Use for the Tesla Switch

[Matthew Jones]

Gotta remember the goal of the setup is to get the output voltage up to the level of the batteries or bit higher so you can compare.
To compare you run the load on all 4 batteries from fully charged down to 12.2 or so.
Then run the switch with the same load and manually rotate the batteries. Charge the tops discharge the bottoms, then rotate. You can even let it rest in between runs a bit.

For example the 4 batteries ran the light bulb for 4 hours at an average voltage of 12.4 and 1 amp of current. Thats 49.6 watt hours total.
Now the switch runs the same bulb but does it at 16 hours total and everything is the same or close to above you now 198.4 watt hours.

The real trick is tuning. Depending on the transformer. If its a normal transformer with plates your gonna make heat if the frequency is higher than about 150 hertz. But the toroid you mentioned might be made out of ferrite or something so its gonna start around 50 khz. Well the batteries aren't gonna take that. And going slower lowers the potential.

Try to set the thing up as recommended and you'll see better results. Just gotta play it by ear until you know for sure how things are gonna act. Thats the point of the simple circuit is to learn to tune it.

Cheers
Matt

[Donald Haas]

Thanks Matt,
I fried the recommended transformer and the ssr's.
Not sure what failed first the ssr's or the transformer.
I've redone the transformer with #20 wire instead of #18 and now I'll try the MJL's.
Your MJL circuit looks like the IC and the MJL's are powered by the TS batts and the diode arrangement in the most basic cct is not functioning.
I'll power the MJL's with the TS batts but the arduino with a USB cable to PC.
Very easy to tune on the fly.

Greatly appreciate your presence on the thread.

Thanks,
bro d

[Donald Haas]

I do not know what a CCT is... Let me know when you hit 50 KW we'll compare.

Cheers
Matt

CTT = circuit to my unejukated mind.

[Donald Haas]

quote:
Try to set the thing up as recommended and you'll see better results. Just gotta play it by ear until you know for sure how things are gonna act. Thats the point of the simple circuit is to learn to tune it.


Hi All,
Got the basic setup working now with MJL21194's.
Tuned it to 15.22VAC.
On times are .006sec (167hz?) The scope says on time is about .005sec. Full cycle would be about 40hz
I put the diodes back in between the batts because when I started tuning, tranny heat was too much. The diodes slowed down the heat issue.
When I got close with the tuning the heat issue became very sinkable.
I have a small window motor that might be a good load. It has good torque at low rpm under heavy load. Will run under heavy load at 12W.
Any tips on load sizing?
Thanks,
bro d

[Donald Haas]

2 pics

Hi All,
two pics.
One is the beautiful square wave that Matt's circuit produces with arduino.
Two is pic of circuit while running.
Thanks bro

[ATTACH]14632[/AT[ATTACH]
scope TS sw.JPG

[Matthew Jones]

I would stick with a resistive load under 10 watt. Inductive loads beyond the transformer tend to heat that thing up a bit.

Matt

[Donald Haas]

[QUOTE=Matthew Jones;261444]

Then run the switch with the same load and manually rotate the batteries. Charge the tops discharge the bottoms, then rotate. You can even let it rest in between runs a bit.

Hi Matt,
4 new charged Garden tractor batts.
Took 6.5hrs to draw them down to 12.35v with a 2 amp load.

Currently running load on switch. (Big fan keeps heat down)

When do I want to manually switch the batts?

Thanks,
bro d

[Matthew Jones]

Then run the switch with the same load and manually rotate the batteries. Charge the tops discharge the bottoms, then rotate. You can even let it rest in between runs a bit.

Hi Matt,
4 new charged Garden tractor batts.
Took 6.5hrs to draw them down to 12.35v with a 2 amp load.

Currently running load on switch. (Big fan keeps heat down)

When do I want to manually switch the batts?

Thanks,
bro d

When the bottom set hits 12.35 +-. Put them in a top spot. Generally I keep them on the same side. But you can rotate them around the input power for switching if your not powering externally.

Matt

PS there is another option too. Let the system run till it balances out if it will. So you let the top batteries get into a charge state IE 14 + volt and let the bottom one go low, but not below 11 v or so. I have had them to a point in which the bottoms stop going down and the tops stop going up and everything just runs. Small batteries may not take to it, but you can try one time and see.
I've seen the system hold stable that way up to 12 hours. Inside an climate controlled room they may go longer. Temperature change will often cause this situation the change the stability of the system.

