Dimming Curves -- How to

[eurbani]

This year, I am replacing a lot of incandescent lights with LEDs.

 

I need some help understanding dimming curves.  The help files don't really help much and I haven't found the how-to of dimming curves anywhere on the forums or google.

 

So, this is what I know. (Or what I think that I know)

 

1. Dimming curves exist.
2. You upload them to a controller (or specific channels on a controller) using the hardware utility.
3. There is one file in the ...light-o-rama>DimmingCurves directory named LOR-Curve01.ldc
4. In a dimming curve file there are 1024 lines that (I think) correspond to dimming from a level of 0 to 1024

Soooooooo, here are my questions.

1. Is there a guide for dimming curves anywhere?
2. If in the sequence editor, we set levels from 0 to 100%, why do the controllers support 0 to 1024?  (i.e.  0%=0, 100%=1024  99.90234375% (approximately)=1023, but I can't set a channel to 99.90234375%.  So, what is the point?)
3. Besides opening up notepad and manually entering 1024 numbers, is there a way to create a dimming curve?
4. I can set my channels to LOR Standard, On Off Only,  OR  Dim Curve01- Dim Curve08.  Only Dim Curve01 is on my PC.  What are Dim Curves 02-08 (I assume these are for user created curves.)

 

Some testing results:

 

On the standard dimming curve, this set of lights appears to be all on at 70%

On the LED dimming curve (01), this set of lights appears to be all on at 83%

 

I don't know what the standard curve is.  I always assumed linear, but I don't really know that.  70% of 1024 is 717 but 83% of 1024 is 850 (which corresponds to 884 on the LED dimming curve).  This leads me to believe that the standard dimming curve is not linear.

 

I created a linear dimming curve from 0 to 887.   Now my lights were all on at 70% which corresponds to 717.  (Again leading me to believe that LOR's standard curve is not linear.)

 

Next, I created a linear dimming curve from 0 to 717.  Now, the lights don't seem to go on all of the way.  So, I'm confused.

 

 

Random notes:

 

You must name any curves LOR-CurveXX.ldc   (where XX is a number from 01-08).

The curve files are recognized when you click advanced configuration.  There is no need to restart the hardware configuration program.

There is a bug when you click "change curve file location"  It causes the hardware configuration to lose focus, but focus is on some window that I cannot find.  I cannot get the focus back to the window and I had to force the program to stop.

Why, when you select a unit in hardware configuration, is everything now related to that unit EXCEPT the console that requires to you select the unit again?  Arrrrrgh

 

[plasmadrive]

Look up Square Law Curve.  That is a standard for regular stage dimmers, but not sure about LEDs..   I am thinking LEDs that run from 120vac.  LEDs are someone linear from a DC controller, but not from an AC controller that uses phase angle control.  However your eye won't perceive them as linear even with the DC controller.

[EARLE W. TALLEY]

eurbani, I tried to figure this out last year and ran into a brick wall, everything I tried failed. Hope someone can figure it out, I have not seen anything from LOR other than "dimming curves exist"

[MSB]

We did some testing and created our own. You need to use the console  to determine how your LED's react as you ramp them up. Look at the  Found out that different manufacturer's have different responses. Write down the intensity values when you first notice the lights coming on and then the intensity of the lights at full power. You will notice it will be significantly less than 100%. Then make a copy of your LOR-Curve01.ldc and name it LOR-CurveXX.ldc where XX is 02, 03 04...etc.
 
Then open the file in a spreadsheet as a text file. You will see 1025 Rows of data. The number represents the intensity level.

 

I have a the curve we're using on our website as I can't figure out how to post it here. http://rotarymakingspiritsbright.org/steering-committee/design-build. I hope this helps!

[EARLE W. TALLEY]

MSB, can you tell me the low value and the high value of the dimming curve that you have.

