I am lighting long passageways in a block of flats. I have ceiling mounted Passive Infra Red Sensors in the ceiling and the 20 individual wall lights turn on 2 at a time (using timer relays set at 0.1 second intervals).
I use a special "ramp up dimmer modules which goes from "filaments just glowing" to "on" in 2 seconds to give a ripple down the corridor and fade up effect. It looks great but there is an issue.
The 20 wall lights turn on in pairs. They are "up/down" lighters with 2 x 35 watt x (UK) GU10 mains halogens. Total load per light =3D
70 watts.
2 . Each "ramp up module" ramps up a pair of lights so the total load per module =3D 140 watts (at 240 volts the load on each ramp up module =3D 0.583 Amps)
The 10 =93pairs of=94 lights turn on at 0.1 second intervals (controlled by DIN timers) so after 1 second the load gradually increases to 5.83 Amps.
The corridor has been running for a few months initially fed by a 6 amp breaker. This breaker tripped out perhaps once very 2 weeks. Obviously this breaker was too small for the 5.83 amp loading
10 days ago a 10 amp breaker was fitted (and I assumed this would solve the problem) but yesterday this 10 Amp circuit breaker also tripped out. (I have yet to find out if a bulb blew as part of this event as I'm on holiday).
My question is can you think of any reason why initial loading with a ramp up module into a cold filament should draw more current than predicted? I know the =93cold=94 filament resistance is much lower that hot resistance but does the slow warming of the filament prolong the initial "inrush" current?
If so I could change the circuit breaker to a 10Amp type C (motor type) breaker.
Without any calculations I had merely assumed that the slow ramp up would reduce the initial surge as the voltage was lower (etc). The ramp up module has a "pedestal" control to set the start point. This has been set for a small initial glow so the initial voltage could be starting at half??
What are the group=92s thoughts? I'm struggling to understand what's going on. ( I can=92t measure the load for a few weeks and only have an old AVO 8 so a short spike may not show up on an analogue meter
The first question that springs to mind, is how much power are you dissipating in the dimmers? That may be enough to push you over the lower MCBs limit.
If a bulb has blown, then all bets are off with mains halogens....
The soft start usually limits inrush quite nicely...
What cable is the circuit wired in? What is the earth fault loop impedance like at the end of it?
Does the trip occur at switch on or after the lights have been running for a bit?
How long are they typically on for?
Perhaps the losses from the dimmers account for the occasional trip of the 6A MCB and a bulb failure took out the 10A one?
I am lighting long passageways in a block of flats. I have ceiling mounted Passive Infra Red Sensors in the ceiling and the 20 individual wall lights turn on 2 at a time (using timer relays set at 0.1 second intervals).
I use a special "ramp up dimmer modules which goes from "filaments just glowing" to "on" in 2 seconds to give a ripple down the corridor and fade up effect. It looks great but there is an issue.
The 20 wall lights turn on in pairs. They are "up/down" lighters with 2 x 35 watt x (UK) GU10 mains halogens. Total load per light =3D
70 watts.
2 . Each "ramp up module" ramps up a pair of lights so the total load per module =3D 140 watts (at 240 volts the load on each ramp up module =3D 0.583 Amps)
The 10 =E2=80=9Cpairs of=E2=80=9D lights turn on at 0.1 second intervals=
(controlled by DIN timers) so after 1 second the load gradually increases to 5.83 Amps.
The corridor has been running for a few months initially fed by a
6 amp breaker. This breaker tripped out perhaps once very 2 weeks. Obviously this breaker was too small for the 5.83 amp loading
10 days ago a 10 amp breaker was fitted (and I assumed this would solve the problem) but yesterday this 10 Amp circuit breaker also tripped out. (I have yet to find out if a bulb blew as part of this event as I'm on holiday).
My question is can you think of any reason why initial loading with a ramp up module into a cold filament should draw more current than predicted? I know the =E2=80=9Ccold=E2=80=9D filament resistance is much lo= wer that hot resistance but does the slow warming of the filament prolong the initial "inrush" current?
If so I could change the circuit breaker to a 10Amp type C (motor type) breaker.
