Best LED MR16?

[Justin Passing]

MR16/GU5.3
I'm sure mine don't last any longer than the halogens which were there before.
They have 3 different LED psus. Non dimmable, about 4- 5W.
Doing the sums, I'm getting way less than the claimed life - maybe a year.
I've used a range from a couple of quid to Philips ones , much the same life as far as I can tell.

I wonder if dimmable ones might control the power better.

Too many are 6400k or 3700K or 6W. I want about 4500k..

Has anyone found it's worth paying more than the low end?
Good make/supplier?

There was a post on this May 2020, many makes have changed

 

[flameport]

Don't bother with any of them.
To use properly, they will require a specific 12V LED driver, rather then the usual type designed for halogen lamps.
Even then it's still a compromise as they all contain a rectifier to ensure they work as the connector is not polarised.
Choice is extremely limited compared to 230V GU10 types.

Replace the fittings with GU10 types, and make sure there is sufficient space above/around them to avoid overheating.

 

[ericmark]

I fitted 4 in bathroom in about 2014 still running today. They say on them 50 Hz and the transformer is 50 Hz.

 

[JohnW2]

Doing the sums, I'm getting way less than the claimed life - maybe a year.

As I always have to remind people, the 'expected lifetimes' of things like lamps/bulbs are invariably quoted as median figures.

That means that if the (median) life expectancy is, say, 10,000 hours, then half of a batch of lamps is expected to last for 10,000 hours or more, and the other half for less than 10,000 hours - but that gives absolutely no indication of 'how much less'.

To take an extreme example, if half the lamps last for 10,000 hours or more, but the other half all die within the first 50 hours of use, that is still a (median) expected life of 10,000. Given that early failures are pretty common with mass-produced manufactured goods, one therefore should not be surprised if there is a substantial number of early, perhaps very early, failures, despite a long 'expected' life!

On therefore has to understand what one can/should realistically 'expect' - and what one experiences in practice is to a large extent down to (good or bad)luck!

I've used a range from a couple of quid to Philips ones , much the same life as far as I can tell.

That's similar to my experience. If anything, I think that I've suffered more 'early failures' with expensive ones than with 'the cheapest and nastiest I can find'.

Kind Regards, John

 

[ericmark]

All the MR16 G5.3 lamps I got were clearly marked 50 Hz and worked fine from a toroidal lighting transformer, if however they are powered from a switch mode device at kHz then one has to expect problems. They were not designed to run from that type of power supply.

Mine was a heavy lump, around 200 VA may have been larger, and the cheap Lidi bulbs worked fine. I now don't have any extra low voltage spot lamps left, all now are GU10's but since about 2010 I have been using LED bulbs, some have failed, around 4 or 5, but considering how many bulbs that's not bad.

For the 230 volt versions really should have a surge protection device fitted, without one can't really blame failures on the bulbs, if could of course be a faulty batch, but can't really blame the bulb if not SPD fitted.

 

[JohnW2]

For the 230 volt versions really should have a surge protection device fitted, without one can't really blame failures on the bulbs, if could of course be a faulty batch, but can't really blame the bulb if not SPD fitted.

You probably would have done ('blamed the bulbs') just a few years ago, when you probably hadn't even heard of SPDs in domestic electrical installations! Maybe you should stick with very cheap LED lamps/bubs, since if all they have is a rectifier and a capacitor or two, they're pretty (I suspect very) unlikely to be harmed by 'surges/spikes'

Unless/until I ever seem some decent 'chapter and verse', I'm probably going to remain sceptical about domestic SPDs. 'Surges' and 'spikes' seem to have become a (theory-based) scapegoat, without evidence I've yet seen that they are, in practice, a significant problem which needs a solution!

