Joule Lamp

[JouleSeeker]

  I'd appreciate some wisdom on what LED bulbs to get.
Below we see a 6W LED bulb for only $3.99 + shipping, a bargain.  But -- only 320 Lumens, that is, 320/6W = 53 Lm/W.  Not so great.

OTOH, I find other LED bulbs at 3W that claim 300 Lm, nearly as much as the one above, but at 3W input instead of 6W.  And that is 300Lm/3W = 100 Lm/W!  It also accepts input voltage from 85 to 260 VAC.  Costs more, around $12.

There is a whole range of LED bulbs available, 12V-DC, 120 V, 220 V, 85-260V, and various configurations and Lm/W also.

EDIT:  Add:  Is DIMMABLE preferable?  these cost more...

Any suggestions on which LED bulbs might  be best for these Lynx-lamp experiments?  What is the Lumens/W of your bulbs Lynx, others?

  I'm hoping to MAXIMIZE Lumens/Watt with the lynx-lamp... not necessarily with the bulb as purchased and run on the grid.

[Lynxsteam]

This Utiltech pro LED warm 450 lumens, is the LED I bought at Lowes for $9 each.  Its not wasted money because they will last a long time.  This is a nice bulb and it looks very much like the one Laser Saber uses in his video.  Lumens is stated as 450 at 7.5 watts.  I think if you can find this one we can all compare.

[Lynxsteam]

I tested the UtiliTech Pro 450 on AC from the wall plug.  This bulb is spec'd at 7.5 watts.  But it pulls 18.5 watts and is very bright.  On my LJL circuit it consumes 4.7 watts. 

[Lynxsteam]

Here is some of what I am learning that might help fellow explorers.

Our secondary is running at very high frequencies.  At this high frequency the transistor is fully on and off so fast that the light is flickering at several thousand times a second.  The eye is fooled into perceiving this as on.  With normal household AC the light is on and off 60 times a second and the eye still sees it as on.  But at 60 hz, power is on fully for longer than with high frequency spikes.  This may be the reason LaserSaber can demonstrate higher lumens on less power than with 60 hz supply, not to mention the losses from the inverter circuitry, voltage and frequency regulators.

The next thing is at these high frequencies even with an aircore coil we can experience core losses similar to a ferrite transformer, because of the "skin effect".  Only the outside part of the secondary wire carries the current.  So in my experiment yesterday with a heavier gauge secondary I saw a higher power draw and slightly less performance.  Smaller wire has greater surface area.

But if you go too small you lose the amount of material needed to provide the necessary inductance.

The primary coil is running at this high frequency too.  Maximizing surface area is important here too in order to reduce the skin effect problem.  I am running 4 insulated primary wires to maximize induction and reduce skin effect.  The use of litz wires or twisted insulated wires could be of benefit.  4-6 twisted insulated wires for the primary may be much better than a single primary wire.  Using copper tubing would greatly increase surface area because of the inner and outer skin of a tube.

Proximity of wire next to wire causes capacitance problems which can't be avoided in the secondary at high frequency, but should be avoided in the primary by spacing the primary turns apart.

Because the aircore coil's inductance can be several thousand times less than a ferrite core transformer the primary to secondary ratio needs to be higher.  I find that a ratio of 10:1 isn't enough for LEDs off 12 volts.  If we use a ratio of 15:1 we should see 180 rms acv without a load.  Once the load is applied voltage will drop across the load to about 60 vac.  Right now with a 10:1 ratio volts across the load (1 bulb) are dropping to 39 vac and the DC component is 3.9 as opposed to the ferrite transformer's DC component at 4.9 v.  Proof that a ferrite transformer has much higher induction.

So what's all this mean?  I am going to try going back to 368 turns 20 awg on the secondary.  25 turns on the primary of twisted insulated wire spaced evenly apart.  The goal is to get closer to the performance of the SJR2.0.  Why bother?  Because if the problems can be addressed, air core coils have an advantage at higher frequencies and power draw could be lower than a ferrite transformer.  And for fun! 

[b_rads]

In researching the "proximity effect" yesterday while looking at stranded vs. solid core wire, I found reference to a way to overcome some of this effect.  Parallel wires cause the effect however crossing of the wires eliminates some of this effect.  It might be interesting to try winding the first half of the primary the full length and then bring the second half back over and wind to the beginning.  This might increase the space between the parallel wires in addition to creating intersections where the first half and second half cross over each other, if that makes sense.   

Brad S