Why are my halogen downlighters cheaper on than off????

[wookey]

I'm pretty sure JohnS has explained the mechanism for the odd readings.

HalyconR: Please look at LEDs again. They _are_ dimmable (just fit a PSU with that functionality - all LEDs are intrinsically dimmable by altering the supplied current). Efficiency and colour rendering has also improved enormously over the last couple of years.

I don't believe there is any situation they are not now suitable for - even the previoulsy unassailable flourescent tube market is coming under pressure.

Replacing GU10/MR16 downlighters is simple, but is not generally good engineering - you can get much better results by putting in fittings designed for LEDs, rather than trying to force them into fittings designed for halogen bulbs. This improves heat dissipation which is the key to long life and high efficiency. You clearly care about your power consumption, and halogens are shockingly inefficient for the lighting level attained - the _only_ thing they have going for them is that the bulbs are cheap. But this is a short-term false economy. LEDs are the same/better in every other way: lifetime, colour choice, start-time, dimmability, colour control, long-term cost, power consumption, luminaire design flexibility, lighting design flexibility.

All new shop lighting as already LED, and all new household lighting will be LED from approximately now onwards. Sitting there saying that they are 'no good' and only halogens will do is ignoring reality.

[stuartiannaylor]

Wookey I have to dissagree about LEDs being inherently dimmable. Leds are current controlled devices and the only way to efficently dim Leds is via PWM.

The performace of voltage / current dimming of LEDs isn't very good.

[clivejo]

I have a wireless monitor called an OWL something or other.  Most of the time its pretty accurate, however when I plug in my Electric Power Saver (http://www.windtrap.co.uk/energy-saving-devices/power-factor-reducers/eps-188/) it goes absolutely crazy!  When the EPS is online certain devices appear to consume vastly more energy and some less than without it.  Just switching it on and off totally changes the consumption on the Owl monitor, with the same physical loads connected.  Finding this strange I started reading about what the device actually does and its designed to modify the power factor of the circuit, which is why I mentioned it earlier on.  I also read that Switched Mode power supplies affect the power factor.

The jury is still out on if the EPS actually works or not!

[SimonHobson]

What I cant work out is why only the halogen cause it to drop when switched on, the TV, dishwasher, kettle, toater all cause it to jump as you'd expect

What type of transformer are they run from. Kettle and toaster are a sizeable resistive load, so current will go up. If you have some reactive load on, then adding a suitable size of other reactive load can result in less current while still taking more power.
Several things have highly inductive loads - an idling motor and a flourescent light are two such devices. With these, the current "lags" the voltage, is if you plot the voltage and current over a cycle, the current lags behind. With a capacitor, the current leads (you need to put energy into the capacitor before it's voltage will rise). If you put the two in parallel, then the two effects cancel out and power factor is improved - this is what the capacitor in a traditional flouro fitting is for.

So it's possible that you have a poor power factor from something in the house, and the load from the lights (or more specifically the transformer) is counteracting that.

Another possibility is that the transformer (assuming it's electronic) is so rubbish that it spews lots of noise into the mains, and this is interfering with the monitor.

I have a wireless monitor called an OWL something or other.  Most of the time its pretty accurate, however when I plug in my Electric Power Saver (http://www.windtrap.co.uk/energy-saving-devices/power-factor-reducers/eps-188/) it goes absolutely crazy!  When the EPS is online certain devices appear to consume vastly more energy and some less than without it.

Yes

Just switching it on and off totally changes the consumption on the Owl monitor, with the same physical loads connected.  Finding this strange I started reading about what the device actually does and its designed to modify the power factor of the circuit, which is why I mentioned it earlier on.

It's a box of snake oil for the gullible.

I also read that Switched Mode power supplies affect the power factor.

In theory, modern SMPSs have to have fairly reasonable current waveforms to meet regulations, however it is well known that a lot of cheap kit from China actually has empty spaces where components were designed in (and probably fitted for certification tests) but are not fitted in production. Note that it is generally load current waveform that is particularly bad, not necessarily power factor. In particular, without any correction circuitry, an SMPS tends to take a large current peak around the peak voltage, but little or no current between those peaks. This creates a lot of "harmonics" (in particular the third harmonic) which unlike the fundamental 50 Hz does not subtract in 3 phase circuits. With nice load waveforms, the neutral current is always less than (or equal to) the highest of the phase currents, but with a lot of harmonics the neutral current can be considerably higher. This was a problem when computers started to become common, and lots were installed in offices that had "small" supply installations to start with. I believe there were quite a few problems caused by overloaded neutrals.
Flourescent lights also have the same problem.

The jury is still out on if the EPS actually works or not!

Not in my mine. Put simply there is no possible mechanism for it to work. It may affect simple current monitors (like the Owl and it's like), but it cannot affect the real power taken by something else.

