Help with calculation

Hello, i'm looking for a formula to figure out, using manufacturer specs, spd or other, for example -
percentage of total ppf contained within 400-500nm.
Thanks

hyposomniac said:

Hello, i'm looking for a formula to figure out, using manufacturer specs, spd or other, for example -
percentage of total ppf contained within 400-500nm.
Thanks

Click to expand...

You sort of answered your own question, SPD stands for spectral power distribution.

Thanks mate.. Is that expressed as % of total ppf? It just says relative intensity on that axis.

Im just trying to correlate what I read (e.g. 10% of total ppf in 400-500 range) to the spd... Somehow i thought a great deal of conversions would be neccesary

You need to graph the spectrum. There's a variety of free software that will allow you to automatically or manually plot the spectral curve. Some of them don't do well with the colored lines used in SPDs but it only takes a few minutes manually. Once you have the relative figures you can convert them to percentages then add the percentages in the 400-500 range for a total.

There will be a difference in your totals depending on whether you use the full range or the 400-700 range.

Thank you.. Ive Been searching but I cant find software or spreadsheet to do what you suggest. I also cant find the process to do it manually

Is there a particular spectrum you are looking at?

hyposomniac said:

Thank you.. Ive Been searching but I cant find software or spreadsheet to do what you suggest. I also cant find the process to do it manually

Click to expand...

Try GetData or DigitizeIt.

JorgeGonzales said:

Is there a particular spectrum you are looking at?

Click to expand...

I'm looking at various clu058s and trying to decide which to try. And trying to make the spd data something i can compare to other stuff im reading.
The first graph in this post of yours is what Im after
https://www.rollitup.org/t/cri-test-and-mcree-weighted-results.919923/page-2#post-12931024
Trying to hit a target of @15% blue

hyposomniac said:

I'm looking at various clu058s and trying to decide which to try. And trying to make the spd data something i can compare to other stuff im reading.
The first graph in this post of yours is what Im after
https://www.rollitup.org/t/cri-test-and-mcree-weighted-results.919923/page-2#post-12931024
Trying to hit a target of @15% blue

Click to expand...

I'd go with 3500K, 80 or 90 CRI depending on what you want to see for green.

Some more crappy graphs:

Delicious!
I have to learn to do those.. Will be playing with the software mentioned.

I use this http://www.arizona-software.ch/graphclick/ for OSX.

Wow, an extra 5.7% (12.5% gain) in red!

Do you happen to have the 3000k handy by any chance?

I've been using getdata and ran a citizen clu058 2700k 90 cri through it and have gotten some weird outcomes...
Manually plotting and using the technique Rahz described is simply not working.. totals way over 100%, even when adding and averaging, and wildly differing results based on number of datapoints (using one point every 10nm to start). I fucking hate math and I've been doing it all night voluntarily. Someone punch my face.

0 390.447761194
0 399.4029850746
0.0035587189 409.845434748
0.0142348754 419.5309927232
0.0569395018 429.915015669
0.1672597865 439.451851065
0.3523131673 448.8771445265
0.231316726 459.5054974239
0.153024911 469.3238434164
0.1423487544 479.7875391725
0.206405694 489.3934243374
0.2918149466 499.7137090349
0.3665480427 509.3036596378
0.4128113879 518.9361024061
0.4555160142 529.3201253519
0.512455516 538.9366335582
0.5622775801 549.3100334626
0.615658363 558.9318531896
0.6619217082 569.3105646147
0.7117437722 578.9376958623
0.768683274 588.5542040686
0.8434163701 598.8904233282
0.9181494662 608.4803739311
0.975088968 618.8431507941
0.9928825623 627.7718170712
0.9715302491 638.9977160461
0.9145907473 648.7841929144
0.8149466192 658.634408031
0.7010676157 669.2521378871
0.5871886121 679.1235990864
0.4697508897 689.7466404632
0.3701067616 699.5968555798
0.2882562278 709.4205130929
0.2135231317 719.9798162214
0.1601423488 729.760981569
0.1174377224 741.018749668
0.0818505338 750.0270887555
0.0604982206 759.760450417
0.0427046263 769.4885005577
0.0320284698 779.2059276571

100 / (spectrum total / data point) = data point percentage.

You don't have data for 390 or 400, (the numbers on the right are the frequencies) but you can add the data points for the rest and use the total in the equation. The more points you digitize the more accurate the results will be, but a 10 point spread will probably give a decent rough estimate.

Spreadsheets make everything easier.

Here's what I got in the 400-500, 500-600, 600-700 ranges.

B- 9%
G- 34.5%
R- 56.5%

Also worth noting, using a 100nm spread for each color might not be entirely accurate since green is including 580-600, which one could argue is more valuable than "green"



If we use slightly different ranges, 400-480, 480-580, 580-700 the picture changes a bit.

Rahz said:

100 / (spectrum total / data point) = data point percentage.

You don't have data for 390 or 400, (the numbers on the right are the frequencies) but you can add the data points for the rest and use the total in the equation. The more points you digitize the more accurate the results will be, but a 10 point spread will probably give a decent rough estimate.

Spreadsheets make everything easier.

Here's what I got in the 400-500, 500-600, 600-700 ranges.

B- 9%
G- 34.5%
R- 56.5%

Also worth noting, using a 100nm spread for each color might not be entirely accurate since green is including 580-600, which one could argue is more valuable than "green"

View attachment 3780381

If we use slightly different ranges, 400-480, 480-580, 580-700 the picture changes a bit.

View attachment 3780382

Click to expand...

That's what I get too, ~7% of that red above 700nm.

Those color delineations are just scientific convention, but conventions do make it easier to compare between different light sources, and plants don't see colors.

I've seen different values for sunlight, but the 400-500nm range is usually 25%+ of the 400-700nm spectrum, and a bit more red than green.

What happens if we factor in the par readings? These results aren't percentages but they are relevant to each other and represent absolute values. I'm using an alternate set of measurements than the ones in the recent par test thread, and should show correct values for the 2700/90 sample. Also, these are 400-480/480-580/580-700 figures.



2700K is putting out more orange/red than the other samples even when par efficiency is factored in. However, it's not putting out a ton more orange/red. Whether a few points more in the target zone and peaking at 630 can make up for the deficiency in the lower range remains to be seen. 3000/70 is revealed to be not too shabby in the orange/red zone. The 3000/80 sample is very similar but the data isn't showing the red shift. It could be assumed that 3000/80 will be a better performer.

I think what this data shows is that it will be a close race all around.