Hmm, so what makes leds have an issue with higher frequencies then, I'd have
thought this was of no concequence , but maybe its to do with the individual
colours of led in the package having different levels on higher
frequencies.

Brian

--
From the Sofa of Brian Gaff Reply address is active
"Arfa Daily" wrote in message
...


"Colin Stamp" wrote in message
o.uk...
Hi all,

I tried out some Ikea LED candle bulbs today, wondering if 90 lumens
would be enough:-

http://www.ikea.com/gb/en/catalog/products/30217927/

I thought I was impressed. 90 lumens turned out to be a little bit on
the dim side, but they look quite reasonable in the fittings, which were
intended for halogen candles. The light comes out in about the right
places IYSWIM, The colour is fine and they're cheap.

Then came the realization. The sodding things strobe really badly. Now
I've noticed it, I can't get away from it. Anything that moves in the
room turns into bloody set of freeze-frames, as does the whole room
every time you move your eyes!

Is this a common thing with cheap LED bulbs? I'm all worried about the
impending Ebay purchase of Chinese replacements now. The only other LED
bulbs I've got are two brands of GU10 and they're fine, so maybe it's
just Ikea ones that are crap...

Cheers,

Colin.


The way to get the maximum brightness and longest life from LEDs, is to
pulse drive them. They're actually not very good at handling high
frequencies, so they tend to get driven at fairly low pulse rates, and
that's why you see the strobing. It's been driving me mad for years with
LED car tail lights where the situation is even worse because they use the
fact that they are pulse driving them, to open up the mark-space ratio to
something like 50% when they are a tail light. This gives the impression
of half brightness. When the car owner treads on the brakes, the pulse
drive tightens right up to give high brightness, but the strobe effect
becomes less noticeable. Those lights that they stick on the top of
traffic cones are also LED based, and also strobe for the same reason.
Possibly, the better quality LED lamps like your GU10s, use better quality
LEDs that can be driven at a higher pulse rate, or maybe, they are DC
driving them ?

Arfa

On 27/02/13 09:19, Brian Gaff wrote:
Hmm, so what makes leds have an issue with higher frequencies then, I'd have
thought this was of no concequence , but maybe its to do with the individual
colours of led in the package having different levels on higher
frequencies.

Brian

They are so bad that they are de facto light source for infrared remote
controllers operating at tens or hundreds of khz..and for driving
optical fibres at GHZ rates.

Man, that's really bad.
..



--
Ineptocracy

(in-ep-toc-ra-cy) €“ a system of government where the least capable to
lead are elected by the least capable of producing, and where the
members of society least likely to sustain themselves or succeed, are
rewarded with goods and services paid for by the confiscated wealth of a
diminishing number of producers.

On Wednesday 27 February 2013 09:19 Brian Gaff wrote in uk.d-i-y:

Hmm, so what makes leds have an issue with higher frequencies then, I'd
have thought this was of no concequence , but maybe its to do with the
individual
colours of led in the package having different levels on higher
frequencies.

Brian


I've driven LEDs (very high brightness types but not full on mental Cree
types) at upto about 10kHz with no apparantly ill effects. At that frequency
there will be no strobe effects that a human could detect.
--
Tim Watts Personal Blog: http://squiddy.blog.dionic.net/

http://www.sensorly.com/ Crowd mapping of 2G/3G/4G mobile signal coverage

Reading this on the web? See:
http://wiki.diyfaq.org.uk/index.php?title=Usenet

On Wednesday, February 27, 2013 9:40:25 AM UTC, The Natural Philosopher wrote:
On 27/02/13 09:19, Brian Gaff wrote:

Hmm, so what makes leds have an issue with higher frequencies then, I'd have
thought this was of no concequence , but maybe its to do with the individual
colours of led in the package having different levels on higher
frequencies.

Brian

They are so bad that they are de facto light source for infrared remote
controllers operating at tens or hundreds of khz..and for driving
optical fibres at GHZ rates.

Man, that's really bad.

It's a long while since I was using LEDs to drive optical fibre links, but I recall that they ran out of steam at around 10MHz for most types, with some specialized ones managing up to 100MHz. Almost all the links I've looked at since (and everything at above 100 Mbit/s) use lasers, and a high proportion of those use a separate (even if possibly integrated in the same package) modulator.
Visible LEDs as I recall were something like an order of magnitude slower than infra-red - but I'd agree that this is still more than fast enough not to be visible.

On Wednesday, February 27, 2013 4:02:19 PM UTC, Tim Streater wrote:
In article ,

wrote:

It's a long while since I was using LEDs to drive optical fibre links, but I
recall that they ran out of steam at around 10MHz for most types, with some
specialized ones managing up to 100MHz. Almost all the links I've looked at
since (and everything at above 100 Mbit/s) use lasers, and a high proportion
of those use a separate (even if possibly integrated in the same package)
modulator.

I have this recollection that the short range optic modules for Juniper
routers (over multi-mode fibre) used IR LEDs at 850 nm. These would be
driven at any speed up to 10Gbps. The longer range modules (up to 60km
or so of single-mode fibre) were all lasers at 1310 or 1550 nm.

