1.5mm2 speaker wire. 8 times doubled up. 19cm diameter means, 5m per plate, 10m total.

I left 22cm for hookup (cut it off). and doubled up the rest till 8 layers.
cut the ends and soldered together.
 next time, use copper pipe!

I twisted the 8 pairs slightly, to get a total length of 59cm.

I think it is best for the windings, to split the bifilar speaker wires in single strands. this means the coils will be unifilar, with less charged capcity, which will result in higher voltages. and frequencies

First strip both ends of the 8 pairs. then tape one end together end solder it.
then clamp the soldered end, and slightly twist the 8 wires, and tape it down each 5 cm or so. then clamp down the other end, and solder it.

I have wound the bifilar speaker wire over the core, and counted 77 windings x2 makes 154 total, but will need to recount after splitting and rewinding.
the total length of the bifilar was 432.6 cm so I will need to double that.
making 865.2cm for the first layer.

The wire will be split into two, for each core one layer. This way they are equal.
I will cut some extra length for hookup. making the total length 885cm (+19.8 so each end has 9.9 extra).

the core measures a diameter of around 12.5mm average

I wound the single first layer, with the 885cm wire, started with 10cm extra. the other end was 90cm extra, measured from the 10cm extra.
So The total wire needed was 885-90cm=795cm for the first winding.

I will keep the extra 90 attached, and will roll the second core later, and see if it the same.
Lots of windings to count also.

the second layer has a diameter of around 19.41mm

Next time I need to make the core more precise. now the capacity with the First coil layer is very chaotic. with high voltages, this will create stress points on the coil insulation.

A copper pipe, or if wire is used, make all the bifilar wires into single ones, thus making 16 wires. those would still be hard to bundle so... why not use, the 4mm2 wires I have as core material? hmm. Yes. and only use 4 wires. no more. so they stay inplace.

but for now I will continue with the cores as they are

77-10 is 67 excess wire from the second core first coil layer. thats 23cm extra on the coil.
lets compare windings and see If I can remove some to make them equal.

turns on core1: 147
turns on core2 was 150, so reduced to 147, giving extra wire, so now the difference is only 6cm between B1 and B2 Fine!
Next layer, needs also 147 turns. but now, diameter has increased from 12.5 to 19.4 mm
thats factor 1,55 so I guess, I need 1.55x 800 (little extra it was 795)=12.416 meter

I used 626cm bifilar for the second layer including equal wire extensions. So 2x=
1252cm. A very close match. I wrapped it loose so I will stick to the 1250cm for the 2 second layers.

measure cut and split for winding. this thing is getting massive

I wound the second layer starting with 20cm extra for hookip
I then ended with a 227cm extra on the end of C1, taking account of 90cm, is an extra of 137cm
so instead of 12.5m, I could have used 11,13 meter. Oh well. good to know for the next time. it probably because the single wire is placed between 2 windings of the first layer, giving less diameter.

because the wires are on top of each other, I gave it 2 less windings then the first layer, so only 145 turns.
Else the second coil would be placed on the core, which isn't good I think.

I used the red wire for this, so it will become blank red blank for one core, and red blank red for the other core

winding the second layers was much easier, enough done for today.

I do thing the diameter, vs windings is way out of control.
I would prefer a much larger core, maybe with copper tape on a sewer pipe. Whith a cooling fan tool cool it. 10cm diameter or something like that. several layers of copper tape could be stacked, for high currents.
or use thicker foil/tape.

that would also save a lot...

cant wait to figure out the switching

nope, will have to redo the hole core, its a cap...
might wind it on a 3.5cm X 35CM white pvc pipe, 19 1.5mm2 bifilars side by side

from chat gpt:

Quote

The lead sulfate crystals that cause battery failure primarily grow on **both the positive and negative plates** of a lead-acid battery during discharge.

While the chemical reactions at each pole are slightly different, the formation of lead sulfate is a consequence of the overall discharge process in both.  So, both poles are susceptible to sulfation, but the negative pole might be slightly more prone to it due to the specific reactions involved.

So instead if impulsing the positive it might be more useful to impulse the negative pole of a lead acid battery to restore it.

Or both.
I looked at the Re-charger and it looks like a fun project.

But I want to invert the project and use a pnp transistor to make negative impulses on the negative pole of the battery.

also wondering if I could use a single toroid core for 2 joule thiefs. one positive (npn) and one negative (pnp)
if the circuits are properly tuned, they should work simultaneously, and produce a positive and negative impulse simultaneously, that might kick the battery back into action.

2 circuits one for the positive, one for the negative side of the 12V lead acid battery.
These can be combined into a single circuit,
whereby the R1 and R2 need to be tuned, so both fire simultaneously
and produce a simultaneous positive and negative impulse on the battery poles.
Kicking the Lead sulfate crystals to pieces.
might work, I see nothing wrong with it.

The single core, is so that the magnetic field that collapses produces a simultaneous positive and negative impulse

EDIT:
oops, no this would set the 12V under a constant 20V... Dont do that!
and I should have switched the NPN and PNP

This could work.
The 20V P1 and 20V P2 are parallel to the 12V battery
so the battery sees 8V from both in parallel, so no problems there.
I swapped the NPN and PNP to make it work

So a dual isolated channel (no earth) power supply is needed. I got one so no problem there.

EDIT
this needs 2 more diodes.
placed at the toroids center taps.
preventing the impulses to enter the coils.

the 12V, thus has each pole with 3 diodes. All oriented the same towards the pole so only positive or negative can enter the pole

I bought a TIP31C NPN and a TIP32C PNP
they can handle 100V 3A 40W with max 5Vbe
3Mc/s  they are a identical pair.
also got 2x 1k ohm trimpot.