Kind Regards, John

 

[ericmark]

I'm probably going to remain sceptical about domestic SPDs. 'Surges' and 'spikes' seem to have become a (theory-based) scapegoat, without evidence

I do have to agree with you, I have seen of late more LED failures, but more likely due to having more LED's. In the days of tungsten I would say on average I would change a bulb a fortnight, today rare, however also today I actually have more bulbs, in last house living room started with 2 x 100 watt, by time I left it was 10 x 6 watt LED, same with this house living room was one 100 watt bulb when I moved in now 8 x 6 watt LED. Plus up lighters and cabinet lights.

I was always against the idea of MR16 lamps, however the design of the house has resulted in pod spot lights to aim light at dark corners, although not technically MR16 as LED's do not have a reflector, they are the replacements for the old MR16 lamps.

The failures are mounting up, first were so very cheap 0.58 watt from PoundWorld, but really they were toys so don't count, so the 5 foot fluorescent tube replacement was rather annoying as only 18 month old, and most of that time the house was empty, the colour changing GU10 was also annoying as it cost so much to start with, the G9 I opened to see what was inside and found a dry joint and returned to service, I have also lost a candle B22d, and one standard GU10 and that is about the lot, which is a rather low count over the years, I have tried to work out when I started fitting LED, I think likely around 2017, so only been using them for around 5 years, I thought it was longer. Originally I was keeping a record, but must have been on my now de funked laptop.

However it does seem either very short life, or a very long life, in the main except for a fire house in USA the old bulbs did not last very long, LED can be over or under driven, Phillips make the 'Dubai' bulb It seems this has been designed with a very long life in mind, but not released for the rest of the world.

 

[JohnW2]

I do have to agree with you, I have seen of late more LED failures, but more likely due to having more LED's.

Indeed - and you have to bear in mind my reminder above of what one's 'expectations' ought to be. To be consistent with a claim of a (median) life expectancy of, say, 10,000 hours, all that is required that at least half of them last for at least 10,000 hours - and that remains true even if the other half all fail early, perhaps very early. It's therefore only if more than a half 'fail early' that one has to start wondering about the manufacturer's claim

In the days of tungsten I would say on average I would change a bulb a fortnight, today rare,

Very much the same day. Given the size of my house, I have a lot of lamps/bulbs - quick mental count suggests probably atound 120 in total (including table/standard/bedside/cabinet/uplighter lamps etc.), many being candle bulbs in 3-biulb or 5-bulb fittings. In the days of incandescents, I bought bulbs in large batches a number of times every year (and used to 'joke' that light bulbs were a significant part of the 'running cost' of my house!). With LEDs, I'm only having to buy a small batch (usually of 'the cheapest I can find') every couple of years or so.

... However it does seem either very short life, or a very long life, in the main ...

As I've said, that is what is expected with mass-produced manufactured products of this type - the same goes for many electronic, and even mechanical, products - although the proportion of early failures obviously varies between types of product..

.... Phillips make the 'Dubai' bulb It seems this has been designed with a very long life in mind, but not released for the rest of the world.

I strongly suspect that, at least for me, such things would not be cost-effective. Indeed, the same is already true of 'ordinary' Phillips LEDs which cost considerably more than my 'as cheap as I can find' ones, but which, in general, don't seem to last any longer!

Kind Regards, John

 

[Justin Passing]

Sorry, I'd forgotten I'd asked this until a lamp started flickering!
To repeat, these are on LED psus - 12V switch mode driver things designed for LEDs.. (Non dimmable) Three different makes/ power ratings, as it happens.
There are suggestions above to change the lamp base fitting to GU10. That would cost a few hundred quid and in some cases be impossible because of hole shape/size.

I'm getting about 1000 hours life from them, maybe less.

I found these dimmable ones https://ledhut.co.uk/products/4-8w-mr16-led-35w-repl-345lm-4000k-d?_pos=1&_sid=4f8b41779&_ss=r
Repeat , I don't want to dim anything
I assume the non dimmable "bulbs" just have a rectifier and resistor - the psu has a regulator so should take care of smoothing, maybe all mine are poorly regulated.
Or, which I thought more likely, the LEDs in the bulbs are overdriven. So, buy 6W ones and take them apart to make them run at less?
I assume the dimmable bulbs are designed to work from a pwm 12V supply , but what's in the "bulb" that's different?

