Do these litle lights need a "starter"?

[winston1]

Unless it's a clever/expensive one which looks at instantaneous voltage and and current very frequently and then integrates the vector VA product to get power, it will be assuming a perfect sine wave current, which I imagine is probably not the case with and SMPS

Maybe.

Well, for a start, I thought you said that the 'heat' was the same. Whatever, how do you know that the heat is being dissipated in the lamps, rather than in the SMPS - did you perhaps measure the lamp temperature?

If it is in the lamps their life will be shorter. If it in the SMPS its life will be shorter. Either way the efficiency is a lot worse.

I have, in the past, deconstructed some LED lamps, and cheap ones often just have the rectifier and a resistor (and often a pathetically small reservoir capacitor). I also suspect that 'dimmable' ones may be the same (regardless of price), since that is presumably the easiest way to facilitate dimming.

I have deconstructed 240 volt ones and some are as you say. But I meant I hadn't deconstructed 12 volt ones. Just consider the size of a capacitor needed to drop 12 volts to 3 volts at 0.45 amps, 50Hz. It is well over 100µF and it can't be electrolytic. So they most contain a SMPS, particularly as most can work at DC or 50/60Hz.

In any event, when there are losses, they result in additional heat being dissipated in the rectifier diodes, not in the load as you suggest (or is that perhaps what you meant when you said "in the lamp"?).

Indeed the rectifier diodes are part of the load.

 

[JohnW2]

I have deconstructed 240 volt ones and some are as you say. But I meant I hadn't deconstructed 12 volt ones.

I was talking about 12V ones. What I said was "just a bridge rectifier and resistor" and, unless there were an awful lot of LED elements in series, if that method were used with a 230/240V supply, a high proportion of of the supplied power would be dissipated as heat in the resistor. With that supply voltage, it makes much more sense (in the cheapos) to use a bridge rectifier and capacitor.

Just consider the size of a capacitor needed to drop 12 volts to 3 volts at 0.45 amps, 50Hz. It is well over 100µF and it can't be electrolytic. So they most contain a SMPS, particularly as most can work at DC or 50/60Hz.

I'm not sure where the 0.45A came from, but I certainly agree that a very large series capacitor would be needed, and I have already acknowledged that this clearly is not the way most (all?) 12V LED lamps work - since, as you say, most (all?) of them will work with a DC supply.

I'm not at all sure that 12V LED lamps really need an internal SMPS. If one puts three LED elements in series, there's not a lot of voltage left to be dropped, so a resistor or very simple constant-current circuit would probably be adequate.

[ By the way, one can use electrolytics with AC, if one puts two in series (opposite polarities) with diodes across each of them - but I'm not certain that would be reliable enough for 'production' items ]

Kind Regards, John

 

[winston1]

I'm not sure where the 0.45A came from,

I'm not at all sure that 12V LED lamps really need an internal SMPS. If one puts three LED elements in series, there's not a lot of voltage left to be dropped, so a resistor or very simple constant-current circuit would probably be adequate.

It was a rough ball park figure for a 5watt lamp at 12 volts. Actually it works out at 5.4w so 0.45A is a little high.

3 elements in series with a constant current source sounds good. Must experiment sometime with a variable voltage source.

 

[JohnW2]

3 elements in series with a constant current source sounds good. Must experiment sometime with a variable voltage source.

Given that the people who design these things presumably are not daft, I had always assumed that such was what they did. If one assumes that it is going to be supplied from a supply which really is around 12V, it doesn't make much difference whether one uses a resistor or constant current circuitry since, in either case, the 'excess power' is dissipated as heat (either in a resistor or a transistor). However, if one takes that approach it makes no real sense to use just one LED element, and hence 'waste' about three-quarters of the supplied energy, when one could put three in series and 'waste' only about one quarter.

To effect a constant-current source, one only needs a transistor, a couple of resistors and a zener diode (or 'zener equivalent' using silicon diodes).

Kind Regards, John

 

[winston1]

I'm not at all sure that 12V LED lamps really need an internal SMPS. If one puts three LED elements in series, there's not a lot of voltage left to be dropped, so a resistor or very simple constant-current circuit would probably be adequate.

Well I have now done my experiments with a couple of 12 v GU5.3 LED spot lamps, a 5w one and a 3w one, both originally from Lidl.
Using a variable DC supply I measured the current over a range of voltage.
With the 5w lamp the current fell as the voltage was increased from 11.5v to 15v. (as far as I was willing to take it.) The brightness remained constant as far as I could tell and quick VA calculations at various points showed constant power at 4.8w. Below 11.5v the current and brightness fell to zero at around 8v. The 3w lamp behaved in a similar was but the "turnover" point was around 11.8v.

So I think we can safely say they contain a SMPS, not a resistor or constant current source.

Just found this. The first part is a 12v 3w SMPT LED driver from about 3 mins 20 secs in:

 

[JohnW2]

With the 5w lamp the current fell as the voltage was increased from 11.5v to 15v. (as far as I was willing to take it.) The brightness remained constant as far as I could tell and quick VA calculations at various points showed constant power at 4.8w. Below 11.5v the current and brightness fell to zero at around 8v. ... So I think we can safely say they contain a SMPS, not a resistor or constant current source.

