Quick hall effect sensor + mosfet conversion to sg wheel

[bigmotherwhale]

i have been doing a few experiments just now, i have to say MUCH better with hall switching to see what's going on, you can pick a point on the magnet to look, and monitor output on the neon and battery input / output
just tried some coil shorting which i have not been able to do until now, and have seen the sec battery charge with it with just a spin of the wheel by hand. might have to put in a reverse diode see if i can get the current loop.
also have seen energy flowing back in little blips on the scope when the power is turned off, dont really know whats going on there, sometimes its high as half the supply voltage for an instant and then drops down to a volt, you put the wheel in a different place and it changes instantly not sure whats going on there yet.
the coil shorting is interesting shows the forces at work "inside" the coil you can see the current flowing on the needle in reverse, and how different switching positions affect it.
Primo

[bigmotherwhale]

well thats the end of the mosfets thought they would last longer than that, not really sure what killed them, they were running cold and the wheel going well untill missed a few beats and then nothing...
i stopped the neon lighting up using a full bridge, it seems without the trigger windings dampening the spike, it was no longer one sided. maybee to late for the mosfet...
has anyone here got knowledge of gate drivers? im currently running strait of a h1d1 optocoupler into the mosfet gate clamped down with a 10k resistor, i have some bdc243c transistors and i have seen those on the back end of a cap dump circuit being used to drive the gates, anyone help?!

[bigmotherwhale]

Improved the mosfet driving circuit with a Darlington paired optocoupler h11d1 and bd243c and heavy resistor pull down loading, now getting square edged hits, replaced the mosfet with 600v rated coolmos pulled from a server grade PSU and installed pulse discharge capcitor on supply rail. its cranking FAST.
200ah battery pushed from 12.7 to 12.9 in 25 mins under 800ma load and a 15 - 20 percent duty cycle, i have all the windings on the core parallel for now before i insulate for HV and run them in series.

[Gary Hammond]

Hi Brain,

Hi Brian,

Give me a few days and I can post a schematic of what I did. I'm currently away from home on a mini vacation and posting this from a motel room using my wife's laptop.

cheers,

Back from vacation now, so here's the schematic of my circuit using two hall switches on two small PNP tranistors to trigger the five MJL21194's. The hall's are mounted separately for adjustability, but both are triggered from the south pole of the same magnet whose north pole is approaching the coil assembly at that same time.

[bigmotherwhale]

Mosfet driver working really nice by the help of krill, got rid of the curve at the bottom of the square on the trailing edge, by using a transistor to short the gate / source
i decided to finish the isolation and build a simple hv psu out of some junk computer bits and a vairac.
made an air coil out of a full spool of copper, and joined it up, not going to give the best capacity either.
turned it on and it worked first time, i can run from anywhere from 0 to 240v. most i have gone is 130 at the moment, mosfet temp is a reasonable 35 degrees C. most ive pulled from the mains is 98 watts!
efficiency is not good, i need either a core in there or some powerful magnets and a flat coil. im going to series link an old 8 filler with a core and should work MUCH better.

http://vid936.photobucket.com/albums...psf0b822bd.mp4

[Brian McNece]

Thank you Gary, I would like to try triggering my sg with a hall, just to see the difference.

Hi Brain,



Back from vacation now, so here's the schematic of my circuit using two hall switches on two small PNP tranistors to trigger the five MJL21194's. The hall's are mounted separately for adjustability, but both are triggered from the south pole of the same magnet whose north pole is approaching the coil assembly at that same time.

[ATTACH=CONFIG]3738[/ATTACH]

[bigmotherwhale]

made a coil out of my old 8 filler by joining all in series, this has an iron core, the machine is going well, i have had it up to 260 volts, and the mosfet stays stone cold, so does the coil, both around 22deg c
the load bulb which is a 100w incandescent is glowing very brightly looks like a 60w bulb,and the pulse width is very very short, im pulling about 50 - 60w from the mains and the wheel is going like the clappers, i think if i connected it strait to the wall through the variac it would proably be not far off in brightness for the same input, i may have a go and see with a cadmium sulphide sensor and meter,
the wheel bearings are incredibly noisy not sure how much life its got left in it, starting to get a bit sketchy.
need to make some decisions as to what sort of rotor to make next really and how or run it at a low speed for now, might change the capacitor on the 555 for even shorter pulse width.

[bigmotherwhale]

i have taken my machine all the way up to 320 volts with no bad effects the coil im driving is 27ohm high capacity wound.
seems very efficient indeed.

