Hi I'm new here

I film things with high speed cameras and am having problems with flickering lights

typical cameras operate at over 1000 pictures a second with each image being 1/2000th of a second or shorter
a standard lightbulb on a 50 hz AC mains supply flickers 50 times a second and that is quite clearly visible
if you film the filament you can see it heat up and glow then cool down

because of this we mainly use either high frequency lights that operate at 1000hz, large tungsten halogen lights 5kw and above where the filament is much heavier and dose not have time to cool down before the next cycle heats it up again
or run lights from DC either battery or DC psu

I would like to use 12/24v LED lights on PWM dimmers that are powered from switch mode PSUs
but need to find high frequency PWM dimmers

What I would like is advise on how to measure the PWM frequency of these dimmers and also the PSU
I think I need an oscilloscope ?
if so which one would be suitable for a beginner
and would also need advise on how to use it !

I have very basic electronics knowledge and I am a qualified electrician

regards

robin

 

How much current do you need for the LEDs?

 

How much current do you need for the LEDs?

Hi crutschow

it very much depends on the scale of what I am filming
I make up lights from 24v LED tape at 18w/m
a typical light might be 200w

best

robin

 

It would seem that you could just use a standard switch-mode power supply with a variable output voltage as that would have no flicker.

 

It would seem that you could just use a standard switch-mode power supply with a variable output voltage as that would have no flicker.

thanks Crutschow

i need to fade the lights up together using the DMX control system
I already have lots of dimmer units some of which work better than others ...
what I would ideally do is measure the output frequency of all of them then could take a far better guess as to how they will behave

wil a scope do this ?

thanks

robin

 

wil a scope do this ?

Yes, a scope would be the preferred instrument for that measurement.

 

The flicker from PWM dimming of LEDs is a very much known problem in photography. I have seen it with my simple light board sometimes when I forget not to use any PWM dimming and I am not using a high-speed camera. PWM is a digital technique but there are also analog dimming techniques (see here).

I'm not an expert on this, but it seems like the best solution for your specialized need is not to use PWM dimming at all, if you can possibly accomplish the dimming with an analog method- by reducing the current, (or voltage as @Crutshow mentioned). Switch mode regulation can operate at a much higher frequency and should be flicker free at all dimming levels, but I understand that it can effect color temperature.

If you have no choice, than the higher the frequency of PWM scheme, the better. Your PWM dimmers that work better than others you have may be because they operate at a faster frequency. I don't know of products that operate above 15KHz though, probably because for our naked eyes, a much slower frequency will not produce a noticeable flicker. Not to say that they may be out there and even common.

I am not smart enough to tell you what the frequency needs to be, but if you take your best units and can find out what frequency that they are operating the PWM dimming, then you know that ideally you need to be faster than that. As Crutshow already mentioned, you should be able to determine/estimate the PWM frequency using an oscilloscope.

From here - "And I mentioned those guys doing ultra speed photography or videography in a lab… they’re also making their own lights as needed."

EDIT: As I read your question again, I realized that part of it is asking about how to use an oscilloscope to measure the PWM cycle time. Take a look at the scope pics in this AAC article...that should give you an idea.

 

Thanks everybody - you guys are great

I shall purchase a scope and get testing

unfortunately I often need very large amounts of light - 2 megawatts is the largest set up so far ! but that was using tungsten halogen light

Ideally I would be able to go to a rental house and test the lights they have available
some of it is adjustable upto 96KHZ but other stuff is fixed at 500HZ the lower the frequency the smoother it visually dims so that is where most of it is optimised
annoyingly the adjustable stuff is generally not marked in HZ but just numbered 1-10 or the like and dose not follow a uniform progression
and trying to find out exactly what it is relative to the software version the light might have gets quite tricky

LED lighting is becoming very popular in film lighting not that they would ever have 2MW of them !
but having the ability to use them close in to temperature sensitive things is very useful

thanks again

robin

 

Good deal. This topic interests me a lot and I hope you can stop back in and tell us how it goes.

