Use for the Tesla Switch

[garrypm]

Hi all,

I've managed to remove the laminations and removed the 2 coils.

As you can see from the pic, there is a seperator in the centre of the spool.

Guessing that this will need to be renmoved before winding the trifilar coil.

Picked up a little trick to make it easier to reassemble the laminations. Instead
of peeling out every individual piece, leave the cross piece adhered to its
neighbour "E" piece.

Thanks, Garry

[Matthew Jones]

Got the Sziklai pair switch up and running. Just followed the directions and it ran the first time I flipped the switch, Having spent all that time fighting to get the relay version to work did me lots of good. I have a much better understanding of how this stuff goes together now
Picture of my relay switch build is at post #3783 and here is a pic of the Szziklai pair build.
I also video'd it on my YouTube channel. 11Turion

Getting to be a pro.

I would like to see the load running and some results though.

But I happy for ya either way.

Matt

[Turion]

Matt,
Just got my headlight, so will get my cap completely charged and then get this tuned this afternoon and spend tomorrow running and post results. Will also make a YouTube video of it running, although the only way you KNOW it's running is by putting a volt meter on it. I can't believe this thing is so dang quiet! My laptop is louder than it is.

Also, I decided to build your manual version just because it interests me. The more I know, the better I like it. Solid state is probably the best way to go, but I am still interested in what you can do the old fashioned way. I am not afraid I might be in danger of actually LEARNING something.

David

[Sobakin]

Hi Matt, hi all,
my results for 0.2A cooling fan:
Matt's 2-switch mosfet TS: 14 hours (7 hours, battery swap, 7 hours)
My old 6-switch mosfet TS: 14.5 hours (no need for swap batteries, as they discharged simultaneously).
6-switch TS was running thru same transformer and same frequency for clean comparison.
I tried for different loads, pulse frequencies and duty cycles, and still no good result - it works, all parts are cold (except bridge rectifier under 21W load and bulbs), so almost no power dissipations here, but no any advantage in discharge time
Tried 55W 12V halogen bulb - it discharged batteries very fast and such very heavy load can fast kill them, remind, I use small 4x 1.3 A-h 12V SLA batteries.
Yesterday I built 4-switch symmetric TS based on Matt's design, it's no needing for battery swap (but won't work w/o transformer, unlike 6-switch), it works good, but results the same.
Maybe I misunderstood something because of my bad English, and this TS scheme is really not for longer battery run but for training purposes, and real Matt's TS is something another?

[John_K]

Sobakin,

In my experience SLA batteries are not good. You only got 2.8Ah out of 5.6Ah worth of batteries.

SLA batteries don't like to be charged with double the battery voltage. They prefer a couple of volts over the battery voltage with some current.

I think you will see better results with flooded lead acid batteries.

The batteries are 90% of the system IMO.


John K.

[Sobakin]

Sobakin,
In my experience SLA batteries are not good. You only got 2.8Ah out of 5.6Ah worth of batteries.
SLA batteries don't like to be charged with double the battery voltage. They prefer a couple of volts over the battery voltage with some current.

There is no double voltage charge in TS if only load is not short-circued, even if so in Matt's TS its impossible because of transformer in output, so even in high load (55W bulb) voltage drop across batteries which is in series not more than 1.5-2V at each battery and voltage rise in charged-side battery this time is max +2 V.

[Turion]

Matt listed a .5-1 amp load at 212-14 volts.

If you're having trouble, here is the closest I could find. Called the Sylvania service center, gave them what I was looking for, and this is what they recommended and what I purchased today
SYLVANIA 211-2 LAMP BULB - .97A 12.8V T3 DOUBLE ENDED

It looks like a fuse and has a bar at each end just made for alligator clips.

Dave

[Bit's-n-Bytes]

Here is my build (so far)

Hi Gang, take a look;

MOV01716.MPG - YouTube

Not hooked up to the transformer yet, but wanted to show a safe way to tune and check out the connectivity.


Thanks

Bit's

[John_K]

There is no double voltage charge in TS if only load is not short-circued, even if so in Matt's TS its impossible because of transformer in output, so even in high load (55W bulb) voltage drop across batteries which is in series not more than 1.5-2V at each battery and voltage rise in charged-side battery this time is max +2 V.

Sobakin,

OK, maybe not double the voltage because you have losses and voltage drops. But with a 1 ohm load as Matt is suggesting (the load is the transformer winding) you should still be seeing 9-10V over the top of the battery. Take the charge battery out of the circuit and test the voltage if you want to.

Anyway, my point is that gel-cells are crap and in my experience are not a good battery choice for this application.


John K.

[Matthew Jones]

Sobakin

Johns probably right about. I guess it didn't dawn on me. I don't use them and never have so I do not know first hand.

One thing you could try is measure the volt from 24 bank through the transformer. I am sure it will be high even after the switch's, maybe 22 volt or something.
You could then look at maybe wind a transformer with a higher resistance and see if that helps out.

One thing though I have seen from ya(I probably missed it), is your baseline? Have you done baseline test so you know how long it takes to discharge under normal conditions?

Matt

[LutherG]

Hi Gang, take a look;

MOV01716.MPG - YouTube

Not hooked up to the transformer yet, but wanted to show a safe way to tune and check out the connectivity.


