Best LED MR16?

[ericmark]

A LED needs some method to restrict the current, the problem is when we buy a bulb we have no idea how this is done in that bulb, it could be a pulse width modulated controller, a simple resistor, or a capacitor, with the exception of the resistor, it may need a smooth DC supply, or a 50 Hz supply, and a pulse width modulated chip supplied by a device also using switch mode regulation or pulse width modulated without so form of smoothing could do all sorts of things.

Radio receivers and transmitters use diodes, coils and capacitors plus the odd resistor, and the LED light has the same components so could easy become a radio transmitter, when doing my Radio Amateurs exam we looked at filters and tuned circuits, and we could be unwittingly be forming these with the high current involved.

So we have two options between the lamp and voltage dropping device, 50 Hz and a lamp designed for 50 Hz or smooth DC and a lamp designed for smooth DC, other combinations may work, but we really have not got a clue what we may be making by combining non matched lamps and power supplies.

The same to some extent with low voltage (230 volt AC) we hope the lamps have no DC component and will not affect the operation of out protective devices (RCD's) we know the type A RCD can handle 6 mA of DC, but we have no idea if the devices we are using produce no DC or 100 mA DC, some devices like EV charging units do monitor DC and will auto disconnect if it exceeds 6 mA, but most of the stuff we use has nothing to monitor the DC let alone trip if it exceeds the safe limit, we know with a type A it can handle 6 mA, and we know with type AC it may or may not handle 6 mA, but only safe option is type B which are very expensive.

So we cross out fingers, we really have not a clue what is going on, with the exception of @bernardgreen using an oscilloscope, so all we can do is avoid the switch mode power supply, and the toroidal lighting transformer is a large heavy lump, so in real terms the GU10 is likely the best option, can still use same fittings.

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.

That does not ring true, as no need to change fitting, however running the earth may cause problems.

I was under the impression now unless better than 75 lumen per watt, or decoration only as with Christmas tree lights old lamps had to be withdrawn from sale, but today walking around Pound Land both G5.3 MR16 and G9 quartz bulbs were still on sale. Maybe MR16 are seen as for decoration only?

 

[bernardgreen]

the LED light has the same components so could easy become a radio transmitter,

Many of the Switch Mode Power Supplies do radiate radio frequency energy that can interfere with wireless communications,

Two examples

An LED desk lamp that blocked comms to a hand held walkie-talkie when the unit was on the desk close to the lamp.

An "electrician" replaced the toroidal transformers supplying tension wire lighting in a picture gallery overlooking a harbour with compact SMPS units. Caused havoc with marine radio when the lights were turned on

 

[bernardgreen]

I reckon that 3.19 V qualifies as ...

Within the limits of measurement errors I agree.

Resister was ( is ) 100 Ohm )

 

[Justin Passing]

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

My first post complained about short life, the reply "Replace the fittings with GU10 types" so far is unexplained.

--

I have scopery, including a Velleman (not from a kit) I could take up a ladder if only I could find it. I really must get one of the £30 ones! I have previously looked at the output of this type of cheap psu - I must say I've not found them to be very bad at all. In smps design it's pretty easy to stop high spike excursions.
Call me a cynic but are you saying a better regulated supply is always squashed into every mains input GU10 led "lamp" base?? Not much in this transparent one:


Maybe all those leds are in series! Rec + cap + ???


The psus don't last very long either, so I made sure I used replaced them with higher rated jobbies - to the extent that I could put one halogen on the same output in addition. That didn't seem to help. This one I think

100W LED Transformer/Driver - Non Dimmable

100 Watt LED Transformer to use with 12V LED spotlights and LED strip lights. Free 5 year warranty and 30 day returns.

ledhut.co.uk

though I note the same rated thing at Farnell is twice the price.


Dibbling with a minimal circuit for a 3W led- not practical values though, C1 is large.
So they must use a controller.

 

[Justin Passing]

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.

That does not ring true, as no need to change fitting, however running the earth may cause problems.

The bulbs are different length, and the "luminaires" aren't mains rated, there's nowhere to put an earth. Rewiring would not be easy either.

 

[Justin Passing]

Helps if you use multiple leds. This is 30 leds in series. Changed to blue to keep the vf up. Should have used zeners!
More reasonable values.

the problem is when we buy a bulb we have no idea how this is done in that bulb

Er well, yes, that's what I'm asking. I have one of mine to saw/grind open.

I've seen circuits for BC sized bulbs, those have components you can alter if you can change SMD resistors. If you derate them a bit they do last longer, apparently.

Wouldn't it be nice if they lasted say 10 years...

 

[JohnW2]

My first post complained about short life, the reply "Replace the fittings with GU10 types" so far is unexplained.

Fair enough. Although that's what your first post was about, I don't think (but may be wrong - he will hopefully confirm) that flameport was really responding in relation to lamp life - I think he was merely expressing his view that LV lamps are preferable to ELV ones, not the least because the latter require external PSUs/'drivers' (which, as we are discussing, can be 'problematic')

I have scopery, including a Velleman (not from a kit) I could take up a ladder if only I could find it. I really must get one of the £30 ones! I have previously looked at the output of this type of cheap psu - I must say I've not found them to be very bad at all. In smps design it's pretty easy to stop high spike excursions.

We may not be talking about the same type of PSUs since, with some, my experiences are similar to bernards. If the output is pulsed (which it often is) then to achieve an average output of 12V ("12V eff") obviously requires thatthe pulses are greater than 12V, increasing so as the duty/mark-space ratio of the pulses decreases. If a PSU produces something approximating to DC with superimposed ripple (i.e. not 'pulses') it is much more likely to be marked as "12V DC" than "12V eff" (or even "12V DC eff").

