To add to your thoughts regarding an LED MR16, to me one of the most significant negative issues is the CRI... which is typically in the mid sixties or worse. MR16 applications typically call for bright focused light on specific objects or areas where CRI should be a significant portion of the overall evaluation. Bottom line is that earth tones look crappy with a CRI so low, although blues and grays would look acceptable.
To add to the posters comments regarding the performance difference in an MR16 versus a GU10... I struggle to understand the point. Would the GU10 not refer to the BASE of the lamp? An MR16 is an MR16 (along with it's varying degrees of photometric availability, typically 10, 35 or 60 degree) regardless of the base - or perhaps I'm missing all the info?
I agree we're talking about the base! If somebody could sell me some convertors (from a GU10 base to a 2 pin MR16 socket) I'd buy them!! I could just locate transformers next to these light fittings. The real solution is to find an identical light fitting which takes MR16's
Dear Meow 2
As my automatic "soft start ramp up" modules start from a dim glow and ramp up within 2 seconds I can't see a lot of difference doing what you suggest (putting 2 in series and switching) and what I'm doing. Part of the effect is the smooth "ramp up" ..........In all honesty circuit breakers are probably tripping when the lamps blow but I'm not getting the feedback from those resetting the breaker.
LED MR16 used as ceiling mounted aisle lights on recent UK railway carriages, supplementing cove mounted fluro I think. CRI can get to low mid eighties with still reasonable efficiency. In a transport application they have obvious advantages of low power consumption and shock resistance , guess they form part of the emergency lighting system as well.
Getting higher CRI by mixing in amber LED for instance works well but drops the efficiency, pure phosphor white LED looks terrible on skin tones.
Its mebbe a UK way of referring to them, yes a GU10 is a type of MR16 lamp but generally MR16 is used to refer to low voltage MR16s typically bi pin GX5.3 12V, where as GU10 refers to 240V twist and lock base MR16.
Swapping lamp bases may not only consideration with fitting though, GU10 base lamps have an aluminium reflector heat goes out front. Most LV MR16 lamps have a dichroic coated glass reflector which dosent reflect heat, most goes out back. The fitting may rely on aluminised reflectors to push heat out front. There is GZ10 240V MR16 lamps with twist and lock base but chamfered shoulders that has a dichroic coating.
For the number of units involved getting something assembled to spec shouldn`t be too difficult, also gets round any approval issues.Quite number of UK companies will handle fitting assembly. Birmingham still make some things.
Assembled rather than manufactured , out of standard lamp parts into an existing outer shell, used to work for a company that did similar things, some for quantity lot less than 200 pieces. Its a light fitting not a space shuttle.
Why dont you ask in sci.electronics.*, they'll explain why although the visual effect is a gentle ramp up, the current waveform begins with about 14x normal run current. Its your problem if you dont address it.
If you understand how triacs work, you'll know the only way they can reduce rms voltage is by switching on later in each half cycle than right at the beginning.
If you're familiar with filament lamps, you'll know the glow level is in the region of half rated voltage, and hence lighting dimmers produce around half rms voltage output at minimum setting.
And if you understand how a dimmer reduces rms voltage, you'll appreciate that this point coincides with switching on power at the very peak voltage of each half cycle. Not only do they maximise inrush current this way, they also lengthen the inrush period by slowing warm up of the filament.
In short, standard mains lighting dimmers change a brief high inrush current to a lengthened and more consistently high inrush current. I squared t is what trips fuses and breakers, and ramping up on dimmers significantly increases i squared t, as explained.
Add a huge number of mains halogens and the OP could not have devised a better way to trip breakers on a frequent basis. Oh, yes one can - run the lighting circuit up to its rated current limit, as done.
The OP made a design oops, and its one that isnt so simple to fix. There are fixes, but they all mean more work. Series/parallel switching the bulbs is a simple low cost way to eliminate the switch-on trip problem, but it wont protect against failed lamps arcing over. Putting the lights on 3 circuits will mean that when it occurs, 2/3 of the lights will still work.
There are certainly better options. Another is to run a 260v lighting circuit with a series impedance to each lamp to bring it back to
240v locally. The impedance acts to reduce magniture and duration of control arcover current, greatly reducing the chance of a breaker trip.
