resistors can be a higher value, like 680 ohms, and still turn your power NPN on all the way. Run the test with one coil and see how it works. Then try to power all three coils.

Peter

Lighty,

As you know, MOSFETs can work with these circuits. The biggest problem with FETs is the capacitive nature of the gate. Getting the device to turn ON quickly is no problem, but getting them to turn OFF quickly is a little harder. The little FET driver ICs help, but that is one more level of complication. I am trying to demonstrate circuits that are simple to understand, as well as work effectively.

Again, for you, I'm sure you can make MOSFET circuits do all of these things, but I am trying to keep it simple for learning purposes.

Peter

Thank you for the answer. You're correct, of course- extremely fast switching times can be achieved with MOSFETs but for the extreme purposes one has to use special MOSFETs, diver ICs with additional gate sink and clamp and several other things. Thinking of it again I can understand your logic in this matter.

Jetijs,

I was thinking that you would be able to use a hall effect switch instead of a reed switch without any modifications to your timing wheel. If you're interested tell me I'll post you a simple schematic for this purpose, although it is quite easy to design such a circuit. But if you've done it with optos then that's ok. Just a suggestion.

Elias

Thanks elias,
but the optical switching works excellent.
Ok, here's the summary.
I finished the circuit with optical switch. Now I can get perfect timings for all four pulses per perriod. I had problems with bearings, when I bought them, they were sealed and filled with a thick grease that would not allow the bearing to rotate by its inertia for some time. When I cleaned the grease out, the bearing would rotate for about half a minute on its own by its inertia, but it also made some noise as if the balls inside were loose. Also some backlash appeared. So I cleaned the grease out and lubricated the bearings with some synthetic oil. I mounted them on the motor and could get speeds up to 900 RPM, but the motor made a terrible noise as if the bearings were broken, so I replaced them by new ones, but that decreased the rotor speed to about 500RPM. Anyway, I soldered the circuit and started testing. At first I run the motor with only one NPN transistor and one strand of the startor coil.
I slowly increased the resistance of the 200 Ohm voltage divider to see how high I can go till the RPM's start to dropp. I went up about 100 Ohms at a time. I used two high wattage pots for this. When I reached 600Ohms, I noticed a decrease in speed of the rotor. So I reduced the resistance to 550 Ohms and got almost the same RPMs as with 200 Ohms, so I decreased the resistance just a bit less than 550 and got the RPMs I need. Now there is a voltage divider at each NPN base with the resistance of 430 Ohms.
So at first I used only one NPN transistor and one coil, I put a charging battery on the output and got 485 RPM with 0.4 A current draw at 24V. Then I run the motor with two NPN transistors and two coils in parallel as in the circuit above. This time I got exactly 500RPM and a barely noticable increase in current draw. And when I put in the third transistor, the RPM's increased to 508. That is not what I expected. I figured, maybe I need a high wattage low resistance resistor between each emmiter of the NPN's and the negative battery lead. Because the transistors are not all exactly eaqual, one may open more than others and so take all the current on itself, leaving only small current for other transistors. So with these resistors between emmiters and negative terminal I could ballance the current for each transistor eaqually. Is this right? I figured that I could make such a resistors form a resistance wire wound to a correct ommage.
Thats it for now. I will test some more. Here's a picture of the optical switching part:


Thanks,
Jetijs

Jetijs,

Everything sounds good. The three parallel voltage dividers are doing what they are supposed to do.

There are two possibilities to explain the performance. The first possibility is that the PNP device is current starved and not opening up all the way. Try some smaller resistors in the voltage divider run by the photo-transistor. Your photo-transistor can handle 50ma, so try two 500 ohm resistors in this bridge.

If that makes no difference, then the motor is telling you that the inductive rise-time of the coil already fills the ON-TIME window at about 500RPM. So, to get the motor to run faster, take 50 turns OFF the coil. Remember I said originally that you should start with your #21 wire at 250 turns, and if the need arises, that we could take some turns OFF. Well, that time may have arrived!

Start by changing the resistors in the P-T section. If that doesn't work, remove 50 turns on your coil. The motor should turn faster, draw more current, and charge the back battery faster.

Also, please draw out the entire circuit as you have it now with the battery being charged, just so I can see that there are no other mistakes.

Great work!!!

Peter

Ok, here is the exact circuit:

all the diodes are rated 1000V and 4A and all the resistors are 2w.

Before removing 50 turns of wire I made another test, just to be sure. This time I used digital amp meter. Here are the results:

Transistors Current Voltage RPM
1 transistor 0,44A 24V 450
2 transistor 0,44A 24V 481
3 transistor 0,45A 24V 486

And these are the results when the 50 turns were removed:

Transistors Current Voltage RPM
1 transistor 0,55A 24V 463
2 transistor 0,56A 24V 562
3 transistor 0,56A 24V 572

Just saw the information in your edited post about the reducing the resistance of the optotrigger voltage divider. I think that you are right about that one of the transistors may not open fully, because when I try to run the motor with only one transistor I see a difference in performance compared with other two ransistors. I try them all out one after another. That means one of them does not fully open.
I will try to reduce the resistance as you told and post the results then

Edit:
One of the NPN's keeps blowing. I already replaced it 3 times. It always blows when I run the motor only with that one transistor and that happens on the very first impulse. There is a flash in the neon bulb acros the collector and the emitter of that transistor and the motor stops. One time the transistor blew so that it was constantly on, drawing 1.5A from the battery, other two blew with a little click sound and just stopped responding, when I removed them, I could measure a resistance of 460 Ohms between collector and base that is not there on a new transistor. I double checked the circuit and can not see why that happens, everything seems to be exactly like in the circuit above. Also that transistor does not blow when I run the motor with all thre coil strands and all three transistors. I will resolder everything on a new plate tomorrow and try it again as it is already late today.

Thanks

Elias,

This is totally OFF TOPIC for this thread. Please delete these posts and start a new thread.

Thanks,

Peter

Peter,
I resoldered everything on a new soldering plate. I decreased tha optotrigger bridge resistance to 500 Ohm. Now I ran the motor with only one transistor at a time, to verify that each one of them has the same performance. First transistor got the RPM's up to 480. The second transistor performed the same. The third transistor was a little weaker and got the RPM's to 475.5 (4.5 RPM's less than other two). Now I connected them in parallel, here are the results:

# of transistors | Current draw |RPM's

1_______________0,55A_______480
2_______________0,55A_______527
3_______________0,55A_______538

Does this mean I need to remove some more turns from the startor coil?
Thank you,
Jetijs