LED dimming CCR vs PWM and driver options?

jrpark22000

jrpark22000

Premium member
It's been a couple years since I've started a supplemental LED build. It involved lots of research to achieve the full spectrum to match a 14K halide and match photo synthesis. I've a long thread on RC going into details.

http://www.reefcentral.com/forums/showthread.php?t=2165340
(I'll copy part of the RC thread here for ease of reading and to understand why PWM dimmed is preferred)


It's been a couple years. With all the LED experience here including guys like Bill [MENTION=144]wld1783[/MENTION]

Are there any PWM dimmed drivers that have come onto the market in since?
 
Old 05/02/2012, 02:42 PM
jrpark22000
Registered Member

LED dimming CCR vs PWM and driver options?
One of the last steps before I start my DIY LED accent build is determine what drivers to use. The two common drivers are Meanwell ELN P&D series and Inventronics EUC. These give me the option to use PWM or 0-10 analog control as a dimming input. I’m using an apex so I’ll need analog. After reviewing the spec sheets for the various drivers from both manufacturers both look to use CCR (constant current reduction) method to dim the lights, not PWM (pulse wave modulation.) If possible I’d prefer to prevent a phase shift which I cannot calculate or predict with the available information.

Here are 2 links for info on the 2 types of dimming.
http://www.lutron.com/TechnicalDocumentLibrary/048360a_PWM_vs_CCR_LED_App_Note.pdf &
http://www.waitingforfriday.com/index.php/Controlling_LED_brightness_using_PWM


My question; Is there a ready to use PWM driver which uses an analog dimmer trigger circuit?



And Thank you to GeorgeMonnatJr for helping. He posted 3 blog entries today to help on LEDs, PAR, and general theory.
http://www.reefcentral.com/forums/blog.php?u=274275
 
jrpark22000
Registered Member

Quote:
Originally Posted by C5flyboy View Post
From my research that I did for my build (currently awaiting parts to come to start!) If you are using the Apex as your controller you have to use the D series if you are going with the meanwell drivers which are apparently pretty awesome… from what I found the analogue drivers do not operate using the PWM, only the CCR…Anyone with better and possibly more accurate info please chime in here and correct me where I'm wrong but this is what I came up with from my research…I also found a a lot of folks actually made their own drivers from scratch….

Agreed the Apex puts out a 0-10V analog signal from the VDM ports. I’ve also read the reviews of both meanwell and inventronics, both are highly recommended. You research matches what I’ve read, anything that has an analog trigger input uses CCR method to dim the LEDs. I’ve looked at the PWM driver components. I’ve again ran into the same issue, PWM drivers use a PWM input trigger.

Quote:
Originally Posted by renisel View Post
Could you clarify your rationale for converting a perfectly good analog signal into a PWM signal? What phase shift are you worried about?

I’m not looking to convert signals, I’m looking to maintain the analog trigger but utilize PWM dimming method.

The phase shift is explained in the OP links in detail. There are 2 methods to dim LEDS, by reducing current(CCR) or by rapidly turning the LED on and off using a duty cycle (PWM.) By reducing the current to dim the LED the light spectrum is changed by an unknown value. Using PWM the current is maintained across the LED maintaining the rated and known spectrum.

1) I won’t be running the LEDs at 100%.
2) I use an Apex and would like to control the DIY LEDs with it.
3) I’m augmenting my existing AI setup which I have spectral analysis for and looking for specific frequencies to fill out the my spectrum. I am dimming my AIs and have yet to find what type of dimming they use, although I would guess CCR based on my research. I’ve only read of one manufacturer to use PWM dimming. With my research the variances between each LED, mixing royal blue and blue LEDS, and their density over my tank I believe CCR dimming the AIs for me would still provide a similar spectral analysis as Dr. Sanjay Joshi published. http://www.advancedaquarist.com/2012/3/aafeature

Each manufacturer produces LEDs with different nm range and even within a single manufacturer a given lot of LEDs properties will vary. Add to that 2 more factors, water shift and CCR dimming driver shift. Water shift can be calculated roughly. CCR Driver shift on the other hand varies per LED color type, manufacturer, lot, amperage and age. The only true way to know what I’m giving the coral is to measure the spectrum under the water level, which isn’t an option for numerous reasons. No manufacturers or any studies I’ve read measure LEDs spectrum at anything less than full rated power level, including Dr. Sanjay. An analog triggered PWM LED driver would solve for some of the variables, let me use my Apex, and make it practical to estimate spectral range at a given water depth.