Changing to 12v halogens, 1 per fitting instead of 2, would wipe the problem out totally, and reduce relamping by a factor of 2 or probably somewhat more. If the OP wants the CCT of mains halogen, its not difficult to trim the V_out of a toroidal transformer to give the required CCT.
Never seen a dimmer soft start or otherwise, seen a few, that starts in middle of half cycle from zero. They all start from conduction point round, this dosen`t geve a very linear charecteristic and big dimmers have dimming law curve correction as part of the feature set.
The voltage is limited to considerably less than half mains at startup combined with resistance of connection and filament and reactance of RF choke inline with load, dimming substanially limits inrush current.
Mains halogens always remind me of Magicubes fro some reason...
Dosent appear to meet OPs brief of fade up either
sub distribution appears to be 10 circuits already so splitting the rail feeding them seems a good answer to complete darkeness.
Future electricians would love trying to fathom why some circuits are well over voltage.
LV halogen seems best answer, see your point about a central lamp in the fitting. Sorry about confusion Vic even I wouldn`t recomend LED for MR16 replacement in an application like this
I cant figure out what you mean by that. Starting at glow level means starting near peak voltage (instantaneous), which is half v (rms)
R of connections is trivial here. Filament R dominates Dimmer chokes only make significant difference at rf frequencies, and dont affect starting current
If you retain the dimmers it doesnt make any difference to the fade up requirement afaics. Fading may be less smooth for each individual fitting, but since the v switching times are spread in time the rippled fade up still happens, and the v switch will occur at fairly low rms voltage.
:)
totally
truth is theyre not replacements, theyre just mrketed as such. With much lower power theyre not equivalents, and the poor CRI defeats one of the main points of the OP's plan, to have nice lighting. CFL or fl would be far better than LED in every respect.
In my experience, most dimmers can trun on at the point that delivers RMS voltage around 25-30% of full, my eyeball estimate from seeing with what color and what brightness the filament glows.
And once the lamps start glwoing, the dimmer can be turned down until RMS voltage is around 8-10% of full before turning off. (My eyeball estimate from color and brightness of the filaments.)
This is for 120 VAC use.
Most 120 volt incandescents visibly glow at 12 volts, especially most
60 watts or more and most 15 watts or less (since the latter nromally have a vacuum).
I'll take your word on that Don. When I was a kid I strung 3 mains filament lamps in series (240v here) and no glow, so I'm not sure why the difference. I don't remember the details of the lamps though.
Me back again!! There seemed to be a bit too much speculation & plain =93wrong =96headedness=94 creeping into some of the posts so I did some empirical testing to try to get to the bottom of this.
There has been talk about "dimmers" (even I used the term -sorry) but remember I am using "soft start" modules where I can control the "start point" (pedestal) and the rate of "ramp up" from 0 =96 60 seconds.
For measurement purposes I started with a 60 second ramp time and noted the point where the filament started to glow (the glow is the start point in the real passageway)
Using an ancient Avo 8 I tried different "start points" and the "glow" happened at 0.04 Amps with one 35 watt bulb. The measured voltage across the lamp a the starting =93glow=94 point was approx 30 volts RMS (a definitely no more than 40 volts) For any international readers the UK line voltage is 240 volts.
Over the next 60 seconds the current rose (reasonably) constantly to a maximum current of 0.12 Amps which (at 240 volts) equates to 30 watts. The voltage across the lamp rose in a completely steady & linear manner over the 60 seconds.
I couldn=92t understand why the max power measurement was only 30 watts and wondered if my Avo had lost calibration (after all its 30 =9640 years old!!) but when I disconnected the =93soft start=94 module the current went to 0.14 Amps which is about right. Hence there is a small voltage drop in the =93soft start=94 module
Finally I slowly sped up the =93ramp up=94 speed to 2 seconds as per the actual passageway.
At no time in the testing were there any initial =93surge=94 currents and even though I=92m using an analogue meter, the way I can slow things down over 60 seconds proves the point (even an analogue meter would flick to the top of the scale if there was a surge).
Hence I defend my design!! I=92d say there are no initial surges. I think the MCB tripping is happening when the bulbs blow but I=92m still waiting for the porter to report back if there is correlation!! An C rated 10 Amp breaker has been fitted but what do the group think??