Call me a cynic but are you saying a better regulated supply is always squashed into every mains input GU10 led "lamp" base?? Not much in this transparent one:

[ the image of you 'transparent one' doesn't seem to have appeared ] but ....

..... Maybe all those leds are in series! Rec + cap + ???

I think that might well be the case. It is a long time since I last bought/used (hence dissected) anything other than the cheapest of LED lamps/bulbs, and they nearly always have just a bridge rectifier and a couple of capacitors - so obviously no question of any high voltage, high-ish rate pulses being generated.

Dibbling with a minimal circuit for a 3W led- not practical values though, C1 is large. So they must use a controller.

What you depict there is what one will usually find in at least the cheap LED lamps/bulbs, except that there will usually be no "R1", just a very low value of "C1". However, I'm not sure wat point you are making. Such a circuit obviously results in the output you illustrate - (with your R1 present) DC with superimposed 100 Hz ripple, which become pretty clean DC, of a voltage equal to Vf of the LED, across the LED. ... obviously no question of 'pulses', high voltage or otherwise.

Kind Regards, John

 

[ericmark]

I have also opened some LED bulbs, only ones which have failed, but some really simple, others more like this and until failed one has no idea which are which. But in all looking at around 6 bulbs in 6 years, I would not say that is a problem. OK when the fluorescent tube replacement went at around £16 I was some what peeved, and had one smart bulb fail at around £6, two were very early versions from PoundWorld, and one I actually found a dry joint and returned to service.

But I would guess around over 40 LED bulbs in the house, so 6 fails in around 6 years not really bad.

However my son has had more failures, so I have considered why, he does not have any SPD fitted, and he bought a batch of cheap bulbs from toolstation, he blames cheap bulbs, it seems the cheap bulbs I fitted before leaving the house are still working, also cheap, Home Bargains and Lidi, and all the 12 volt power with a toroidal lighting transformer.

I looked at the link @Justin Passing posted, and it is DC, and the bulbs I got from Lidi were clearly marked 50 Hz so designed for AC. I know sons house is low voltage just 230 volt, as it caused problems with the old 65 watt fluorescent starting, not really measured here.

I feel @Justin Passing is touching on the problem we are all having, lucky for me it was only one fitting, I have posted this picture many times, both G9 but the large one had a large smoothing capacitor inside, nearly as big as the small bulb, the small bulb was failing to switch off until this fitted, and then there was a shimmer when switched on, it did seem connected to the switch, which is electronic as we have no two way switching in bedrooms, so use these remote controls on bedside table. These have been a problem, did have 5, now down to 2, for many of my lights moved to smart bulbs, also used smart relays, but the problem is we are buying bulbs forced on us by the government, but there is no control over bulbs offered for sale.

On Tuesday I went to Poundland, and they are still selling G5.3 and G9 quartz halogen bulbs, I thought now bulbs had to be 75 lumen per watt or better, seems that's not the case.

But at work we have festoon strings of 110 volt LED bulbs with it seems no problems, same with replacements for fluorescent tubes.

One point is all my M5.3 bulbs are in pods, so good cooling, but think just 15 x GU10 for whole house, and 2 failures, so the M5.3 are not doing as well as other LED bulbs.

 

[bernardgreen]

Golf ball LED lamps are intriguing.

The village Christmas decorations use about 1000 golf ball 1 Watt lamps which have to be tested before being installed

So far I have found 3 types that all have the same circuit with the same components values but a different number of LED elements.

No image of the 6 LED version which has 6 small surface mount LED elements

 

[JohnW2]

But I would guess around over 40 LED bulbs in the house, so 6 fails in around 6 years not really bad.

That might bring us back to what I said about 'expectations' early on in this discussion.

It obviously totally depends upon the use of the bulbs, but if I pull out of the air a (numerically convenient!) assumption that the bulbs were used, on average, for about 2.3 hours per day, over 6 years that would amount to about 5,000 hours.

If the claim for the bulbs was an average (median) life expectancy of, say, 10,000 hours (another assumption), then one would expect half of your 40 bulbs (i.e. 20 bulbs) to fail within 6 years.

I obviously have no idea as to how close my assumptions are to the situation in your house but, if it were remotely correct, you would have done pretty well to have only 6 (rather than 20) failures in 6 years.

Kind Regards, John

 

[ericmark]

I also looked at the inside of the bulb, and realised the GU10 and the candle bulb were nearly the same, both had a flat plate with the chips on, so hardly any light sent back towards the base.

With the bases on the bottom and light reflected off the ceiling they seem to work well, but reverse it and it needs some thing to defuse the light, plus the heat goes into the bulb, the capacitor here is going to get a lot hotter if the base is above the plate.

 

[Justin Passing]

I'm not sure wat point you are making. Such a circuit obviously results in the output you illustrate

I simulated it to check ripple really, and component sizes with altered designs, and see if anything else showed up. Leds aren't ohmic of course so calulations aren't trivial. The simulator does do a reasonable job.

As Bernards cct shows, there are other ways to do it.

I'm not convinced that spikes are causing problems - they're too easy to kill in the cct.
None of which really helps with my short-lived bulbs tbh, other than to try another psu. They were the old 12Vdc supplies for halogen, which flashed. I still want to know the difference in the dimmable bulbs!
Looking for an inductive 12V transformer supply, nothing obvious leaps out.

 

[bernardgreen]

They were the old 12Vdc supplies for halogen,

One of those supplies was rated as being 12V DC (eff).

The output was 100 cycles per second, Each cycle started at 0 volts and rose to about 30 Volts before switching down to 0 Volt for the rest of the cycle. Averaged out to be about 12 Volts

Not a problem for an incandescent lamp but fatal for electronic modules looking for a 12 volt supply.