LED Strobe light circuit

[Bob-M]

I am repairing an old 1950's automotive distributor testing machine, that had a rotating neon strobe light. The strobe would flash once, every time the distributor's points would close.

I would like to replace the neon strobe with an LED. The LED needs be about as bright as a 4 cell Maglight flashlight, and be able to flash at a maximun rate of 300 flashes a second. There is a 6VDC power source on the machine. Any help with a schematic would be very much appreciated.

 

[crutschow]

Was the original strobe light a xenon lamp, rather than neon? Those put out a lot of light in a short pulse, so you may require quite a few LEDs to match that.

You could use high brightness white LEDs. You may need to do a test with a short pulse to see how many you need to get the desired brightness.

With 6V you could likely put two LEDs in series with a resistor to limit the LED current to the desired value. You could parallel as many of these two LED circuits as needed. All the LEDs can be switched by a transistor controlled by the points.

Do the LEDs need to flashed for a short period (shorter than the time the points are closed) to stop the motion of the distributor? If so, you will need a one-shot, such as with a 555 or 74121, to generate a short pulse to the transistor each time the points close.

If you want to vary the time position of the strobe light to see the points in various parts of the cycle, then you will need a second adjustable one-shot for that. For example, to view the points just before they close would require a one-shot time of up to one revolution (or 3.3ms for 300 RPM).

 

[Bob-M]

The original WAS a Neon...kind of orange in color, and not near as bright as a Xenon. Original parts are available but cost over $500.00 for a transformer and neon tube!
The strobe lights a pointer on a degree wheel so you can watch the mechanical advance in the distributor work. The flash should be short, to stop the motion of the pointer....like an automotive timing light. I don't need to vary the time position of the strobe. I was hoping that one of the 8 or 10 MM LEDs would be bright enough?

 

[crutschow]

At 300RPM one revolution it takes 3.3ms for one revolution. To freeze the pointer within one degree, for example, would require a light pulse width of .33ms, thus you would have to test the LED with that short a pulse at 300Hz to see how bright it is. One degree is only a 0.28% duty-cycle so the LED perceived brightness will be much less than the steady-state brightness. You may be able to pulse the LED with a higher current to improve it's brightness (the maximum pulse current is often given in the LED data sheet).

 

[Bob-M]

The machine turns at a maximum of 2000 rpm which = 33.333 revs per second = 12000 degrees per second = .0000833 second per degree. So I guess I'll need to push a lot of current through the LED to get it bright enough to see? From what I've read, you can push a lot of current if is a short enough pulse. I'm way more mechanical than electrical so I need some help here.

 

[Dr_Doggy]

......plus the points prolly wont go to 0 ohms....... what about if you used like a 555timer on each point as a trigger for a one shot?

 

[Bob-M]

I have very limited knowledge, but I did experiment, over 30 years ago, with 555s. Are you able to to get them down to .08 ms? Here is a little more info. The LED will only need to "fire" every 45 degrees for an 8cylinder engine, 60 degrees for a 6 cylinder, and 90 for a 4 cylinder.

 

[Dr_Doggy]

idk you hafta chk the datasheet, a capacitor could drag out the input pulse, if you have more current going in than out(of the cap)

 

[crutschow]

Yes, in my calculations I was confusing RPS with RPM. 300RPM is only 5RPS. Your calculations are correct. You would indeed need a 83.3µs pulse at 2000RPM for 1 degree of movement during the pulse. You would likely want to pulse the LED at its maximum rated current to get the maximum brightness at such a low duty-cycle. You will probably have to experiment to know how many LEDs you will need to get an acceptable brightness.

Note: You probably don't want to use a regular "white" LED since they generally use a phosphor to provide some of the light, and the phosphor may have a significant persistence which would interfere with a short light pulse output. Likely better to use a single color LED such as a high brightness blue.

You might also consider the LEDs they use for a cell phone camera flash. They put out a fairly bright flash. You could possibly use an LED flash driver such as this. But it's not clear to me if those LEDs are fast enough for your requirements.

 

[Bob-M]

Good thinking on the "white" LED...color is not important. Is a 555 capable of a 83 us pulse? Once again, I'm not an electronics guy, but, could I build a 555 pulsed led circuit using a large 10 MM LED, and then just keep increasing the voltage untill it was either bright enough, or, it burned out, at which time I could lower the voltage and add additional LEDs untill they became bright enough? Kind of like trial and error??

 

[crutschow]

I believe the 555 can go down to 10µs. The 74121 can go below 1µs,

You should not increase the current above the LEDs maximum rated pulse current otherwise it may fail. The light output will probably not increase significantly above that point anyway.

Incidentally you don't want to operate an LED with just a voltage. They alway should have a current limiting resistor in series with them. Without a resistor you will likely zap the LED.

 

[Bob-M]

OK, So now I plan to use a 555 timer, and 1 or more 10MM blue LED (40,000 MCD. Operates 3.2V to 3.8VDC, current is 20mA). Can some one come up with a schematic for me? Please include component values and part numbers if possible. I'd like to use over rated components, where I can, so I won't have to keep changing transistors , etc. if I need to keep adding LEDs for more brightness.

 

[Dr_Doggy]

google: one shot circuit; and use a "555timer calculator" for RC values

 

[Bob-M]

Thanks doggy_ The "Doctronics" site has everything I need. I sure appreciate all the help you and Crutschow have given me. I wish there was something I could do for you guys...Bob

 

[Ubergeek63]

Good thinking on the "white" LED...color is not important. Is a 555 capable of a 83 us pulse? Once again, I'm not an electronics guy, but, could I build a 555 pulsed led circuit using a large 10 MM LED, and then just keep increasing the voltage untill it was either bright enough, or, it burned out, at which time I could lower the voltage and add additional LEDs untill they became bright enough? Kind of like trial and error??

get a power LED or a flash LED. lens it to a tight focus and discharge a capacitor into it with just enough resistance in line to stay under absolute max