[eurbani]

We did some testing and created our own. You need to use the console  to determine how your LED's react as you ramp them up. Look at the  Found out that different manufacturer's have different responses. Write down the intensity values when you first notice the lights coming on and then the intensity of the lights at full power. You will notice it will be significantly less than 100%. Then make a copy of your LOR-Curve01.ldc and name it LOR-CurveXX.ldc where XX is 02, 03 04...etc.

 

Then open the file in a spreadsheet as a text file. You will see 1025 Rows of data. The number represents the intensity level.

 

I have a the curve we're using on our website as I can't figure out how to post it here. http://rotarymakingspiritsbright.org/steering-committee/design-build. I hope this helps!

Thanks MSB.  I guess that helps, but I still don't understand why 1025 rows of data.   

 

0% in a sequence  = 0 on the dimming curve.  (I get that.)

 

Does 100% = the 1025th row of data?  

 

And why 1025 rows of data if we can only set 100 different levels?

[a31ford]

Knowing a bit about dimmers and stage lighting, I can say this:

 

the AC waveform (60 Hz, or CPS) says that since the MOST "punch" in a waveform is in the moving to a higher voltage from the current voltage (be it positive following or negative following, gives you 120 points of "punch" the wave has to move through both + and - portions of the wave to equal 60 CPS, therefore, 1/120th to the 30/120th of a second, and the 60/120th to the 90/120 of a second are the points within the wave, that will give the brightest portion of the ramp, the other portions (30 to 60, and 90 to 120) will be the dimmer portions, and on a "downward" slide, so the lamps (LED's being almost instantaneous in lighting up) will appear to be darkening during this time, we CANNOT see 60 flashes a second (really 120) BUT we can see 30 (really 60) remember both + and - of the AC cycle.

 

When fluorescents are cold we can see the "pulsing" (30 aka 60 flashes a second) We cannot see a standard incandescent because of the "afterglow" of the filament.

 

THEREFORE: the first and third portion of a dimming curve will have to be a slower rise than the second and fourth. Does this make sense ? I'm having a hard time following my own message, SHEESH !

 

Greg

[a31ford]

Working with MSB's "Rotary Club's curve", and plotting it in excel, I get this:

 

 

Notice that the first bit is a vertical line, BUT then it levels out and goes on a linear uphill ramp ?

 

Forget all the mumble jumbo I spoke about in my last post,,, that was the TIME in a waveform, NOT the curve in a dimmer..... NO wonder, I got so confused 

 

Greg

 

 

We did some testing and created our own. You need to use the console  to determine how your LED's react as you ramp them up. Look at the  Found out that different manufacturer's have different responses. Write down the intensity values when you first notice the lights coming on and then the intensity of the lights at full power. You will notice it will be significantly less than 100%. Then make a copy of your LOR-Curve01.ldc and name it LOR-CurveXX.ldc where XX is 02, 03 04...etc.
 
Then open the file in a spreadsheet as a text file. You will see 1025 Rows of data. The number represents the intensity level.

 

I have a the curve we're using on our website as I can't figure out how to post it here. http://rotarymakingspiritsbright.org/steering-committee/design-build. I hope this helps!

Edited October 25, 2013 by a31ford

[EARLE W. TALLEY]

All I want is a dimming curve that goes from 0 to 80, very simple, if anyone has one please PM me.

Thanks 

[rhino]

also looking for a simple dimming curve for LEDs, that actually start ramping up when the program says to, not 3 secs later.  Using HLE LEDs.  I have the Curve01 (from LOR) and Curve02 (from MSB above)

[plasmadrive]

With LEDs and 120vac input, you will have a small lag where the voltage has to exceed the turn on voltage of the LEDs and then you start drawing current.  Remember they are diodes and have a specific voltage drop per LED and they are current devices not voltage devices per se`, so that needs to be taken in to account for your dimming curve. 

 

With the phase angle control of the standard dimming circuit you have to get far enough in the phase to create that voltage to even start to turn them on.  Then you get the relationship of V/R to equal the current to the LED.  V being the input RMS voltage minus the drop of the LEDs in series, R being the inline resistor.

