LED Flash Design - 4A, 5V design

[frankqt]

Hi I like to design a LED flash that can pump up to 4As.

I have looked at the existing flash drivers from Exar and similar and there are great ICs for Flash (including 4.5A) but the problem is the control. Their on/off time cannot be as small at 10-20usec. (Usually 100mA or more) I am looking for a solution where I turn on a powerful LED (or LEDs) for 20usec and pump the maximum current. This is for high speed photography.

A reference schematics would be great.

Thx,
F.

 

[amu_cpln]

https://www.instructables.com/id/Hack-a-flash-camera-into-a-emergency-strobe-light-/#step1

that shows a camera flash construction .. the time constraint is a matter of choice of switching transistor

 

[frankqt]

I need a professional reference design.. this is a good hobby site.

 

[mister_rf]

If you don’t find some dedicated IC driver, you may improvise using discrete components.
But in order to do that need to define design specification:
- number of the LEDs and their parameters
- the power supply voltage to be used for.

 

[frankqt]

I will drive series of three LEDs. Their respective max currents are 1.5Amp and voltage drops are 2V. I will provide from a 1.8V IO the start and stop signal. My supply power is 3.3V, I can use a booster. The critical thing is timing and constant current through the LEDs.

 

[sky_123]

Hi Frank,

(Apologies for the digression from your requirement):
I wonder, at 4.5A for 20us, is that enough light to expose the image? A solution is to use a high-power laser which fires for fractions of a second, and use lens to illuminate the object. It sounds interesting what you're doing, do report back if an LED works.

 

[amu_cpln]

I will drive series of three LEDs. Their respective max currents are 1.5Amp and voltage drops are 2V. I will provide from a 1.8V IO the start and stop signal. My supply power is 3.3V, I can use a booster. The critical thing is timing and constant current through the LEDs.

What is the current capacity of your source .. if your led drops only 2 volts .. (and your voltage source is capable of sourcing 4.5A ) you simply need a constant current source circuit
the 10 - 20 usec timing is large enough for mosfet with turn on time in order of 5 to 10 nsec

you can control the on/off at the gate of mosfet

 

[frankqt]

Thanks this circuit is really helpful however what I need is to build a charge pump that can pump this type of amp from a 3.3V source. How can I build a charge pump such at that one.

 

[amu_cpln]

this could help

 

[poorchava]

Problem is that ordinary LED won't work that great. I've experimented with this kind of flash circuits (actually a strobe light). I used to dump a large loaded capacitor into 10W white LED. The problem is that white leds are in reality blue leds with phosphorous layer utilizing Stokes Shift of wavelength (significant difference between absorption and emission maxima). Even with very rapid turn on (dumping through BTJ, N-MOSFET and SCR) the phosphorous has quite significant time constant which defeats the purpose. I think LED technology still can't beat good old "dump bazillion farads capacitor into transformer primary with secondary hooked up to discharge lamp" approach.

 

[frankqt]

Problem is that ordinary LED won't work that great. I've experimented with this kind of flash circuits (actually a strobe light). I used to dump a large loaded capacitor into 10W white LED. The problem is that white leds are in reality blue leds with phosphorous layer utilizing Stokes Shift of wavelength (significant difference between absorption and emission maxima). Even with very rapid turn on (dumping through BTJ, N-MOSFET and SCR) the phosphorous has quite significant time constant which defeats the purpose. I think LED technology still can't beat good old "dump bazillion farads capacitor into transformer primary with secondary hooked up to discharge lamp" approach.

Would you mind sharing your circuit? I really like to try this. I don't have this anti stokes problem you are referring to.

 

[godfreyl]

Thanks this circuit is really helpful however what I need is to build a charge pump that can pump this type of amp from a 3.3V source. How can I build a charge pump such at that one.

It would be a lot easier if you drive the LEDs in parallel, rather than in series. Then they can be driven directly from 3.3V by a single MOSFET, with resistors to set the currents.

If we assume 2V across each LED, and (say) 0.1V Vds across the MOSFET, then there will be about 1.2V across the resistors, so they will need to be about 0.8 Ohms each, for 1.5A current.



---------- Post added at 12:15 ---------- Previous post was at 12:09 ----------

The problem is that white leds are in reality blue leds with phosphorous layer utilizing Stokes Shift of wavelength

Only a problem if you use a white LED. Wouldn't a combination of red, blue and green LEDs be good enough for illumination?

 

[poorchava]

Yes, it would be but that would increase the cost. Also, you can't really ged a good colors representation in camera with this kind of flash, because if has noncontinuous spectrum. Ever wondered why all the better camera brands (Niko, Canon, Minolta etcetc) have discharge tubes instead of LED?

As for the project there isn't really a one. It was simply wired because i needed a strobe light for some measurement. Then I've experimented a little and torn this stuff apart.

IIRC circuit was composed of some random pic micro generating pulses (extremly simple code), a capacitor, some power resistor, bench power supply, some large BJT, MOSFET, SCR with whatever gate/base drive circuit that i could assemble in few seconds.

 

[frankqt]

My idea is to use 4-6 of these LEDs. I think it will be possible, I will let you know.

 

[JoyceHerstad]

- number of the LEDs and their parameters
- the power supply voltage to be used for.




-----------------------------------------------------------------------



**broken link removed**
**broken link removed**
**broken link removed**