"Enhanced Generator" from JPKBook

[Yaro1776]

Okay Gary,

Hi Yaro,



The hall Jules is using is positioned with the back face toward the north poles of the rotor and the sensing face toward the south poles of the magnets. I suspect the north poles are turning the FET on and then when the coils are conducting the south poles of the coils are keeping the FET turned on too long.

The distance from the magnet does have some effect on how long the hall is triggered on. But it has to be close enough to trigger reliably each magnet pass. The width of the magnet more greatly affects the on time than the distance away does.

Good clarification of the Hall operation and distance issue. From a simplified approach the position of the Hall between the coil and magnet appears to be problematic. If as you suggest that the Hall/magnet field is the trigger, then the coil magnetic field is redundant and skews the control aspect in this instance - makes sense.

I will note that the Hall device position is not centered on the magnets (Page 9 of the SPG.pdf). This vertical position, if varied up or down, should influence the total on time, when taking in consideration the magnetic field shape from a circular magnet (cone) versus a block magnet (rounded rhomboid or something). Why else does the author show the pic and the off center Hall position in the .pdf? Point of max field strength is always the magnet center. All very engaging, but of uncertain value to Jules at this point.

I believe I am slowly being sucked into this project. Thanks for the diversion.

Thanks guys,
Yaro

[Faraday88]

Hi Yaro,

I can confirm that my cores are all producing inward facing South poles so 'attraction' mode is the way to go. However my Hall sensor seems to have quite a wide angle of response in that it triggers about 15 degrees before TDC and switches off about 15 degrees after. I am therefore going to need to move it quite a bit towards the approaching magnet as indicated in the two attached diagrams I have drawn to show the issue.

Alternatively, I am quite drawn to the electronic option of setting a time delay from the moment it switches on, and then an off time (hence a pulse width) using the 555 timer chip. Could be a neat option but I will need to bounce some design attempts of a few people

Jules


[ATTACH=CONFIG]7068[/ATTACH][ATTACH=CONFIG]7069[/ATTACH]

Hey,
You can use a Mono-shot or Mono-stable multi-vibrator circuit using the 555 Timer. and set the required delay.
Rgds,
Faraday88.

[JulesP]

Gary,

That's an interesting point - that the rotor facing side is turning the current on and then the South pole being created as a consequence is keeping the Hall triggered longer than it would otherwise. I guess this would not happen if the Hall sensor was of a type that only had one main sensing face as I perhaps yours do?

I am going to try first to move the one sensor to the right, as per one of the two position diagrams I sent yesterday, and see if I can improve matters. If that doesn't work and I can't get the refinement of 'on' time and duration, then I will explore the double sensor arrangement. My coil positions are 72 degrees apart (to an accuracy of about 0.5-1 degree) so I could place the sensors by two different coils and build adjustment into them. I will first draw another diagram depicting that as it helps me visualise the sequence of events and, if I go down this route, will bounce the revised circuit of you. I note from what you say that this method is independent of rpm.

Jules

[JulesP]

Hey,
You can use a Mono-shot or Mono-stable multi-vibrator circuit using the 555 Timer. and set the required delay.
Rgds,
Faraday88.

True but it has been pointed out that this will work only at a specific rpm as it is time based. Ideally one wants a distance or angle based method so its independent of rotor speed.

Jules

[JulesP]

2 Hall Sensor operation

Hi Gary,

I have put together three drawings that simulate your 2 sensor setup and operation so I can figure out the best positions. It seems that my wide 'Trigger zone' will interfere with the clean cut off of sensor 1 at the appropriate moment so I have had to move it quite a lot further round to work. I hope the drawings and text on them are clear and I look forward to any thoughts.

Jules

[JulesP]

Gary, When using an inductive strobe won't that only pick up on the HV pulses (coil outputs) rather than the coil current switch on event?

Jules

[Gary Hammond]

Hi Jules,

Gary, When using an inductive strobe won't that only pick up on the HV pulses (coil outputs) rather than the coil current switch on event?

Jules

Yes that's correct and it would be at the shut off point when the magnetic coil collapses.

Better yet, you can use a LED (with dropping resistor) in parallel with the gate of the FET to find both the turn on point and turn off point. It will stay on for exactly the same duration as the FET.

[Gary Hammond]

Hi Gary,

I have put together three drawings that simulate your 2 sensor setup and operation so I can figure out the best positions. It seems that my wide 'Trigger zone' will interfere with the clean cut off of sensor 1 at the appropriate moment so I have had to move it quite a lot further round to work. I hope the drawings and text on them are clear and I look forward to any thoughts.

