General DIY LED thread
- Thread starter Kampo
- Start date
- Tagged users None
AbjectMaelstroM
Valuable Member
Hmm might be something on my end with pop-up/script blockers.
AbjectMaelstroM
Valuable Member
Confirmed, it was my browser. Works fine on my cellphone. Also played around with my expected build and got a poo poo result....
Anyone know how to import custom/updated led emitter data?
Yea it's fairly easy once you get the hang of it..
Some is directly importable from spectrum analyzers.. Others for us mere mortals involves work..:
first import a spectrum chart into this:
Plot every 5nm
Note that when doing the import into SPECTRA you MUST start at 380.
so
380, 0
385, 0
390, .015
ect..
but you'll fix that in step 2.
After generating a data set go to it and format: fixed
Then one needs to clear all zeros (and fix and oddball like if you plotted 427 (425, or 430)
Which you will also do in step 2
step 2 .. Use this (btw this is the only data set you need for the viosys violets SPECTRA will fill in the rest later:
Notepad ++ (trust me windows notepad sucks..)
Needs to look like this:
380,0.005
385,0.005
390,0.013
395,0.046
400,0.114
405,0.255
410,0.551
415,0.798
420,1
425,0.721
430,0.397
435,0.228
440,0.108
445,0.06
450,0.034
455,0.016
460,0.008
465,0.003
Save it as txt (you need to save somewhere b4editing btw. Notepad++ won't let you edit the imported data immediately.
Import file in SPECTRA CLEAR.. DATA..FILE: (add the 2 parameters) SAVE
Preview (have broken bits in the sepectum of spikes ect. Just needs to look like your orig spectrum will tell you if data is formatted correctly .
Only added info is beam angle and lumens. SPECTRA will generate from there. Trick is 1)knowing your chips lumens and 2)understanding mW isn't lumens..Diode power listed in mW gets a little tricky.
What SPECTRA does w/ the raw numbers:
you can export and fix if you like
NOTE the 4 lm = 933 mW w/ 420nm peak diodes
brandon clow
Active Member
- Joined
- Oct 22, 2017
- Messages
- 121
- Reaction score
- 155
This is really helpful, I was wondering about this also.Yea it's fairly easy once you get the hang of it..
Some is directly importable from spectrum analyzers.. Others for us mere mortals involves work..:
first import a spectrum chart into this:
Align axis
Plot every 5nm
Note that when doing the import into SPECTRA you MUST start at 380.
so
380, 0
385, 0
390, .015
ect..
but you'll fix that in step 2.
After generating a data set go to it and format: fixed
Then one needs to clear all zeros (and fix and oddball like if you plotted 427 (425, or 430)
Which you will also do in step 2
step 2 .. Use this (btw this is the only data set you need for the viosys violets SPECTRA will fill in the rest later:
Notepad ++ (trust me windows notepad sucks..)
Needs to look like this:
380,0.005
385,0.005
390,0.013
395,0.046
400,0.114
405,0.255
410,0.551
415,0.798
420,1
425,0.721
430,0.397
435,0.228
440,0.108
445,0.06
450,0.034
455,0.016
460,0.008
465,0.003
Save it as txt (you need to save somewhere b4editing btw. Notepad++ won't let you edit the imported data immediately.
Import file in SPECTRA.. Preview (have broken bits in the sepectum of spikes ect. Just needs to look like your orig spectrum will tell you if data is formatted correctly .
Only added info is beam angle and lumens. SPECTRA will generate from there. Trick is 1)knowing your chips lumens and 2)understanding mW isn't lumens..Diode power listed in mW gets a little tricky.
What SPECTRA does w/ the raw numbers:
you can export and fix if you like
NOTE the 4 lm = 933 mW w/ 420nm peak diodes
no guarantee there aren't easier ways..
Webplot is tricky at first and some have got auto extraction to work (not me) but I found manual works best for me..
second biggest stumbling block for many is the data formatting for import
small errors of a comma or decimal place or a non-5nm step is fatal..
.
Webplot is tricky at first and some have got auto extraction to work (not me) but I found manual works best for me..
second biggest stumbling block for many is the data formatting for import
small errors of a comma or decimal place or a non-5nm step is fatal..
Suppose I should respond to you directly:
"MY" data set.. Used 70 degrees (right or wrong) and 4 lumens.(had to try to match mW data)
So there is some artistic license involved w/ this but, hopefully, kept it realistic..
YMMV.
Data set's too long to do a c/p for the mint atm
Kampo may have it handy.. Pretty sure they did some chart construction..
380,0.005
385,0.005
390,0.013
395,0.046
400,0.114
405,0.255
410,0.551
415,0.798
420,1
425,0.721
430,0.397
435,0.228
440,0.108
445,0.06
450,0.034
455,0.016
460,0.008
465,0.003
AbjectMaelstroM
Valuable Member
***NOTE*** K16s simulated with 10 each RebelES RB.
So, playing around with high CRI COBs trying to get the number of LEDs down, any thoughts on something like this?
I still haven't gotten around mapping the Sunplus CW, but I may drop one of the fresh fish and replace it with 4-6 Sunplus cools whites.
This is with 2 fresh fish.
K16 is an awesome RB puck, but I fear it will be too purple once whites are added in, this is where the blue, mint, cyan, and Sunplus CW (since they take care of some of the cyan) would come in.
So, playing around with high CRI COBs trying to get the number of LEDs down, any thoughts on something like this?
