"P.S. I just was closing down some of the browser windows that I used for my quick research and it occured to me that if the stage LEDs are in fact being powered by regular dimmers, then since the DO increase and decrease in brightness (somewhat) as the voltage applied to the LED/resistor combo changes, then during each half-cycle (every 1/120 of a second) the LED is going to get brighter and then dimmer, and without rectification is going to be off completely during the second half of the power cycle."

They do make ac led's. They're actually 2 led's in reverse direction to each other housed withing one package. One led is on for a 1/2 cycle, and the other is on for the other 1/2 cycle. You obviously will however get flicker with them as they cross the zero mark.

Plenty of the good oil on LED lighting implementation here:

http://www.philipslumileds.com/pdfs/AB20-3.pdf

Thats only one of many technical papers from Luxeon.

Bob.

There's also a really great utube example of PWM lights here:



Notice how the flickering stops on occasion. I would imagine that these are the points where shutter speed and flicker are a best match for each other.

That's an LED matrix display. No two LEDS are on at once.
Most seven segment LED displays are the same and they can be a PIA to get video or even still images of,

All these problems become much worse with cameras with a rolling shutter. Shutter speed in these cameras makes little to no difference as it's the readout time that causes the problem.

Bob.

**But led lighting is made up of more than one led. A typical led stop light on the back of a car for example is composed of maybe 15 led's so it doesn't take a large variance in current to cause a brightness shift.**

LEDs are never connected in parallel, except maybe in some cheapo Chinese flashlights. Certainly not in any commercial application without individual resistors. It's just too dangerous considering the miniscule amount of excess current needed to send one LED (the warmest one) into meltdown. We seem to be getting lots of speculation in this thread gleaned from resistive load theory that is entirely out of line with LED technology.

"LEDs are never connected in parallel,"

The Christmas lights on the front of my house are connected in parallel (every 6 leds)

**The Christmas lights on the front of my house are connected in parallel (every 6 leds) **

I'll let that one speak for itself.

A very wise person once told me, "you need to pick your arguments." I'm finished with this one.

musicvid:

I don't consider this an "argument" and I certainly hope you don't either!
This is all just information, alternative, and opinion designed to assist.... nothing more.

Well first let me say that Christian is probably on the right track to answering the poster's question. PWM would explain why some cam operators experience flicker and others not... depending on the frequency setting of the PWM vs the shutter speed. Furthermore, any time the PWM setting is changed then shutter speed will have to be tested again.

Blink, I get the impression that you may not really understand what PWM is. PWM stands for Pulse Width Modulation. In a normal design, the frequency of operation is selected by the designer and should not change during operation as the intensity is adjusted (or modulated). The output normally would be a square wave switching between 0 volts and some V-on voltage. For an LED drive, there would still be a resistor (or some other method) to limit the maximum current through the LED at the full V-on voltage. The LED is always fully on or fully off.

The control occurs because the ratio of on-time vs off-time in the cycle can be varied. The maximum is 'always on' and the minimum is 'always off', which are odd cases, because there is no measurable switching frequency. The middle case is 50% where the on and off times are the same -- a normal square wave. At the low end there is a short on pulse with a lot of off, and top end there is a long on with short off. The ratio changes but in a normal design the frequency, or total time for one on/off cycle, would not change.

Christian pointed out that a higher switching rate is good for video because there will be several cycles during any standard shutter rate, so the camera won't be confused. If the PWM frequency is high enough vs the shutter speed the changes in the light source should not create the flicker beat in the camera. If the PWM frequency is low, the shutter may see a changing portion of the cycle and this will cause the flickering. Selecting some particular long shutter time may help.

He also pointed out that for the LED source, if you merely change the current to the LED, the light output will change, but the color may also shift. That's why PWM is good in the source, because the on-current is constant so the light color will be too.

The posted UTube video seems strange to me. It may not be normal PWM or the frequency is low enough that the shutter is responding to changes in the modulation. Hard to say which without knowing what the LED control software is doing.

"Blink, I get the impression that you may not really understand what PWM is. PWM stands for Pulse Width Modulation."

I know full well what PWM is all about. I also know what it stands for. You can see that if you read my posts which clearly states PULSE WIDTH MODULATION. You can create one by squaring up a sine wave and it can be operated over a zero crossing, above a zero crossing, or below. Pulse with modulation is a square wave with an adjustable time scale. It can also contain an adjustable voltage but usually the voltage is constant and it's just time that is adjusted.

I do own a scope (albeit an older analogue scope) and do know how to use it.


