240 Volts 100 Hertz 29 LED lamp

[RODALCO]

This is the improved version of the LED lamp posted about 1 year ago which ran on half wave rectified dc directly from the 240 Volts mains.

As I test my projects thoroughly and want to ensure a good reliability of the project a long time 2000 hours duration run is done prior to posting the details on the web.
Again, I have choosen for the resistive option because it is more reliable than a capacitor voltage dropper.

Parts required:

29 x White LED's
6 x 2k2 1Watt resistors
4 x 1N4007 diodes
small piece of vero board
discarded CFL lamp base

Photo's attached with assembly details.

Calculations in part 2 of this thread.

 

[Hero999]

What about using a transformer?

I have some small transformers that would probably fit in a small space like that.

From memory they're 1VA, 24V but the open circuit voltage is higher. Adding a voltage quadrupler should give a high enough voltage to power 29 white LEDs with just the capacitors limiting the current.

 

[RODALCO]

The amount of LED's and placement is up to the user.
I have built 8 of these lamps and all are working satisfactory over the last ½ year with no faillures or dimming.

4983. Old CFL, be carefull removing the lamp tube (just in case the glass breaks). I wrap a rag around it and try to wriggle it off the base, and often the tube and cap pop off in one piece.

4984. Base, drilled out with 4.8 mm drillbit, I use a small file to ream the hole out a little so that the led fits in tightly.
I made up a simple template to divide the circle into 12 equal sections.

4985. Top circle from PVC, also drilled out with 4.8 mm drill bit, for assembly it gets hotglued to the base.

4986. Lamp on, bottom view.

4987. Lamp on, side view.

4989. Current on mA meter at 240 Volts ( 11mA ).

4990. Resistor stack, 6 x 2200 ohms 1 Watt resistors.

4992. Bridge rectifier consisting of 4 x 1N4007 diodes.

White LED about 3.0 Volts drop.
29 LED's is 87 Volts
To dissipate 240 - 87 = 153 Volts in series resisitor.
153 divided by 13200 ohms = 11.6 mA's (LED current)

Power dissipated I²R
therefore
0.0116² x 13200 = 1.78 Watts.

With a 6 Watt total rating the resistors are well within their design rating to ensure a long reliable life of the lamp and minimal heat dissipation.

This lamp can easily be made to run on 120 Volts US mains by recalculating the series resistors.
120 -87 = 33 Volts
at 11 mA
therefore:
a 3000 ohm series resistor is required.
I would opt for 4 x 680 or 820 ohms 1 Watt in series.
power dissipation would be less than 0.5 Watts in case of the 120 Volts version.

IMPORTANT !!

This is a mains voltage lamp. Work safe and don't have exposed bare wires sticking out of the lamp.

 

[RODALCO]

You are fast in your reply, Hero999 ?

Transformer is fine, I haven't got a very small TX laying around which fits.

It adds extra weight but could be used.

The full wave bridge reduces LED flicker as the LED's run at 100 or 120 Hz.

 

[RCinFLA]

You did not post schematic but from the parts list the resistor stack is going to waste a lot of power and generate heat.

You can do a similar approach but the ballast is a capacitor instead of resistor stack.

Only catch on this approach is there should be a protection shunt across the LED stack to prevent a situation were someone quickly turns off and on the light. This may result in the capacitor ballast charged to its maximum peak voltage and when the turn on event happens to coincide with the peak voltage of the AC mains a large voltage spike can occur.

The protection circuitry is a small resistor (100-300 ohms) in series with ballast capacitor, or off the rectifier output, and either as shunt cap that is > 10 times the ballast cap value or a zener diode that is slightly higher in voltage then the LED forward bias stack. I don't recommend using an electrolytic cap for the protection shunt across a large LED stack since there will be a sizable ripple voltage on the electrolytic cap which reduces its reliability.

This approach will work with a stack of a single LED up to a LED stack of half the A.C. mains voltage.

Just about every LED replacement bulb you can buy uses this approach.

This is an example but lacks the protection circuit.
http://www.discovercircuits.com/H-Corner/AC-Powered.htm
http://www.discovercircuits.com/DJ-Circuits/aclinepilotled1.htm
**broken link removed**

 

[tcmtech]

Nice. I have built many of those too!
I use the capacitor ballast method myself. I have trouble with the white LED's always fading over time. The first six months to a year they are good but after that they always drop off and are about 1/4 their perceived brightness after two years or so. Brand, supplier, and price dont seem to make much difference.

I have a pair of under cabinet LED lights that are fairly expensive factory made and they to faded within a year of running continuously.

I still have my first one I ever built that uses about 50 green LEDs and has been on since around the fall of 2004 and is just as bright as the day I built it!
A buddy of mine has a yellow one and a red one that have been on for several years also without fade problems also.