Going to use a single ferrite iron core, for both windings, made of two equal length wires.
so they will behave identical.

diodes are mur460 which have a 1.25V drop.
so 2.5V total.Plus the transistor which I guess has a 1.5V drop so in total I loose 4V.

13.5V for the pulsed dc charge plus 4 volt is 17.5V for each power supply could work.

going to give the 1k trip pot a series resistor to limit the max voltage to 5V

I am winding the toroid with 2x1m 0.5mm wire
1m for each half circuit
While turning I realised, that the coils need to be counter wound, else the induced magnetic field in the toroid would be opposite, due to the opposing voltages.

So... I will move one of the two center taps, to the other end of the windings.

so the wires are all wound in the same direction, but opposite attached, for the opposing voltages

I'm first going to make the two separated joule thief circuits (PNP and NPN),
 without all the diodes and the load battery,
to see how they perform, and if I can tune them to work together on the single toroid core

OK the PNP version works.
 this is with only 0.7V on the psu
1k on the base, so I will need more resistance to keep the 5V limit (10V/div)
orange is the negative impulse on the  collector
blue is the signal at the base.
frequency is 9.6 kc/s so I can hear it. might need to reduce the windings.

The NPN also works. Obviously with the intended polarities swithced.
But this time I needed to reduce the base resistance to get to the same freqeuncy.
this time 9.5 kc/s
The blue base voltage, now is limited.
I suspect this is because for both the tests I earth grounded the negative terminal.
but it works, so I can continue.

I will add 2k in series to the 1k trimpots, that should bring the frequency up out of the audible range

2k2 ohm in series, with the same setting on the npn, with 0.07A and 2.1V psu
gives 25.4kc/s

2k2 in series with the base of the PNP isnt enough to bring the freqeuncy high enough.
I  guess the windings and the transistors have enough difference to make this work differently.
F is now 21 kc/s with the 1k pot maxed out.
So I need to see if I can get the npn to the same frequency with the 1k pot. else I need to make the series resistor smaller at the npn base

I grounded the positive this time. now signals look more ore less the same (I left the psu alone on the same setting)

I replaced the 2k2 of the NPN base, with 1k5 ohm, and now I can tune it to the same frequency of 21.0kc/s as the PNP had.
again with the same voltage on the PSU, but now, due to the smaller base resistance, the current is up to 0.09A. but that is all fine.

this was again, done with the negative on earth ground.

So the 2 joule thiefs work, and are tunable to the same frequency.
now the question is, Can they work together on the same core.

How do I test this? well, I can connected them up to my two channel psu, remove earth ground, and see if it works, since I already tuned them to the same frequency.

I did this and it works. I now can use the voltage of both channels to influence the frequency.
Orange and blue now show the collectors of the PNP and NPN transistors. So it works!
how cool is that :cool:

Now I can continue to add the diodes, so I can add the 12V lead acid battery.

https://www.st.com/resource/en/datasheet/tip32c.pdf PNP
https://www.st.com/resource/en/datasheet/tip31c.pdf NPN

very nice  work  to like  the cores

I combined the two power supplies into a single one,
and put MUR460 diodes (can also be another fast diode) on the commons of the PNP and NPN transistors.
I can easily get 70V positive and negative voltage impulses, at 21 kc/s
from 3.3V with 0.22A= 0.73W (unloaded)
the trimpots at the base, dont need precice tuning. it just works, so I set them to max resistance,
and controlled voltage and freqeuncy of the spike with the PSU.

Next, is connecting a 12V battery, and see how the spikes will behave.
I expect a need for much higher power, we'll see

probed after the diodes

So the battery will get simultanious positive and negative on its terminals. shocking the inside lead sulfate with both polarities on both poles.

I completely left out the pulsed DC, it can be added later again if needed

I connected a lead acid 12V battery with 12.5V (charged) and looked at the base of the PNP (blue) and NPN(orange)
The voltages there are clipped at 7,5V which is 2.5V above their max 5V rating. I dont like that.
This was at a low PSU (ungrounded) setting. of less then 2V
more power= more clipping= more risk of damaging the base to emitter
Maybe a zener could help controll those voltages?

Now I measured at the battery poles and gave it more power.
0.52A 7.0V PSU =3.5W
F=35kc/s
as expected, the impulses now are clipped by the low impedance battery load.

polarities go up and down simultaneously.

I can bring the power up higher, but no doubt the signal keeps getting clipped, while the pulse widens.
earlier I had the NPN PNP producing more heat then the PNP. NPN
 lets check again.

0.75A 7.3V psu =5.5W
frequency goes up to 67 kc/s (is that right? look at the cursor. looks more like 40kc)

Earlier not the npn but the pnp became hot.
So I reduced the resistance at the base of the npn.
now both transistors get hot (no heatsink yet).
since each can handle 40W (with heatsink)  is probable from the base clipping.
I could also reduce the base windings of the joule thief.

Signal on the battery as expected has again the clipped off impulse, this time with a relatively wider pulse width

I don't expect benefits from this... unless I crank the power up higher. but... first I need to take care of the heat produced by the base.

a capacitor could temper those voltages also. so... a zener between base and emitter with capacitor and resistor for discharging the excessive energy, while also reducing the base windings?
I can increase the series base resistor, but that will only take care of the current, not the voltage.
I could also insert a choke to limit the voltage. but that is again frequency depended