[clivejo]

The jury is still out on if the EPS actually works or not!

Not in my mine. Put simply there is no possible mechanism for it to work. It may affect simple current monitors (like the Owl and it's like), but it cannot affect the real power taken by something else.

I have to agree on the 'power saving' claims of the device.  It was given to me and I run a few tests, I didn't notice any savings over about two months.  I currently use it with a square wave inverter as it seems to smooth out the waveform and reduces the 'buzzing' sound from appliances.

[SimonHobson]

I currently use it with a square wave inverter as it seems to smooth out the waveform and reduces the 'buzzing' sound from appliances.

Well yes, a capacitor would help with that.

[wookey]

<quote>Wookey I have to dissagree about LEDs being inherently dimmable. Leds are current controlled devices and the only way to efficently dim Leds is via PWM.

The performace of voltage / current dimming of LEDs isn't very good.</quote>

pish. I don't know where you got that idea.

Yes LEDs are current-controlled. If you control the current then they go dimmer: it's very simple. PWM is one way of doing that (actual current control is the other). Power LEDs have a max lumens/W at some particular current, which is typically around 1/2 to 1/3rd output. So if you _really_ care about efficiency you buy 2.5 times as many LEDs and run them all at 1/3rd power (obviously this is expensive so isn't often done).

I don't know what you mean by 'performance of voltage/current dimming isn't very good', but as explained above, the max efficiency point is not the max ouput point so the efficiency improves as you turn down the LED to say 1/3rd power, then drops off again, and dimming has to be done by current control. LED 'voltage dimming' is a nonsense (it might 'work' due to the characteristics of the driver circuit and other (resistive) components mounted around the LEDs, but that just because the current is changing in the right way when the voltage is varied). Trying to control an LED by changing the voltage is unstable and thermal runaway (i.e melting your LED) is easy to do.

So, I repeat: All LEDs are inherently dimmable, using current control.

Now of course you may buy an LED _unit_ that is not just one or more bare LEDs and thus is not easily dimmable because the unit has some extra electronics in it. Usually this is so that it can be driven by a voltage-controlled PSU. This is more poor engineering, but people are used to thinking in terms of power supplies delivering volts so it's easier to sell them a '12V' power supply (which delivers 12V and between  0 and 1A) then it is to sell them a '1A' power supply (which delivers 1A at between 0 and 12V) (both supplies having a maximum 12W output).

So yes if you buy a '12V' LED device which has a built in driver to make it run at 1A when you supply 12V then it won't dim nicely. But if you buy a plain '1A nominal' LED device then you can drive it (or indeed as many of them as your PSU will run in series) at any point you like between 0 and 1A and they will dim perfectly.

Sadly sellers are just as hopeless at this as your average person so it's incredibly hard to find out which of these things they are selling you, and similarly a '12W' PSU rarely makes clear if it is constant current or constant voltage (or current-control, or voltage-control if dimmable). The whole thing is a mess, which is one reason I'm making my own LED house lighting from bare LEDs, aluminium and current-controlled supplies so I know what I'm getting (it's also cheap, and very efficient, but rather fiddly).

[HalcyonRichard]

Hi Wookey,
                 It's not as it seems. I have 8 35 watt halogens in my kitchen. One is permanently off as it is above a small T.V. The rest are switched as 3 off above a dining table. The other 4 off are above work surfaces. The lights are only used in winter and only when we are preparing or eating. The rest of the time a 5 watt CFL above the fridge is on. I calculate that this usage is an average of  0.2 kWh per day. My total average daily use is 4 kWh including working at home and using a reflow oven for making PCB's. So the lights are 5% of my usage (not allowing for useful "heating" in the winter months").
I assemble PCB's as part of my work - the vast Majority are by daylight. But some rush jobs don't allow me to organise this properly. So placing 0402 components 1mm x 0.5mm just does not work with CFL's or LED's It's funny I don't know if this is colour rendering index or my old eyes. So if daylight is not available then it's a kitchen table job with the halogens. I have tried CFL's and LED's but they are very different. I intend to look again next year when I will be moving home.  But at a running cost of £10/year it's low priority in relation to other energy saving projects. i.e. 1 mile in a car is probably equivelent to 1 kWh of electricity use. (my current useage of my partners car is 3000 miles per year including work mileage) or maybe that wood burning stove has a better chance.

Regards Richard

[wookey]

Interesting that you found placing small components harder with LED lighting. I don't know about you but I can't do 0402 without a microscope these days. My previous place of work had two nice binocular microscopes for soldering. One had a halogen light (and a noisy fan to keep it cool). The other had an LED light. Both worked just fine.

