3.7 VOLT BATTERY POWERS 56 WATTS

[gyulasun]

   USB input to setup  5.03 volts at 1.15 amps. (5.7845 watts)  Output from setup capacitor to bulbs 
5.93 volts at 1.03 amps  (6.1079 watts)  I checked this many times and came up to the same results.
A plus .3234 watts unknown ?? power.  All bulbs lit brightly during the test.

Hi magnetman,

Many thanks for doing those measurements.

I would like to ask: when you said the "output from the setup capacitor to the bulbs", did you mean the 5.93V
across capacitor C1 or across the positive OUT and the negative IN and OUT points? In your schematic these
two latter points are indicated as OUT+  and  IN AND OUT-.   But I do not know where these two points are
in the snapshot photo I attached in my previous mail, I put two ? onto two wires I guess they are. 

For me the "setup capacitor" designation would mean your C1 capacitor and that the 5.93V was measured
across it, is this correct? If yes, then please measure the DC voltage output across the OUT+ and
the IN AND OUT- points too.
And if you meant the latter two points having the 5.93V across, then please measure the voltage across C1
at the terminal strip where I included a voltmeter symbol in the snapshot photo attached to my above post.

In this schema below I included capacitor C2 I would like to suggest you apply for test purposes, this capacitor
would help filter the DC output coming from C1 via the thyristor. The value of this filter electrolytic cap,
C2 could be any uF like say 470uF and higher, voltage rating could be say 40V DC (you are working at
5V input now). And between the OUT+ and IN AND OUT- points the peak voltage level (due to the thyristor
switching) would not be higher than say 15-20V peak value, (this could nicely be seen on a scope).

Sorry for my quest for such details on your setup, my aim is to help explore such circuits exhibiting unusual
behaviour i.e in your case giving "a plus .3234 W unknown?? power". Thanks again for your cooperation.

Greetings
Gyula

[gyulasun]

Hi all, this is not exactly like magnetmans setup, though has practical use value.
It is essentially a joule thief oscillator type, using a ferrite TV flyback c-core.
It is using 4 watts of input power from a 12 volt tractor battery.
Here is short video showing lighting level. 5 led bulbs are lighted, 2 are together towards the back.
https://youtu.be/C3vl1w62aaU
peace love light

Hi Skywatcher,

Indeed very good lighting level in comparison to 4W input, thanks for the video too.  I guess this good impression
might come also from the good distributive nature of positioning the 5 LEDs to illuminate the most space in the room.

Because you use 120V LEDs, there is perhaps not much sense using a thyristor or any similar device to switch
energy out from your 150 uF puffer capacitor because the needed 120V is comparable to the amplitude of the
created flyback pulses that average out to that value when see that kind of load. 
To use a switch to feed a load from the puffer cap I think the puffer cap should be kept at 2-3 times 120V at least
to get usable power level for the LEDs. 
I would appreciate if you would be so kind and check the DC voltage across the 150uF cap and the total current
going into the paralleled 5 LEDs.  This data would surely help seeing more clearly the light... 

Thanks,
Gyula

[magnetman12003]

Hi Skywatcher,

Indeed very good lighting level in comparison to 4W input, thanks for the video too.  I guess this good impression
might come also from the good distributive nature of positioning the 5 LEDs to illuminate the most space in the room.

Because you use 120V LEDs, there is perhaps not much sense using a thyristor or any similar device to switch
energy out from your 150 uF puffer capacitor because the needed 120V is comparable to the amplitude of the
created flyback pulses that average out to that value when see that kind of load. 
To use a switch to feed a load from the puffer cap I think the puffer cap should be kept at 2-3 times 120V at least
to get usable power level for the LEDs. 
I would appreciate if you would be so kind and check the DC voltage across the 150uF cap and the total current
going into the paralleled 5 LEDs.  This data would surely help seeing more clearly the light... 

Thanks,
Gyula

I took out one of my eight bulbs and measured 3.54 volts inside the lamp socket?  All lamps were lit during this test.  My room looks like  bright  daylight at 12 o'clock midnight with all lamps lit.

[SkyWatcher123]

Hi all, Hi gyulasun, i will do that later today.
Yes, that was the idea, i feel it is best to utilize the natural higher voltage from the flyback spikes, though i'm sure there are efficient methods to transform it down to lower voltages.
The particular ecosmart led bulbs i have, are able to light at lower voltages, like 40 volts at the capacitor.
Whereas, some of the newer bulbs i have need higher voltage, like 85 volts or higher.
I'm picking up some 40 watt equivalent philips led bulbs today to compare, non-dimmable.
I did check the voltage at the puffer capacitor yesterday, using the 5 older model, ecosmart bulbs, it was around 43 volts.
I will check the amperage later today.
peace love light

[gyulasun]

   I took out one of my eight bulbs and measured 3.54 volts inside the lamp socket?   All lamps were lit during
this test.  My room looks like  bright  daylight at 12 o'clock midnight with all lamps lit.

Hi magnetman,

That was a good idea to access directly to the paralleled lamps if that place seemed the easiest. 
I think the 3.54V DC (meter was in DC range, right?) you measured with DMM means that the switching
frequency of the thyristor may fool the meter.  This may also mean that if you set the meter to AC range,
the shown voltage would still be a false value, not only of the frequency but the possible sawtooth waveforms.

When you connect the capacitor C2 as I indicated, perhaps it would be better to start with say 4.7uF and test
higher and higher values to find an optimum.

To tell the truth, some filtering at the 5V input side may also be needed, I am not familiar with that USB device
you happen to use, nevertheless it is expected to give a relatively clean 5V DC output which then feeds your setup.

Thanks,
Gyula