Led lights

wld1783 · May 17, 2011

Hahnmeister said:
Spectrum Shift? I have not experienced this myself. Do you have some more info or a link? What is the frequency range that this would apply to? I tend to stay in the 1kHz range and up so maybe that has something to do with it...???

Wow, Door Bell Wire you say. Thats interesting. I suppose we are talking about currents in the nA range though. One wire with low current is no problem, but if you had to control banks of enable pins with an analog signal, your signal source might not provide enough current. Once we move over to totem-pole/CMOS architecture, now the current required for the dimming input drop a good amount from what I have seen. If you are only doing a few channels on analog, no problem, but having a dozen or so would start to be a problem.

Door Bell Wire It works for me... I measured less than a .5 VDC drop across the wire that length and thats conservative given how I measured it. Once you meet the mA threshold for controlling the driver its the Vdc that count. Apparently the APEX has the out put mA for a long reach. The Reef Keeper is less capable and I have not tested the Profilux though I doubt it would have a problem.

A dozen Drivers would provide a problem I'm only controlling 4 on each on my personal system (out put V1&V2 only 8 drivers total). Every time you add a driver you divide the output current so eventually you will reach the threshold of the drivers not responding due to the lack of current. Of course you are aware of this but I've found practicable testing best for our recommendations. Otherwise we would simply default to the PDF Spec Sheets like the other guys.

As far as PWM.... We do believe it works well and the drivers you mentioned are efficient. However we prefer the 1-10Vdc over PWM for the following two reasons...

-Major Reef Controllers (APEX, Profilux & Reef Keeper) use the 1-10Vdc control over the PWM. Also there are numerous drivers just as efficient using the 1-10Vdc.

-PWM does cause a spectrum shift here is one paste of an article. My apologies as I could not readily find the older research we've come across but a google search should give results after numerous pages... Here is one link...http://vbn.aau.dk/files/39948127/Led_spectral_and_power_characteristics_.pdf

We've documented a difference in the Spectrum of Cree Royal Blue and even provided a specific spectrum of the Cree Royal Blue to the University of Heidelberg for their students physics thesis. They needed our Premium Royal Blue 450-455nm...This spectrum when dimmed under PWM may change.

For us, its all about the specific premium spectrum for the reef and to have PWM dimming change that spectrum makes us very cautious. Not that you will have problems with growth but the looks may not be the best...Please let us know about your research.

Bill

http://vbn.aau.dk/files/39948127/Led_spectral_and_power_characteristics_.pdf

LED SPECTRAL AND POWER CHARACTERISTICS UNDER HYBRID
PWM/AM DIMMING STRATEGY

Hahnmeister · May 22, 2011

This is the place:
LED Group Buy | CREE LEDs for DIY Aquarium Lighting

Milad even has optics for the XM-L's now, and he gets the coolest color bins. Right now the blue, royal blues, etc... are on sale.

I buy or build my own buck regulators to convert from a DC power supply, sometimes I opt for these...
3023 Wired BuckPuck - LED Supply.com

A009 BuckBlock? - LED Supply.com

And run it all off of a meanwell HRPG-300, 450, or 600 at 24volts.
HRPG-300-24 Mean Well Linear & Switching Power Supplies

The heatsink is best purchased from here:
Extruded Aluminum Heatsinks from HeatSinkUSA

Hahnmeister · May 22, 2011

wld1783 said:
Door Bell Wire It works for me... I measured less than a .5 VDC drop across the wire that length and thats conservative given how I measured it. Once you meet the mA threshold for controlling the driver its the Vdc that count. Apparently the APEX has the out put mA for a long reach. The Reef Keeper is less capable and I have not tested the Profilux though I doubt it would have a problem.

A dozen Drivers would provide a problem I'm only controlling 4 on each on my personal system (out put V1&V2 only 8 drivers total). Every time you add a driver you divide the output current so eventually you will reach the threshold of the drivers not responding due to the lack of current. Of course you are aware of this but I've found practicable testing best for our recommendations. Otherwise we would simply default to the PDF Spec Sheets like the other guys.