 

hi All

looking at relatively cheap oscilloscopes I can get delivered i've come across the rigol brand
the DS1054z and the DS1052e both about £300 here in England

do you think these would cover my needs ?

or can you recommend another unit for that sort of price

thanks

robin

 

hi All

looking at relatively cheap oscilloscopes I can get delivered i've come across the rigol brand
the DS1054z and the DS1052e both about £300 here in England

do you think these would cover my needs ?

or can you recommend another unit for that sort of price

thanks

robin

@Mark Hughes you wrote a nice couple of articles on scopes, including this one. Can you please weigh in here? I *still* have not purchased one and in addition to the above query, do you have an opinion on this cheaper one (for myself) .

 

Hi @Raymond Genovese and @robinb,
I believe I understand the problem with the lighting flicker as a problem of sampling, and it's not one that I think you can solve with a PWM driver of any frequency. Unless you happen to record at a frame rate that is an exact integer divisor of the drive frequency, you will eventually have positive and then negative lighting coincidences in your photography. I am not aware of a pwm-driver that allows you to carefully adjust the drive frequency. You need a constant power source for your LEDs and adjust your scene brightness with neutral density filters on either the camera or the lights.
Your only other option is to exactly synchronize the light source with the camera shutter -- which is much more difficult to do, and might not be possible with your camera -- I don't know enough about them to provide an answer. The closer you get to exact synchronization, the slower the lighting will appear to change, but it will still change.
As for oscilloscopes -- the RIGOL is a feature packed oscilloscope for the price -- but I would venture a guess that even with more information, you'll not solve the problem until you create a suitable power supply to drive the LEDs.
Mark

 

You really need to connect a fast photo-diode to your oscilloscope and characterize the light output of the lights in question.

Only then can you really understand what is going on.

 

Unless you happen to record at a frame rate that is an exact integer divisor of the drive frequency, you will eventually have positive and then negative lighting coincidences in your photography.

I'm inclined to agree with this statement. I think you either need an exact match in frequency, or you would need the pwm frequency to be a LOT higher than the frame rate, which may not realistic. (Or you should use an analog dimming scheme to avoid pwm pitfalls altogether.)

As for oscilloscopes, I haven't worked with too many alternatives for comparison, but I use the DS1054z at work and I've been quite happy with it so far. I think it would allow you to see what you want in terms of frequencies.

Finally, I'm not sure how will this would work when dealing with lots of rental gear, but if you're interested in creating your own pwm control circuit(s) there are dedicated pwm chips that allow specific frequency settings. I've worked with the nxp chip linked below, and found it to be quite useful for my needs, but I don't know how accurate the frequency settings are. It also mentions accepting external clock sources and a few other features that might help with synchronization, but I can't vouch for those.
http://www.nxp.com/products/interfa...bit-pwm-fm-plus-ic-bus-led-controller:PCA9685

 

You really need to connect a fast photo-diode to your oscilloscope and characterize the light output of the lights in question.

Only then can you really understand what is going on.

thanks Sensacell

I have now bought an oscilloscope so am very interested in trying the fast photo diode idea
how would I go about doing this and what would a good diode be

thanks

robin

 

Thanks for all the replies everybody

I now have an oscilloscope a Hantek DSO5102p
and am having fun trying to understand what it is showing me/how to use it

I have connected it to a a 24v meanwell HLG 320 "led driver" and it seems to have a waveform smaller than I can measure
which is probably a good thing or operator error !

hooking up one of my LED dimmers powered by the meanwell HLG 320 it can show 30khz which is what it is supposed to be so thats good as well

I now just need to work out how to test other devices that have an unknown PWM frequency
the photo diode idea may well be the one

best

robin

 

@robinb,
This article might help you better understand what you are looking at with your scope. You are working with relatively low DC voltages (12 V or 24V), you should be able to simply connect the probe to the positive wire of the LEDs. Until you learn more, do not use the secondary alligator clip on your probe -- disconnect it and put it in the desk drawer.
Good luck,
Mark

 

@robinb,
This article might help you better understand what you are looking at with your scope. You are working with relatively low DC voltages (12 V or 24V), you should be able to simply connect the probe to the positive wire of the LEDs. Until you learn more, do not use the secondary alligator clip on your probe -- disconnect it and put it in the desk drawer.
Good luck,
Mark

great thanks Mark