Luxeons and low-light detection: Turner, Groth & Long (2

Links to relevant web sites, copies of patent documents and newspaper and magazine articles. Feel free to post relevant items here. This forum is moderated and non-relevant postings will be deleted.

Moderators: Steve Anderson, Dave Moll, Andrew Davie

Luxeons and low-light detection: Turner, Groth & Long (2

Postby chris_vk3aml » Tue Jul 08, 2008 2:05 am

The .pdf text attached here was written by Clint Turner (KA7OEI), Michael Groth (VK7MJ) and myself (Chris Long VK3AML). The title "Dollars Versus Decibels" was chosen to describe the way in which we had to balance costs against communication system benefits on the design of a long-distance (150 km+) atmospheric optical communication outfit. The paper was presented at the San Jose (Silicon Valley, California) conference "Photonics West 2008" by Chris Long VK3AML on 21 January 2008.

The most important conclusion that we reached was that non-coherent beams present considerable practical advantages over coherent (laser) beams for long-distance optical comms. work. The reasons for this - principally related to atmospheric turbulence and the loss of wave-front coherence - are discussed at length in the .pdf paper, which was published by the SPIE (Society Of Photometric Instrumentation Engineers).

The principles that we use for low level detection and Luxeon modulation could gainfully be applied to NBTV systems using mechanical scanners.

Several parts of Clint Turner's website are particularly pertinent for users of the new high-output Luxeon sources:

(1) Linear Luxeon modulation: ... lator.html

(2) Luxeon overcurrent protection: ... ction.html

(3) Sources for lenses and optical components, esp. mail order over Internet: ... urces.html

In addition, these pages of my own authorship are relevant:

(4) Discussion on advantages of Luxeon sources for modulation: ... uxeon.html

(5) On broadband transimpedance photodiode amplifiers: ... ghtrx.html

We would welcome comments, additions or corrections to these various papers and postings.

- Chris Long VK3AML (for Clint Turner KA7OEI, Mike Groth VK7MJ, and himself).
(764.84 KiB) Downloaded 178 times

Postby Steve Anderson » Tue Jul 08, 2008 4:14 pm


Very interesting! If you recall I put together a PWM schematic for Steve O's Luxeon mechanical colour monitor. Steve chose to use the K2 flavour of Luxeons and within the datasheet it does use the expression, "Driving these high power devices at currents less than the test conditions may produce unpredictable results and may be subject to variation in performance. Pulse width modulation (PWM) is recommended for all dimming effects." Note they use the word 'may' twice.

So I was quite surprised when one of the above links refers to Luxeons being used with linear current modulation from 0-100% with no ill effects. Thinking that the LuxeonIII is a different beast in some way I downloaded the two pdf's for these and they too have a very similar phrase within.

So either you were lucky and/or Philips are being very conservative.

Very useful links.

Steve A.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
Posts: 3800
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Postby chris_vk3aml » Tue Jul 08, 2008 7:30 pm

Thanks for the complements, Steve.

The answer is that the RED (AlInGaP) 'truncated inverted pyramid' chip Luxeon III is a very different device to the WHITE (InGaNP) Luxeon III. The RED device is more linear, much faster in rise time, and it can be driven to much higher current densities than the WHITE Luxeon. Every doubling of current through the WHITE Luxeon, by contrast, produces much less than double the light output. It will therefore reproduce a picture with very low gamma, unless the WHITE Luxeon is placed in a non-linear driving circuit such as the one shown by Klaas Robers here:

All WHITE Luxeons can be dimmed by varying their series current, but the dimming will not be linear (this is probably the reason that Philips recommends the pulse-width dimming approach) and that probably applies to high-output standard WHITE LEDs more than it does to Luxeon sources.

For very long distance (100km+) atmospheric communication, simple amplitude modulation of the Luxeon is, in general, to be preferred to PWM, FM or any other system of modulation requiring broad receiver bandwidths and/or receivers employing Schmitt trigger systems and a subcarrier demodulator. Refer the section of this page of Clint's for an explanation (fairly well down the page) under the heading "Why not FM subcarriers?": ... index.html

The elimination of atmospheric scintillation by optical, rather than electronic means is probably to be preferred. This can be done by increasing the comms beam diameter to straddle several atmospheric turbulence cells and 'aperture average'. It can also be done by using multiple beams and spatial diversity - easily expedited by wiring several transmitter Luxeons in series, and collimating them on the receiver with separate, preferably spaced Fresnel lens collimators. If, however, you're interested in pursuing the PWM approach - although we hardly found it worth the extra complexity - you can refer to Clint Turner's PWM design web page here: ... lator.html

The K2 Luxeon is not quite capable of the same luminous output as the Luxeon Star or Luxeon Emitter. The Luxeon "emitter", lacking the metal-backed board of the "Star" can be crowded more closely onto a heatsink to provide a Luxeon bank for a bright scanning disc receiver source, but a large and efficient heatsink will be needed. These "emitters" should be glued to their backing heatsink with heat-conductive epoxy of the automotive type - I think this is retailed in America under the trade name of "JB Weld" but there are probably similar proprietary heat-conductive epoxy adhesive products available in other countries.