[Brian McNece]

Gary
Can I ask, why two hall sensors? What is the benefit vs. one? Also if it is possible to just eliminate that top hall? Curious I guess, I hadn't ever seen two used like that.


Back from vacation now, so here's the schematic of my circuit using two hall switches on two small PNP tranistors to trigger the five MJL21194's. The hall's are mounted separately for adjustability, but both are triggered from the south pole of the same magnet whose north pole is approaching the coil assembly at that same

[bigmotherwhale]

the two hall sensors are configured so you can adjust the pulse width otherwise pulse with will be determined by the magnet length, which will waste alot of current.
i will put up the schematic i used with a 555 for the pulse width, it can easily be altered for driving transistors, either will work just fine, mine uses the hall sensors found in brush-less fans, pretty easy to get hold of.

[Gary Hammond]

Hi Brian,

Gary
Can I ask, why two hall sensors? What is the benefit vs. one? Also if it is possible to just eliminate that top hall? Curious I guess, I hadn't ever seen two used like that.

Like BMW said, I wanted to be able to adjust the pulse width as well as the timing. Two hall circuits in series lets me do this. Also as BMW said, a single hall will be on for the entire width of the magnet as it passes.

I have three different 26" bike wheels. One with 18 magnets (4" spacing), one with 21 magnets (3.5" spacing), and one with 24 magnets (3" spacing). All the magnets are 7/8" wide, so that calculates different percentage on times for each wheel. The 4" spacing gives 22 percent on time, the 3.5" spacing gives 25 percent on time, and the 3" spacing gives 29 percent on time.

What I wanted to achieve is the exact amount of on time it takes to fully charge the coil each magnet pass. No more, no less. Speed of rotation, magnet spacing, pulse width, and base current all affect the amount of fill time of the coil. So being able to adjust both the timing and pulse width are imporant factors in fine tuning for maximum effect.

As it turned out, I am using the 18 magnet wheel with 4" spacing which would give 22 percent on time with only one hall switch. It seems to perform best at 20 percent on time with this particular combination of wheel and coil. It runs at 260 RPM and draws about 500Ma without any added mechanical load, and runs 235 RPM with an added genny coil lighting 20 LEDs for the same or slightly less current draw.

[bigmotherwhale]

Hi Brian,



Like BMW said, I wanted to be able to adjust the pulse width as well as the timing. Two hall circuits in series lets me do this. Also as BMW said, a single hall will be on for the entire width of the magnet as it passes.

I have three different 26" bike wheels. One with 18 magnets (4" spacing), one with 21 magnets (3.5" spacing), and one with 24 magnets (3" spacing). All the magnets are 7/8" wide, so that calculates different percentage on times for each wheel. The 4" spacing gives 22 percent on time, the 3.5" spacing gives 25 percent on time, and the 3" spacing gives 29 percent on time.

What I wanted to achieve is the exact amount of on time it takes to fully charge the coil each magnet pass. No more, no less. Speed of rotation, magnet spacing, pulse width, and base current all affect the amount of fill time of the coil. So being able to adjust both the timing and pulse width are imporant factors in fine tuning for maximum effect.

As it turned out, I am using the 18 magnet wheel with 4" spacing which would give 22 percent on time with only one hall switch. It seems to perform best at 20 percent on time with this particular combination of wheel and coil. It runs at 260 RPM and draws about 500Ma without any added mechanical load, and runs 235 RPM with an added genny coil lighting 20 LEDs for the same or slightly less current draw.

I was playing with the machine last night and the good thing with adjustable voltage and pulse width is that you can do exactly what you described, you can see on the scope when the coil is full by when it tops out, increasing voltage changes how much energy can be stored in the coil, its very cool, also you can see on the scope vertical lines where the on time is on the superimposed waveform, which means you can select the up ramp or the down ramp using the hall position

[Brian McNece]

Thank you, Gary and BMW.

[Gary Hammond]

Hi All,

Well, I was in error in my last post, as I was remembering an incorrect figure. Started playing with it again (hall switched SSG) and measured the current draw at close to 800ma not 500ma like I thought and posted previously. If I can improve the output of the genny coil just a little I think it will power the hall switching circuits and reduce the input current draw by the same amount.

[Volty]

Hi Gary,

Instead of shorter Hall On-time with two in series, this way or a slanted bar magnet can make longer On-times using 3175 Latching Halls. For SG, most find the bike shaft is not meant to rotate, but magnets on the covered spokes one side allows placing small magnets at almost any radius. The second side of this one was to be for FWBR + to Battery, but that i will mechanically commutate