[Donald Haas]

Hi Matt,
Read through most of the thread, pg #120 at present.
My switch has been doing small watt load now for 2 days.
Have everything for a bigger build.
Don't want to be in a hurry.
Very grateful for your work.
Be talkin',
bro d

[Donald Haas]

Been studying Arduino to switch at microseconds.
Have to get into C++.
Did a test with small load and it didn't work.
Direct from batts was longer than TS.
I'm testing again with larger load.
Will scaleup as soon as I can get 8 times more with TS than direct from batts.
Been looking into Benitez patents also.

bro d

[Donald Haas]

Test results

TS - under load, down to 12.215 = 18hr42min

4 charged Batts - under same load down to 12.215 = 16hr35min

I switching at 1ms per on time per side. 2ms per cycle.
This gives highest voltage at AC out and the least amount of heat in the transistors.

Didn't notice any heat in the transformer.

Obviously I'm missing it some where.

I'm using clip leads on the transformer connections, cap bank connections and load connections.

Garden tractor batts. (about 12 Ah )

The load in this test was a 20W 12v halogen bulb.

Previous test was much smaller load and the results were a little better but not close to 8 times more work with the TS compared to batts.
On previous test it took about 3 days for the discharge target for each half.
It seemed like it would have been about the same if there was no load on the TS.
I will follow instructions.
Thanks for any help,
bro d

[Matthew Jones]

Don't be discouraged. Thats not bad on the first few runs. You should use solid wire connections.
More runs will tune the batteries a bit. 1 ms seams kind fast but you have to be the judge. If the transformer can handle it then I guess that may be working. Most iron EI cores like 60 hertz.

You rotated the batteries? Did you try to hit an equilibrium. See if the bottom batteries stop discharging?

Matt

[Donald Haas]

Don't be discouraged. Thats not bad on the first few runs. You should use solid wire connections.
More runs will tune the batteries a bit. 1 ms seams kind fast but you have to be the judge. If the transformer can handle it then I guess that may be working. Most iron EI cores like 60 hertz.

You rotated the batteries? Did you try to hit an equilibrium. See if the bottom batteries stop discharging?

Matt

Thanks Matt,
I'll upgrade the quality of the build and keep testing so I can report on all your points and questions.

Thanks,
bro d

[Donald Haas]

Thanks Matt,
I'll upgrade the quality of the build and keep testing so I can report on all your points and questions.

Thanks,
bro d

Hi Matt,
I redid the setup with no clip leads and #12AWG.
9W load gave me 36 hours with the TS.
I did rest periods between manually shifting batts with TS.
It seemed like I hit an equalibrium where the lower batts seemed to drop a lot less that expected over night.
I expected to seem then in the low 11v range in the am but they were both at 12v.

Same load direct from batts gave me only 15 minutes less run time.
I drew the batts down to 12.06v in both cases.

I used a new arduino board because there was a bad pin on the other.
I would like to scale up to the "big one" with 105Ahr batts but hesitate until I can see better "small one" results.

If I go longer than 1ms pulses the output is less, a few ms more and the tranny's began to heat up.

No heat on the transformer.

I picked up a Tektronix 200Mhz refurbished, warrantied scope on ebay for $335 from a local company. Glad about that.

I could feed the tranny's with a signal gen instead of arduino or could switch to some other micro board.

I should probobly at least do the same 20W load as before with the new wiring. And see how that does.

What stands out to you the most on my efforts as far as need for correction.

Thanks

bro d

[Matthew Jones]

What kinda specs do you have on the transformer. Size and windings.

One thing at first you might be shooting for best output, later you might start looking at recovery. If you run the arduino off a battery in the system you may try taking it off and running it from a power supply or something separate see how much difference that makes. After all its a grounded load and kind defeats the purpose, but I put in that way for simplicity.

Maybe you have some pictures?

Many,Many factors, sometimes the answer is just try things. Remember 60 hertz which is what most transformers are designed to run at is 8.3 ms per side for a total cycle is 16.6. So you are really high speed. Maybe not saturating enough and loosing power in the mix.
You may wanna add just a little off time between half cycles.

Start looking a 1 hour runs and the recovery rate, to see what speeds and changes make a difference.
You do not have to run the thing 36 hours.

If you have another set of batteries try them just to compare.

Several people worked in this thing for many months before it started working well, that was because at that point they new what adjustments made what difference and they could tune right into it. Mine worked the same way, it took a little while to get results.

Don't be scared of a little heat, I heat things up a bit. Heat from current passing through 2 potentials is not the same as heat on a grounded load. 100 watt light bulb only consumes about 4.7 watts between potentials. How great is that?

And trust me thing will not run forever. Might only get 2-3 times out who knows. I have seen as high as 8 times out in the small system several times now. The bigger systems the clock is still out on. Buts its up to you to work it, I wish I could stand there and help but....

Biggest thing is document what you have tried and the results so you can look for patterns to to use to tune.

Cheers
Matt