My current options are to not use my Apex, buy a few LEDs for every manufacturer and send them off for testing, or use CCR and guess at what nm ranges I'm providing.
 
Old 05/04/2012, 08:51 AM
jrpark22000
Registered Member

I think I’ve covered the bases on everything but the lights. When you compare the AI spectrum to photosynthesis for chlorophyll A, B C and Beta Carotene, the violet and cyan range are lacking or missing. When the AI are compared to the 250MH spectrum again violet and cyan are lacking along with green and red. I am planning on adding colors from each range on individual dimmer ckts to balance color to match a 250mh to start. I’ll go from there with long term experimentation.


Old 05/04/2012, 09:20 AM
jrpark22000
Registered Member

Agreed on the red range and thanks for the link. There are also other studies which show coral have adapted to the water shift of light and no longer use red, actually it can inhibit coral growth.

For reference some of the photo shop spectra I've compiled during my research in determining what to try next. They are very messy but after enough staring the different lines are visible. I don't have one with cyan in it, but the cyan will cover the hole left by the other LEDs.

references;
Dr. Sanjay for the AI-MH underlay
Each manufacturers quoted 100% current spectral outputs
Numerous internet references for the photosynthesis range


spectragraph of AI SOLs and Photosynthesis
View attachment 7762

spectragraph of AI SOLs, Photosynthesis and Violet LEDs
View attachment 7763

spectragraph of AI SOLs, Photosynthesis, Violet LEDs and whites
View attachment 7764
 
Old 05/24/2012, 09:27 AM #30
jrpark22000
Registered Member

Join Date: Nov 2010
Posts: 592
A deeper search into the color shift turned up a few really indepth links and article. There are some really intense guys in the led flashlight world who spent a lot of time on the color shift in PWM vs CCR.

http://www.candlepowerforums.com/vb/showthread.php?70073-Constant-Current-vs-PWM-dimming-Revealed
http://www.molalla.net/members/leeper/pwm_cc.htm

and

http://www.candlepowerforums.com/vb/showthread.php?216904-PWM-vs-Constant-Current


Attached are a couple of the photos from the links to show the color shift of CCR dimming as compared to PWM.


http://www.molalla.net/members/leeper/bw01pdim.png
View attachment 7765


http://www.molalla.net/members/leeper/bw01cdim.png
View attachment 7766


http://www.molalla.net/members/leeper/bw01c_p.png
View attachment 7767
 
Old 01/09/2013, 07:46 AM #44
jrpark22000
Registered Member


http://www.instrumentsystems.com/fi...ads/Products/LpR10_Instrument_Systems_web.pdf

It’s a good article on the burn in effect and aging of white LEDs in relation to spectral shift. I haven’t been able to find any similar credible research on non-phosphorus dope LEDs. With that I don’t know if narrow spectrum LEDs will share a similar age shift or not.

For an update on my solution; I ended up going with PFC Meanwell drivers instead of the ELN series. They came out with the new series in fixed current for lower wattages and are more efficient.
I am using;
lpf-16D-48 qnty 3 (whites)
lpf-16D-42 qnty 1 (cyan)
lpf-25D-54 qnty 1 (violet)



The visio layout for how the LED build on my system will look.
Green circles - cyan LEDs
Purple circles - Violet LEDs
Red circles - White Tri-star LEDs
View attachment 7769
 
jrpark22000 said:
It's been a couple years since I've started a supplemental LED build. It involved lots of research to achieve the full spectrum to match a 14K halide and match photo synthesis. I've a long thread on RC going into details.

http://www.reefcentral.com/forums/showthread.php?t=2165340
(I'll copy part of the RC thread here for ease of reading and to understand why PWM dimmed is preferred)


It's been a couple years. With all the LED experience here including guys like Bill [MENTION=144]wld1783[/MENTION]

Are there any PWM dimmed drivers that have come onto the market in since?
Click to expand...