"If the PWM frequency is high enough vs the shutter speed the changes in the light source should not create the flicker beat in the camera. If the PWM frequency is low, the shutter may see a changing portion of the cycle and this will cause the flickering."

A PWM is made adjustable so you can dim lighting that's what the term "modulation" is all about and it is accomplished by turning a pot. A rather simple PWM can be built from any 555 timer that you can find. (I think they even still sell them at Radio Shack), a couple of resistors, a trim pot and a cap .. shutter speed is also adjustable. In order to dim lighting you must slow the modulation timing... in other words you have to LOWER the frequency and if you do that then you will have to find a new shutter speed that doesn't clash with the new PWM setting.

"He also pointed out that for the LED source, if you merely change the current to the LED, the light output will change, but the color may also shift."
This is only true for a particular LED... not true for most others. Forward biasing a normal every day that you find in a flashlight for example with 10ma will result in the same color at 20ma

I think Rex is correct in most everything he says. The phrase, by the way, that is used to describe what is happening when you vary the on time vs. off time by changing the timing resistance in your 555 timer (or whatever) is duty cycle. Here's a quick ASCII version of what it looks like:

         ______           ______               

        |      |         |      |              

        |      |         |      |              

________|      |_________|      |_________     





         __               __               

        |  |             |  |              

        |  |             |  |              

________|  |_____________|  |_____________     

The bottom waveform would make the LED less bright, even though the frequency of the two waveforms is identical.

Everything I have read says that most LEDs will change brightness AND color temp as current is changed. Unless you drive them with a constant current source (which is ultimately the correct way to do things, and a voltage source and dropping resistor are just crude ways to approximate this), you will never get the light and color exactly the same from one LED to the next. However, if you drive them fairly hard, you'll get them up into a range where those variations are small (it doesn't change much in the upper ranges), this doesn't matter as much. However, at the point just above where they first turn on, all bets are off. That's why you don't want to dim them by varying current.

[Note: like all PN junction devices, you have to think about current, not voltage when designing with LEDs, just as you have to think about voltage, not current, when designing with tubes or MOSFETS. You certainly have to know the forward voltage drop and PIV spec, but it is current you are trying to control in order to get uniform brightness and color. In fact, the voltage across two "identical" LEDs will always be slightly different when driven with the same current.]

However, the main reason I posted is that in reading Rex's excellent post, it made me think of another reason why a video camera might flicker. I was mostly thinking about the interaction of the PWM frequency and the shutter speed, and forgot that the camcorder sensor chip is actually being scanned. I've seen descriptions on how this is done, and it is pretty complicated. Perhaps there may be some interaction with that frequency. Someone mentioned camcorders with "rolling shutters" and I think this is a special case of what I am talking about.

I wish I had some sort of LED controller, because it would be fun to play around with and get to the bottom of this, especially since this sounds like something many of us (including me, since I tape stage productions) will have to deal with.

In a CCD camera all photosites are effectively read at once.
I say effectively because the charge on all photodiodes is transfered to the Charge Coupled Device at the same time. The voltage stored in the CCDs is then sequentialy transfered to one or more A>D converters.

In a CMOS camera each row of photosite (still photodiodes) are read one at a time. No charge storage involved, less noise. However the time to read all rows is typically fixed regardless of shutter speed. For the EX1/3 the time is 1/60th of a second, even if the shutter speed is 1/1000th of a second. I agree this is hard to grasp, it really took me a while to get my head around it. With the EX1 at 60p the last row of frame 1 is read just before the first row of frame 2. That's why with something moving fast accross the frame you can get a lot of skew.

So with a CMOS image the first rows are say exposed for 1/1000 sec, they all 'see' the LEDs as On. 1/100 sec later the middle rows of the imager are exposed and the values read. They see the LEDs as Off. By the time the camera gets to the last rows the LEDs are On again. Result is a dark band in the middle of the frame. As the fps and the PWM clock driving the LEDs aren't locked the dark band will 'roll' up or down the frame.

If you've got other light source in the frame you can get some really funky effects. I've got some great footage where the walls of the venue look like there's wave of light travellin up them.

Getting back to LEDs. I think the change in CT with current only relates to white LEDs. The monochramtic LEDs would not shift wavelength with current given how the light is generated. I'm pretty certain the Zylight does not use PWM, instead it uses current control however you cannot dim it to off, only around 20% brightness. I suspect the design was done this way specifically to avoid any possible problems with flicker.

Bob.

Great explanation, Bob. Thanks.

Litepanel has a new version of the 1x1 with one knob to control the brightness (w/o changing the CT) and a second knob to adjust the CT from daylight to tungsten continuously.

