DIY LED Bulb Project PBO April 2011 P88

[nigelmercier]

I was interested to read the DIY LED Bulb Project in PBO April 2011 P88. Interested is perhaps the wrong word, astonished may be better.

The article is quite specific, this is intended for a TriLight. Nowhere are any calculations done on the brightness of the lamp (note: it is not a "bulb") and whether it meets the required range.

Moreover, the author does the voltage calculations based on when the engine is running, which is precisely the time when the use of a TriLight is prohibited!

I often wonder if these projects are checked by PBO staff.

 

[Elemental]

I've not read it myself. But it was in the APRIL edition was it?...

 

[nigelmercier]

I've not read it myself. But it was in the APRIL edition was it?...

Yes, but I'm pretty sure it wasn't a joke - at least not intensionally.

 

[VicS]

Were the LEDS used the correct ones to match the colours of the colored sectors? If not then its not worth the paper its printed on.

After umpteen years as a regular reader I dont buy PBO any more

 

[nigelmercier]

Were the LEDS used the correct ones to match the colours of the colored sectors? If not then its not worth the paper its printed on.

No they are warm white, which is OK.

 

[Salty John]

It has to be the joke! It doesn't have the usual references to April 1, but the picture of the finished article is the give away. It looks like a childs first attempt at glass blowing!

 

[VicS]

No they are warm white, which is OK.

They emit significantly at frequencies that match both the red and the green lenses?

I wonder why they are not more generally used then instead of the special coloured ones that are.

 

[webcraft]

They emit significantly at frequencies that match both the red and the green lenses?

I wonder why they are not more generally used then instead of the special coloured ones that are.

I've fitted one. Red is fine, green is slightly blue-ish but still definitely green. Had to remove it because it interfered with the VHF, which was a shame.

I would have no hesitation in using a warm white in a coloured lens even if the colours were slightly off. The light police are rare visitors.

- W

 

[Rigger Mortice]

I would have no hesitation in using a warm white in a coloured lens even if the colours were slightly off. The light police are rare visitors.- W

I'll second that. Slightly blue, but mine is also a little brighter than the original bulb.
At sea other boats react as I would expect; it must be working!

 

[Dipper]

I've just fitted a 'proper' tricolour LED and a warm white anchor LED. When I shone the anchor light LED through the green lens it looked distinctly blue to me. That was in the house. Maybe the effect is less noticeable from a distance.

 

[William_H]

LEDs

LED bulbs run on a specific current with a volt drop typically 2.5 volts. There is a simple way to control the current using a series resistor. The value of the resistor sets the max current through the LED (or usually 3 or 4 in series). Any more current will destroy the LED. So you have to calculate and set the current based on max battery voltage. ie engine running. Even though normal use is at more like 12v (battery not on charge). You could calculate for 12v supply and get max brightness and current for the LED but if it were switched on inadvertently when battery is under charge you would melt the LEDs.

The better way to control the current is by a switching regulator. This will give precise current under all battery voltage conditions. However by switching the current at high speed it will/may cause interference to radios. Webcraft you may be able to rewire the LEDs for resistive current limiting no interference but less efficient. An alternative is a linear regulator ie 7809 giving stable 9v to then have a stable current from a resistor. But this gives worse efficiency.

The problem with building a tricolor lamp is getting the sectors clearly defined. Most bulbs have a defined arc of light emission. However this is not sufficiently clearly defined.
I would use individually coloured LEDs not a coloured lens and I would use baffles to improve definition. LEDs emit specific colours because of the material used to emit the radiation not a lense. I would think colour was OK. It is actually more difficult to make a white LED.
good luck olewill

 

[electrosys]

I've fitted one. Red is fine, green is slightly blue-ish but still definitely green. Had to remove it because it interfered with the VHF, which was a shame.

Is it possible to screen the switcher ? Or maybe wrap the supply leads a few times around a ferrite toroid (as recycled (nicked) from old computers) as near to the switcher as a possible ?

