Thanks to advice from Gyula and SeaMonkey, see at

Gyula: http://www.overunity.com/13523/has-anyone-seen-lasersabers-new-motor-runs-on-1000uf-cap/msg364365/#msg364365
SeaMonkey: http://www.overunity.com/13523/has-anyone-seen-lasersabers-new-motor-runs-on-1000uf-cap/msg364301/#msg364301  ,

I could do some more tests with the DadHav circuit.

The new circuit drives the little ring magnet spinner very efficiently (in comparison to all other drive circuits I ever used) but the circuit is very prone to self-oscillations. The value of the capacitor C1 has to be chosen very carefully and also depends on the supply Voltage.

May be the experts have some ideas of how to stop the self-oscillations in a better way (other than with the 10 K resistor in series with the capacitor C1, which seems to be an insufficient way of doing it).

Greetings, Conrad

P.S.: a feedback diode over the coils cuts off nicely the back EMF spikes but slows down the rotation of the ring magnet significantly.

I'm curious...do you have any information or help for replicating this device from lasersaber ?

Conrad,

Your scope shots of the MOSFET Drain to
Source signals are quite interesting.

It appears that the MOSFET is not turning
off sharply and remains in the partially
conductive linear region for some amount
of time after turn off is desired.   This may
be responsible for its tendency to self
oscillate.

Presently, the MOSFET turn off is accomplished
by discharge of the Gate capacitance through
the 100K resistor connecting Gate to Source.

Reducing the value of this resistance will speed
up turn off and possibly curtail self oscillation.
You may want to try a 10K resistor there and
also a 1K resistor to note differences in circuit
operation with those.

@SeaMonkey: I tried a 10 K resistor between Gate of the MOSFET and Source.

Scope shots are essentially the same.

Self-oscillation has gone away below 11 Volt but seems to be higher above 11 Volt.

I still need the 10 K resistor in series with C1 to make the circuit work.

The problem seems to be the part: base of 2N3906 - 100 K - positive rail - coils - C1 - 10 K resistor - base of 2N3904.

Why is the 10 K resistor in series with C1 necessary? Note, the coils have a rather high impedance (280 Ohm DC resistance).

Greetings, Conrad

@Mags: Your rotor, ring for the coils and base plate look professionally cut or turned? Did you have them made or do you have the right equipment?

Greetings, Conrad

Everything is made by me and no special tools really. An older Dewalt jigsaw using laminate blades. A cordless drill and some bits made for plastics and a couple sizes of Unibit.  And a dremel tool with an assortment of bits as needed.

The laminate blades come in scroll or straight cut. If you take your time, you can get some amazing cuts with nice edges. If it takes me 1/2 hour to cut 15in rotor from 1/2in plexi then Im happy. And I have it now.  Well, in 30 min.  Then with some sand paper finish the edges and Im good. The laminate blades do well with not melting most plastics.

When I have to make a rotor, I make the piece larger than my rotor size and find center then drill to size. If the material is thicker than 1/4 in then I find center on both sides and drill pilot holes half way through then finish of each side with Unibit to fit hobby shop bearings. This helps to keep your hole straight when drilling by hand or you might get a wobbly rotor because the top and bottom bearings are not always dead center if you didnt hold the drill perfectly straight just drilling all the way through from one side. Then I mount the piece to the bearings then on the base to check level. Then having a mark for the outer diameter I spin the rotor and mark the circle to be cut. The reason I do this at times is if I cut the outside precisely, if that center hole isnt dead nuts, all that time fine cutting the outer diameter was wasted. From there if trying to sand or cut to correct, the dia will be smaller than what I originally wanted.

I use superglues with great success, but I dont trust it for everything.Light sanding and clean of the surfaces helps a great deal for longevity. Sometimes Ill use superglue just to fix a piece in place and then apply either marine goop or epoxy. Goop and superglue are my favorites. There is a superglue called ' Wonderlock em Tite Chairs"  It has a couple fine nozzels and it is the thinnest superglue Ive experienced. A tiny bit can be applied and cures quickly. Pvc board loves superglue. The bond is almost always stronger than the material itself. Excellent building material. I like to work with plexi a lot also.

Mags

I'm curious...do you have any information or help for replicating this device from lasersaber ?

Hey Forest

Laserhacker.com

There is a forum there for the 'EZ Motor"

Laser uses a 3D printer to make just about everything other than the wire, magnets and the needle rotor shaft with jewel as bearing. If the thing can be printed cheaply the rest is 6 magnets, about 15 bucks in wire and what ever the needle and jewels cost. he has a list of sources in the thread.    With 3D printer file for download, in case the is a 3D printer shop in town or if someone has a printer.


Im doing my own version with more coils and magnets. Also my coils are thin from front to back. I think with air core particularly, the field is weaker further away from the magnet for generators, so I want most all the wire in the coil being affected to be be in the same field strength as the rest. So the longer or deeper the coil, the windings further from the rotor will experience less field density than the windings closer to the magnets. With cores, such as seen in commercial motors and gens, the fields are controlled as to where they go, where they 'snap' when 'cutting' windings.    So I believe for air cores, the length of the coil for motors doesnt mater that much, but for gen, the shallower the better. Here with Lasers motor, it is motor and gen with same timing. As its running if you give the rotor a bit of a speedup by hand or blow air, the cap voltage rises to the speed the rotor is spinning, and will then be running at that voltage/speed. Very cool little motor.



Mags

Hi Conrad,
.........