Jules

What you show looks good, but your comment on "stage 3" has one incorrect statement regarding pulse width. Moving the sensors closer together will increase the pulse width. Moving them farther apart will decrease the pulse width.

The leading sensor sets the turn off point and the trailing sensor sets the turn on point.

[JulesP]

Hi Gary, I already have an LED (shown in my current circuit) that comes on when the Hall sensor triggers the transistor which then triggers the FET so I know at what point the current, in relation to the rotor/magnet position, should come on and this will be useful to find the right starting position for the sensor(s). Small adjustments after that will produce the optimum rotor speed for the lowest current drawn on the battery. I gathered that the value of the strobe is to see the ‘frozen’ position of the magnet when the coil collapses and which I presume will be somewhere near the middle of two cores.

Ok re sensor spacing. Seems counterintuitive but if one works it through s l o w l y . . . . .

Jules

[JulesP]

Motoring at last

Hi Gary and Yaro,

I have it working! . I set the sensor on an adjustable slot so that it just turned off at TDC and, after a manual spin, it stayed running much to my delight. However, it was only about 200 rpm and best speed and lowest current (1.2 A) were achieved when the sensor was moved about 4mm towards the approaching magnet. I have yet to measure the rpm but I guess 450-500. I think that the few ms delay in the collapsing field meant that in the first position there was still too much pull from the temporary S pole.

You can see a video of it running called 'Moving at last.mov' at: https://www.dropbox.com/sh/0y15dybr5...n2q6QhHCa?dl=0

I have no idea yet what is happening to the pulses or if indeed the battery is charging. I mentioned the reverse polarity to the 'developer' but he has not observed anything like that in his case and his battery was charging well. I am also aware of the thoughts that discharging and charging at the same time is not the best way to operate the battery but I am keen to try and keep just the one battery. If necessary I could envisage a system where, for a few ms, the battery was only being discharged and then for the next few ms only being charged in a repeating cycle. No doubt the good old 555 chip could come in handy here.

Re measuring the HV spikes, can I put my scope straight across the capacitor, in attached circuit, without any special attenuation and would increasing the capacitor to say 10,000 uF make any difference.

I want to thank you and others in the forum for helping me get over these hurdles. I'm sure the next stage of discovery is only just beginning.

Jules

[Faraday88]

True but it has been pointed out that this will work only at a specific rpm as it is time based. Ideally one wants a distance or angle based method so its independent of rotor speed.

Jules

Tweeking the mono-shot multivibrator should get you to the sweet spot that you are looking for your set up of the rotor..
Rgds,
Faraday88

[Gary Hammond]

Hi Jules,

You can see a video of it running called 'Moving at last.mov' at: https://www.dropbox.com/sh/0y15dybr5...0last.MOV?dl=0

Nice work.

[Yaro1776]

Hello Jules,

Very good to see your machine in operation, sweet! Congratulations to you on your persistence and focus. Now the real adventure begins! Keep us posted on your progress... BTW the .mov came across in the vertical position, not horizontal - it may be my player...

[JulesP]

Thank you.

I took it on an iphone in a normal way. Guess you had to watch it at 90 degrees!

One technical query: Why don't Bedini type systems discharge the coil pulses straight into the battery rather than charge up a capacitor first and then discharging that? Wouldn't the result on the battery be the same? Agreed it is much easier to measure the voltage, and hence energy changes, of the capacitor but you need the additional switch to dump the charge. I expect that can be done electronically instead so that when the capacitor reaches a certain voltage value it discharges - like those water buckets in a playground; when they get full they tip over and empty the contents.

The next stage of my journey is to see what's happening to the pulses . . . . . . .

Jules

[JulesP]

Revised Circuit

Hi Gary,

Yes I thought that the moment I wrote it.

What I have come up with as a 'temporary' solution is to incorporate an isolating FET switch between the battery and the top of the coil. I don't know if the FET needs to have its Source at 0V but I'm sure you get my idea. Maybe this can be done better with a regular BPJ such as the BC549 I am using now or a power transistor but this is how I see it working with most of it what is working now. I refer you to the attached pen modified drawing 'Revised Circuit with Isolator' and the 'BEMF Generator V3 (Revised)'.

When the Hall output pin goes low the collector on the tranny goes high which sends a high to both FETs. This allows current to flow from the battery with the additional FET acting as a one way valve in the battery + line. When the passing mag field switches off the sensor and both FETs are off then the now HV positive at the bottom of the coils can only flow to the battery through the three diodes.

As I say, it might be simpler to use a regular power transistor (I have a spare T13009) but then it has to allow up to 4 amps through it.

If this works then at least I can get the battery responding to the pulses and tackle the question of discharging and charging simultaneously to later.

Regards,

Jules