 

[pteron]

Is it possible to screen the switcher ? Or maybe wrap the supply leads a few times around a ferrite toroid (as recycled (nicked) from old computers) as near to the switcher as a possible ?

The better way is to design the switcher to operate at a frequency that will not interfere.

Using anything other than a switcher to drop the voltage is going to significantly reduce the efficiency gain of using an LED in the first place.

BTW 2.5V is seriously optimistic for a white LED - they are typically 3.5 - 4.5V forward drop.

The reason the white LED looks blue when viewed through the green lens is that the light is actually a blue LED exciting a yellow phosphor - not much green in it to start with and the filter is taking out much of the yellow.

 

[Conachair]

The better way is to design the switcher to operate at a frequency that will not interfere.

Using anything other than a switcher to drop the voltage is going to significantly reduce the efficiency gain of using an LED in the first place.

BTW 2.5V is seriously optimistic for a white LED - they are typically 3.5 - 4.5V forward drop.

The reason the white LED looks blue when viewed through the green lens is that the light is actually a blue LED exciting a yellow phosphor - not much green in it to start with and the filter is taking out much of the yellow.

What about 1 of these .. http://www.maplin.co.uk/l200c-adjustable-voltage-and-current-regulator-8064
Only goes down to 0.1a but that should be low enough?

Not sure what drivers are inside, but the internal LEDs i have are very quiet re radio noise, just tried ssb radio tele aerial right next to them, no noise until almost touching.

 

[Rigger Mortice]

Why bother with regulators or switch mode supplies?

If you are making up an array, fixing a number of led's in series would work. (5 would do it....I think)
Make up as many strings (each string connected in parallel with the other strings) as you need to get the overall brightness.
Why not?

 

[Conachair]

Why bother with regulators or switch mode supplies?

If you are making up an array, fixing a number of led's in series would work. (5 would do it....I think)
Make up as many strings (each string connected in parallel with the other strings) as you need to get the overall brightness.
Why not?

Well I'm no electronics designer but have blown cheap led replacements before. You need something which will work equally well from below 12v to 16v. From a flat battery offshore to when you are equalizing the batteries and forget to turn them off. I don't think connecting them in series or using resisters as current limiters really cut it.
I'm off to maplins now

 

[Rigger Mortice]

Fine, but a regulator is more inefficient. If you are dropping from, say, 12v to 3v, most of your power is lost as heat. As always these things are a compromise, but connecting in series would avoid that heat loss.

LED's have a remarkably long life. They won't normally blow when used properly.

 

[nigelmercier]

What about 1 of these .. http://www.maplin.co.uk/l200c-adjustable-voltage-and-current-regulator-8064...

Gulp! Maplin is not the best place to buy stuff like this, see Farnell instead.

This would be fine, but the problem is that it will only drive a string of 3 or 4 LEDs from 12V, and if you parallel up the strings you defeat the object of using a constant current driver.

 

[Conachair]

Gulp! Maplin is not the best place to buy stuff like this, see Farnell instead.

Seriously cheaper!! But you can't walk there on a saturday afternoon from my boat

 

[Conachair]

Fine, but a regulator is more inefficient. If you are dropping from, say, 12v to 3v, most of your power is lost as heat. As always these things are a compromise, but connecting in series would avoid that heat loss.

Well, still not convinced. I'm not the most educated about these things, so went to maplins and bought couple of L100 regulators to play around with. With 2 x 9v duracells in series to get up to 18v, then down to about 11v open circuit through the L200 into a replacement g4 from bedazelled, the current going into the chip was 175Ma ( with a bit of current limiting as the cheapo meter only does 200ma) , the current out of the chip into the led was 145Ma. Which takes it to above 80%.
Not strictly of much use as the LED I was using has it's own limiting circuit, but seems to suggest that the L200 chip is actually quite efficient.

But so much to learn - how much to LED's care about the voltage if you can keep the current down? Is it even worth bothering about setting the voltage or just do the current?

the chip did seem very quiet as far as RF goes as well.

Such fun.