Safe Speed Forums

'Linear' in context: the persisting image superimposes atop of what is being received at the time, the contribution being in proportion to the brightness of what caused the persisting image, regardless of the real time brightness received.

TBH, I’m not an expert in eye physiology so I can’t give you a definitive reason why strobing doesn’t help (beyond flicker frequencies), but I am an expert at testing for such characteristics.
When strobing at frequencies beyond that of noticeable flicker (with equal average current), any perceived extra brightness resulting is small and likely to be a result of wavelength shift because of the higher drive current (you can actually see the wavelength shift with green and blue LEDs). The subsequent higher voltage across the LED (at the high current) roughly cancels any increase of photometric efficiency, so the overall perceived power efficiency remains more or less the same.

Ahhh. So you're saying that high speed switching / duty cycle strategies DON'T provide for power saving at equal perceived brightness?

That IS interesting. It's something that I'd always accepted as true, but not something I've ever actually tested.

_________________
Paul Smith
Our scrap speed cameras petition got over 28,000 sigs
The Safe Speed campaign demands a return to intelligent road safety

I have not noticed this effect with car lights but there is a section of the A12 that has electronic cat's eyes that stobe. I find my eyes getting locked back onto them as they move into my peripheral vision.

I had started my testing on the premise that the persistence effect was significant, but it really isn't. I've done very little testing at noticeable flicker frequencies, I might have a go at this later.

My previous testing (which I have revisited just for you) was done on various LEDs (including Cree XLamps), at various average currents, at frequencies from 60Hz to 100kHz, to duty cycles down to 5%. There was no point going beyond 5% as the peak current would be beyond the linear region of the intensity/current transfer characteristic for our application.

I realise this flies in the face of what is expected, but all us SSers know we should not accept what self-proclaimed experts (especially 'so-called' expert witnesses) state - unfortunately that would include me


This leaves the question as to why such LED clusters are strobed at these low frequencies. I’m guessing the designers wanted to make them more attention grabbing, which they are, but that’s not much use if you end up focussing upon them wondering ‘why does that look so odd?’

Abso-bloody-lutely! Thanks mate, never heard of it before, but it does explain a lot of peculiar phenomena experienced by trained observers such as I am party to.

_________________
Aquila



Licat volare si super tergum aquila volat...

Actually it's just bloody wrong, isn't it?

In a voltage regulated system there would be big power savings in the regulation process, but in a current regulated system that doesn't work (the devices aren't ohmic; Vf is effectively a constant). It seems to me that the assumption of regulation power saving has been carried over into an area where it doesn't work.

There may also be the false idea (which I shared) that power could be saved for equal perceived brightness above flicker frequency.

_________________
Paul Smith
Our scrap speed cameras petition got over 28,000 sigs
The Safe Speed campaign demands a return to intelligent road safety

I think there are switchers in the system getting you nearly to the right voltage to recover efficiency.

If you use a switch without a transformer you'll need a charge pump to get to 'nearly the right voltage', and if you use a charge pump the flicker is gone.

Is anyone really going to specify in an expensive wound component to save a few watts in a vehicle?

Anyway, you can get somewhere close to the 'right voltage' by putting LED groups in series strings across the supply with one current limiting resistor for each series string. This strategy reduces component count as well as power.

_________________
Paul Smith
Our scrap speed cameras petition got over 28,000 sigs
The Safe Speed campaign demands a return to intelligent road safety

There is that. The effort needed to save those one or two watts when you've already saved so many by replacing the extremely inefficient incandescent lamps, used with a colour absorption filter, when so much electrical power is available almost for free, seems pointless. A cheap FET/resistor PWM circuit is the simplest method, but a switched current source would offer a more controlled solution (input voltage/temperature variation).

A switcher may be ideally suited for those LED beacon posts (these also visibly strobe) you see at the side of some country roads at night – assuming they are solar/battery powered.

I just happened to notice this:

http://www.osram.com/products/led-syste ... light.html

_________________
Andrew.

I too suffer from being able to see strobed LEDs (I also can't work on a monitor set to anything below 75Hz, and own a 100Hz TV)

For those who don't suffer from the problem, here are some experiments to try to demonstrate how annoying it is.

Set your computer monitor to 50 or 60 hz. Sit at a right angle to it and look at a point on the wall such that the monitor display is just visible in the corner of your eye. You should notice the flicker.

Drive through the variable speed limit section of the M42 when the opposite carriageway has VSL active. Try to reach a speed where your car is being vibrated by the road (it helps if you drive a complete shed) then observe the VSL signs in your wing mirror. You should see the white speed sign bouncing or swirling about inside the red circle since the white is in a different phase to the red. Try not to rear end people when doing this.