Without any calculations I had merely assumed that the slow ramp up would reduce the initial surge as the voltage was lower (etc). The ramp up module has a "pedestal" control to set the start point. This has been set for a small initial glow so the initial voltage could be starting at half??
What are the group=E2=80=99s thoughts? I'm struggling to understand what's going on. ( I can=E2=80=99t measure the load for a few weeks and only have an old AVO 8 so a short spike may not show up on an analogue meter
Best regards - LightingMan
au contraire. It limits temperature on filament hotspots, and 12v supplies may be current limited, but not mains dimmers.
ok, some points.
Youre using 20x 240v halogens, which is a design thats going to cause problems whatever you do. IIRC this link explains why such a design is likely to trip the breaker repeatedly:
formatting link
dont see a simple easy way to fix it.
Power loss in dimmers is negligible.
As a temporary option, fitting a fuse rather than a breaker may help. These are 16th edn compliant, but very tolerant of current inrush. Unfortunately theyre not a great option for long term use in your app.
Reducing the bulb powers is also an option that may just work.
I'll admit to wondering why you'd choose mains halogen for the type of lighting youre using. A zillion of those is going to gobble energy and money, and cause breaker hell.
A type C breker or less likely D might help, but you'd need to work out your disconnection times and see if they comply.
The ripple on effect is going to stretch the usual 0.1s-ish surge out in time over what, 1 or 2 seconds. 12v halogens on current limited supplies would have kept things within spec at all times, and would slow the lamp warm up as you want, without needing dimmers. Would still gobble electricity though.
would reduce the initial surge as the voltage was lower (etc).
The lowest power setting a triac dimer can manage is to turn on right at peak v, thus creating max possible current surge. With halogens this will give you in the region of 14x run current. By running the bulbs at around half power, they will stay at lower resistance longer, making matters worse.
Do you mean instantaneous voltage or rms voltage? Instantaneous will be peak, or 1.414x 240v =3D 340v. Rms will be 170v.
An avometer won't even begin to register current spikes. You need to calculate rather than measure.
The short version is this issue is probably the result of the design. Whatever you do, if you keep using mains halogens you will sometimes get breaker pops when bulbs blow, and 20x 1500hr bulbs is 1 dead bulb per 75 hrs. I really would consider dumping the mains halos, theyre about your worst option.
To fix it will mean either changing the lighting, or getting fairly creative on the electrical side. Lets explore.
An electrically much better option than dimmers would be to swtich the bulbs on in series pairs, then switch them to full power. This is easily done with a time delay relay, and you dont need any dimmer modules. Whether that would get you under the tipping point... I dont know, maybe. Series pairing them would reduce the on surge by a factor of 4, which sounds like what you need.
Another option that would definitely fix it is motorised variacs, but theyre not cheap - and 10 of them is even less cheap.
Really I'd look at a more suitable type of lighting. If you really want to keep the existing fittings, there's always the option to reduce the bulb power in them and add some 3000K fl lighting, using hidden eledctronic ballast fittings.
Depends on the style of dimmer. If it is like a normal lamp dimmer in series with just the live feed to the bulb, then there can be no nett loss since it can do nothing to lower the total resistance (and hence raise the current flow) through the bulb.
However if it has its own independent supply and acts as a switching PSU for example, then there can be some additional loss in the dimmer.
I get the impression that the aesthetics would not work with a low energy bulb. A low voltage halogen would however save the problem of CPD trips on bulb failures.
yup
A link to the dimmer in question may be handy here
I think the first thing is for me to describe the =93ramp up=94 modules
They are the SSC25 made by United Automation in the UK
formatting link
answer to John=92s questions
I don=92t think much power is dissipating in the =93ramp up=94 module as=
they =93dim up=94 in 2 seconds after which the lamps are at full power. The modules are not mounted to a heatsink and run cool
Each pair of lights is fed by 1.0mm T&E. The corridor is about 40 metres long. The longest cable is hence about 40 metres and feeds a pair of up/down firings (4 x 35 watt bulbs). The cables get closer and closer to the control panel and the shortest cable is 3 metres.
I don=92t know when the MCB trips (you asked did it happen right at the start or after they had been on for a while) as the failure is once every few weeks and in a communal passageway =96 it is reported to the porters and it=92s even difficult getting a report of the bulb blew at the time never mind such a detailed observation!
