Light bulb query

[JohnW2]

I didn't say anything.

OK, you 'asked' rather than 'said' - but I was trying to be diplomatic (at least, trying to avoid criticism!) by avoiding just saying that I had already answered your question!

I was asking because I wasn't sure if the effect was the same as a strobe when the light didn't appear to be flickering.

Yes, I realised that, and (assuming I understood your question correctly) in that previous post I had indicated that the answer is 'yes' - by pointing out that you can/will see stroboscopic effects even when the frequency of fluctuation of the light source is far to high for a human being to perceive the light from that source to be 'flickering' ....

... continuing the example figure I gave, if the light source were 'flashing' at, say 1000 Hz (obviously far to high a frequency to be perceived as flickering) and illuminating a rotating object, that object would appear stationary if its speed of rotation was some exact multiple (or 'sub-multiple' of 1000 revolutions per second.

Kind Regards, John

 

[JohnW2]

I'm not too well versed in the operation of LED/LCD TV's either, from the little I have picked up I believe the persistence is electronically generated, not least of which to maintain brightness/contrast.

I'm not sure what you mean by 'electronically-generated persistence'. If you mean that each, say, LED element remains illuminated until it changed to what was needed for the next frame, that is what I had always assumed, and the rate at which that happens is the 'refresh rate' (which, in turn, determines whether there is any perceived 'flicker').

It has never occurred to me that there could be engineered persistence which overlaps with the next frame, and I somehow doubt that is done.

Kind Regards, John

 

[JohnW2]

Interesting question about white LEDs - something I've never thought about. I've just had a quick look around and (as one would expect) the answer seems to be 'yes' - for example, one of the first hits was the design for a DIY stroboscope which uses white LEDs (click here), to which an editorial note has been added which says ...

At switching frequencies above 100 Hz (6000 rpm) the persistence of the phosphor of the white LEDs will make the circuit ineffective.

However, I haven't yet found any particularly authoritative information about persistence of the phosphors used in white LEDs

 

[SUNRAY]

Interesting question about white LEDs - something I've never thought about. I've just had a quick look around and (as one would expect) the answer seems to be 'yes' - for example, one of the first hits was the design for a DIY stroboscope which uses white LEDs (click here), to which an editorial note has been added which says ...

However, I haven't yet found any particularly authoritative information about persistence of the phosphors used in white LEDs

Ah white LED's were not available so the solution was to use a colour (usually blue) to fluoresce a coating, just like a fluorescent tube. The result was a tinted white light. I found it difficult to wire panels under LED lights and wire colours were hard to identify. Things have changed now.

 

[JohnW2]

Ah white LED's were not available so the solution was to use a colour (usually blue) to fluoresce a coating, just like a fluorescent tube. The result was a tinted white light. I found it difficult to wire panels under LED lights and wire colours were hard to identify. Things have changed now.

I don't things have changed, other than that the (very fancy) 'phosphor' is, I believe, within the LED rather than a coating.

The alternative approach of trying to 'synthesise' white light with R, G and B ones has never been very successful.

Kind Regards, John

 

[SUNRAY]

I'm not sure what you mean by 'electronically-generated persistence'. If you mean that each, say, LED element remains illuminated until it changed to what was needed for the next frame, that is what I had always assumed.

This is my assumption too and I described is as electronically generated persistence as with a CRT a pixel would have only been illuminated at the point of scan, if that was applied to LED without any physical persistence it would be a very dim image and possibly very flickery.

It has never occurred to me that there could be engineered persistence which overlaps with the next frame, and I somehow doubt that is done.

Kind Regards, John

Me neither and I didn't mean to infer it is done.

 

[JohnW2]

This is my assumption too and I described is as electronically generated persistence as with a CRT a pixel would have only been illuminated at the point of scan, if that was applied to LED without any physical persistence it would be a very dim image and possibly very flickery.

Yes. but since an LED screen does not involve physical 'scanning' in the same sense as does an electron beam in a CRT, I see no reason why it cannot be arranged ('electronically') that the output of each LED element remains unchanged for all of a frame. I think (but am not certain) that LED/LCD 'elements' on a screen are individually addressable - so, to use a disk drive analogy, one can have 'random access', rather than the 'sequential access' which is inevitable with the physically moving electron beam of a CRT.

That, in itself, obviously would not necessarily avoid flicker (assuming that the refresh rate was low enough for flicker to be perceived by humans) since, although LED element output remained unchanged for all of an individual frame, there's no law saying that it could not alternate between being 'on' and 'off' with alternate frames.

Kind Regards, John

 

[JohnW2]

Me neither and I didn't mean to infer it is done.

Fair enough - but in the case of CRTs, I presume that there was at least some frame-to-frame overlap of phosphor persistence, since (given manufacturing tolerances and other sources of variability) that would be the only way of ensuring that the persistence was always going to be long enough to do the desired job.

Kind Regards, John

 

[ban-all-sheds]

Nothing to do with flickering and persistence per se, but this is an interesting example of how "problems" can arise because of the way we perceive things.

