LED Torch Bulbs

[Onoff]

Where to get LED replacement? This one's 12V/0.7A. I'm seeing Cree ones at a tenner a pop! 

 

 

Anyone else upgraded their old torches? Cheap Chinese ones by the bucket load on eBay but SO many sellers!

[Jeremy Harris]

I've just fitted a cheap 6V LED bulb to a very old torch, with surprisingly good results.  The beam has a bright spot in the middle, but there is a decent spread of lower intensity light that makes the torch better than it was with the incandescent bulb.  The bulbs I used were these:  https://www.ebay.co.uk/itm/P13-5S-LED-Focus-DC-Flashlight-Replacement-Bulb-Torch-Work-Light-DC3-12V-4-5V-6V/322336571927?hash=item4b0cc1c217:m:mVeG4zafc5mofuWLfrboPMw

 

They are OK on 12 V, too, although the torch I fitted one to was a 6 V one.

[MikeSharp01]

Make your own by popping down to Maplins, getting an ultrabright white LED, work out a suitable resistor, unsolder the base of that bulb to remove the glass and fit the LED and resistor onto / into the holder - bingo.

[Jeremy Harris]

18 minutes ago, MikeSharp01 said:

Make your own by popping down to Maplins, getting an ultrabright white LED, work out a suitable resistor, unsolder the base of that bulb to remove the glass and fit the LED and resistor onto / into the holder - bingo.

 

 

The snag is that the efficiency of such a set up would be dreadful at 12 V and the resistor will get very hot.  The bulbs I listed have a high efficiency wide voltage range, constant current,  DC-DC converter in the base, so barely run warm.

 

A typical white LED run hard will have a forward voltage of around 3 to 3.3 V, and for 0.5 W the forward current will be around 0.15 A to 0.17 A (from I = P/V).  The series resistor will need to drop between 8.7V and 9 V, so the power dissipated in the resistor would be around 1.3 to 1.5 W (from P = V*I), three times more power than the LED is using, dropping the efficiency to around 25% before taking into account the LED efficiency.  Also, a resistor running at around this power will be far too large to fit in the bulb base and if it could be squeezed in it would get very hot.

 

Better to get the bulbs with the relatively high efficiency DC DC converter in the base, I think.  The bulbs I linked to barely get warm when running, certainly they are a lot cooler than the 0.5 W incandescent bulb they replaced.

Edited December 22, 2017 by JSHarris

[MikeSharp01]

2 hours ago, JSHarris said:

Better to get the bulbs with the relatively high efficiency DC DC converter in the base, I think.  The bulbs I linked to barely get warm when running, certainly they are a lot cooler than the 0.5 W incandescent bulb they replaced.

Makes perfect sense when you put it that way. Making a small DC-DC converter would be no mean feat. Alternatively just get a cluster of small LEDS and drop the voltage with them in series - we did this with PWM brightness on roadside type approved LED vms signs in the 1990s worked a treat. Not one LED failed in 10 years of operation. 

[Jeremy Harris]

1 minute ago, MikeSharp01 said:

Making a small DC-DC converter would be no mean feat.

 

Indeed it is, I'm amazed that they managed to fit one in the base of one of these small bulbs, but they have, as when tested the current reduces as the voltage increases, exactly as expected for such a device.

[MikeSharp01]

27 minutes ago, JSHarris said:

 

30 minutes ago, MikeSharp01 said:

Making a small DC-DC converter would be no mean feat.

 

Indeed it is, I'm amazed that they managed to fit one in the base of one of these small bulbs, but they have, as when tested the current reduces as the voltage increases, exactly as expected for such a device.

 

Wonder what the design is. 

 

[dpmiller]

tend to be a buck converter.

[Jeremy Harris]

6 minutes ago, dpmiller said:

tend to be a buck converter.

 

Interestingly, these bulbs appear to have a buck/boost converter.  The current drawn increases as the voltage reduces from 18 V, as you'd expect for a buck converter, then the current stabilises between about 5 V down to about 4 V, then it jumps up when the supply voltage drops below 4V and starts increasing again.  Given that the Vf of the LED will be around 3 to 3.3V, and these bulbs work right down to about 2.5 V, below the Vf of the LED, I think there may well be a boost element in there as well.

 

If I had a spare one I'd be tempted to try and take it apart to see what's in there, just out of curiosity.

Edited December 22, 2017 by JSHarris

[Jeremy Harris]

I've just done some testing on these bulbs, to see how they work over such a wide voltage range.

 

The bulb starts to illuminate at 2.8 V, and increases in brightness until it reaches a peak at 4 V, where it's drawing around 116 mA.  As the voltage is increased, the brightness dips slightly at 4.05V, then stays exactly the same (as far as I can see) from there up to 12 V, with the supply current reducing very roughly linearly with increasing voltage. At 12 V the current drawn is 36 mA.

 

So it looks like the total power is a maximum of around 464 mW at 4V, reducing to a pretty constant 0.430 mW or so over the rest of the voltage range.  Although these are advertised as "0.5W" bulbs, in typical Chinese fashion the actual rating is significantly over-stated.  Still, the one I have fitted to an old 6V Duracell torch seems to work OK, and draws around 57 mA, so with 4 off C size industrial Duracell alkaline batteries the torch should work for around 80 hours on a set of batteries, which seems pretty good to me.

Edited December 23, 2017 by JSHarris

[MikeSharp01]

You wonder where they hide the inductor - must be the biggest part. What does the ripple look like on the output as BUCK inverters are not good at dead flat without extra components. (If you have not dismantled it yet I guess you might get a look at the ripple using a photodiode reciever.)

[Jeremy Harris]

I've not taken it apart, as it's in my torch and I doubt I'd get it back together again!   As above, I think it must be a buck/boost configuration, as it lights the LED up below the normal Vf of a white LED, especially taking into account the inevitable internal voltage loss when operating at low voltage.  It may be that they just use a low Vf white LED, and rely on the DC resistance of the inductor in the buck converter to limit the current, when the voltage is too low for the buck converter to oscillate.  That might work, and might also explain why there is a current and brightness peak at about 4 V - that could be the point just below the start up voltage of the buck converter.

 

I suspect it is really pulsing the LED at some high frequency when operating above about 4 V, and won't be attempting to smooth the output.  Standard constant current switched mode LED drivers work like this, and it's one reason why finding ones that don't cause a lot of EMI is challenging.

Edited December 23, 2017 by JSHarris