Need to have LED lights "flicker" at 60 times per second

Thread Starter

laughingcamera

Joined Apr 27, 2022
11
I have a machine (a high resolution image scanner using a lens and CCD) that uses fluorescent tubes to light the object that I would like to change over to LED tubes. The machine is calibrated to see the flicker of the florescent tubes and when just replacing them with LED tubes it freaks out. So I need to make the LED tubes flicker at 60 times per second.

A plug and play solution would be best but also a DYI circuit is an option. The current LED bulbs I have use 120v AC input, I believe the AC to DC conversion is taking place in their built in drivers. Any ideas?

Thanks,
Mark
 

Ian0

Joined Aug 7, 2020
5,158
I think you mean that they flicker at 120 Hz, not 60Hz.
Fluorescent tubes haven’t flickered for many years, you must have a device that was intended for use with fluorescent fitting with wound ballasts. If so, you can probably take the LEDs out of the replacement fitting and connect them to the old ballasts. Not as efficient, but same old flicker!
 

Audioguru again

Joined Oct 21, 2019
4,690
One rectifier diode in series with a string of series LEDs with a current-limiting resistor causes the LEDs to blink at 60Hz.
I haven't measured it but maybe a fluorescent tube flickers at double the 60Hz (120Hz) since it has no series diode.
Google says I am correct.

A full wave rectifier circuit (has 4 diodes in it) will cause a string of series LEDs with a series current-limiting resistor to blink at 120Hz. 120Hz will also be caused when one string of LEDs is connected parallel to a second string but on string is connected in reverse so that the strings light alternately.

The fluorescent tube keeps shining light but the light is dimmed a little at 120Hz. A string of LEDs will abruptly turn on and turn completely off at 120Hz which might upset the machine calibration.
 

Thread Starter

laughingcamera

Joined Apr 27, 2022
11
I think you mean that they flicker at 120 Hz, not 60Hz.
Fluorescent tubes haven’t flickered for many years, you must have a device that was intended for use with fluorescent fitting with wound ballasts. If so, you can probably take the LEDs out of the replacement fitting and connect them to the old ballasts. Not as efficient, but same old flicker!
Yes, you are correct, the flicker must be at 120 times per second, my bad.

So what you are saying is get the replacement type of LED tubes that require a ballast, and connect it directly to the ballast? Or am I confused?
 

Ian0

Joined Aug 7, 2020
5,158
Yes, you are correct, the flicker must be at 120 times per second, my bad.

So what you are saying is get the replacement type of LED tubes that require a ballast, and connect it directly to the ballast? Or am I confused?
You’ll probably find that the LED luminaire has a string of LEDs and a driver. the operating voltage of the LEDs is probably about the same as the tube (about 100V peak) - need to check.
If so, you could connect the LEDs to the old ballast, but it might be too powerful for them (what is the actual power?). You might need a smaller ballast.
This information is from taking a small number of LED luminaires apart, so might not apply to all of them!
Then you would get the 120Hz flicker, and @Audioguru again is correct that the LED flicker would be more abrupt than the fluorescent flicker due to the persistence of the phosphor, but white LEDs also have phosphor which has some persistence.
 

Thread Starter

laughingcamera

Joined Apr 27, 2022
11
And you can't get the correct bulbs?
I can get the fluorescent bulbs from a few places in Germany that still stock them but at some time they will not be available. Plus I have to wait every time I use the machine (it is an image scanner) 20-30 minutes for the bulbs to color stabilize. LEDs would be much better. But the scanner freaked out and imaged with ultra thin black lines in the image when I ran tests with LED tubes connected to 120V wall plug. My guess is because the CCD is expecting a flicker, I really am not sure.
 

Ian0

Joined Aug 7, 2020
5,158
My guess is that it wasn't expecting a flicker!
Can you work out from the scan speed and the distance between the lines whether they are due to 120Hz flicker?
Does the LED driver have a wide-range voltage input? If so, you could try rectifying the mains and smoothing it, and feeding DC into the driver. Then there should be no flicker.
Also, don't forget that cheap LEDs have a poor colour rendering index (CRI=60). Some of the tri-phosphor tubes had CRI>90. You can get LEDs with CRI>90 but they are a lot rarer.
 

dendad

Joined Feb 20, 2016
4,136
Does the machine run ok in just sunlight?
As @Ian0 says above, you can most probably run the LEDs on DC without the flicker so there will not be any strobing to take care of.
 

Thread Starter

laughingcamera

Joined Apr 27, 2022
11
Does the machine run ok in just sunlight?
As @Ian0 says above, you can most probably run the LEDs on DC without the flicker so there will not be any strobing to take care of.
The machine wits over 1800 lbs and is bolted to the cement floor, so no option to try it "in the sun". I want the flicker, not trying to get rid of it.
 

bassbindevil

Joined Jan 23, 2014
422
Try looking at the fluorescent light through a digital camera (or phone or tablet or camcorder). If the original tubes strobe, you should see some weird effects in the camera's LCD viewfinder. Try it again with the LED tubes.
 

bassbindevil

Joined Jan 23, 2014
422
You could measure the amount (and frequency) of flicker with a light sensor (phototransistor?) and an oscilloscope. Compare the fluorescent tubes with the LED substitute. I still wonder if the LEDs are turning off briefly and that's what causes the black lines; I've come across LED drivers that don't stay on 100% even on full brightness for some reason (bad firmware?).
 

Thread Starter

laughingcamera

Joined Apr 27, 2022
11
bassbiddevil, I think you have misread this thread, I WANT the LEDs to flicker at same rate a fluorescent tube does. I did purchase a flicker meter and the LEDs showed 0, while the fluorescents showed plenty.
 

Ya’akov

Joined Jan 27, 2019
5,666
Since the fluorescent lamps have persistence and the LEDs have none (or far less) it might require that you use a capacitor of the right value to emulate the fluorescent flicker. It would be helpful to characterize the flicker in terms of both frequency and amplitude so you can properly emulate the tubes. An oscilloscope and even a sensor, even an LDR or.a small PV would probably get you enough information and you could tweak the LEDs by comparing the tubes’ waveform to them.
 

Ya’akov

Joined Jan 27, 2019
5,666
Yaakov,
What about using flicker meter? Like this
It probably doesn’t provide the information you need. You want to emulate a particular waveform, it might coincidentally correspond to the percentage it indicates but it’s not because it is a necessity.

If you don’t have a scope, you could try to use the flicker meter and it might work, but it is my impression is that if the scanner is using the frequency, and that meter is only measuring the effect of the duty cycle where the light crosses a threshold (not zero), then the LED might have to flicker at a different frequency than the tube to get the same reading.

That’s why the scope seems much more useful since you will have to hold the frequency constant while adjusting the average intensity by reducing or increasing the minimum output from the LEDs.

This is all surmise, though, since I don’t know how the meter is determining percentage nor how the scanner is detecting the flicker.

You might get more information if you try both light sources at once and see if the scanner behaves.
 
Top