You may wish to use a 10 kOhm emitter resistor: just connect the emitter of the pnp to one leg of a 10 kOm and the other leg of this resistor would go to the battery positive. And also use a 10 kOhm between the gate and source instead of the 100 kOhm so that the voltage gain of the pnp stage would be 1 (10/10).  Now omit the 10 kOhm in series with the 100 nF and maybe use a 100 kOhm potmeter instead of the other 100 kOhm resistor between the base of the pnp and battery positive, to have some means for further controlling (reducing) the AC gain.

To answer your question why the series 10 kOhm is needed: it influences the AC (and the DC) gain of the circuit (its role is similar to an op-amp amplifier circuit whose output is fedback by a resistor, say R2, to its input and there is a series resistor, say R1,  at its input to receive the input drive, so that the ratio of the two resistors, R2/R1 defines the gain of the amplifier stage).

Also, when you increase supply voltage, the voltage gain normally also increases, this explains why oscillation returns above a certain input voltage level where loop gain can become just enough aagain to cause self oscillations.

........

rgds, Gyula

Progress, I am talking about an improvement of this circuit (PNP and MOSFET pulse motor drive circuit without sensor):
http://www.overunity.com/13523/has-anyone-seen-lasersabers-new-motor-runs-on-1000uf-cap/msg365075/#msg365075

@Gyula; you gave exactly the right advice.

I am referring to the attached circuit:

R4 (between emitter of 2N3906 and positive rail) seems to stop self-oscillation. 1 K is a good value, 10 K is too high (changing R3 does not help).

R2 (between capacitor C1 and base of 2N3906) is very critical, needs to be 20 K, but should be higher for more than 10 V supply Voltage.

R1 and R3 (pull up / pull down resistors) seem to be good at 100 K, changing them has only decremental effects.


The following seems to be the right procedure to come up with a reliable and efficient circuit:

1) roughly define the supply Voltage (e.g. 7 V to 10 V, or 12 V to 15 V)

2) adjust R2 from 20 K to 200 K to find best efficiency

3) adjust R4 around 1 K (probably not necessary, 1 K will do the job of stopping self oscillation)

Will do further tests, but the critical components for a "PNP and MOSFET pulse motor drive circuit without sensor" are identified (R4 = 1 K, R2 from 20 K to 200 K depending on supply Voltage).

C1 = 100 nF and R1 = R3 = 100 K seem to be good values.

Greetings, Conrad

P.S.: At supply Voltages higher than 15 V the Gate of the MOSFET has to be prevented from going higher than 15 V!

.......

A) This morning I was lurking at another forum and noticed a schematic I totally forgot about and it exactly shows a MOSFET switch with a pnp transistor, drawn and tested by JoeFR (also tested by Romero). See the schematic here:
http://www.underservice.org/index.php?topic=3.msg584#msg584   (click on the schematic to blow it up)
JoeFR showed scopeshots on the waveforms two posts down wrt the post the link points to.  He compared the performance to a pnp - npn switch shown earlier by Romero, this latter is here: http://www.underservice.org/index.php?topic=3.msg579#msg579 
A piece of advice from JoeFR on the tuning of the MOSFET-pnp circuit: http://www.underservice.org/index.php?topic=3.msg590#msg590

B) Now that I have seen and recalled the JoeFR schematic (which evolved out from Romero's findings of a pnp-npn circuit to get rid of Hall or reed switches as sensor elements) I think it would be worth testing by you too. I suspect the drive coil (L1) in his test has a low impedance because the spikes are in the some hundred Volts range and it must involve much higher coil current when the MOSFET is ON. This lower coil impedance also means the circuit is less prone to self oscillations (versus high impedance coil(s) you use). I do not know whether JoeFR has tested lower value coupling capacitors for C1 (shown as 2.2uF versus your 100nF or 200nF). Diode D2 clamps the positive base-emitter AC peaks coming from the coil (either by induction or by spikes), effectively saving the base-emitter junction from overloading in the reverse voltage direction.

.......

@Gyula:

Ad A): Great find, it helps a lot to see JoeFR's circuit. In principle similar but some more ideas for tuning. The diode from the base of the PNP to the poisitive rail is interesting. May be the diode protects the PNP from a  too high Emitter/Base Voltage (V-EBO), which is critical. It is good to see that the circuit works well for a more hefty pulse motor (low DC resistance drive coil).

Ad B) I suspected that the circuit has to be adjusted not only to the supply Voltage but also to the particulars of the drive coil.

The Voltage at the Gate of the MOSFET has to be considered carefully. I destroyed an IRLIZ44N because I forgot that its Gate should not be subjected to more than 16 Volt. I thought it is 20 Volt and the 20 Volt were already too much. To all who experiment with this circuit: buy more transistors, you will need them.

Cleaner switching: the scope shots over Drain /Source of the MOSFET have not changed much. A sign for a good adjustment of the circuit to the supply Voltage seems to be a 50% ON-time of the MOSFET (with my ring magnet spinner and two 280 Ohm coils in parallel). With a too high supply Voltage the ON-time becomes up to 70% and the motor just consumes more but does not turn faster. The ON-time of the MOSFET is tricky, it can also become longer with a too low supply Voltage. I do not yet fully understand this. It is pretty evident that the circuit has to be adjusted to the supply Voltage, I see that clearly.

Greetings, Conrad

Here is a pic of one of the 24 coils, single wire. 3300 turns till just about full. 649 ohms.  24 coils=   79200 total turns  15.5kohm all in series.

Will make 24 more wound bifi after.


Mags