Muller Dynamo

[konehead]

Hi Ron

thanks for the help - I was hoping you would take a look at it and give your read on it...

that circuit you have up is exactly what I was doing before throwing in the opto isolator...

wouldnt the H1D11 opto "isolate"the leg 3 of the halleffect (its "output)  to the "intput:" of the 4421 (now 4422) driver via its phototransistor??

I see that it inverts the signal, thats why I changed to 4422 driver from the original 4421 that wa sbeing used before otpo is installed...(jsut like your circuit)

but doesnt it as well as inverting, also isolate the signal form the hall effect to the driver and so it works like and isolated-relay (opto controlled) between the halleffect and the driver???

"RS" who send me these "H1D11 opto between hall effect and driver circuit" a few months ago,  says that it will dampen any ringing occuring form coil-shorting or whatever might cause some ringing....and I used the opto since I was getting "ghost' triggering from adjacent halleffects when I had 5 mosfets firing sequentially, all sharing same ground and source leads....(thatgs reaons for doing this)
I have a RomeroUK-variant machine working and testing that I thought would be fun to fire each coil position, (5 facing coils - so 10 coils total - coils wired in facing-paralell)
and with the 4 all-N rotor magnets, I could actually fire the coils 20 times in one revolution, so it would hav lots of toque plus lowered the pulse widht down to 1/2ms too so it draws hardly anny current (around 50ma shows on analog meter)  then after I got that going reliable, then next step is toshort the coils right after the motor coil swtihc OPENS and that would fill up caps, and caps dump to run cap (or stright to run cap infrist place)  and hopefully this would capture the recoils psike andbackemf too if I can get the coil shorting close enough in time to the switch opens....so idea is it would be "shorting" the actual backemf as ell as the induced power the magnets going past the coils make too upwards in inwards intot the caps and non-reflecitve too.....anyways that the plan but had problems with the motor-coil circuit in first place before moving onto 2nd step becasue of the ghost/bleed-through problem of adjacent halleffects..... this opto seems to fix it - maybe jsut a transistor would too that inverts the signal...maybe its somethign else altogether causing the ghost triggering too

....I worte about this over on EVGRAY - Mike Ross said same as you that its not isolating since its common pos suplly both sides but I am looking at is like its a light-contorlled relay, the light does the isolating, the relay does the "inverting" and switching...

and you do want to have the pos supplies not on different sides of the H1D11. since it being like relay you need to supply both sides anyways...could be all wrong as usual but jsut trying to make it work right.

[i_ron]

Hi Ron

thanks for the help - I was hoping you would take a look at it and give your read on it...

that circuit you have up is exactly what I was doing before throwing in the opto isolator...

wouldnt the H1D11 opto "isolate"the leg 3 of the halleffect (its "output)  to the "intput:" of the 4421 (now 4422) driver via its phototransistor??

I see that it inverts the signal, thats why I changed to 4422 driver from the original 4421 that wa sbeing used before otpo is installed...(jsut like your circuit)

but doesnt it as well as inverting, also isolate the signal form the hall effect to the driver and so it works like and isolated-relay (opto controlled) between the halleffect and the driver???

"RS" who send me these "H1D11 opto between hall effect and driver circuit" a few months ago,  says that it will dampen any ringing occuring form coil-shorting or whatever might cause some ringing....and I used the opto since I was getting "ghost' triggering from adjacent halleffects when I had 5 mosfets firing sequentially, all sharing same ground and source leads....(thatgs reaons for doing this)
I have a RomeroUK-variant machine working and testing that I thought would be fun to fire each coil position, (5 facing coils - so 10 coils total - coils wired in facing-paralell)
and with the 4 all-N rotor magnets, I could actually fire the coils 20 times in one revolution, so it would hav lots of toque plus lowered the pulse widht down to 1/2ms too so it draws hardly anny current (around 50ma shows on analog meter)  then after I got that going reliable, then next step is toshort the coils right after the motor coil swtihc OPENS and that would fill up caps, and caps dump to run cap (or stright to run cap infrist place)  and hopefully this would capture the recoils psike andbackemf too if I can get the coil shorting close enough in time to the switch opens....so idea is it would be "shorting" the actual backemf as ell as the induced power the magnets going past the coils make too upwards in inwards intot the caps and non-reflecitve too.....anyways that the plan but had problems with the motor-coil circuit in first place before moving onto 2nd step becasue of the ghost/bleed-through problem of adjacent halleffects..... this opto seems to fix it - maybe jsut a transistor would too that inverts the signal...maybe its somethign else altogether causing the ghost triggering too