 

[ban-all-sheds]

Or this:

Pick any point on the "spiral" and trace it with your mouse pointer - you'll find that it isn't actually a spiral.

 

[ericmark]

We are looking at replacements for a fluorscent tube, since a fluorescent tube with a electronic ballast will give around 95 lumen per watt, the LED replacement has to be better, so has to look at least 100 lumen per watt or simply no point in making LED replacements for fluorescent tubes, to get 100 lumen per watt you are not going to use a simple capacitor to limit current, it will need a pulse width modulated or switch mode power supply.

OK the PWM or SMPS can clearly produce pulses, but it will not be linked to 50 Hz, although a LED bulb may have a 100 Hz output, it is unlikely a LED tube will. It will likely have a strobe effect, however at what speed this would show its self is another question, and since HF fluorescent is permitted because the frequancy is so high, why should an LED tube not also be allowed.

 

[JohnW2]

... to get 100 lumen per watt you are not going to use a simple capacitor to limit current, it will need a pulse width modulated or switch mode power supply.

Whilst you're probably right about the PSU, could you perhaps explain that statement a bit?

Kind Regards, John

 

[ericmark]

If you look at a LED "bulb" the lumen per watts vary around 75 to 85 for domestic use, look at DC versions for boats and caravans and they have around 100 lumen per watt and also a massive voltage range, typical 10 - 30 volt, to get that range clearly it has to use some form of switch mode power supply.

With the fluorescent tube replacement you can if you wish leave the magnetic ballast in place, or feed the tube directly with 230 volt, the ballast for a 58 watt tube will reduce the voltage across the tube to around 100 volt once fired up, the LED tube is often a reduced wattage so will not need to run on 100 volt but it will be running at somewhere around 150 volt so again these tubes although they do not state it must be able to run with a massive voltage range, again pointing to the use of a switch mode power supply, also the lumen per watt is again around 100 so both things combined it seems unlikely it could be anything other than a switched mode power supply.

We are likely still looking at a pulsed supply to the LED's so it could still produce a strobe effect, however likely in the kHz range and with multi tubes it is unlikely they would synchronise with each other so like the HF fluorescent there is no need to split lighting across phases.

I do question the use of LED tubes, as today we must use electronic ballasts which in turn means longer tube life and more lumens per watt, it is hard to get an accurate figure as the tube lumen is still quoted using magnet ballast and the electronic ballast make claims as to how much extra light out put and also how much drop in watts consumed which varies, but it would seem with an electronic ballast looking at 90 - 95 lumen per watt. Which is only just under the LED lumen per watt, and the quoted life for both are around the same, however the LED I fitted had a much lower life to fluorescent. So the only gain is when lighting corridors where you need the spread of light but not the lumens so reducing the lumen output will not matter, although there are replacements for 58W fluorescent which are a similar wattage, most are around half the wattage and half the output typical 2400 lumen against 5400 lumen for fluorescent tube with electronic ballast, and with an electronic ballast it only draws around 56 watt. Since a LED tube is over four times the price of a fluorescent it simply does not make sense to use them. OK the electronic ballasts do have a limited life, specially if failed tubes are not changed quickly, but economics wise fluorescent is still king for long tubes.

This is not reflected however for folded or coiled fluorescent tubes, when made into the shape of a bulb the fluorescent was rather poor both life time and lumen output, and the electronic control gear is in the main built into the bulb, so they are expensive, there are some exceptions like the 2D but I have now replaced nearly all my bulbs with LED, which asks the question what do I do with the boxes of unused tungsten bulbs. In the main could not use even if I wanted to, as new fittings are mainly SES and old bulbs were BA22d.

 

[JohnW2]

If you look at a LED "bulb" the lumen per watts vary around 75 to 85 for domestic use, look at DC versions for boats and caravans and they have around 100 lumen per watt and also a massive voltage range, typical 10 - 30 volt, to get that range clearly it has to use some form of switch mode power supply.

Well, yes, but we were talking about 230V AC ones with a narrow voltage range, not DC ones with a wide voltage rane.

Are you saying that it is the use of a SMPSU which results in one getting 100 lumens/watt, rather than 75-85?

Kind Regards, John

 

[ericmark]

Yes, with an LED you will typically get between 100 and 120 lumen per watt, but the control gear also uses power, and the LED will only give maxium output if the current is spot on, so if you build a bulb you have a number of things which bring down the lumen per watt, first is you need some current to work items like dimming switches, without it causing the bulb to glow dim or flash every so often, so there is always a bleed resistor, this is why larger bulbs show a better lumen per watt.

The next is they have to be able to run at 254 volt so using a simple capacitor means at 230 volt it is running under voltage so again another reason for low lumen per watt.

I am sure there are other factors as well, but when I bought a 7 watt tube replacement for my quartz security lamp it was marked 85 - 250 volt and it was rated at 100 lumen per watt, but 7 watt bulb works out at 85 lumen per watt, so it seems clear SMPS means better lumen per watt.