I think you hacked into our dedicated research bunker underneath a nondescript building in a nondescript location near a Starbucks...

For starters good if not great SPS growth has been proven with a Cree Cool White, Royal Blue and Blue mix as the Cool White for the most part offers a Full Spectrum less the Red which is mostly blocked past 10 meters.

Having a bit of Red not only 660nm but the 620-630 or up slide of the Red Photosynthesis Curve has some benefits with some corals gaining from this added shallow spectrum.

Anyway stay tuned for the release of this new light which has been labeled the Radion Killer by our Beta testers.

jrpark22000 it looks like you already come close to the LED Mix based on your research. The Mean Well LDD series is still the most efficient and cost effective PWM driver on the market. There are a few others out there that are a couple % more efficient but at a much higher cost.

Given the right circuit board you can simple snap in a 700mA and then later for less than $6 swap it for a 350 or 1000mA version.

PWM with older generation LEDs did cause a spectrum shift but the Latest Cree 5 watt LEDs are more resilient and with given the apex is 1-10VDC analog dimming there are simple PWM converters available.

Having true multi channel control with the right mix of LEDs can easily duplicate any MH and with the right drivers dim down to less than 1%, Our low voltage measurements of the Neptune VDM give it a 5-11% minimum dimming capability based on measured VDC.
 
wld1783 said:
I think you hacked into our dedicated research bunker underneath a nondescript building in a nondescript location near a Starbucks...

For starters good if not great SPS growth has been proven with a Cree Cool White, Royal Blue and Blue mix as the Cool White for the most part offers a Full Spectrum less the Red which is mostly blocked past 10 meters.

Having a bit of Red not only 660nm but the 620-630 or up slide of the Red Photosynthesis Curve has some benefits with some corals gaining from this added shallow spectrum.

Anyway stay tuned for the release of this new light which has been labeled the Radion Killer by our Beta testers.

jrpark22000 it looks like you already come close to the LED Mix based on your research. The Mean Well LDD series is still the most efficient and cost effective PWM driver on the market. There are a few others out there that are a couple % more efficient but at a much higher cost.

Given the right circuit board you can simple snap in a 700mA and then later for less than $6 swap it for a 350 or 1000mA version.

PWM with older generation LEDs did cause a spectrum shift but the Latest Cree 5 watt LEDs are more resilient and with given the apex is 1-10VDC analog dimming there are simple PWM converters available.

Having true multi channel control with the right mix of LEDs can easily duplicate any MH and with the right drivers dim down to less than 1%, Our low voltage measurements of the Neptune VDM give it a 5-11% minimum dimming capability based on measured VDC.
Click to expand...



Thanks for the reply and info Bill! It took many many hours of looking by google let me right into your bunker :D

I agree, I get good growth with the SOL spectrum, it's just some of the corals that can have better colors are bland. I'm hoping the extra spectrum can help them reach what they can, but didn't have any red planned. I'll keep it in my back pocket for the future as there is still more heatsink that can be used.

I don't follow this bit, can you explain more?
"Given the right circuit board you can simple snap in a 700mA and then later for less than $6 swap it for a 350 or 1000mA version."

All the colors of my build each have their own channel. It will be fully tune able. The limit is my eye. It isn't accurate like yours likely is, or anywhere close to a meter that can read the spectrogragh. Is there a reasonable instrument that I can rent from equipment suppliers? Or is it corporation level money to do so?

Thanks for the input on the apex vdm. Now I know it's limits I can plan for the low end control.
 
jrpark22000 said:
Old 01/09/2013, 07:46 AM #44
jrpark22000
Registered Member


http://www.instrumentsystems.com/fi...ads/Products/LpR10_Instrument_Systems_web.pdf

It’s a good article on the burn in effect and aging of white LEDs in relation to spectral shift. I haven’t been able to find any similar credible research on non-phosphorus dope LEDs. With that I don’t know if narrow spectrum LEDs will share a similar age shift or not.

For an update on my solution; I ended up going with PFC Meanwell drivers instead of the ELN series. They came out with the new series in fixed current for lower wattages and are more efficient.
I am using;
lpf-16D-48 qnty 3 (whites)
lpf-16D-42 qnty 1 (cyan)
lpf-25D-54 qnty 1 (violet)



The visio layout for how the LED build on my system will look.
Green circles - cyan LEDs
Purple circles - Violet LEDs
Red circles - White Tri-star LEDs
View attachment 7769
Click to expand...