The ceiling mounted PIR=92s are mounted outside the front door of every flat=92s front door and they are set to come on for 4 minutes (long enough to walk up the passageway)
In answer to Meow2 comments
Why do you say that I can=92t use a fuse as a long term solution? I must admit it is very inconvenient if it blows!
You can see the ramp up modules I have are already starting the bulbs softly, sorry I didn=92t provide a link to the module which may explain things better now =96 sorry I didn=92t make things clear - but I think I am already doing what you suggested.
General comments
The choice of mains halogens was dictated by the style of fitting, if I could find a version with low voltage halogens I would probably change them =96 It took 3 years and hundreds of hours to find a decent light fitting which looked good and gave out the right light.
The system doesn=92t =93gobble energy=94 as it=92s only on for literally=
minutes each day. It=92s in a block where people tend to go to work at
7 am and come back in the evening) In fact it takes less than the old system of low energy wall lights which gave a horrid harsh white and bland uniform light.
I have chosen a short on time so as many people can observe the lights switch on as the ripple and fade up effect is a =93work of art=94
The mains halogens have a lower colour temperature than 12 volt halogens and give a lovely warm light which complements the carpets (please don=92t get me on to a rant about how disgusting the colour temperature of virtually every low energy lamp is!!! =96 One day I know there will be warm LED=92s that can look like halogen lighting but I=92m still waiting!!)
I could supply my control panel with a 30 Amp breaker and as I have ten 1.0mm cables coming back to the control board I could easily fit ten x 10 Amp MCB=92s after the dimmers and timers to protect each 1.0mm T&E. Fusing downstream would at least stop total failure - it would just knock out a pair of lights. It is just slightly inconvenient as there are no spare ways on my control board. I would have to fit another board and rewire the lot again to keep things neat.
I guess you can if you want, but large buildings normally use breakers. Maintenance will take a minute longer with a fuse if it pops - OTOH it will pop less often.
which thing that I suggested do you mean?
=BDs only on for literally
fair enough. Low energy lamps are a mixed lot, and if you specify the right tubes you can get a good quality light - dont specify and you probably wont, as you've found.
3000K fl tubes are available, but better to avoid the cheaper halophosphate ones, they have poor CRI at 3000K.
Or you could split the lighting circuit into three.
Well, it might - or the mcb might trip first. But it would improve the frequency of total trips.
Well, you're going to keep what you have, and have found it wont run on the MCB you have.
Begs the question does the "ripple" start from the PIR that has been triggered in both directions or does it always start from the same end of the corridor? If the latter I'd hate to be at the far end from the start.
The spec sheet on the dimmer modules suggests a consumption of the dimmer at about 5W. So we are talking about about 200mA extra load overall for ten of them. Enough to push just past the 6A point - but unlikely to be far enough past to cause a trip on a 6A MCB
The spec sheet has the dimmers consumption at about 5W - so not huge but a couple of hundred mA overall.
So you have about 1.5 ohms round trip cable resistance P to N or P to E at the longest point. So as long as your earth fault loop impedance at your supply is 0.8 or better, you should have a prospective fault current of at least 100A which will open even a 10A Type C breaker on the "instant" part of its trip.
Yup, I can see that making it difficult!
Don't worry about it... some folks round here have a devotion to low energy lighting that verges on the religious. ;-)
How about splitting the lights into two sets - with the lamps organised such that adjacent pairs are in different sets. With a 6A Type C protecting each set. You could then arrange for the ripple to start from both ends of the corridor and run toward the centre - crossing over in the middle. If you do get a trip, it will only affect half of the lights leaving the whole corridor lit at a lower level.
Thanks for the contributions, this is the internet at it's best!
Thanks to John for the calculation & confirming a C rated breaker would be OK & yes I agree about the religeous devotion to low energy lighting, most stairwells have 2D lights and give a hideous, white, bland and dingy effect. I know LED lights promise to be excellent in the end but we're probably a few years off and the colour temp of even warm white is not good.
Thanks to Meow2 about the colour temperature of low energy lighting (why do you always see mixed colour temps of lamps in large buildings- like at Gatwick airport!)
