Dimming Curves -- How to

[a31ford]

This is mainly for Earl, BUT can be used by everyone.

 

Earl, here is your custom dimming curve, you will have to rename the LOR curve to something else (so you still have a copy of it) and THEN, place the curve I've supplied into that folder (the light-O-rama dimming curves folder) and rename the file (AND extention) to LOR-curve-01.ldc

 

 

NOTE: !!!!! this Curve will produce only an 80% light output, and should NOT be used with normal (incandescent) bulbs....

 

 

 

see the link below:

 

LED-80-Manning-Curve.txt

Edited November 2, 2013 by a31ford

[EARLE W. TALLEY]

Greg, I got it, will change it over and test it out in the next few days and post the results.

Thanks 

Earle

[viennaxmas]

Is there any more advice on how the dimming curves work? I looked through the LOR manual today but there was nothing in it. How do the values in the curve file relate to the 100 dimming steps of LOR? Is it 25 values per step to allow for long ramp times?

[EARLE W. TALLEY]

Greg, couldn't wait, the file had and error: "invalid curve file: two few fields"  I added another "818", file loaded ok, but in testing, it looks as if things might be backwards, the lights only dim down to about 20% and no futher, this was the problem I ran into last year when trying to figure it out.

[a31ford]

Earl, I tried to get 1024 but might have missed a couple.The problem is I'm only going on assumption, as my stuff will not handle curves.

 

20 % huh.... HUM... will try something, let me get back to you....

 

Greg

 

Viennaxmas,  still working on this, it's too bad that I do NOT have a CTB-16-D controller........

 

G.

 

I'm going to change tactics here, instead of working from 1023 and back, I'm going to find 80% and work to the end (1023) with the 818 amount...

 

G.

Edited November 2, 2013 by a31ford

[EARLE W. TALLEY]

OK, I am standing by.

[a31ford]

OK, NOW a new version of the Curve (working from 80% and holding, instead of 100% and down)

 

NEW-LED-80-Manning-Curve.txt

 

This file is compared in the Excel graph below, the dark line is the OEM LOR curve, the pink line is the new 80% curve

 

 

here is the NEW version 

 

NEW-LED-80-Manning-Curve.txt

 

Sorry about the confusion.

 

Greg

 

[EARLE W. TALLEY]

Anyone with the new version that Greg posted hold off, it does the same thing, we are working on it and doing some testing.

Earle 

[eurbani]

OK, So I'm "Out" can't do anything for my boards, HOWEVER, being a "Electronics & Computer & Electrical & Software Type Guy"  I think I've figured out eurbani's and earl's problem.

 

Eurbani, the dimming file needs 1024 number POSITIONS (0 plus 1023) if you look at LOR's OEM file ( LOR-Curve-01.ldc ) you will find it produces a graph plot in Excel like MY POST #13 HOWEVER, if you look at  MSB's curve (Post # 8) you will notice that it still has 1024 positions.

 

LOR's curve, has 1023 in the last 65 or so positions (so that is 99.8 %) for 65 "ticks in the ramp" so to speak.

 

Don't think of the curve as a small set of small numbers, think of it as place holders (with value) that have to cover a file length of "X" (X being 1024 place holders).

 

I can't do it with my controllers BUT looking at the files contents and looking at the graph in post #13, speaks for itself, AND this is also why your 717 curve, didn't work, there whern't enough place holders.

 

Double up the numbers if you have to eg: 20,20,20, 23,23,23,25,25,25 etc. to get 717 at the end of 1023 positions.................

 

Greg

I get that you need 1024 slots in a dimming curve.  The thing that really confuses me is how those relate to 0-100%.  If we can only set from 0-100%  (101) intensities why to we need 1024 lines in a dimming curve.  

 

Is there anyone from LOR that monitors these forums?

This is mainly for Earl, BUT can be used by everyone.

 

Earl, here is your custom dimming curve, you will have to rename the LOR curve to something else (so you still have a copy of it) and THEN, place the curve I've supplied into that folder (the light-O-rama dimming curves folder) and rename the file (AND extention) to LOR-curve-01.ldc

 

 

NOTE: !!!!! this Curve will produce only an 80% light output, and should NOT be used with normal (incandescent) bulbs....

 

 

 

see the link below:

 

LED-80-Manning-Curve.txt

You can actually name your curves LOR-curve-XX.ldc where XX is a number form 01-08.  You do not have to rename and save the default LOR curve.

[TonyD]

I get that you need 1024 slots in a dimming curve.  The thing that really confuses me is how those relate to 0-100%.  If we can only set from 0-100%  (101) intensities why to we need 1024 lines in a dimming curve.  

 

Remember in computer terms 1024 is 1K (one thousand) and often in hardware "data" is stored at the bit level. 1024 values uses 10 bits. So you can squeeze 1024 different values in only 10 bits of storage.

 

"Words" are a collection of bits usually related to hardware design (perhaps register size?) and computers often are optimized (in low-level software (firmware) especially) to deal with data the size of the hardware's "word". So moving data around and comparing it is really easy when done at the word level. To move/compare data the size that could store 1025 possible values might take 2 operations or twice as long for example.

 

I'm guessing that's why there are 1024 different values possible in a dimming curve. So when figuring out the percentage you do it with 1024 possible values not 100 (or 1000). 80% of 1024 would be 819.

[eurbani]

You can actually name your curves LOR-curve-XX.ldc where XX is a number form 01-08.  You do not have to rename and save the default LOR curve.

Correction LOR-CurveXX.ldc

[eurbani]

Here is a excel dimming curve creator.  It is pretty simple.  (nothing like BobOs stuff).