If you are trying to do it without a microscope then it's probably just down to light levels. I've noticed that fine work requires a very bright light. White LEDs have gone from 7 lumens/W to 130 lumens/W over the last 13 years, and so far no sign of finding a limit (although one is expected around 250 l/W), so things really have improved dramatically. There are colour temp issues and colour variation which depend on the phosphors used, and some LEDs are hugely better than others in terms of colour rendition. I don't know how much this spectrum affects fine work, as opposed to simple intensity.

[stuartiannaylor]

all LEDs are intrinsically dimmable by altering the supplied current

Sorry but I got confused at the above quote. As you don't alter the current. They are only efficient at there rated current.
You could say you alter the current over time, but actually PWM is turning off and on the LED at different mark/space ratios and the current whilst on remains the same.

I am only being pandantic as I have come across quite a few installs where they have been fitted to dimmer banks or switches. Dimable with a PWM driver doesn't mean it will be compatible with restive loads as quite often you need either a 0-10 or PWM signal.
Leds are a little more complex than your explanation and the above quote is totally wrong as current stays static the only thing that changes is the ratio of the on/off period. In the on period the current should be at the rating of the LED if you expect anything near its rated output.

[HalcyonRichard]

Hi,
    I think there are two main factors with the lights one is Colour rendering index and the other is shading i.e. you need multiple light sources to prevent shading effects. Placing the small components comes with practice. I have found tweezers and reading glasses work fine. It's strange but you have to be in the right mood and breath "propery" if that makes sense - you need to be in the zone.

Out of interest in the current electronics magazine they are looking at laser lighting systems. They use 4 colours to produce a warm light. In tests people preferred the laser lights above standard halogens and incandescants. So maybe the Turin light test for non halogens will be passed soon.

Regards Richard

[stuartiannaylor]

There is another thing about colour and thats just because we have become used to the colour index of incandescants. Actually the cooler white LED colours are much nearer the daylight spectrum.

I think eventually we will get that LEDs are not incandescants in operation, colour and distribution. Its strange that the majority of LED lighting follows retro-fit methods whilst its a completely different beast altogether.

I made some LED fixtures out of some Luxeon K2s and fitted them into a room. It was very strange as it sort of felt dark but you could see everything with more clarity.
I guess this is something to do with color sensitivity of the human eye. I quite liked it but most of my friends instantly made comments like its weird. No it wasn't it was just different but that difference was to much for many.

I found it was great for reading books and sh*** for watching TV, I thought it was brilliant  

[SimonHobson]

Sorry but I got confused at the above quote. As you don't alter the current. They are only efficient at there rated current.
You could say you alter the current over time, but actually PWM is turning off and on the LED at different mark/space ratios and the current whilst on remains the same.

You need to differentiate between "LED" meaning the actual light producing device, and "LED" meaning the same light producing device packaged up with some electronics etc.

An LED chips is very dimmable and is a CURRENT controlled device - IIRC light output is roughly proportional to forward current, while the voltage-current and voltage-output curves are very non-linear. That's why all quality LED systems control the chip current and let the voltage be determined by the chip.
To give some idea, I put up some LED lighting in the server room at work. This is the stuff where it comes in long strips, with many LED chips in groups with a resistor for each group. Feed it with 12V and it lights at roughly it's design brightness. I had a good play with a variable power supply (variable voltage/variable current). Varying the voltage was tricky - almost nothing happened until about 7 volts, and the brightness goes up rapidly. However, setting the voltage limit to 12V (so as to prevent over-driving and destroying the LEDs) and varying the current limit, that's a different story - brightness easily variable from nothing up to full by setting the current limit from 0 up to full.

There are two reasons why LEDs tend to be dimmed by PWM - one is colour rendering (the colour varies with temperature and current), one is ease (in many ways, a PWM device is easier to build and control - especially if the control source is digital (eg a microcontroller)). When using PWM, the effective current is the arithmetic mean of the actual current over time, and the light output follows suit - ie if you run at 25% duty cycle, the effective current is 25% of the on current and brightness will be close to 25% of full output.

Of course, what gets presented to the outside terminal once an LED has been packaged up with some electronics, a housing, possibly some optics, and so on is different. Normally that ends up being voltage controlled for backwards compatibility, and for much of what's under discussion here it will be some approximation of an existing standard package (eg replacement for a 12V A halogen bulb). That may or may not be dimmable with a standard dimmer and that depends on how the electronics have been designed.

A log of "cheap" 12V "LED Bulbs" simply use a rectifier (to get DC) and resistor to set the current. For these, both PWM and voltage dimming will work (though the latter will behave quite differently to incandescent in dimmer setting vs brightness curve).

Better quality devices will have active control (ie a small controller chip) and control the LED current over a range of input voltages. These probably won't work properly with either type of dimmer.