Tesla Switch need help

[Groundloop]

[EDIT] Deleted.

[Laserrod]

Hi Groundloop,

Groundloops are a BIG problem but I like your schem of bat FET opros and PIC!

I was inspired to critic your circuit so please take this as help, not DIS.

Power each optos collector through a single res with a cap across collector to to that FETs source from a single PS of 15v plus bat voltage. The resistor will isolate opto collectors when they all use a common above bat volt supply that insures FET gate drive. You want the FET to turn completly on by using 15v gate drive above FET source. Your schem won't work good as is.

It's like each Fet is opto driven from a isolated source. Have each FETs driver floating so it's driver ground stays with its FET source. Hope U get it now. Some FETs turn on at w/ gate volts others more...


Add a more opto driven FETs instead of diodes to prevent .7v x Amp power loss , as I said B4 hook it up backwards!  That's called syncron. rectifi... If the FET is off it's a lossy diode.

Sorry I can't sugarcoat it all for U.  It's a great idea and I think it will be rewarding!

Cheers

DGM

[Groundloop]

[EDIT] Deleted.

[plengo]

Ok guys,

I have some surprising data to explain. I decided now to test my 4 switch this time using a simulator where I would have 4 caps of a certain value arranged as if they were 4 batteries. They are all arranged where the switchs are done manually, simulating off course the 4 swtich that Brandt used but more in a conceptual format.

I performed 4 type of tests where in each test 2 caps would be fully charged to 12v (I used here 22000uf caps giving a total of 264mC for 12v). Let's call each cap by C1, C2, C3 and C4. Then connect two group of caps (C1, C2) to the other two group of caps (C3, C4) changing only the arrangement and letting the exchange of "charges" complete until no current is flowing among them.

Each test is basically this:
test 1)  C1 and C2 fully charged to 12v being equal to 264mC. Connect them in SERIES. Caps C3 and C4 are uncharged (so 0volts) and they are connected in SERIES. Now connect C1 and C2 to C3 and C4  as C1~C2 --> C3~C4 ( ~ means series and -> means connect).
   Result was:
      C1 = 12v and 264mC
      C2 = 8v  and 176mC
      C3 = 4v  and 88mC
      C4 = 4v and 88mC

test 2) C1 and C2 fully charged to 12v being equal to 264mC (Same as before). Connect them in PARALLEL. Caps C3 and C4 are uncharged (so 0volts - same as before) and they are connected in SERIES. Now connect C1 and C2 to C3 and C4  as C1 ! C2 --> C3~C4 ( ! means parallel and -> means connect).
   Result was:
      C1 = 9.6v and 211.2mC
      C2 = 9.6v and 211.2mC
      C3 = 4.8v and 105.6mC
      C4 = 4.8v and 105.6mC

test 3) C1 ~ C2 --> C3 ! C4 (C1 and C2 in SERIES and C3 and C4 in PARALLEL)
   Result was:
      C1 = 8v and 176mC
      C2 = 8v and 176mC
      C3 = 4v and 88mC
      C4 = 4v and 88mC

test 4) C1 ! C2 --> C3 ! C4 (C1 and C2 in PARALLEL and C3 and C4 in PARALLEL)
   Result was:
      C1 = 6v and 132mC
      C2 = 6v and 132mC
      C3 = 6v and 132mC
      C4 = 6v and 132mC

As you can see by the numbers we have the following conclusions:

SERIES going to SERIES
When having C1 ~ C2 going to C3 ~ C4 result is an increase in 33% final sum of voltage and mC. In other words, I started with a sum of 24v of both C1 and C2 and a sum of 528mC from both C1 and C2. C3 and C4 having 0v and 0mC. Final results were an increased total overall C1, C2, C3 and C4 to 28v and 616mC being then a 133% total or 33% increase.

PARALLEL going to SERIES
When having C1 ! C2 going to C3 ~ C4 result is an increase in 20% final sum of voltage and mC. In other words, I started with a sum of 24v of both C1 and C2 and a sum of 528mC from both C1 and C2. C3 and C4 having 0v and 0mC. Final results were an increased total overall C1, C2, C3 and C4 to 28.8v and 633.6mC being then a 120% total or 20% increase.

SERIES going to PARALLEL
When having C1 ~ C2 going to C3 ! C4 result is an net of 0% increase or loss on the final sum of voltage and mC. In other words, I started with a sum of 24v of both C1 and C2 and a sum of 528mC from both C1 and C2. C3 and C4 having 0v and 0mC. Final results were and net equal to total overall C1, C2, C3 and C4 to 28v and 528mC being then net of 0% increase/loss.

PARALLEL going to PARALLEL
When having C1 ! C2 going to C3 ! C4 result is an net of 0% increase or loss on the final sum of voltage and mC. In other words, I started with a sum of 24v of both C1 and C2 and a sum of 528mC from both C1 and C2. C3 and C4 having 0v and 0mC. Final results were and net equal to total overall C1, C2, C3 and C4 to 28v and 528mC being then net of 0% increase/loss.

Repeating the experiment using the simulator gave me the same results every single time. As one can observe the simulator correctly stated that when transfering the "charge" from two cap in parallel to another two caps in parallel it only distributed the "charges" equally among them. So I assume that the same is true for the other tests above.

I will tonight do a real lab experiment for these scenarios so assert those number above are correct. I must say already that some of my tests using the 3 swtich setup with the real batteries kind of reflected some of the tests above more specifically the test number 1 (Series to Series) where on of my batteries would not fluctuate in voltage as you guys can see in some of my previous posts.

I am totally astonished by theses results because in 2 tests the net result of voltage and "charge" was higher than the initial setup. I can barely wait to test this in real lab and to replicate this in higher frequency than I can do by one instant. This kind of demonstrates the claims done by who replicated this 4 swtich device succesfully. If done properly it is possible , at least mathematically, to use the current of the exchange of the "charges" for a load and still keep the original voltages and total net "charges".

I have available to who wants the liveWire file that I setup for this experiment. It is possible to also use SPICE.

Fausto.

ps:[Edit] I just updated the Caps orders on the tests. So it should always be C1 and C2 GOING to C3 and C4.

[Groundloop]

[EDIT] Deleted.