 

Use it this way:

 

1. Download and save to the LightORama>DimmingCurves directory
2. Plug in a set of LEDs to a controller
3. connect the controller to your computer
4. start LOR Hardware
5. Click Refresh
6. Click Console
7. select the controller number that you have connected
8. dim up the channel with the LED lights
9. when they come on mark down the %  displayed in the console (min level)
10. dim up more 
11. when they don't get any brighter mark down the % displayed in the console (max level)
12. open the spreadsheet below 
13. enter the two numbers above
14. click control-l (Crl-l)  (that is a lower case L-- not a 1)
15. It will create an LDC file named LOR-CurveXX.ldc
16. rename the file changing XX to 01-08 (note 01 is that default Curve if you still need it, start at 02).

 

Assumptions

1) You have MS excel 2010 installed (It will probably work in 2003 or 2007, but I haven't tested).

2) You used the default LOR directory (C:\Users\<USERNAME>\Documents\Light-O-Rama)

3) I don't need to explain that the controller needs to be plugged in to the wall and a USB 485 and the USB 485 into your computer.

 

Lastly, if you make the mistake of not saving your new dimming curve to a new name, the spreadsheet macro will create ONE back up and lable it "Backup of LOR-CurveXX.xlk"  Same for the too.  You cold save it each time that you use it with a new name like "Curve for cheap kmart lights"  OR  "Curve for expensive LED lights"....

 

 

I can't upload this file.  So I posted it here:

 

http://not4profit.net/Files/Linear%20Dimming%20Curve%20Calculator.xlsm

 

If you have any problems send me a PM and I will try to fix them.

[LightORamaJohn]

The reason there are 1024 brightness levels is to make sure we have enough resolution to do smooth fades and make the luminosity linear. The CCF2D flood works this way as well as the upcoming CF50D 50w flood.

 

For example, consider a LED color that when using a linear 'curve' comes on at 15% and is fully bright at 75%. This means we are spanning only 60% of the available brightness levels to go from off to full on. We've given up 40% of the available levels because those outside 15-75% don't do anything. With 1024 brightness we still give up 40% of the range, but 60% still leaves us with 600 levels, 4 times the number we had with the older scheme. This makes slow fades much smoother.

 

Because the floods use specific LEDs, we can also use this scheme to make the luminosity linear from 0 to 100% by taking a bunch of luminosity measurements at various intensity settings. We can then use this to extrapolate a set of values in the 600 brightness levels available that result in a linear luminosity increase. You could do this for your various LEDs, but it would be a lot of work since there are so many LED string types/colors. It would be overkill for LED strings, but is noticeable when using high power LEDs like those in the floods.

 

Once a curve is loaded, the 0-100 LOR percentages simply index into the curve table to get the correct value to produce that output level. Hope this helps.

[Ken Benedict]

... as well as the upcoming CF50D 50w flood.

 

 

 

Oooooooooooooooooooohhhhh.    

[eurbani]

The reason there are 1024 brightness levels is to make sure we have enough resolution to do smooth fades and make the luminosity linear. The CCF2D flood works this way as well as the upcoming CF50D 50w flood.

 

For example, consider a LED color that when using a linear 'curve' comes on at 15% and is fully bright at 75%. This means we are spanning only 60% of the available brightness levels to go from off to full on. We've given up 40% of the available levels because those outside 15-75% don't do anything. With 1024 brightness we still give up 40% of the range, but 60% still leaves us with 600 levels, 4 times the number we had with the older scheme. This makes slow fades much smoother.

 

Because the floods use specific LEDs, we can also use this scheme to make the luminosity linear from 0 to 100% by taking a bunch of luminosity measurements at various intensity settings. We can then use this to extrapolate a set of values in the 600 brightness levels available that result in a linear luminosity increase. You could do this for your various LEDs, but it would be a lot of work since there are so many LED string types/colors. It would be overkill for LED strings, but is noticeable when using high power LEDs like those in the floods.

 

Once a curve is loaded, the 0-100 LOR percentages simply index into the curve table to get the correct value to produce that output level. Hope this helps.

 

Let me see if I understand this using examples.

 

For an incandescent string when I set a level of 10% that corresponds to 10% of 1024 on the dimming curve (or 102).  Right?  And then 20% would be 204.

But when I am  dimming from 20% to 10%, the controller is actually doing 204, 203, 202, ...  105, 104,103,102.  Right?  

 

So, there really is no way to set an intensity of  203, because I cannot set an intensity of 19.82421875% in SE, but 203 is used to smooth out a dim from 20% to 10%.

 

 

If there is linear dimming curve set with on at 25% and full on at 75%.  It would look something like this....

 

0

256

256

257

257

258

258

...

767

767

768

768

 

So when I set intensity 10% in this case,  this would correspond to 307 on the dimming curve and 20% would correspond to 358.    But when I dim, from 20% to 10%, the controller would  now send 358,358,357,357,356, ... , 309, 308,308,307,307

 

(BTW, John, when you answer this, I have another question, but I wanted to make sure that I understood this much first.

[EARLE W. TALLEY]

I did some testing today with a couple of dimming curves that I was sent. To make a long story short, the best dimming curve that I found is the LOR Curve01, but along with that the controller Intensity Range was set to a max of 86%, the power range was from 0-82.9 with very smooth fades.

Earle 

[LightORamaJohn]

eurbani,

 

There are non-linearities at 0-1% and 99-100% for historical reasons, so the selection of the count for 10% intensity would be a little different, but would be linear between 1-99%.With that said, the rest of your example is correct.