....I worte about this over on EVGRAY - Mike Ross said same as you that its not isolating since its common pos suplly both sides but I am looking at is like its a light-contorlled relay, the light does the isolating, the relay does the "inverting" and switching...

and you do want to have the pos supplies not on different sides of the H1D11. since it being like relay you need to supply both sides anyways...could be all wrong as usual but jsut trying to make it work right.

Doug,

It is good to see you using a proper gate driver after the hall. Kudos

The A1120 is a robust little unit with its own voltage regulator and good clean output with Schmitt trigger, so no problem there.

What can cause ghost triggering is if the hall 'sees' the coil firing but if you have a separate mag wheel for the hall(s) then the firing should be clean.

This only leaves the gate driver as a source. Note on the spec sheet there is a 4uF7 cap common to pin 1 and 8 and a .1uF on each  pin 1 and 8. So if the supply to the gate driver is clean then there should be no problem. It is always best to run your logic chips on there own supply. Using the coil plus rail is not a good idea because of the spikes and weird voltages present when the thing is running and generating as well as being pulsed.

So if you can use a separate supply for the hall and gate driver?  9 volts incidentally is just borderline for the 44XX, the suggested gate voltage for the IRFP460 is 10 volts so a 12 volt supply would be the ticket.

Note that the max rating on this fet is 220 watts, not the 500 X18 (9000 watts) Also watch the build, the all plastic encapsulated fets are usually much harder to cool so are de-rated

The problem with opto's is their slowness and the ugly wave shape out of the transistor when you try and run them fast. The 44xx's are 'sort of' a Schmitt trigger input, so take this and clean it up a bit, like they switch around 2.5 volts to make the 12 volt output pulse.

If you are still getting false pulsing change the resistor between pins one and three on the hall. It will work anywhere from 3k to 20K but the lower values might be cleaner.

Ron

Edit:Most opto-isolator datasheets quote the rise and fall tmes of the opto-isolator outputs rather than a maxmum frequency. How can we use these values? The diagram below shows a more realistic signal from an opto-isolator.

In this diagram, the rise and fall times are shown equal, but they are often not - especially with open-collector or open-drain type optos where an external pull-up resistor controls the rise time.

[chalamadad]

You should read this interesting article about how to build a negative inductance and a time reversed magnetic field. Please give comments.

http://4gang.wordpress.com/how-to-build-a-negative-inductance/

[joefr]

Hi all forum members

Please check bottom video I made about Pulse motor speedUP with BEMF recovery to Light Bulb:

http://www.youtube.com/watch?v=C65JJg2bacA

This is the schematic I used:

JoeFR

[mondrasek]

A note about rotor aerodynamics:  I recently took off the top stator coil mounting plate from my rig in order to make another one to mount some test coils.  While that stator plate was off and the rotor was exposed I went ahead and made some adjustments for better aerodynamics.

My rotor is a 1/2 in thick piece of plywood.  I created rotor magnet "pockets" with a flat bottom only 3/8 into the rotor with a Forstner bit.  So the bottom of my rotor is all wood and smooth.

The magnets I placed in the "pockets' are only 1/4 inch thick.  So that left an unfilled pocket of approximately 1/8 inch deep above each rotor magnet on the top side.  Those pockets would be causing drag on the rotor due to the turbulence they create in the surrounding atmosphere.

So while the rotor top was exposed during the disassembly I covered all the exposed magnet pockets with cellophane tape.

The results?  Before taping over the pockets the max RPM for a specific input power was ~2330.  After covering the pockets the RPM at the same input power is ~2850.

M.