And - if the frequency is so high that one cycle is shorter than the flash duraction, then the flash is just on all the time. The signal or device that is triggering the flash or manipulating contacts or whatever has two states. Call them on and off, high and low, open and closed, 1 and 0, whatever. What are the max and min times for each state? 4 numbers, each in milliseconds in milliseconds.

I don't think a capacitor-dump circuit of any kind will work here. You'll need a power supply that can make the required voltage and current constantly, and then switch its output to the LED load.

ak

 

Thanks for your reply. I was planning to use a 110v ac power source that converts to 12v dc. So it would have a power regulating ability. I was thinking about the capacitor-dump circuit you mentioned and I tried a small sample one, but even using a push button switch the LED did not turn off quickly after it turned on. This seems to be a key problem, turning off the LED when the switch hasn't been turned off. I didn't experiment with a bunch of capacitor/resistors. It turned off only when I pushed the button. At the max, one point will open probably 83 times per second near the max and probably 1.63 times per second at the minimum. You are right in that at the quicker rate the flash will appear to be lit up constantly. The LED on time therefore looks to be longer at the higher number, but I think in actuality it is no longer than the minimum time, it is just less frequent. So what I'm trying to do is get the bulb to flash at the instant the switch opens, turn off and remain off until the switch opens again, and on and on. I hope my math is correct. I believe the duration of the flash will remain the same throughout, it's only the frequency that changes. I'm not really certain about the duration of the flash. I would assume the duration of the flash would need to be similar to a timing or strobe light. Thanks for reading. I hope this makes sense.

 

Impossible! The human eye can see only 50 flashes per second at best!
You cannot even see the off periods at 83 per second.

http://www.100fps.com/how_many_frames_can_humans_see.htm

Televisions make pictures using one dot on the screen...

Light bulbs turn on and off 60 times per second.

 

Impossible! The human eye can see only 50 flashes per second at best!
You cannot even see the off periods at 83 per second.

http://www.100fps.com/how_many_frames_can_humans_see.htm

Televisions make pictures using one dot on the screen...

Light bulbs turn on and off 60 times per second.

Even the article you linked disagrees with you. It clearly states:

Tests with Air force pilots have shown, that they could identify the plane on a flashed picture that was flashed only for 1/220th of a second.

I know I can see the flicker of fluorescent tubes at 120Hz when looking out of the corner of my eye especially if I move my head while looking (remember, the voltage oscillates +170V/-170V in a single cycle, so the bulb is energized at a rate of twice the line frequency). This is a documented effect, and some individuals are more perceptive of it than others.

If you used an AC transformer to drop line voltage to 3V, put a bridge rectifier on it, and attached a white LED to it, you'd definitely see the flicker. It'd be even more noticeable if you didn't use a rectifier and just let the LED serve as a half-wave rectifier (so it's only flashing at 60Hz instead of 120Hz).

The only reason incandescent bulbs don't flicker at 120Hz is because they're incandescent; the light is produced as a product of the filament heating up, and at 120Hz the filament never cools down and thus produces continuous heat (and light).

A factor in this particular case is duty cycle. If the light is flickering at >50% duty cycle on a sine wave (as is the case with fluorescent lights), then yes, it will be very difficult to see due to afterimage, which is why I don't notice the flicker of fluorescents while going about my daily business. However, at say 25% duty cycle on a square wave the flicker will be more noticeable.

 

Even the article you linked disagrees with you. It clearly states:


I know I can see the flicker of fluorescent tubes at 120Hz when looking out of the corner of my eye especially if I move my head while looking (remember, the voltage oscillates +170V/-170V in a single cycle, so the bulb is energized at a rate of twice the line frequency). This is a documented effect, and some individuals are more perceptive of it than others.

The only reason incandescent bulbs don't flicker at 120Hz is because they're incandescent; the light is produced as a product of the filament heating up, and at 120Hz the filament never cools down and thus produces continuous heat (and light).

A factor in this particular case is duty cycle. If the light is flickering at >50% duty cycle on a sine wave (as is the case with fluorescent lights), then yes, it will be very difficult to see due to afterimage, which is why I don't notice the flicker of fluorescents while going about my daily business. However, at say 25% duty cycle on a square wave the flicker will be more noticeable.

Well then take a picture of it...100 times per second and react to it....I'm not arguing the LED rate, you just cannot view it and react.
The only thing that you may perceive is that the LED may look brighter or dimmer.

 

Impossible! The human eye can see only 50 flashes per second at best!
You cannot even see the off periods at 83 per second.

Not correct, but also not important. The TS has already stated that at high speeds the LED would appear to be on continuously, and that was ok. But even if you cannot perceive individual on and off cycles, you can tell the difference between a LED flashing at 50 Hz vs 100 Hz. They just look different. The limiting factor is the persistence of the retina, which increases with age.