Certainly not a resistor or constant-current source - so, as you say, maybe a "SMPS".

I suppose it depends on what you feel qualifies as a SMPS. It could just chop the supply (on/off), with a varying 'mark-space ratio', and I would not be inclined to call that an SMPS. As has been observed in the past, it's far from unknown for LEDs to be 'overdriven' in a pulsatile fashion.

I don't own any 12V LED lamps, so cannot currently do any experiments (or murders+autopsies!) myself - but maybe I'll get some!

Kind Regards, John

 

[Roger465]

Thread resurrection… so I originally replaced 17 halogen bulbs (in ceilings, bathroom mirrors, kitchen cabinets) with LED bulbs. Despite everyone’s misgivings here (which I respect), up to now they’ve all worked fine – so that’s for about a month so far.


However, one of the ceiling units, and the two in the kitchen cabinets start to flicker 5 or 10 minutes after firing up. So I swapped bulbs between a flickering and OK unit, and the flicker stayed with the unit. OK I thought – that only leaves the transformers, which have been here since 1994 as far as I know.


Now the ones I’ve got (see the white one in the pic) are marked ET60T-2, with various other hieroglyphics – so I thought I’d get a couple of brand new ones off eBay – but there weren’t any when I looked. However someone was selling a bundle of new ET60T-3’s for very little, so I bought them.


Unfortunately, when I connected one up and switched it on, the LED went pop instantly. I thought maybe I’d fried the transformer, but I put the halogen bulb back in the fitting, and on she went.


So it would seem that there’s some kind of subtle difference between the ET60T-2 and -3, possibly along the lines some of you have already suggested would pop LEDs. But, although I don’t understand all of the markings on the transformers, the only one which seems to be different is “11.6V RMS” on the original (white) transformer, but “11.6V EFF” on the new (green) one.


Is that my problem?


I know I’m not following best practice here, but it all seems to work apart from that, so I'd be grateful if you’d indulge me

 

[Roger465]

PS just read up a little and now know that EFF means "effective" (I knew what RMS was, without really understanding it), but can't quite divine anything useful from that...

 

[JohnW2]

PS just read up a little and now know that EFF means "effective" (I knew what RMS was, without really understanding it), but can't quite divine anything useful from that...

"Effective" means roughly what it says - essentially that (probably because the output is a bizarre waveform) some sort of average of the output voltage is 11.6V, although it may oscillate between figures much higher and much lower than that. This probably results in much the same power/light as with an 11.6V RMS sine wave with an incandescent (e.g. halogen) bulb/lamp, but what it means in terms of an LED, or whether the high voltage excursions would harm an LED, is anyone's guess. That's one of the problems of using something not designed for LEDs!

Kind Regards, John

 

[Roger465]

"Effective" means roughly what it says - essentially that (probably because the output is a bizarre waveform) some sort of average of the output voltage is 11.6V, although it may oscillate between figures much higher and much lower than that. This probably results in much the same power/light as with an 11.6V RMS sine wave with an incandescent (e.g. halogen) bulb/lamp, but what it means in terms of an LED, or whether the high voltage excursions would harm an LED, is anyone's guess. That's one of the problems of using something not designed for LEDs!

Kind Regards, John

Thanks John. Getting close to just putting all the halogens back in here. Seemed so simple when I started out... save a little money by using frugal LEDs everywhere. But if I have to replace the transformers to make them work properly, it'll cost more than I could ever save by using less electricity!

 

[JohnW2]

Thanks John. Getting close to just putting all the halogens back in here. Seemed so simple when I started out... save a little money by using frugal LEDs everywhere. But if I have to replace the transformers to make them work properly, it'll cost more than I could ever save by using less electricity!

I can understand your feeling. I'm afraid that it's a problem which results from having 12V lights - since, as you've discovered, "there is 12V and 12V" as far as electronic power supplies ('electronic transformers') are concerned. Had you had 230V halogens, you could have simply swapped them for 230V LEDs, since 230V 50Hz AC is always 230V 50Hz AC!

Kind Regards, John
Edit: Missing digit inserted!

 

[winston1]

As you were told in the original thread they are not transformers (ignore the lies written on then) they are switch mode power supplies operating at 10s of kHz. LED spots invariably say DC or AC50/60Hz. So why did you think changing like for like would work? You either need DC power supplies, a genuine transformer, or best of all change to 240volt LEDs.

 

[Roger465]

So why did you think changing like for like would work?

Because 14 of them DO work, as stated above.

 

[winston1]

Because 14 of them DO work, as stated above.

They ALL worked initially. Just wait a little longer and more and more will fail. Use the correct DC supply or transformer.

Both your old and new supplies clearly say on them FOR LV HALOGEN LAMPS. It does not say say for LEDs.

 

[333rocky333]

since 230V 50Hz AC is always 20V 50Hz AC!

You sure