(The bizarre thing is that I have another stairwell with 8 of these GU10 fittings and no =93soft start=94 modules. It has been running for 2 years. Bulbs blow but the B rated MCB has never tripped!! Although I know that some brands of MCB are more prone to tripping than others- the big old black bakelite Crabtree breakers never trip when bulbs blow!!!)
But with things slowed down the thermodynamics of the filament are different. Having said that I don't know if they would be significantly different...
Yes I wondered about that but I started testing at 60 seconds and tested again and again, setting the "ramp up" faster and faster. There were no changes to the "roughly" steady current increase. Certainly no surges at the 2 second ramp up point.
Not sure what happens with no soft start module (probably large start up currents) but I don't care about this as I have soft start modules on every light fitting.
So you dont understand that the mechanism of action of the 2 is the same. That implies you dont know how they work. If you dont know what they do with the waveform, how are you going to calculate what happens?
How exactly did you get your moving coil avo to tell you the rms voltage and current? Are you aware that moving coil meters dont read rms values on non-sine waveforms, or that you're working with non-sine waveforms? The numbers you read off the avo are not factual numbers, because its not an rms meter and you're not reading sine waveforms. They wont even be close.
So you're trying to measure a phenomenon that takes at most 0.1 seconds using a very old moving coil meter with a response time in the region of 2 seconds. And you interpret the failure to detect any short lived current flow this way as proof that it doesn't exist! If you checked the cold resistance of your lamps and did some basic calcs you'd begin to understand.
The problem has been explained, but youve not understood it. If you do the cold resistance check and look at the curves of your MCB, things might begin to become clearer.
If you'd crossposted to a sci.electronics.* group they could have explained the characteristics and behaviour of filament lamps, triac control modules, MCBs etc. Instead youre asking DIYers.
LightingMan back again! - In response to Meow2=92s comments
I do understand the =93ramp up=94 module is like an automatic dimmer
I do understand it uses triacs or similar and the wave forms which come out
Having said that I don=92t want a flame war with you - I actually need your help and your comments are some of the most useful on the thread!
I agree I had totally forgotten about the chopped up waveform when I connected my Avo 8 and carelessly used the term RMS (as that=92s what it usually measures!) Point totally accepted
Where I disagree with you is the response time on the Avo is quicker than 2 seconds. You can play audio into an Avo and watch it jump up and down like a VU meter in a recording studio. I=92m guessing you could see quite short spikes.
I agree with you about measuring cold resistance and I know how it varies.
What I don=92t really understand it that if I slow the whole process down to 60 seconds I don=92t see current increase above the norm and drop down when the filament is hotter. (As you say just because I can=92t see it doesn=92t mean it=92s not happening)
However, I was assuming that the Avo and a circuit breaker both operate with coils, soft iron cores, magnets etc. I had assumed if I couldn=92t observe the effect on the Avo then it wouldn=92t be enough to trip a circuit breaker (other wise any GU10 lamp would trip a breaker on switch on).
To conclude I don=92t discount that my design hasn=92t made things worse but I get the gut feeling (from observations and not calculations) that the soft start modules make it better.
I had a 10 minute chat with the "soft start" module manufacturers. (Interestingly, their modules are triacs controlled by a little integrated circuit - I thought it might be a simple RC circuit controlling the triacs)
He agreed that my Avo 8 was not likely to be "all that accurate" due to the non sinewave output but said with a largely resistive load like a GU10 lamp that there was no way there would be initial current surges. He mentioned the whole point of a "soft start" module was to give a "soft start"
He also mentioned that the module was "backed off" to cope with inductive loading which is why the 35 watt GU10 only drew 30 watts (as indicated on the Avo as Meow-2 points out, I was not reading RMS).
We discussed a "special" production run of 160 modules with a 0-3 second "ramp up" range as I was the potentiometer for "ramp speed" was right at the end of it's range and was tricky to set. The manufactuer said that if there was a special production run that they could make sure that there was less voltage drop in the module as it was a non- inductive load.
After that chat I chased up the porters again and it was finally revealed that a bulb had "gone" when the 10Amp MCB tripped. Interestingly enough I left them with a 20 pack of bulbs and they have used 3. (The MCB has also tripped three times!!)
So thanks to Meow-2 (even though I felt mildly offended!!!) and the person who calculated I could use a C type MCB. I think Meow -2 (and others) in the group are right - a 12 volt MR16 version would be far better but I can't find a fitting like that!
HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.