LED DC boost circuit

[Johnny147]

Thanks ! Now I understand and I know what to do
I'll use fixed resistors with lm 317

 

[Cyparagon]

He may fry the diode at setup if something is unaccounted for.

How is this exclusive to a dropper resistor setup? I hope you're not suggesting a lm317 is immune to user error.

If you're going to insist on making a circuit idiot proof, increasing the complexity is most often counter productive.

 

[FireMyLaser]

A low ohm resistor at a low current is acting more like a wire than current limiting. For that setup to work, then the driver, resistor and diode voltage drop must be in harmony, or else the current will be way too low or way to high. It is very unlikely that anyone will get this right on the first try. And whichever way it goes is a coin flip. Not good odds to bet an expensive diode on, when actual voltage drop is unknown. I am helping this gentleman not to mess up.

With a constant current setup, what may happen is either the voltage supply is too low and the diode don't get full brightness, or voltage is excessive and waste some energy, but the diode will be fine. User error was not the issue here.

 

[Johnny147]

I received my components and 488nm 90mA diode.

And did a quick try : adjusted the boost circuit to get 90mA in a 50 Ohm R.
Then I added heatsinked laser diode + Ampmeter to that loop.
Now there's 90mA going in the diode and it increase by only 2 or 3mA as the diode warms. It seems very stable.

Only the resistor get too hot to touch and I'm scared it fry and become more conductive.

Next going to try 130mA (not higher) with
Maybe adding just larger resistor and cap

Is it risky ? As diode is stable from current and temp.

( I will make a pointer of this and 3D print a circular holder to center board and spring inside host tube )

 

[FireMyLaser]

What is the voltage over the resistor? If you get 90mA then it should be 4.50V, and 0.405W. Is your resistor rated for that? Also where are you putting the Ampmeter? It should be in series with the diode, not with the input of the boost.

 

[Johnny147]

Thanks !!
7,62v x 0,110mA
0,83 W
Old 3mm thick resistor seems 1W

Can laser stays ON 100% duty cycle as long as brass heatsink doesn't go higher than 50°C ? ( datasheet says 60°C max )

 

[Johnny147]

It's 5,7V so 0,6W that should be fine for a such large resistor. That R is hot but not boiling
What is looks like now

 

[FireMyLaser]

7,62v x 0,110mA
0,83 W
Old 3mm thick resistor seems 1W

Those measurements don't add up if you are using a 50 ohm resistor, but I'll just guess you are in the general range anyway. Maybe very high temperature coefficient, and/or the resistor need to be larger. If the datasheet says max 60°C, then anything below that should be ok.

 

[Cyparagon]

It is normal for resistors to get very hot at their dissipation rating. 150C is a common spec.

 

[Johnny147]

My 488nm diode fryied, went very dim.
I increased the power to 200mA as diode heatsink wasn't hot ( saw a graph here where that diode was pushed double higher )
It seems to happened when switching ON-OFF-ON but it was also lukewarm ON for 3 minutes before.
So the 'driver' was just the voltage boost + resistor.

I wonder if it fryied because of :
- current spikes
- capacitor lack
- heat of resistor, conducted right into the diode input pin (that scared me from the beginning and wondered if that would also happen if used LM317)
- just pushed too hard (diode well pressed in brass, was not so warm)

Is it possible to guess ?
Thanks !

At a point, the 1W resistor unsoldered itself by heat. I replaced it by 4x 1/4W parralel resistors which were getting as hot but stayed in place.

 

[FireMyLaser]

If it happened on start up of the laser then it is probably the driver overshooting, if no other factors was involved, such as ESD. A larger cap would help damping this but it may depend on how bad the overshoot is. I believe a constant current regulator would help this further. Do you mean ON-OFF-ON in the timespan of a few seconds?

 

[Johnny147]

Yes multiples quick ON OFF in a few seconds, looking for bad contacts in the switch.
It was shielded in a tube so probably not ESD.
Thanks. Will do better next time

 

[Buffo]

If it happened on start up of the laser then it is probably the driver overshooting

I don't think he is using a driver, and yeah, that's why the diode failed.

It was pointed out earlier in the thread that using the voltage boost circuit alone was insufficient to protect the diode from over-current spikes and that he should include a constant-current driver after the voltage boost circuit. (Or simply purchase a constant-current driver that also included the voltage boost circuit.)

In the operational picture posted above I only see the voltage boost unit - no driver. Running in that configuration is a sure way to kill the diode from over-current. (Especially on power-up, but also on power-down, as the inductor on the voltage boost board will induce a downstream spike when the field collapses.)

It doesn't take a lot of inductance in the circuit to create deadly current spikes. I killed several A140 diodes (which were remarkably robust otherwise) because there was a bad run of Flexmod P3 drivers that included a wire-wound current sensing resistor instead of the specified metal film resistor. This resistor was tiny - just .2 ohms - but the fact that it was wire-wound meant that it added enough inductance in the circuit to create severe current spikes in the output. Here's a 'scope trace. And the full post on imgur if you want the details.

Bottom line: Either this driver from DTR or the low power Simpledrive I from BBE (no link - have to contact him direct) would work just fine and would also omit the need for the voltage boost board.

Adam

 

[FireMyLaser]

In this context I choose to call his boost a driver but yes, the warning was made.

 

[Johnny147]

Thank you very much everyone ! Everything makes sense now