Incidentally, an illustration of the improved performance of a large-area Luxeon beam over that of a laser may be obtained from this 3.7 MByte downloadable .mpeg clip from our San Jose presentation, which shows the instantaneous intensity of the beam-cross section after passing through a 15 km atmospheric path, on a 20-cm diameter optical receiver lens: ... _2a1b1.mpg

At visual optical frequencies, it is impossible to maintain temporally and spatially coherent wave fronts over any appreciable distance through even moderately turbulent air. There is therefore virtually nothing to be gained by using visual lasers with links in excess of a few kms. Coherence and heterodyne detection are possible with IR beams, but the sources, detectors, and the optical heterodyne system is difficult of alignment and VERY expensive. Better - and cheaper - to use noncoherent Luxeon sources with cheap large-aperture molded Fresnel collimators.

I hope this expands on the subject - but be assured that we've found the Luxeons very tolerant of high-current abuse. The only one I ever blew was a result of me placing my multimeter across the Luxeon's current-limiting resistor. Usually I read the voltage across this resistor to calculate the Luxeon current. However I forgot, and connected the multimeter across the resistor while it was switched to its 10A range(! - duuhhhhh!) The red Luxeon III probably withstood a 4 Amp surge before it blew its internal gold connecting wire - it was quite a spectacular flash, though also a depressingly expensive one!

Incidentally, we have found that the RED Luxeon III's can be linearly modulated around a standing current of 1 Amp, so they can be intermittently pushed upwards to about 1.7 Amp with impunity. The WHITE Luxeons are less sturdy, and slower in response.

Best wishes,

Chris Long VK3AML.

Postby Klaas Robers » Tue Jul 08, 2008 9:10 pm

Can it be Chris that white LEDs are non linear and slow because of the fluorescent material that converts part of the blue light into yellow light? Is that process linear also at high light levels?

I wondered always that white LEDs are simply used for NBTV clusters. However some non linearity will not be observed and can easily be compensated for by some more gamma correction.

But lack of speed would be seen immediately by a yellowisch after tail. More or less to my surprise I never read this as reported.
User avatar
Klaas Robers
Frankenstein was my uncle.
Posts: 1431
Joined: Wed Jan 24, 2007 8:42 pm
Location: Valkenswaard, the Netherlands

Postby chris_vk3aml » Tue Jul 08, 2008 9:43 pm

Well, the red, red-orange and amber Luxeons all use the sturdy AlInGaP "TIP" chip. Of these three, the AMBER Luxeon gives (subjectively) the brightest output, because its output wavelength is higher on the eye's colour response curve, but their actual flux output is roughly equal. AMBER would be best for NBTV, for visual purposes. However, the RED Luxeon provides an output closer the the maximum response of silicon photodiodes, and so we use that for communication purposes. The rise time of these AlInGaP devices have been measured to around 10 to 15 nsec.

The quoted rise time for the WHITE Luxeons is greater than 100 nsec (still quite usable for NBTV) which almost certainly is dependent on the white phosphor, but may also be a function of the InGaN chip, which is used for the violet, blue, green and (with phosphor powder over a violet-UV chip) the white Luxeon. I really haven't enough information to definitively indicate the reason for the slower response, or the curved current/light output current characteristic, but the white chip is almost always much larger than the red-orange-amber chip for the same light output.

If anyone can find more definitive data, please let me know!

Chris Long VK3AML.

Postby Steve Anderson » Wed Jul 09, 2008 1:58 pm


Thanks for all that data. I must admit I wasn't aware that the white Luxeons were not as linear as the single colour ones. I have no intention of using them so I just glossed over that data.

The notice in the Philips datasheets implies application of PWM to all flavours not just specifically to the white. Had I known about the linear characteristics you have discovered I wouldn't have gone down the PWM path for Steve O's colour monitor.

Sorry Steve.

Steve A.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
Posts: 3800
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Return to Resources

Who is online

Users browsing this forum: No registered users and 1 guest