Are those 1" or 1.5" heatsink?
 
jrpark22000 said:
Thanks for the reply and info Bill! It took many many hours of looking by google let me right into your bunker :D

I agree, I get good growth with the SOL spectrum, it's just some of the corals that can have better colors are bland. I'm hoping the extra spectrum can help them reach what they can, but didn't have any red planned. I'll keep it in my back pocket for the future as there is still more heatsink that can be used.

I don't follow this bit, can you explain more?
"Given the right circuit board you can simple snap in a 700mA and then later for less than $6 swap it for a 350 or 1000mA version."

All the colors of my build each have their own channel. It will be fully tune able. The limit is my eye. It isn't accurate like yours likely is, or anywhere close to a meter that can read the spectrogragh. Is there a reasonable instrument that I can rent from equipment suppliers? Or is it corporation level money to do so?

Thanks for the input on the apex vdm. Now I know it's limits I can plan for the low end control.
Click to expand...

The Coralux Board with a snap in adapter

DSCN0799_zpsa887ecca.jpg


Bill
 
wld1783 said:
The Coralux Board with a snap in adapter

DSCN0799_zpsa887ecca.jpg


Bill
Click to expand...

THANK YOU BILL! You just solved the problem I've been fighting, how to mount the huge enclosed meanwell drivers. These will fit into the LED enclosures I have. You da man!

You now have me completely rethinking how I'll build the driver and wiring. It's shaping up to be much simpler.
 
jrpark22000 said:
THANK YOU BILL! You just solved the problem I've been fighting, how to mount the huge enclosed meanwell drivers. These will fit into the LED enclosures I have. You da man!

You now have me completely rethinking how I'll build the driver and wiring. It's shaping up to be much simpler.
Click to expand...

Yep, instead of using one driver per channel requiring all 5 enclosures to be wired together, this will allow for each enclosure to have it's own driver per channel. Much easier to wire. Thanks again Bill.

Now I just need to find a ac/dc transformer once I spec out the wattage per enclosure.
 
jrpark22000 said:
Yep, instead of using one driver per channel requiring all 5 enclosures to be wired together, this will allow for each enclosure to have it's own driver per channel. Much easier to wire. Thanks again Bill.

Now I just need to find a ac/dc transformer once I spec out the wattage per enclosure.
Click to expand...

This 48VDC powersupply may work. Its 150 watts and can bu tuned to 52VDC for 15 LEDs per driver. One Should handle 50 Cree X series LEDs driven at 700-1000mA no problem

https://reefledlights.com/shop/mean-well-rs150-48vdc-power-supply/
 
wld1783 said:
This 48VDC powersupply may work. Its 150 watts and can bu tuned to 52VDC for 15 LEDs per driver. One Should handle 50 Cree X series LEDs driven at 700-1000mA no problem

https://reefledlights.com/shop/mean-well-rs150-48vdc-power-supply/
Click to expand...

Thanks Bill. Can you also get the LDD drivers in 300 500, and 600ma?

Right now I'm looking at this parts list, but want to spend some more time planning before I confirm. (I'm underdriving all the LEDs for added life.)
Meanwell LDD-600H qnty 5
Meanwell LDD-500H qnty 10
Meanwell LDD-300H qnty 10
Coralux 6up board qnty 5
 
jrpark22000 said:
Thanks Bill. Can you also get the LDD drivers in 300 500, and 600ma?

Right now I'm looking at this parts list, but want to spend some more time planning before I confirm. (I'm underdriving all the LEDs for added life.)
Meanwell LDD-600H qnty 5
Meanwell LDD-500H qnty 10
Meanwell LDD-300H qnty 10
Coralux 6up board qnty 5
Click to expand...

We stock the 700 and 1000ma drivers. with the Storm Controller we found that that 256 levels of dimming down to 1% works well. If you really want to limit the current a 1000mA driver in two parallel strings is 500mA +/- a few percent for current imbalance.

Bill
 
You must log in or register to reply here.
Back
Top