Are you sure they were LEDs? I'd have expected them to be VCSELs (vertical cavity surface emitting lasers); I've seen a lot of gigabit ethernet modules for short distances that use 850nm VCSELs.
Mike

In article ,
Tim Streater wrote:

I have this recollection that the short range optic modules for Juniper
routers (over multi-mode fibre) used IR LEDs at 850 nm. These would be
driven at any speed up to 10Gbps. The longer range modules (up to 60km
or so of single-mode fibre) were all lasers at 1310 or 1550 nm.

Yep, the cisco optics I was spec'ing today were 850nm for short range,
1550nm for the 40km ones.

Eye watering prices...

Darren

In article ,
Tim Streater wrote:

Yep, the cisco optics I was spec'ing today were 850nm for short range,
1550nm for the 40km ones.

Eye watering prices...

Well quite :-)

But upthread there seemed to be an assertion that you couldn't drive
LEDs at the sort of frequencies I mentioned. So I'm not sure whether
that's just wrong or that I missed something ...


Ah, no idea what's in them. Something very valuable (or maybe that's the
special expensive Cisco sticker)

http://www.cisco.com/en/US/prod/coll...0801bc698.html

are the things I've been using. DS-SFP-FC8G-ER to be precise. No clue on that
page that I can see as to what they actually are...

Darren

On Wednesday, February 27, 2013 10:24:18 PM UTC, Tim Streater wrote:
In article , dmc@puffin. (D.M.Chapman)

wrote:
snip

http://www.cisco.com/en/US/prod/coll...0801bc698.html



are the things I've been using. DS-SFP-FC8G-ER to be precise. No clue on that
page that I can see as to what they actually are...

I think the sticker on them should warn you that's it's a laser device.
I can't remember if the short-range ones actually say LED on them or
merely that they *don't* say laser.

I think you will find that they do say 'Class 1 Laser Product' somewhere on them - e.g http://www.ciscotransceivers.com/ds-sfp-fc8g-er (described as 100% Cisco compatible, though not an actual Cisco branded product)
If it's a Class 1 laser it doesn't need the yellow laser warning triangular sticker.

On Wednesday, February 27, 2013 11:36:34 PM UTC, wrote:

I think you will find that they do say 'Class 1 Laser Product' somewhere on them - e.g http://www.ciscotransceivers.com/ds-sfp-fc8g-er (described as 100% Cisco compatible, though not an actual Cisco branded product)

If it's a Class 1 laser it doesn't need the yellow laser warning triangular sticker.
Completely forgot to add - that one is a long wavelength module, I just picked it becase it's the one mentioned previously, but if you look at the 850nm ones from the same site they also say Class 1 Laser Product

"Tim Streater" wrote in message
...
In article , dmc@puffin. (D.M.Chapman) wrote:

In article ,
Tim Streater wrote:

I have this recollection that the short range optic modules for Juniper
routers (over multi-mode fibre) used IR LEDs at 850 nm. These would be
driven at any speed up to 10Gbps. The longer range modules (up to 60km
or so of single-mode fibre) were all lasers at 1310 or 1550 nm.

Yep, the cisco optics I was spec'ing today were 850nm for short range,
1550nm for the 40km ones.

Eye watering prices...

Well quite :-)

But upthread there seemed to be an assertion that you couldn't drive LEDs
at the sort of frequencies I mentioned. So I'm not sure whether that's
just wrong or that I missed something ...

--
Tim


It wasn't so much that you couldn't drive *any* LEDs at high frequencies,
but I seemed to remember reading that it was difficult to drive power LEDs
at high frequencies, because of the capacitance of them due to the die size,
and the amount of metal that they were bonded to. However, LED technology is
moving so fast now that it is probably no longer true anyway, if it ever was
and it's not just my poor old memory getting the better of me ... :-)

Arfa

On Thursday, February 28, 2013 2:35:40 AM UTC, Arfa Daily wrote:
snip

It wasn't so much that you couldn't drive *any* LEDs at high frequencies,
but I seemed to remember reading that it was difficult to drive power LEDs
at high frequencies, because of the capacitance of them due to the die size,
and the amount of metal that they were bonded to. However, LED technology is
moving so fast now that it is probably no longer true anyway, if it ever was
and it's not just my poor old memory getting the better of me ... :-)

It really depends what you mean by high frequency ;-)

You're quite correct, that the die size and therefore capacitance of power LEDs will make driving them fast more difficult; there is also a more fundamental limit given by the carrier lifetime, since even if you switch the current off instantaneously the diode will still continue to emit light as the minority carriers left recombine. The speed of recombination will vary depending on the material and doping but figures I've seen suggest somewhere around a 10ns minimum lifetime.
In a laser, of course, where the majority of the light comes from stimulated emission, the carrier density will drop much faster when the supply of new carriers is turned off - you're not waiting for spontaneous recombination of the carriers.

In article ,
wrote:
On Wednesday, February 27, 2013 11:36:34 PM UTC, wrote:

Completely forgot to add - that one is a long wavelength module, I just
picked it becase it's the one mentioned previously, but if you look at
the 850nm ones from the same site they also say Class 1 Laser Product



Yep, you are right. They do say class 1 laser on them. No triangle as you
say. Short(er) haul 850nm ones the same.

Darren