Thanks Dave, but don't worry, I own one of the flats right at the end. The Danlers ceiling mounted PIR takes a fraction of a second to operate but covers the inside of the flats (when the doors are opened) even if you're quick, you are just coming out of the door to observe the delightful display! The time delay is no different to reaching inside any room in a house to flick on the light switch.
Usually most of the consumption of dimmers is in the form of series v drop rather than parallel i use. 5w is high though, so that might not be the case with these ones. But the secified 5w will be at max specified load, and this load is far below that, so they wont diss 5w irl.
What matters to appearance is CCT and CRI. If both are right, low energy looks as good as filament. But most buyers want the cheapest, which dont match filament lighting and dont look too good.
LEDs are junk performers, but they sure are hyped to death.
Mixed CCT and CRI lighting are usually down to buying whatever's cheapest. Sometimes mixing 2 types is a way to get passable lighting using low end tubes. Many tube buyers know nothing about CCT or CRI, some do but dont care, etc.
Remember these are not simple series droppers - the power consumption of the dimmer itself appears to be fixed regardless of the load. The actual total dissipation due to the load itself is a different matter. The specs include a fair amount of detail on the heatsinking requirements for the dimmers when used at the higher loads (they can run at up to 25A and can control not only lighting but inductive loads such as motors etc).
Unfortunately, some of us do, but are unable to locate CFLs with the combination they require - at least when constrained to use existing fittings, etc. Or if you do not wish to pay a high price (all 'supermarket' ones seem to be 2700K).
I suspect that the mixed colour temperature effect at large sites might (partly) be down to the replacement at different times - whether or not they are of the same spec. It seems to me that lamps vary in look even from one manufacturuer to one spec - especially if supplied at different times. And lamps change in look over their lives.
Isn't the problem simply that mains halogens blow every few weeks, and tend to trip the breaker when they go? It's been discussed on here many times before, and seems to be the experience of many people. I've never found a decent solution to the problem but the more expensive lamps tend to blow a bit less often. I don't have any mains halogens now for that very reason. LV seem to last for years rather than weeks/ months, and don't trip the MCB when they go.
Posting from google groups and added sci.engr.lighting as crosspost, now with added captcha, so not sure how far it now reaches
LED is probably advancing faster than anyone expected, decent output and colour is already attainable its just a matter of cost which can only go down.
Worth differentiating current limited supplies as referred to earlier in relation to LV lamps and constant current supplies that may be encountered with LEDs, they are differnt types of low voltage power supply.
For the quantity, getting the fitting shells assembled with your choice of lamp holders and possibly internal transformers may well be worthwhile. LV MR16 has a different beam pattern to GU10 but there is a much wider choice.
I don't have access to the original message in this thread, but part of it seems to be about LESD-based MR16 lamps.
I would disagree that LEDs are advancing faster than anyone expected. LEDs are just about as good as, or slightly worse then, the predictions made for them at the first high brightness LED conference back in February 2000.
As for LED-based MR-16 lamps. These are currently limited to about 5 watts input power due to the limited heat sink area and the dual facts that LEDs do not radiate any of their waste heat and they must also be kept relatively cool. They therefore can't produce the same light output as a
50-watt halogen MR-16. However, if you can live with significantly lower light output, LED-based lamps are much more efficient than halogen versions.
18 floors x 10 fittings is enough to ask the mfr for what you want, and you may get it.
If not successful, replacing the dimmer modules with relays and starting the halos in series, then have the relay switching to parallel would much reduce the startup surge, resolving one of the trip causes. And the resulting partial soft starting would increase the lamp life significantly as well.
If you stay with mains halo, you have still to address the trip on blow problem somehow. If you don't, resetting breakers will become a job in itself.
With LV lamps, again you could simply run them at reduced V then switch to rated V to ramp up the output, thus avoiding the cost of 180 dimmer modules. I'd pick more than 50% of voltage, and for other than 50% you can set the startup v using a capacitor.
I'd also question the use of reflector lamps there, in a steel or chrome pipe you'd only need a single non-reflector lamp in the centre to get a twin spotlight type beam pattern. Use of capsule lamps would halve your bulb failure rate as well as trimming costs.
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.