I've looked online of course - nothing clear and some explanations are laughable. "Your wall dimmer is a variable resistor which sends a lower power voltage to the LED so it's dimmer."

 

[bernardgreen]

To repeat, these are on LED psus - 12V switch mode driver things designed for LEDs..

A driver designed to drive LED elements will have a current rated output that matches the current rating of the LED element(s).

What you have is a 12 volt DC power supply and lamps that each have an integral driver that controls the current through the LED element(s) in the lamp.

If a "driver" with a 12 volt DC output was connected to an LED element then the element would be destroyed by the unregulated current passing through it. If the power supply is regulated at 12 volts DC then a simple resistor can be used to control the current through the LED element(s),

 

[Justin Passing]

You end up with the designed-for conditions, either way. Giving the same result. So where does extra life come from?
There's nothing I know of in the physics of LEDs which makes precise current control better, as long as they're not overrated. It's easier to use to dim one, but there's no analog dimming done... . There's no inrush like a wire bulb, no thermal runaway...
The current through the led is set by the incorporated resistor.

In a "bulb" There would be a full wave rec with a Vf of traditionally about 1.4V, the Vf of the LEDs - depending how many are in series, and a resistor and maybe nothing else. I'd like to know how dimmable bulbs are effectively different. I know they can flash etc, so there must be a difference. But what?

(The circuit above is incorrect somehow - no reverse polarity protection may be ok, but the current would be 15mA.)

 

[JohnW2]

You end up with the designed-for conditions, either way. Giving the same result. So where does extra life come from?

I'm a bit confused - who has said what about "extra life" ?

There's nothing I know of in the physics of LEDs which makes precise current control better, as long as they're not overrated.

I think you'll find that a lot depends upon what is coming out of your "DC" power supplies. It's generally far from true DC, which is why you will often see their output quotes as, say, "12 effective". That may sometimes mean that although the average of the output is 12V, it actually consists of pulses of appreciably more than 12V, so that the instantaneous current (during the pulses) may (if current is controlled just by a resistor) exceed the 'rating' of the LED element.

I'd like to know how dimmable bulbs are effectively different. ...

I suspect that (somewhat ironically) it may mean that they need to pretty basic/cheap, with just a resistor doing the current limiting. More expensive ones can have 'proper' current regulation within them, which would obviously frustrate attempts to dim the bulb by varying the voltage fed to them.

(The circuit above is incorrect somehow - no reverse polarity protection may be ok, but the current would be 15mA.)

I agree. Mind you, since bernard wrote that the 25 mA was "measured", I wonder whether the Vf of his LED elements was perhaps not (as he has indicated) 3.5V but, rather something just under 3.2V ?

Kind Regards, John

 

[bernardgreen]

I agree. Mind you, since bernard wrote that the 25 mA was "measured", I wonder whether the Vf of his LED elements was perhaps not (as he has indicated) 3.5V but, rather something just under 3.2V ?

The Vf was taken from the data for the devices. I still have that strip. will check the current later.

"12 effective". That may sometimes mean that although the average of the output is 12V, it actually consists of pulses of appreciably more than 12V,

Have seen 30 volt pulses on a "12 volt eff" supply, ( using an oscilloscope )

 

[bernardgreen]

I still have that strip. will check the current later.

Voltage supplied 12.11 Volts
Current taken 25mA
V across resistor 2.58 Volts
V across each LED element 3.19 Volts

 

[JohnW2]

Voltage supplied 12.11 Volts
Current taken 25mA
V across resistor 2.58 Volts
V across each LED element 3.19 Volts

Thanks. I reckon that 3.19 V qualifies as ...

... something just under 3.2V ?

Kind Regards, John