I still haven't gotten around mapping the Sunplus CW, but I may drop one of the fresh fish and replace it with 4-6 Sunplus cools whites.
This is with 2 fresh fish.
K16 is an awesome RB puck, but I fear it will be too purple once whites are added in, this is where the blue, mint, cyan, and Sunplus CW (since they take care of some of the cyan) would come in.
Sunplus has a LARGE beam angle..Other than that it specs close to ff in spectrum..
Really needs secondary optics..if light is more than a few inches off the water line..
Royal blue and added whites (royal blue plus phosphors) is really never err "purple".
Yea you can get into magenta a bit but takes a lot of red..
If you want more "sea aqua" you need to up the cyan/reg blue quite a bit..
These spectrums show the idea..
Sadly, they didn't include a nm scale but the inferences is heavy blue/cyan broad phosphor w/ the violet pump.
Really needs secondary optics..if light is more than a few inches off the water line..
Royal blue and added whites (royal blue plus phosphors) is really never err "purple".
Yea you can get into magenta a bit but takes a lot of red..
If you want more "sea aqua" you need to up the cyan/reg blue quite a bit..
These spectrums show the idea..
KYOCERA Develops World's First Full-Spectrum LED Aquarium Lighting; Innovative technology mimics sunlight at specific underwater depths; designed to optimize growth of corals and marine plants | News Releases | KYOCERA
Kyocera announced that the company developed the world’s first full-spectrum LED lighting for aquariums.
global.kyocera.com
Sadly, they didn't include a nm scale but the inferences is heavy blue/cyan broad phosphor w/ the violet pump.
AbjectMaelstroM
Valuable Member
Hmm add 4 blues for each K16? Maybe do 4x 3up: 1 HyperViolet 2 Blue? Hopefully that with 4 mints brings it in a bit.
Ugh, hard to judge in the abstract. I have a feeling I might just have to build one and modify from there.
Ugh, hard to judge in the abstract. I have a feeling I might just have to build one and modify from there.
See post #69....
Pretty pricey though...
May want to just buy a whole lot of e-bay cheap "egg" cyans..
One thing.. and not sure it is still common.. Cheap cyans were actually double dies w/ a yellow/blue diode to "make" cyan..
They work but it's 2 nm's.
10/$10....
IF the are close to correct, the cheap ones lean blue where the Lux lean green..
The budget bruiser LED build... - Page 3 - Reef Central Online Community
Page 3- The budget bruiser LED build... Do It Yourself
www.reefcentral.com
Pretty pricey though...
May want to just buy a whole lot of e-bay cheap "egg" cyans..
One thing.. and not sure it is still common.. Cheap cyans were actually double dies w/ a yellow/blue diode to "make" cyan..
They work but it's 2 nm's.
10/$10....
1W 3W 5W White Red Blue UV RGB Green High Power LED COB Chip Light With PCB USA | eBay
1W 3W 5W Warm White Red Blue Full Spectrum RGB UV High Power LED Chip With PCB.
www.ebay.com
IF the are close to correct, the cheap ones lean blue where the Lux lean green..
Last edited:
As an eBay Associate we earn from qualifying purchases.
AbjectMaelstroM
Valuable Member
Oh man, I like it! Nice find. K16, fresh fish and hyper violet might be a really good base. Then keep adding some blue/mint/cyan till I see the light I like.
It's too bad Steve stopped selling cyan Lux ES. May have to get 2/3 3ups and add reflow cyans on them. Or be lazy and go with individual stars.
AbjectMaelstroM
Valuable Member
Yeah that was one of the places I found the Cyans on a star already. What do you think about doing 3-ups of Blue instead of individual stars? Maybe do the HyperViolets as 3-up (2 blue 1 hyper 4.0) and do 4 of them for 8 blues total?
AbjectMaelstroM
Valuable Member
Might shoot them an email, if you could mix, that would be nice. Not sure what the minimum order would have to be.
The only issue I see if you can call it an issue, is that the board is wired in series so your entire Channel would need to be modeled on to 1 PCB.
Though I suppose one could always remove the resistors to break the series.
Last edited:
Well I was thinking more like 6 cyan 1 reg blue.. 
www.color-hex.com
You could mix all 3 blues to get the base color you prefer.. Violets will just rich-en the color
I'm sort of catering to what you seem to prefer and that leads to more or less changing from k12 centric to a preferred shade.
Kampo and Lingwendil have a lot of practical experience w/ the k12's or Toshiba "clone"..
You'll be dimming the k12's soo much that I'm not sure how practical that is..
From what you liked:
Sooo.. main "pucks" are preferred blue tone w/ RB 3 ups/violets if wanted.
I see little difference in rb centric or broad blue centric
#33e5ff Color Hex
#33e5ff color hex, #33e5ff color chart,rgb,hsl,hsv color number values, html css color codes and html code samples.
You could mix all 3 blues to get the base color you prefer.. Violets will just rich-en the color
I'm sort of catering to what you seem to prefer and that leads to more or less changing from k12 centric to a preferred shade.
Kampo and Lingwendil have a lot of practical experience w/ the k12's or Toshiba "clone"..
You'll be dimming the k12's soo much that I'm not sure how practical that is..
From what you liked:
Sooo.. main "pucks" are preferred blue tone w/ RB 3 ups/violets if wanted.
I see little difference in rb centric or broad blue centric