Televisions make pictures using one dot on the screen...

Writing on a phosphor that has a relatively long persistence. It is this persistence that lowpass filters the scan rate to present the appearance of a stable image.

Light bulbs turn on and off 60 times per second.

If you mean incandescent bulbs, no. The thermal lag of the filament is many times longer than the period of the incoming energy. Also, that energy is a sine wave, so the bulb is "off" only at each zero crossing, which is a theoretical instant happening 120 times per second in the US.
Many LED bulbs have tiny switching power supplies inside, so they are flashing at kilohertz rates. Fluorescent bulbs with traditional inductor ballasts flash at twice the line freq, or 120 Hz (US).

ak

 

Tests with Air force pilots have shown, that they could identify the plane on a flashed picture that was flashed only for 1/220th of a second.

True, but not applicable here. That identification is relying on image retention in the retina to hold the image long enough for the brain to process it. But in the task of perceiving flicker, retinal retention is the last thing you want. It is a lowpass filter, not a good thing for detecting high frequencies. A leftover survival aid is the fact that as the image source moves farther and farther off the center axis line of the eye (more and more out to the peripheral vision), retinal retention decreases. This is why we can see fluorescent lights flicker out of the corner of the eye, but not when looking right at them.

ak

 

As usual this thread has wandered off on a largely unrelated topic about persistence of vision! This doesn't help the OP.

However, the OP doesn't help the people trying to assist him by failing to give a clear explanation of exactly what it is he is trying to make! It's as though someone wanted help to make an object, but just says "Well, I need a device, possibly with two round items, one at either end, which can take a load of about 75kg. It will need some form of crank-driven arrangement located below and behind a transverse bar which can be used to modify the direction in which the device will move..." Technically correct, but no damned use to people trying to help him design a bike!

WHAT IS THIS THING FOR? If you start by answering that question, maybe you will get some more useful answers. Why is it that we have to drag this information out of people who come onto this forum for assistance. Is it a government secret or something?

 

Hmmmmm. What I tend to see here on this site is a lot of hyperbola and very few ideas. This is so simple of a task that any of you "engineers" should be able to come up with a simple effective circuit in a matter of minutes. And NOT use a microprocessor! This one took 4 minutes to come up with and ten minutes to mock up on a breadboard. It gives a flash of the LED each time the contacts open. The FETs can be any small N channel enhancement FETs like some of the TO-192 case ones. They don't dissipate any power. The LED can be any common type since the 12 volts will easily forward bias it. Select the resistor R to get the length of the flash you want, and the 100 ohm resistor to brighten or dim it.

 

Finally in post 17 was the requirement for rapid turn-off. This eliminates not only a capacitor-dump type of circuit, but also anything based on an R-C timer that does not have a comparator or some other snappy circuit to assure a fast turn-off edge. The output from an R-C ramp followed by a transistor will not have a very fast edge, no matter how high the transistor gain or transconductance is. This brings me back to post 21, and a true monostable followed by a driver transistor.

After an embarrassing miss in another thread (it's the flu drugs, really!), I'll re-read this thread before posting a circuit.

ak

 

As usual this thread has wandered off on a largely unrelated topic about persistence of vision! This doesn't help the OP.

This is a discussion forum, not a service bureau, so discussions happen. Also, the OP isn't complaining. Also, I learned two new things. Also, 5 posts is not much of a detour in a thread that took 22 posts to get all of the requirements.

And, since the original topic is about making a light blink very fast, I think that persistence of vision and its ability to obscure any perception of blinking is entirely on topic. After all, why build a circuit no one can see?

ak

 

Ronsoy2, thanks for your input and circuit design. I will definitely check this out.

 

Hey Ronsoy2. I put together the circuit you listed. I appreciate you putting it up. I must be doing something wrong because I can't get it to work correctly. The components I tried two Mosfets IRF540N and then two Mosfets FQP30N06L. I tried various resistors in the place of the R* resistor 150 ohm, 2.2k , 4.7k 22k and 47k. The diode was a 1N4148, capacitor .01uf 50v polyester film and tried .01uf ceramic disk.

Here's what happens. When I turn on the power and with a normally open switch the light is on. When I press the switch the light goes off. What I need it to do is light up only when the switch first opens (on a normally closed switch). So the sequence would be light off -> circuit opens and light comes on (flashes) and goes off regardless as to whether the circuit is still open -> circuit closes, light is off. I need to be able to adjust the flash duration also. I only have one wire that I can run to the LED bulb. I do have the ability to ground the other side to a case, but can't run two wires to the bulb because of the physical restriction. So if it was possible to turn on the LED through some sort of ground circuit, say run a wire to the + side of LED then the negative side to ground that would be great. I appreciate everyone's efforts. Any ideas?