Muller Dynamo

[romerouk]

thanks for the info, Romero

sounds good about the smoothing cap - what capacity did you use & what did the Vout & Iout change to?

i guess with a little more windings on the gen coils that it would be possible to generate sufficient o/p voltage headroom to use for charging the i/p battery directly, without conversion (apart from rectification, of course)


a technique that i'm finding useful (with a different experiment) is to supply a load circuit direct from a battery as usual - then connect the o/p buffer capacitor to the battery via an inductor which is large enough to block switching transients between cap & battery

the cap would be ok to receive any transients from the o/p and it can develop a slight voltage increase above the battery voltage to keep a constant 'trickle' charge into the battery thro' the inductor

thanks
np


http://docsfreelunch.blogspot.com

I have used a 4700uf/25v and the voltage increased to 13.8 with the load

[neptune]

@Romero UK . Thanks for answering my questions . I am not sure why you are using multistrand wire in your coils . How many strands? or is this litz wire . Surely multistrand is used for high frequency AC? To the best of my knowledge , the charge/discharge efficiency of lead acid batteries is about 80% . I s the charge discharge efficiency of caps 100% ? I think it was nul-points who suggested some coil rewinding to remove the necessity of a DC-DC converter . A simpler solution would be to rewind only the drive coils to work on a lower voltage . If you do this , include taps in the winding to optimise things .
        There is a simple way to measure complex waveforms . I think it was originally Gustav`s idea .Make a box with 2 compartments and a lid of frosted glass or greaseproof paper . Inside , fit 2 identical bulbs . Feed one from the waveform , and one from a variable voltage DC power supply .Adjust supply until both bulbs are equally bright . Read the amps and volts from the meters of your power supply . That will give you the watts . An added refinement is to use a homemade light meter consisting of a small solar cell and a milliamp meter .

[nul-points]

I have used a 4700uf/25v and the voltage increased to 13.8 with the load

excellent! -  thanks, Romero

13.8V (on-load) sounds like it would sustain some feedback from your smoothed o/p to your battery

maybe you could try a couple of diodes (say 1N4007 or similar?) in series to get a forward volt drop approx equal to the difference between 13.8V and the  i/p battery voltage (approx 12.3V on-load iirc?)  -

try first with lamp load switched in - monitoring DC current thro' feedback diode(s) (starting on high Amp range & working down, of course)

if you have any schottky diodes (eg 1N5817 or similar) you could try different arrangements of 1N4007 (approx 0.7V) & schottky (approx 0.4V) to get different forward Vdrop between buffer cap & battery

if you can't sustain 13.8V or near, with lamp load, try without lamp - but start with higher diode Vforward values first


you must be very excited about this?  (your generator build)

In bocca al lupo! 
np


http://docsfreelunch.blogspot.com


[Edit: add clarifications]

[nul-points]

[...]
To the best of my knowledge , the charge/discharge efficiency of lead acid batteries is about 80%.
[...]
Is the charge discharge efficiency of caps 100% ?
[...]
I think it was nul-points who suggested some coil rewinding to remove the necessity of a DC-DC converter . A simpler solution would be to rewind only the drive coils to work on a lower voltage . If you do this , include taps in the winding to optimise things.

thanks for the LAB efficiency, Neptune

cap discharge *should* approach 100% (if cap/cct has low leakage, low internal series R, and you don't cause significant heat loss by the discharge currents)

however, there are losses associated with trying to get charge-separation into a cap in the first place (i had a whole thread going on this issue a couple of years back - confirming experimentally that around 50% of the supplied energy was getting dissipated in external resistance of supply load + wiring) - even when using a series inductor

some academic experiments have shown that increasing the number of pulses to switch current into a cap tends to reduce the losses

[oops - late edit!]  good idea about modding 2 driver coils, rather than 8 gen coils!   this might help if you used a stack of NiMH cells to give 8.4V or 9.6V, say, rather than the existing 12V i/p battery

if you keep the 12V i/p battery then you'd really need to bump up the output voltage  to get sufficient feedback volts - but Romero has achieved this anyway by adding the buffer cap, so no need to alter any coils - neat!


hope your cycling's been less eventful recently - bet it's been great weather to be out on a bike!
np


http://docsfreelunch.blogspot.com

[neptune]

@Romerouk .Can you please tell us more about your magnets . Size , grade , source of supply ? and could you please comment on my question re multistrand wire . Re the coil cores . Are these just pieces of ferrite rod from an old radio , or are they commercial cores ? Sorry about all these questions , but we need as much info as possible to replicate .It would seem from the video that your battery is 17 Amp hour . So if you dispensed with the machine, and connected that lamp directly to the battery , one could reasonably expect a run time of about 9 hours until the battery voltage drops to ,say , 11volts . That is if the battery is in good condition . However you say that the machine has run for several days without substantially discharging the battery .If this is the case ,you can forget about analysing complex waveforms .In my book , this shows indisputable proof of overunity .
@nul-points . Yes fine cycling weather , thanks but very windy . On the flat lands of Lincolnshire , the winds are our mountains ...