"I think the change in CT with current only relates to white LEDs. "

And even with that, it depends on which white LED that you're talking about. In reality there is no such thing as a white led. Because of this the white has to be sort of 'manufactured'. Two types of led will do this. The first is one that contains a red, green, and blue led's which are combined to form white. These leds are not that sensitive to current changes since the individual colored led's within are fed the same current and therefore subjected to the same intensity change. If however each color is fed different currents then you can get some pretty wild color changes since there must be a specific balance in mind between all the colors. If that balance is at all disturbed then a color change will appear.

The second type are a phosphor coated led that operates much the same way as a floro. These are extremely sensitive to current changes and WILL noticeably change in color if you vary the current.

You control the LED brightness by controlling the forward current through it. Hard to do with just DC components, like a resistor, because as John pointed out, the resistance of the LED drops considerably when it turns on.

Enter the Charge Pump. An inexpensive IC that monitors and balances the forward current through the LEDs.

Here's a typical data sheet: http://www.fairchildsemi.com/ds/FA/FAN5607.pdf

Hi,

I would like to summarize here, since this thread starts to divert too much from the subject. If someone wants to discuss electronic details, there are other (better) fora for that...

1. Professtional LED lighting armature aimed for just flood lighting (not dimming) do not not flicker - cheap non-professional ones do (at double the line frequency).

The professional one have electronic PSUs that provide a nice constact current to the white LED strings (if there is ripple, it is at the switching freuency well above 100KHz). The cheap ones have just a number of ridiculously long strings of LEDs in series (or multiples of them) connected to a rectified mains via some dropping resistors, making them even worse (flicker-wise) compared to fluorescent lights.

2. You should never try to dim (with a triac dimmer) such LED lights, the will progably not like it at all. Or the dimmer does not work properly since it does not see a resistive load.

3. Professiona LED lighting armature AIMED for dimming do the trick themselves (they have DMX512 inputs). Neiter do they flicker on video, since professional equipment use rather high PWM frequencies internally for dimming the LEDs.

4. White LEDs (blue leds with a yellow phosphor coating) will change their color temperature quite dramatically, if you vary the brightness just by varying the DC-current. It is completely detectable by the human eye. The professional way of modulating the brightness of white leds is by PWM. This is de-facto the way its done in the majority of such dimmable lighting equipment. If the PWM frequency is high enough there is no flicker even on video.

5. Light armatures producing white (or any other color out of the spectrum) using RGB LED matrixes are as sensitive for current variations in the RGB led strings. To produce a controlled white you must stabilize the current in the leds very precisely. Even here the LEDs are normally run on constant current, and the dimming is done by PWM - in all decent quality professional lighting systems.

Even if it sounds unbelievable, colored leds have also a shift in the peak produced spectrum, as a function of the current. To produce uniform colors (and pure white) you must use a well defined current and chop it with PWM. Even a bigger problem is the intensity versus temperature variation, that causes even bigger headaches for the control system. But that discussion belongs to other forums...

Summa summarum:

- the only way of finding out if your video will flicker with LED lighting is to test it in real conditions (unless the manufacturer provides you with proper data).

- use the lowest possible shutter speed, if you see flicker.

- never dim LED armature with dimmer packs, if they are not inteded for this purpose.

Christian

Christian,
Wonderful explanation. Thank you.

Christian,

Yes, great explanation with lots of excellent electronic technical details.

Hi guys,

Nice to hear my non-native english writing is readable.. ;)

And I'm pleased if I can give something in return for all the valuable advice I have got from all of you more experienced Vegas forum members/users. Such a great "community".

I have a quite heavy elecronics background, having designed both LED lighting PSUs, CCD-controllers, and a plethora of both analog and digital electronics´, so you can throw even quite nasty questions at me, and probably will get an answer or two - hopefully useful... I promise to shut up if I have nothing to add to the subject... It's a good to follow principle, don't try to act wiser than you are - you will be caught sooner or later. Hope all would stick to that principle....

Cheers,

Christian

Off-topic except philosophically, in response to Christian's advice about not trying to seem wiser than one is...

To the famous Taoist adage that "He who says does not know; he who knows does not say", I coined a corollary a few years back: "He who does not say doesn't necessarily know."
And a further corollary to THAT is, you run a greater risk of being mistaken for wise by being silent, than you do by talking.

These are variations on the quote which I always thought was from Abraham Lincoln, but I just checked and various sites say that it was Mark Twain, and others say Samuel Johnson. Anyway, the version usually attributed to Lincoln is:

It is better to remain silent and be thought a fool than to open one’s mouth and remove all doubt.

My mother always used to say this to me when I would say something arrogant or stupid, in other words, pretty much every day.