Thanks

Bit's

Thanks Bits - love the visual feedback. Excellent work as always... thanks for sharing.

I'm trying to decide on a stamp to get but in the interim I might try doing what a friend recommended - using my function generator to do a flip-flop by wiring in one of the H11D1's in reverse... If it works out I'll post a video of it.

regards,

Luther

[Sobakin]

Sobakin,

OK, maybe not double the voltage because you have losses and voltage drops.

My setup has almost no losses and voltage drops because I use mosfets with only 0.04 Ohm resistance in switched-on mode.

But with a 1 ohm load as Matt is suggesting (the load is the transformer winding) you should still be seeing 9-10V over the top of the battery. Take the charge battery out of the circuit and test the voltage if you want to.

I mean that there is no double voltage charge because between 12V battery and 24V bank no short circuit but load with resistance much bigger than internal battery resistance even if I load is 55W bulb.
Transformer winding is about 1 Ohm is active resistance then You put DC to it, but usually we are using transformers for AC, and if You properly calculate wiring, using such transformer at no lower than designed frequencies, is would have reactive resistance which is much higher than 1 Ohm.
With no load (idle) good transformer consumes very low current.
In my case transformer wires is 0.4 Ohms each primary and 0.3 Ohms secondary - it uses a bit thicker wire.

Anyway, my point is that gel-cells are crap and in my experience are not a good battery choice for this application.

About this I can't say anything, maybe gel lead-acid batteries is worse than fluid, need to check this.

[Sobakin]

Sobakin

Johns probably right about. I guess it didn't dawn on me. I don't use them and never have so I do not know first hand.

Did You never use gel SLA batteries? But 4-7 A-h batteries, which mostly using by us (in UPS etc) and which You are recommending for small TS in your PDF are all gel SLA, my 1.3 A-h too, they only smaller.

One thing you could try is measure the volt from 24 bank through the transformer. I am sure it will be high even after the switch's, maybe 22 volt or something.

I didn't understand - how should I measure 24V bank THROUGH TRANSFORMER? I measured it when running by osilloscope thru "-" of lower and "+" of higher battery, and I saw 24-25V on charged batteries.
As I said, my setup has almost no losses because of MOSFET switching, I didn't use BJT transistors or any SSR in switches because of high losses.

You could then look at maybe wind a transformer with a higher resistance and see if that helps out.

I think that higher resistance in transformer would only raise losses in setup, and this is no good.
My transformer has low active resistance, and this is good, high resistance wires would only heat up transformer and You couldn't put big load.
Thats doesn't mean that lower winding turns is better - no, turns count need to be calculated (there are formulas for it) depending on transformer core size and material, desired frequency range and maximum output power, and almost always thicker wire is better, it only must fit into transformer. I think that way, do You disagree?

One thing though I have seen from ya(I probably missed it), is your baseline? Have you done baseline test so you know how long it takes to discharge under normal conditions?

I've done this test halfyear ago then tested my 6-switch TS, and it was no difference.
But now after your ask, and allowing that batteries now may be in worse condition, now I repeating this test - 4 batteries in parallel discharge to 10W bulb w/o any switches and transformers. Now waiting for discharge.

[Turion]

Carroll,
Did you have to take some windings off the AC side of your center tap (Radio Shack) transformer to get it to work? I tried hooking one up to my (working) setup and couldn't compensate for wire length with just the timing alone. My stamp "exceeded" the 16 bit chip when I got up too high in my pulsout command. And that barely got me a glow on the lightbulb. (Starting from "0" voltage in the caps) Could have gone with milliseconds instead, but that has its own issues.

Just curious. I've taken it off my board and gone back to one I wound myself for tuning, testing and publishing some results, but I've almost got my third setup built with just the MJL's, and thought I would try and use it for that one. I have one other one I wound myself, but it's hooked up to my relay setup, and I FINALLY got it working, so don't want to TOUCH it again except to run it.

Here's a YouTube video of my system running under load. Charging my batteries to full, and then will do my test run and post data. Finally.
MJTS - YouTube


David

[citfta]

Hi David,

No I did not take any windings off the modified Radio Shack transformer. I had a small 6 watt 120 volt bulb as my load. I also had a 15 watt 120 volt bulb but it wouldn't light to full brightness unless my batteries were really up in voltage. I also found that I had to go down to 50 herts or so to get the best output. Actually between 47 and 50 hertz was the best for the highest output. I would have thought that 60 hertz would have been better but it wasn't. I did have to use milliseconds for my pulses. The code is not too bad. It can go something like this:


Main:
high 1
pause 50
low 1
high 2
pause 50
low 2
goto main.

Then you adjust the value where I have 50 to get to what you want. At this low of a frequency I don't think you really have to worry about the slight delay caused by the program jumping back to the top. If you feel that is a problem just look at some earlier posts where we discussed how to fix that.

That is interesting that you had better results from your rewound transformer than from the modified one. I have been wondering if Matt's design wouldn't be better but I haven't really had time to try and rewind one. I am glad to see you trying different things to see which work better. That way we all can learn.

By the way congratulations on getting some TS's up and running. I am working on my second one now with mosfets instead of bipolar transistors. I am also remodeling a house so I only get to work on the TS when I am resting from the house work. Got to keep the wife happy too you know.

Later, Carroll