TTL Output by Arduino

Hi,

I am really new to electronics and programming. For a fluid dynamics project I would like to create a pulsed LED light source inspired by this paper:

http://www.researchgate.net/publication/228668240_Pulsed_operation_of_high-power_light_emitting_diodes_for_imaging_flow_velocimetry

I attached an image of the circuitry.

Talking to our electronics workshop, I have been told, that this setup would need a TTL signal with 5V.

Is it possible to create something like this with an arduino?

I might need to create pulses in the μs range (approx. 2 μs duration with a 15 μs delay). The LEDs need a rise time of greater than 0.5 μs from 10% to 90%. Can this be programmed?

I know these are a lot of specific question, but I hope you have some advice for me. :slight_smile:

Yes.

Look at analogWrite(). I believe it will get close enough to that timing specification. If not, you can write pins directly at that frequency.

Rise time is not programmable but it easily exceeds your spec.

Yeah - I don't think analogWrite() itself will do it, but you can definitely do it if you take over the timers.

You know, you probably won't need the gate driver if you use a modern (ie, logic level) MOSFET to begin with, though you'd need to put a little thought into it. How big a LED is this that you're switching, anyway?

In that circuit the capacitors are drawn the wrong way round. The open box is always positive and the black filled in box is always negative. There is no need for the diode D1 unless you are switching an inductive load, which you are not. You need a resistor between the output of U1 and the gate of the FET, to limit the initial switch on current to a safe level for U1.

Hi,

thank you very much for all your answers! :)

@Grumpy_Mike I will look up, what a FET does, as I am new to electronics, there is a lot of reading ahead. Here is some more information about the setup written by the "inventors":

"The operation of the current pulser is very straightforward: a bank of capacitors (C1, C2) with low internal resistance (ESR) is first charged through the power supply VS (typ. 20–50 V). The MOSFET power transistor (T1) is triggered via a MOSFET driver circuit (U1) which has a TTL compatible input. The LED (LD1) cathode is connected to the drain of the power transistor whose source is connected to ground via a low resistance power resistor (R2, typ. 0.01 – 0.1 Ohm ). In this configuration the circuit acts as a switched constant-current source for the load, that is, the LED. The voltage drop across this resistor VR2 is used to estimate the current flowing through the LED. Diode D1 protects the LED from potentially damaging reverse currents that arise during the rapid switching transients of the circuit. The purpose of resistor R1 is to limit the charging current of the capacitor bank in between the current pulses and it acts as a rudimentary method of overcurrent protection in case the LED is accidentally operated in continuous mode."

@DrAzzy I will use a Luminus PT-121-G-L11 http://www.luminus.com/products/Luminus_PT121TE_Datasheet.pdf 3,3V @ 150A in pulsed mode (2 μs duration) Can you explain, what a gate driver is? Is it in the schematics I attached? What do you mean by "taking over the timers"?

@MorganS What do you mean by exceeds my specs? Can I make it rise slower than 0.5 μs? This is an excerpt from the paper that details how the rise time can damage the LED:

"Even though the LED can withstand a continuous train of 200 A pulses of τp = 7.5 μs with rise times of about τrise ≈ 1 μs, it will be immediately damaged when the rise time is shortened below 0.5 μs. The data sheet for the device also recommends rise times of greater than 0.5 μs from 10% to 90% of forward current."

Thanks for all your time and advices :)

You could control rise time with drive current/gate resistor. I don't see why you want such a high supply voltage (20-50v). The datasheet states an absolute max forward voltage of 5.9volt for the green LED. A more common (open frame) 12volt supply would be fine. U1 VCC can also tap from that. Leo..

You are right about the LED supply voltage, I mixed it up with another one.

I will try a 12V supply.

Is it possible that the mentioned U1 is capable to “soften” the rise of the TTL signal and thereby allow the setup to be run with TTL signal, which would usually destroy the LED?

All in all, I think the consensus is, that this setup can be controled by an arduino. Which one should I get? Arduino Uno?

Hi,
The UC37322P is a high speed driver for MOSFETs, its output configuration is designed for high gate currents that will flow with fast rise times.

The IRFB3206PBF is also a high speed device.

I’d leave the circuit as is, just make sure you use decent sized conductors in the wiring between the LED and the storage caps to maximise energy transfer and flash speed.

Tom… :slight_smile:

IRFB3206PBF-datasheetz.pdf (446 KB)

Diode D1 protects the LED from potentially damaging reverse currents that arise during the rapid switching transients of the circuit.

That line shows you that the people writing this had little real knowledge of electronics. It is rubbish.