White LED's just dont hack it on the long term from what I have seen so far.
Blue and ultraviolet are still in the test run phase. I will let you know in about another year what they do.

 

[RODALCO]

Schematic

RCinFLA , sorry I forgot the schema, here it is.

As said I before choose the resistor option because the reliability is greater than for capacitor voltage dropper circuits.

The power wasted in the series resistors is less than 2 Watts for the 240 Volts 29 Led version.

Of course more LED's can be put in series for more light output.

I mainly use these lamps for night lights or outside porch lights where reliability is important.

 

[Hero999]

The problem with capacitors is that they're more expensive and are easily damaged by high voltage spikes. Another hazard is that capacitors have a lower impedance to higher frequencies so if the supply is noisy, the LEDs could burn out; although using a small series resistor and derating the LEDs is normally enough to mitigate this.

 

[tcmtech]

If the capacitive type ciricuit is built right that stuff is not an issue. And I have yet to find a LED that cant take pulse currents several times higher than their rated continuous current.
If a small capacitor is in parallel to the LED string it takes any spike power and greatly reduced it any way. Plus the LED's see less ripple too!

If you cant afford a few 50 cent capacitors you probibly shouldnt be playing with your hobbies. But instead should be looking for a job!

 

[Hero999]

Resistors also burn much more power: I wonder how long it would take for the capacitors to pay fro themselves?

 

[RODALCO]

Led

tcmtech, At what current are you running your white LED's ?

I agree that the blue, yellow, green, red and amber leds are having a better long term light output than the white led's.


Haven';t got enough long term data for white led's.
time will tell.

 

[Hero999]

The problem with white is the phosphor fading with time.

 

[tcmtech]

The LED's I am using are rated for 30 Ma. I run them at about 25 Ma.

As far as the capacitor cost. All of mine are just used take offs from old TV and power supply boards. Same with the diodes, resistors and electrolytic capacitors!
Many of my LED's are from old user control panels from old office equipment. A large commercial copier control board from the 90's can have over 100 LED's and most of them are the same!

We have 120 volt mains. Running 40 LED's with a average forward drop of 3.3 volts at 25 ma means the capacitor only sees about a 35 volt peak across it. The average RMS will test as even lower.
So if you factor you are not running the full line voltage across them the actual working voltage across the capacitor is not going to be as high and thus doesn't need the normal high voltage rating associated with mains powered devices.

It may sound odd at first but the numbers do in fact work! My first green LED set up has a 100 volt non polar poly capacitor and its seen years of continuous run time without problems and thats proof enough for me that it works without reasonable concern.
I know my factory made LED cabinet lights are never going to make the same run time without fading right out of existence.

As far as power cost I am normally not one to nit pick the power ants but realistically if you change from a 2 watt loss using resistors to a .02 watt loss using good poly capacitors or similar in the 100K hour rated life span of the components thats still a few tens of dollars worth of electricity.

Just my theory.

 

[RODALCO]

white leds fading

Thanks for your detailed reply tcmtech

Always great to discuss matters and reasons why a certain method is choosen.

I do the same and often use old power R's or caps from old PSU and the like.
Photo copiers are an excellent resource of parts. (Also for power TX's and stepper motors).

I monitor my white LED lamps with hour meters and have clocked about 2000 hours with no noticeable dimming at this stage.
I will actually make a set up and mark the light cones on the ceiling and take an exact photo and do some comparison as time progresses.
My outside LED porch light does 14 hours a night at the moment.

I must say that most leds I use are underrun, current wise.
The white leds I use are rated at 25mA and are ran at between 10 and 15 mA's in most of my led lamp projects.

It is true that with a Cap. the power loss is negilgible as the phaseshift is 90° and will add to some costs.
2 Watt x 8760 hrs = 17.5 kWh (at 20cts/kWh about $3.50NZ or $2.35US)

Hero999 Good point, I forgot about the phosphors which are probably the cause of the fading of the leds. The white leds are blue leds with a special coating, as I am made to believe.

 

[RODALCO]

To add 8760 hours is one year continuous running.

Below is a video clip I made from another led lamp a while ago.

YouTube - 240 Volts LED pilot lamp -- project 4

 

[Hero999]

Here's a picture of my miniature transformer next to an 8 pin DIL for size comparison.

 

[RODALCO]

Thanks Hero999 for posting the image of the TX.

Thats looks like a 3, 5 or 7.5 VA.

That should fit ok in a CFL base.

 

[Hero999]

It's much smaller than a 3VA transformer.

The transformer on the left is 6VA, the one in the middle is 3VA and the one on the right (the transformer I'm talking about) is porbably about 1VA, mabe 1.5VA at the most?

 

[calfox]

how about the angle of illumination and the brightness of that lamp?.Is it much advantage to the CFL lamp in terms of the intensity of the illumination?..

 

[calfox]

Because you didn't consider here the effect of the illumination using LED bulb..