 

Then you have the perceived brightness compared to the actual current going to the LEDs.  The two are not a linear relationship for voltage/current control like what you are using with the dimmer.  However, they are a linear average using a DC PWM controller.  That is because you always have the same current turning on the LED and you are varying the duty cycle only.  This creates your average current and your perception of the brightness.  PWM is not as efficient below 100% as constant current dimming, but far better control if you can't program a dimmer curve to match the LED output you have. 

 

So, either you need to figure out a better curve overall, or you can control them via PWM and use a linear curve.    BTW, once over 50% the PWM and constant current control are close to the same.  However, constant current is not a linear relationship to the AC dimming cycle because the voltage output is not linear relative to the cycle on time.

 

Here is a very old artical that may interest you.. if you really want to adjust a curve to work as you intend. 

http://www.candlepowerforums.com/vb/showthread.php?70073-Constant-Current-vs-PWM-dimming-Revealed

 

However, we are doing Christmas lights Flashy Thingys... Not sure the audience would know the difference..   Just a thought. 

[EARLE W. TALLEY]

I am still looking for a dimming curve that goes from 0 to 80. Reason - LEDs are full on at 70 - 80%, my controllers are set at 86% but would like them to be 80%.  I have cut my failure rate on LEDs down to 0 in the last two years by lowering the amount of current applied to the lights. It works for me !!!

[a31ford]

OK, so I finally had time to do some research, got the OEM LOR curve (the one supplied by LOR) and processed it in Excel as well, here is what they supply (which even makes more sense.

 

 

I have a light meter and will ACTUALLY do a common set of LED lights in the Hardware utility and note the percentages of level on the channel fader

IN the LOR utility, below.

 

Once I have the numbers, I will post  the Excel Plot, AND upload the Curve.... Stay Tuned, (follow this topic) if you want the curve.

 

Greg

 

[plasmadrive]

Can't wait

[a31ford]

Working on the dimming curve stuff now, as I type........

 

Greg

 

 Well, Well, well... Earl, If I hadn't seen it with my own eyes, I would have not believe it, 80 %, BANG on.... no "extra" brightness..... That is one for the mental note pad.....

 

going to do a quick video of what I'm doing, back in an hour or so....

 

G.

Edited November 2, 2013 by a31ford

[a31ford]

OK <    So it appears that the CTB16PC (DIY Kit type board) does NOT support dimming curves, I'm going to call LOR later today (hoping they are open) and ask about this (it simply could be that my pea brain simply is missing something).

 

hardware utility, configure button (bottom of utility) advanced configuration (right side of window, near bottom) goes to read board configuration, comes up with error, "Controller does not support advanced configuration"

 

still working on the curve stuff for LED's but guess I won't be able to use it myself.

 

Greg

 

 Opened a help desk ticket just now, will have to wait until then.

 

G.

Edited November 2, 2013 by a31ford

[EARLE W. TALLEY]

Greg, the gen 1 & 2 CTB16pc boards do not support dimming curves it came out with the gen 3 controllers. The DIY kits were gen 2 for a while after the gen 3 came out, I don't know if they still are or they are now gen 3, have not checked on that.

Earle

[a31ford]

I think I've kinda figured this out, the CTB16 (PC) (Planet Christmas) DIY cards do NOT have the extra memory on the processor chip for dimming curves, HOWEVER, the CTB 16 (D), (Deluxe)  Do.

 

Based on this, I would assume the Pro stuff has the extra EEPROM do do dimming curves also.

 

NOW, I'm looking for other options in the software, that WILL produce a max of 80% on each channel.... (Per Earl's & my needs.....)

 

Greg

[a31ford]

Hi Earl,

 

 my cards are from this spring (2013), and are CTB-16-PC V1, with firmware revision 4.32 and do NOT support curves. BUT, hold on.... I think I found something you might like....... Trying it now....

 

Greg

[a31ford]

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