Akula0083 Light No3 - Dual TL494

[MarkE]

It looks like you got yourself straightened out with respect to the coil orientations.

[d3x0r]

I've run into another conundrum... (solved; skip to ----------- )


I extended the leadin time on M1/M2 clocks and extended the charge time, to no effect.


first; M3 is enabled to conduct, which drops 4KA (no resistances) instantaneously to charge the power rail to a little more than 16 volts (the voltage enable is at 20V so whatever the mosfet threshold is subtracted from that is the total voltage)


at 100us M3 is disabled, but instantly voltage built up against the M1/M2 mosfets disappears (V(M1Source),V(M2Source)... but I can't find where it goes; there should be an equivalent current at that point for that slope of voltage change that would be KA; but it's nowhere... It doesn't go through the load (I(load)); it doesn't go through any of the diodes... it doesn't go through the source of M1 or M2; during the initial charge, there should be a current in the coils; but none shows...
L1 can only be interacted with by induction
zooming into the start... okay so M1 and M2 start conducting, because I drop the power rail below 0; had to adjust clocks to cover negative voltage on the gate...




The operation of this isn't quite what I thought to begin with; because I thought you could mix 2 currents (you can with matched inductances; but it's still only 1/2 through each but anyway)


In this case; M2(L3) is pulsed at a higher frequency (TL494 with both output tied together)... this really pulses L1 which draws current from the power to the ground...


M1(L2) is a lower frequency; it ends up having something like a DC bias that's actually a fairly high current which should be increasing voltage on the power rail (it doesn't)


------------------


when M1(L2) conducts; it results in a bias that causes a diminishing DC(ish) flow...


power rail and lowrail move up and down together; but their delta is always about the same... over time of course the overall power bleeds to nothing....


so; I don't know where the scheme can be improved...
this is L3:12uH,  L2:50uH (1:4, double windings, 7 and 13 for instance)  and L1 200uH (another 1:4, double windings again to ... 26 windings)...  L1 might be more like 100 since it's only (19 windings?) )


There's not a lot of 'free' capacitance for resonances...

[MarkE]

You should not have the collapses that you see.  I suspect a crossed connection that is not apparent in the schematic.  Disconnect MF2 and see if the circuit operates reasonably.  Then reconnect MF2 and disconnect MF1 and do the same thing.

I have redrawn the schematic here.  He only has a clamp diode on the MF2 drain, even though he should have one on both drains.  If your goal is to model what JB built then I would start with his schematic. 

He's using 1N4007s where their long reverse recovery are going to cause problems:  D3, D4, D5.  The rig in his video looks like it has a lot of undesirable stray inductance that is good for blowing up MOSFETs.  Q1 and Q2 in the schematic are damn it alls as shown.  I have reversed the emitter and collector connections so that they will do something useful.  I don't know how his board is wired up.  With Q1 and Q2 wired as in his schematic, MF1 and MF2 turn off much more slowly than as I have drawn them.  Turning M1 off really fast without protecting it with a clamp diode is a risky affair.  I recommend adding a HFA08TB60 from MF1 drain to V+ the way that D6 is connected to MF2.

[d3x0r]

well; I isolated ground behind a diode; so my 'lowrail' can go below 0... the last picture I posted - other than the sliding power/low rails, looks pretty good....


(references with regards to MarkE's schematic )With coils in the right directions, they don't cause ... actually before I finished that, I did check;  Q1 doesn't get a very high low charge; so the mosfet body-diode effect is probably sufficient... Q2 source gets huge spikes of both high and low voltage.  Since the coil on Q1 is the least inductance; if it gets driven it's a stepdown from the voltage driving it... Although not a horrible idea to include high current diodes to shunt negative flows from the mosfets.


I was most interested in the characteristics of the inductor portion of this...


On this other thread http://www.overunity.com/6763/energy-amplification/msg402534/#msg402534
someone posted about bearden's step-charge caps... http://jnaudin.free.fr/html/tbfrenrg.htm


It strikes me that the intent of the high frequency clock is to step charge the cap so when the low frequency hits it has ample charge to dump, and drive 19T coil to get the real power; (to build most momentum)


Both 7T and 12T coil drive 19T coil in the same direction; but moreover, when 12T coil is energized, it's collapsing (forward emf) is a large draw on positive potential points; and the circuit could benefit from a diode similar to D6 on 12T coil being added to the 7T coil to power.


so revised hypothesis on timing:
So low frequency driver should drive near to saturation for a length of pulse...


and the high frequency driver should be ... as short and rapid as possible... although lots of cycles causes lots of power usage....


... although the high frequency probably shouldn't be less than the pulse width of the low frequency driver... do not want these to collide... because they do not have an additive property.



; might replace power sources with TL494 blocks just to have a complete model; already have series capacitance and resistance on coils... still using ideal diodes...

[MarkE]

well; I isolated ground behind a diode; so my 'lowrail' can go below 0... the last picture I posted - other than the sliding power/low rails, looks pretty good....


(references with regards to MarkE's schematic )With coils in the right directions, they don't cause ... actually before I finished that, I did check;  Q1 doesn't get a very high low charge; so the mosfet body-diode effect is probably sufficient... Q2 source gets huge spikes of both high and low voltage.  Since the coil on Q1 is the least inductance; if it gets driven it's a stepdown from the voltage driving it... Although not a horrible idea to include high current diodes to shunt negative flows from the mosfets.


I was most interested in the characteristics of the inductor portion of this...


On this other thread http://www.overunity.com/6763/energy-amplification/msg402534/#msg402534
someone posted about bearden's step-charge caps... http://jnaudin.free.fr/html/tbfrenrg.htm


It strikes me that the intent of the high frequency clock is to step charge the cap so when the low frequency hits it has ample charge to dump, and drive 19T coil to get the real power; (to build most momentum)


Both 7T and 12T coil drive 19T coil in the same direction; but moreover, when 12T coil is energized, it's collapsing (forward emf) is a large draw on positive potential points; and the circuit could benefit from a diode similar to D6 on 12T coil being added to the 7T coil to power.


so revised hypothesis on timing:
So low frequency driver should drive near to saturation for a length of pulse...


and the high frequency driver should be ... as short and rapid as possible... although lots of cycles causes lots of power usage....


... although the high frequency probably shouldn't be less than the pulse width of the low frequency driver... do not want these to collide... because they do not have an additive property.



; might replace power sources with TL494 blocks just to have a complete model; already have series capacitance and resistance on coils... still using ideal diodes...

The diode D1 in your simulation kind of scotches things when you make your measurements relative to Node 0.  I would get rid of that.

I went through the exercise of drawing out the circuit so that I could see how the TL494s are being used.  Assuming that the inductor orientations are correct, then huge currents flow through the 19T coil when either MOSFET turns on.  I have updated the schematic with what I think is required.