Joule Ringer!

[conradelektro]

I rewound my ETD49 3C90 core, see the attached picture. The taps give me options when connecting different bulbs.

I also have trouble with the Philips LED bulbs, 5 Watt and 9 Watt, specified for 220V - 240V. With the newly wound transformer they use 6 Watt and 12 Watt at 1:8 step up rate (the 5 Watt also likes 1:11 and 1:14, but the 9 Watt needs 1:8.

My best bulb is a 2.5 Watt LED from China (see attached photo), it uses about 2.2 Watt at 1:6 step up rate and it looks like I can connect some more with little power increase (I will buy a few more).

5 Watt and 7 Watt unmodified CFLs work very well with a little less power than specified at 1:16 step up rate (they are specified for 220V - 240V).

12 Watt and 15 Watt unmodified CFLs work but behave strangely. They have difficulties to start up and then draw about 8 Watt at 1:11, the brightness is not optimal. I guess that the transformer has trouble to transfer more than 10 Watt or it is my crappy power supply. I have to get a good 12 Volt Accumulator.

It looks like one has to carefully select a bulb and then to optimise the transformer for that specific bulb (also when using more than one bulb, one has to proceed in this way).

The circuit is very simple but also very sensitive to

- supply voltage ( I want to use a 12 Volt accumulator, charged by a solar pannel),
- transformer type and number of windings
- and type of light bulb.

Greetings, Conrad

[Lynxsteam]

Conradelektro,

It is important to use a heavier insulated wire like bell wire prescribed by LaserSaber.  The heavier insulation acts as a capacitor and helps form the LC circuit.  I have tried the magnet wire primary and it doesn't work well.  I have good performance using magnet wire for the secondary on the air-core, but with the E-Core even the secondary needs to be the insulated wire. 

There is also the issue of what type of material the E-Core is made of.  I know with torroids there is a big difference in materials and I assume the same is ture for E-Cores.  At high frequencies the core may saturate.  Its the difference between banging on a drum and banging on a pillow.

[NTesla]

Has anyone tried to replicate this circuit using a bi-filar tesla pancake coil? I have thought of making such coils - perhaps both primary and secondary bi-filars but with the primary a thicker wire and appropriate ratio of turns. One could use a stack of ferrite toroids in the middle to couple the coils if required.

[conradelektro]

Conradelektro,

It is important to use a heavier insulated wire like bell wire prescribed by LaserSaber.  The heavier insulation acts as a capacitor and helps form the LC circuit.  I have tried the magnet wire primary and it doesn't work well.  I have good performance using magnet wire for the secondary on the air-core, but with the E-Core even the secondary needs to be the insulated wire. 

There is also the issue of what type of material the E-Core is made of.  I know with torroids there is a big difference in materials and I assume the same is ture for E-Cores.  At high frequencies the core may saturate.  Its the difference between banging on a drum and banging on a pillow.

Material as I understand: high saturation level; Ferrite grade 3C30, 3C34, 3C90, 3C92, 3C96; or very big iron cores. One wants "ferrites for power applications".

This document explains a lot: http://www.ferroxcube.com/prod/assets/sfappl.pdf (storage of energy page 18)

This page http://www.ferroxcube.com/appl/applicmain.htm leads to a lot of info about ferrite materials and products. Also this page helps http://www.ferroxcube.com/prod/prodmain.htm .

The heavy insulated wire poses a problem in case one wants a step up ratio of around 1:20 as required in Europe for bulbs designed for 220 Volt. In the US one gets away with a step up ratio of around 1:10 because the bulbs are built for 110 Volt. To make things complicated, there are bulbs built for 110V - 240V, which work with a 1:10 step up ratio.

I had to go back to enamelled wire because the number of turns required for a 1:20 step up ratio would not fit on my E-Core ETD49/25/16-3C90.

I also have a bigger E-Core FERROXCUBE ETD59/31/22-3C90 FERRITE CORE (Farnell order code 3056430). May be I can wind 15 and 300 turns of bell wire on it. The primary needs some turns to reach a useful saturation range.

Since I realised that most of the bulbs in Europe are made for 220 Volt (in contrast to 110 Volt in the US) it started to make sense and I raised the step up ratio to at least 1:16 (1:20 would be better). Now the 5 to 10 Watt CFLs (unmodified, as bought) work very well with less power draw (about 80% of the specified wattage). The 5 Watt Philips LED and smaller LED-bulbs work with about nominal power draw.

Only power hogs (more than 10 Watt power draw) still behave strangely. This might be, because my core becomes saturated at this Wattage (would need a bigger core and more turns of wire for primary and secondary).

We move into transformer design area. And this has little to do with wire insulation. It is all about core material (saturation level) and power transfer (number of turns).

The photo shows two 5 Watt CFLs driven with about 8,3 Watt. Of course, they might be a little less bright than driven with 220 V AC. It is difficult to judge by eye sight, but I suspect they are slightly brighter with the Joule Ringer.

I play with the CFLs because I have a lot of them at home and I only have a few LED bulbs (because they are still very expensive where I live).

Even the old incandescent light bulbs now show some light (step up ration 1:16) and one sees the saturation of the core at about 12 Watt. The incandescent bulbs draw 25 Watt upwards, but the core can only deliver up to 12 Watt.

My opinion:

The reason why Lasesaber's transformer does not draw more Ampere when more LED bulbs are connected is because it reaches saturation and therefore can not deliver more Wattage. This effect seems to be useful with some LED bulb types sold in the US. This effect seems also to be useful with bare CFL tubes (electronics taken out of the socket) and certain air core coils. The air core has a very low saturation level (also because of the very high frequency of a few 100KHz) and therefore limits power draw to a few Watts.

Once the right Voltage (220 V in Europe, 110 V in the US) and the right amount of power (Amperage, saturation of the core comes into the equation) is delivered to the light bulbs, all starts to be normal and as expected. The higher frequency of the Joule Ringer (a few Kilohertz) seems to induce a slightly higher brightness but might brake the bulbs sooner than specified (because they are designed for 50Hz or 60Hz).

Greetings, Conrad

[Lynxsteam]

I am finding the turns ratio is less important than the material the wire is made from.  My best LJL so far has a turns ratio of 1:5. 

The E-Cores I made ignored turns ratio.  I just wound one layer of heavy insulated primary first and then as much telephone wire as I could.  The ratio was  less than 10:1.

Now if this were just a transformer then the voltage on my LJL 5:1 should be 60 vac.  But it is lighting 110 v LEDs very well.  If you look at scope output for the sawtooth wave the peaks can be quite high and at high frequency.  It may not be important what the rms is.  In fact a lower rms might be a good thing.  If we just light the bulb with our peaks at high frequency the bulb spends more time off and uses less power.  The eye can't see the flicker at high frequency.