As far as PWM.... We do believe it works well and the drivers you mentioned are efficient. However we prefer the 1-10Vdc over PWM for the following two reasons...

-Major Reef Controllers (APEX, Profilux & Reef Keeper) use the 1-10Vdc control over the PWM. Also there are numerous drivers just as efficient using the 1-10Vdc.

-PWM does cause a spectrum shift here is one paste of an article. My apologies as I could not readily find the older research we've come across but a google search should give results after numerous pages... Here is one link...http://vbn.aau.dk/files/39948127/Led_spectral_and_power_characteristics_.pdf

We've documented a difference in the Spectrum of Cree Royal Blue and even provided a specific spectrum of the Cree Royal Blue to the University of Heidelberg for their students physics thesis. They needed our Premium Royal Blue 450-455nm...This spectrum when dimmed under PWM may change.

For us, its all about the specific premium spectrum for the reef and to have PWM dimming change that spectrum makes us very cautious. Not that you will have problems with growth but the looks may not be the best...Please let us know about your research.

Bill

http://vbn.aau.dk/files/39948127/Led_spectral_and_power_characteristics_.pdf

LED SPECTRAL AND POWER CHARACTERISTICS UNDER HYBRID
PWM/AM DIMMING STRATEGY

I saw that article as well as a couple others in the IEEE library. One for public viewing is also: http://www.lightingresearch.org/programs/solidstate/pdf/Gu-SPIE6337-17.pdf

The thing is... the spectrum shift with the bluer LED's is more towards the shorter (bluer) wavelengths than the warmer, which would actually be a good thing, and it looks like its only one nano-meter of wavelength shift with PWM anyways. The constant current dimmers seem to have more of a shift, and into warmer frequencies.

There is another thing to consider:

We must seperate PWM dimming from PWM drivers. Just because you use PWM dimming, chances are that the driver is still putting out an amplitude modulated signal. So even though you are feeding a PWM signal into say, the EN pin on a National Semi LM3409, the output is not PWM... its smoothed out with an output inductor and maybe a shunt cap on the output. The pFET may not even be turned on and off all the way on the output, but rather have its output varied in some analog range. Most driver chips out there dont seem to actually have PWM at the output... its just the dimming signal. The only way you could get an actual PWM dimmer signal resulting in a PWM output is to wire the chip up as stock and add an external p-channel MosFET on the output of the driver (after inductor) to be fed with and act as the PWM control. Why anyone would opt for this though is beyond me...

But yeah... PWM dimming =/= PWM LED driving.

wld1783 · May 22, 2011

Hahnmeister

Thanks for the PDF, caught a couple of interesting things but with only a few pages will have to check the sited sources for more detail. Again thanks its a great lead to more knowledge.

Just remember LEDS can "be like a box of chocolates...you never know what your going to get." Its best to have the bins labeled on the stars to know for sure your getting the premium spectrum as they do cost more.

Bill

Hahnmeister · May 29, 2011

Just curious, do you have any means of viewing IEEE library articles, there are a good deal in there about spectrum shifting LED's. I could send you pdf's direct otherwise.

wld1783 · May 29, 2011

Hahnmeister

That would be outstanding. The primary reason I'm hesitant to fully embrace PWM is the spectrum shift. If you could send the PDFs to [email protected]. We are working on a PDF Reference Page and would like to list as much info on it as possible. Who knows you might change our preference

Hahnmeister · Jun 1, 2011

But once again, what you are talking about is where the output is a PWM signal. Many drivers dont even follow this architecture since any pulsing they do is smoothed out with an inductor at the output or a capacitor in series with the LED's. So the likelyhood of actually finding a driver that actually has a PWM driven current to the LED's is low. I think its because of efficiency reasons. Even in those articles we have looked at so far... the PWM is concluded to be less efficient when it comes to lumen output. Makes sense to me I suppose. I guess that the discontinuous current means the voltage can drop below the threshold voltage of the diode, resulting in part of the voltage being applied to the output resulting in NO light when the voltage drops low enough, of course hurting the average output as well.

Ill send the pdf's in the next few days.