High Current High Speed LED Driver?

I presume you intend to switch at 10Mhz using very short pulses at high current with a relatively long off time to avoid LED die self heating and consequently avoid decreased efficiency and light output.

If so, high current drives (100mA-1000mA) are fairly easy to implement with discrete components and a buck converter IC configured as a constant current source. With constant current sources, an appropriate means to pulse the LED is to connect a FET across the LED terminals and pulse the gate. Another means (10mA - 200mA) is to use a constant voltage supply and pulse generator with fast logic coupled to the LED using a switching transistor or FET. The only worry here is to ensure all components used can switch at 10Mhz and have appropriate dV/Dt to ensure clean switching with an appropriately short pulse width.

But keep in mind, discrete components or common ICs can easily be used to design a driver for almost any LED switching application. So if the above link is not suitable search the internet for LED switching circuits.

Your research is intriguing since light is not a typical means of transmission in water due to exponential attenuation, scattering effects and attenuation due to particulate concentrations to name a few. However, applications of light for this purpose are heavily researched in an attempt to overcome these obstacles and some interesting results are being found.

If your approach is to use modulation techniques with the LED, then the above advice does not apply directly (closely related). You'll have to search for LED modulation circuits on the net.

Good luck with your research.

Don't bother trying you will find the practical means are far to problematic as described by the last person there are so many ways light is made to reduce in intensity under water. Even high power lasers are prone to unacceptable losses. This is why sound and very low frequency RF have to be used to communicate under water. LED's lose out at very high frequency because they can not react well enough at above a few hundred Khz. The only way they work is to leave the LED lit and use special optical means to modulate the light.

High School and 10MHz LED driver. Hum !

How far do you think your LED will transmit the information? Try some LASER diode and perhaps that may go through if its wavelength is proper for type of water in which you are going to place it. Perhaps 10MHz ultrasound is a much better idea. You may handle signals to few hundred meters.

Which LEDs you are using? Most of them won't oscillate light at 10MHz properly. Did you check the specs. What is the current you intend to drive?

If you use simple transistors like 2N3904 or 2N2222 then you may easily get that 10MHz drive capability. You can also use 74F06 or 74F07 ICs to drive with 10MHz frequency easily.

If you still find that you can't drive the LEDs at 10MHz then write to me. I can make one circuit for you and can send it. You need to provide me DHL or FedEx or UPS account of your institute on an official letterhead for free shipment. You will get 10MHz driver for free if requested officially.

Hey gallamine,

if you want to experiment with cheap drivers you should have a look at application note AN1122 (fiber optics) of avagotech (former agilent former hewlett packard). You find a simple and fast schematic with the cheap paralleled 74ACT00 which you still get in DIL14 or SO14 packages.
Regards Uwe

Walk before you run.

1. Does your water environment provide adaquate optical characteristics for your applications?

2. How brite does the light from the LED have to be? How large does the LED have to be and how much current does it require?

First walking step. Seperate the variables. Try to communicate at a much lower frequency, say 100 KHz, in order to remove frequency as a problem. Just set up a simple experiment to check if the communication from the LED to the detector works. Try to seperate the two variables "optical characteristics" and "light intensity".

If the optics are so poor that "it just doesn't work" or "light intensity required is not practicle" then abort this idea.

If the optics function with a reasonable "light intensity", try to determine the amount of light intensity required including some saftey margin.

Second walking step. Determine the LED configuration required, keeping in mind some of the following considerations:

A. LED light conversion efficiently (light intensity vs. current) degrades with "time and use" and should be factored in to the saftey margin.

B. Most likely, the speed, for switching the light on/off with an LED, decreases when the "LED die size" increses and when the "LED current" increases. You might try using multiple LEDs instead of a single LED.

Think about this, 2 smaller LEDs might equal the same light intensity as a single LED with each smaller LED only requiring approx 1/2 the current of the single LED. And guess what, if the 2 smaller LEDs are connected in series, the required current is only as much as the smaller LEDs which is approx 1/2 the current of the single LED. Since the voltage across the LEDs is small, current is what is inportant and the driver circuits to drive current throught the LEDs typically can be faster at lower currents.

Third walking step. If you are still standing at this point, then you have determined that the optics are functional and the LED configuration. All that is left is the electronic current switching circuits, which I consider the easy part because this kind of circuit can easily be made to run at 10 MHz with the right design. Probably the most important feature of this circuit is not to use a driver transistor or FET that operates in the saturated mode but make sure it stays in the linear operating range. If this part is a little deep for you and you are still interested, let me know how to contact you in another message.

Now you can try to run.

Dennis

I'd like to come back to this threat, because the actual question keeps me occupied and it has not been answered properly (at least from my point of view).

So the question is, what would be good high current (from 1 to 5 amps) and high speed (1MHz) LED driver. I prefer simple drivers, which need as few external parts as possible.

My task is little bit different than the original but the question is same. I need short (1-100 microseconds) light pulses to "stop" moving target to take a picture. Because pulses are short, the light should be as bright as possible and I am overdriving leds (even 5A).

I am aware of problems the leds have in photographing, but in my case I can ignore them. So the main problem is to find (or make) high speed constant high-current source, wherein the current should be constant with different loads. So even if I change a white led to red one (and threshold voltage is very different) the current should be same.

The closest one I have found is TLC5922, but with that I only have ~1200mA. Is there better driver somewhere?

Thanks already in advance.

Well, you can have that 1-10A current switching for 1-100uS duration. What is the operating voltage for you? Are you switching an array at high voltage? You can switch 100 in series also. You need to have few thousands of dollars for it. I can design one for you. Idea is not all that bad.

can you tell me how to increased the brightness of dot matrix?

Hi!!

I've a problem a little bit differente, it's about a 3d-vision system based on laser leds.

More in particular I've to drive a current in a serie of 7 ir leds (osram sfh4232), in cw modality, modulating it with a sinusoidal signal characterized by a frequency of 7.5MHz.

The voltage forward of my leds is 1.5V, the max dc current 1A, so I'd like to drive a sinusoidal current that swings from 0 to 600mA.

I hope I've used an undertstandable english.

Regards.

Nino

Sorry the question is: do you know an IC driver or a schematic that could let me realize the application I described before?

Thanks again.

Nino