I hadn't thought about the epilepsy issue. There are plenty of people out there with undiagnosed photosensitive epilepsy. This is well documented what with games consoles in the 90s and the infamous Pokemon episode in Japan that set off a significant percentage of the population who were previously undiagnosed (and then again in the evening when the offending clip was broadcast on the national news *doh*). It's a good point that these lights could also trigger an attack in a previously undiagnosed epileptic. I think the last think you want is for you to slam on the guy behind to have a fit!

Has anyone bothered to raise this issue anywhere? It's something that really ought to be done before the technology becomes more widespread.

OR - get a digital camera with video mode - take a picture (in video mode 0 of your screen - and see how much the screen flickers - my son did it and i could not believe the results

I'm pretty sure that photo-sensitive epilepsy trigger frequencies are in the range 3Hz to 15Hz, centred around 7Hz.

This seems to be a long way from the sorts of frequencies in use on the roads.

Unless anyone knows different?

_________________
Paul Smith
Our scrap speed cameras petition got over 28,000 sigs
The Safe Speed campaign demands a return to intelligent road safety

That one is down to the video camera sampling at a different rate to what the screen is refreshing at.

My one about looking at the screen through the corner of your eye has more useful general purposes applications, especially if you value your eyesight.

Basically if you can see the flicker when doing this it means that when using it normally it will be causing eyestrain even if you haven't noticed a visible flicker. Therefore you should increase your refresh rate.

Of course none of this applies to LCDs, leave them at 60hz. The pixels stay on between flashes.

Out of interest, is there any mileage in using backlit LCDs for car lamps, you could use a standard bulb and use the LED to toggle if it's visible or not, thus the bulb would last a long time and you'd get the instant on benefit of LED.


As for the frequency ranges for epilepsy, I honestly don't know. I'm fairly sure the offending Pokemon episode triggered it by alternating red/cyan between frames which would put it at 29.97Hz (NTSC)

Edit: Wikepedia claims it was 12Hz for 6 seconds (and 650 children hospitalised), though I'm not quite sure how you manage to get a 12Hz flash out of an NTSC TV. I'm guessing it was more like 14.985.

You are assuming that the Red and Cyan were displayed for an equal number of frames per cycle - in order to get a flashing/strobe effect, this is unlikely to be the case, with the "flash" colour probably displayed for a lower number of frames.

_________________
Too much of our money is being wasted on supposed road safety improvements that are actually making driving more dangerous.

I was reading an article recently, though I can't remember what in as I think it was a newspaper in the doctors waiting room, anyway, apparently the LED cats eyes that are becoming increasingly common have caused at least one epileptic fit in someone who was previously consisdered unepileptic enough to drive safely, one council has stopped installing them until further research is done.

I can't see why LED taillights would be much different.

Here's a link to a relevant thread

I had the misfortune tonight to follow a car with LED red lights, and where the council has replaced the cats eyes with the LED ones.

I had to pull over as I started to get disorientated and a headache coming on. I felt, I supposed the best description is drunk and dizzy. I thought I was going to be sick.

I followed a car a few days prior with LEDs and felt the same, though not as bad, and just thought at the time I was coming down with a cold, but now I am worried about the increase in usage of these devices!

On a long journey over the weekend I got to see my first 'disco' cats eyes and stuck in slow traffic got the chance to ponder these LED rearlights.

To me the cats eye only 'strobe" when I'm right on top of them, like at a car's length, which is a fat lot of use as the road was NSL and at that speed they'd have grabbed your attention too late - I take it that's the thinking behind them. I also found them a bit distracting but I guess when they're all like that I'll get used to it - or more likely have not driven in years! BTW it there a limit on how long the stretches of these thing can be - the ones I saw were marking a hatched area at a junction.

As for the LED rearlights I got the impression I was drawn to them - amongst all the other rearlights - because they aren't a single big block of colour but lots of individual ones in a pattern and eyes get drawn to patterns. I couldn't see the strobe effect, or feel any of the more extreme sensations some seem to get but they definately do pull you in.

Anything that is a light source at night that distracts or takes your attention away from the road ain't good. But i very much doubt anyone who has accident will ever be able to prove that they were unduely diverted from the task at hand. It ironic really given the endless numbers of attempts to make us drive more carefully that devices are being introduced that potentially reduce safety.

Are these things never tested in real situations with real people? And what is the criteria for rejecting something, anything more than a few percent declaring a problem should see it shelved.

Barkstar

_________________
The difference between intelligence and stupidity is that intelligence has limits.