Strouhal: Is it possible that the mentioned U1 is capable to "soften" the rise of the TTL signal and thereby allow the setup to be run with TTL signal, which would usually destroy the LED?

Question is, what the writers imply with TTL signals. As a switching level description the Arduino output pins supply TTL signals. As current source/sink capabilities, a TTL output hardly can kill a LED, as can't Arduino ouputs - more probably the output circuit will be killed before the LED. This is why frequently a power transistor (FET) is added for driving the load.

PWM operation reduces power dissipation, in contrast to continuous (analogous) regulators. For that purpose the rise and fall times of the output signal must be as short as possible, smooth slopes will increase the power consumption and tend to kill the output power transistor, be FET or bipolar.

A gate driver allows for fast signal slopes on the gate (short switching times), where power FETs have a high input capacity, that requires high currents during signal level changes. These high currents can damage the driving output pin, so that often a current limiting resistor is placed into the digital-output-to-gate connection. That resistor again will increase the switching time, what's unwanted until dangerous. So the perfect solution will use another driver stage between an digital output and the gate of the power FET.

Finally I'm not sure that the LED requires longer rise times than 0.5µsec. If so, you can omit all circuitry between the output pin and the gate of the power FET, and replace it by the current limiting resistor mentioned before. When that resistor is too small, the digital ouput may be damaged (270 Ohm should be a safe value). When too high, the FET may become hot when it lights the LED.

DrDiettrich: Finally I'm not sure that the LED requires longer rise times than 0.5µsec. If so, you can omit all circuitry between the output pin and the gate of the power FET, and replace it by the current limiting resistor mentioned before. When that resistor is too small, the digital ouput may be damaged (270 Ohm should be a safe value). When too high, the FET may become hot when it lights the LED.

The fet listed has a fairly high gate capacitance. About five times the ~30A fets we usually use for Arduino projects. Gate resistor for this fet for ~0.5usec risetime falls into the 22-33ohm range if my calculations are correct.

A default 1Khz PWM output, set to 1/256, should output a 2usec needle pulse. This is a fixed 1000 pulses/sec. Do you want the pulse frequency to be adjustable? Leo..

@Wawa

I will need two pulses of light to create two images with my camera, which will be cross correlated to create a vector field of particle movement. To obtain the best results with the least amount of stress on the LEDs I would like to experiment on both the duration of the light flashes, as well as on the delay between each pulse.

You mention the "gate resistor" has approx. 22-33ohm, I guess this is part T1 in the schematics, what does this imply?

@DrDiettrich

I can only tell you, what the LED manufacturer writes in his datasheet: "In pulsed operation, rise time from 10 to 90% of forward current should be larger than 0.5 microseconds"

This is not about overall power consumption, Tjunction is a different topic, which I will try to deal with appropriate cooling, long enough delays between pulses and lowering of LED current.

I am new to this but I think you would suggest a resistor of approx. 270 ohm between the TTL signal and U1, right?

When you follow the article, use exactly the indicated parts and values, and everything should be fine.

When you replace the U1 MOSFET driver by a gate resistor, you'll have to determine the best value yourself. My 270 Ohm guess may be too high for the required current, you also can try the beforementioned 22-33 Ohm, or something in between. A scope will be helpful in fine tuning, by inspection of the resulting pulse on the LED.

Mosfets are voltage devices. They don't need gate current to stay on or off. Gate current is only needed to change from one state to the other. Special mosfet drivers are used to provide a lot of gate current to switch the mosfet as fast as possible. The thing that slows the switching is gate capacitance. The "gate capacitor" has to be charged/discharged with the driver. Fatter mosfet = bigger gate capacitance = more drive current needed for the same switching time. So you can control switching time with gate current (gate resistor). That sometimes is not wise, because midway the switching cycle the mosfet has to dissipate all the power. That could send it to silicon heaven quickly. Leo..

Strouhal: I can only tell you, what the LED manufacturer writes in his datasheet: "In pulsed operation, rise time from 10 to 90% of forward current should be larger than 0.5 microseconds"

Right, it looks like too steep slopes can kill a power LED :-(

If so, you should be skilled and equipped to check and adjust all critical items in the signal and circuit. Just the board layout is crucial then (see the hints in the article), breadboarding is not helpful, only dangerous.

Depending on how much money you want to invest, the cheapest solution looks like getting a ready-made LED module, perhaps available from the manufacturer (evaluation board). Even then you may happen to kill a couple of LED or Arduino boards, until you have acquired enough knowledge about "high current electronics in practice".