Constant Current LED Driver?

[sdfasdfsdfsd]

Hello,

I'm trying to build a constant current boost converter to drive an LED array with Vf = 40V at a current of 200mA from a 12V supply. I couldn't find any off the shelf components which met my requirements so I thought of using a standard boost converter and adjusting it to work as a constant current driver.  Are there any things I should take into consideration when attempting to do this with a low-cost LMR64010 boost converter IC. I have so far done this by choosing an R2 value which generates the 1.2V feedback voltage at 0.2A current.

I have simulated this circuit below but do I need to put a Zener diode somewhere in here to prevent voltage runaway if the load is disconnected? Would I need to use a feed-forward compensation capacitor too since there seems to be some instability with the feedback loop?

Thanks

[Zero999]

Yes, it should work.

A zener diode would have to dissipate more power than the LEDs, so is a bad idea.

A crowbar (zener+SCR) would be much better, but a fuse is required to protect the wiring and the circuit might still be damaged, if the LEDs are disconnected. Make sure the Schottky diode is big enough not to fuse, before the fuse.

[sdfasdfsdfsd]

Yes, it should work.

A zener diode would have to dissipate more power than the LEDs, so is a bad idea.

A crowbar (zener+SCR) would be much better, but a fuse is required to protect the wiring and the circuit might still be damaged, if the LEDs are disconnected. Make sure the Schottky diode is big enough not to fuse, before the fuse.

Thanks for the suggestion. What Zener voltage should I specify in the crowbar circuit, is 40V sufficient or is that too high given the absolute maximum output voltage of the IC is 40V? Also, can you achieve the same by using a PTC fuse in series with a MOV/TVS diode perhaps?

Lastly, are there any other simpler ways to do this or are there any off the shelf ICs you recommend?

[Zero999]

Yes, it should work.

A zener diode would have to dissipate more power than the LEDs, so is a bad idea.

A crowbar (zener+SCR) would be much better, but a fuse is required to protect the wiring and the circuit might still be damaged, if the LEDs are disconnected. Make sure the Schottky diode is big enough not to fuse, before the fuse.

Thanks for the suggestion. What Zener voltage should I specify in the crowbar circuit, is 40V sufficient or is that too high given the absolute maximum output voltage of the IC is 40V? Also, can you achieve the same by using a PTC fuse in series with a MOV/TVS diode perhaps?

The zener should be a little lower than the maximum voltage rating of the boost converter, but significantly above the maximum forward voltage of the LEDs.

A MOV/TVS is no good, because it doesn't latch on, with a low on voltage, like an SCR triggered via a zener diode does. A PTC fuse is far too slow. Use a real one-shot fuse, or an electronic fuse.

Lastly, are there any other simpler ways to do this or are there any off the shelf ICs you recommend?

I wouldn't be surprised if there are off the shelf ICs, but I'm not aware of any. I have made a similar circuit with a blocking oscillator and another transistor shorting the base, when the voltage across the sense resistor exceeds V_BE.

[kjr18]

I guess there are some of them, for example MAX25612. You will need some more components for this one.

Also you could try parametric search on your preferred site, digikey, farnell, whatever you want to use. That way I found this one example. Maybe you will find something else you could use.

[John B]

I have simulated this circuit below but do I need to put a Zener diode somewhere in here to prevent voltage runaway if the load is disconnected? Would I need to use a feed-forward compensation capacitor too since there seems to be some instability with the feedback loop?

Thanks

Instead of a zener conducting the full 200mA, it could be tied between the output and the feedback node, then the current sense node would be connected to the feedback pin by a high value resistor (>1k) to limit the current through the zener. You would specify a zener to clamp above the operating voltage of the LED string.

[Zero999]

I have simulated this circuit below but do I need to put a Zener diode somewhere in here to prevent voltage runaway if the load is disconnected? Would I need to use a feed-forward compensation capacitor too since there seems to be some instability with the feedback loop?

Thanks

Instead of a zener conducting the full 200mA, it could be tied between the output and the feedback node, then the current sense node would be connected to the feedback pin by a high value resistor (>1k) to limit the current through the zener. You would specify a zener to clamp above the operating voltage of the LED string.

That's a good idea. It would add a voltage limit to the regulator, on top of the current limit.

[sdfasdfsdfsd]

Thanks for the suggestions. On another glance, the absolute maximum SW voltage of the IC is 40V. Since the LED array has a Vf of 40V, the Zener doesn’t have much room to protect the output when connected to the FB pin and might cause instability.

Alternatively, to do LED boost constant current conversion with the lowest number of components, could I just use a standard step up converter topology in series with a linear constant current regulator IC? Thanks for the Maxim part suggestion, ideally looking for a synchronous switching regulator to not have any external FETs.

[Zero999]

Thanks for the suggestions. On another glance, the absolute maximum SW voltage of the IC is 40V. Since the LED array has a Vf of 40V, the Zener doesn’t have much room to protect the output when connected to the FB pin and might cause instability.

Alternatively, to do LED boost constant current conversion with the lowest number of components, could I just use a standard step up converter topology in series with a linear constant current regulator IC? Thanks for the Maxim part suggestion, ideally looking for a synchronous switching regulator to not have any external FETs.

If the forward voltage of the LEDs is near the maximum rating of the switch, then it's too close. Components should never be operated close to their maximum ratings.

You could add another, higher voltage MOSFET, forming a cascode, with the internal switching transistor. Gate to +12V, source to the SW pin and D1 & L1 to drain, rather than SW.

[sdfasdfsdfsd]

Thanks, I'll also look for boost CC drivers with slightly higher output voltage capability.

[Zero999]

Thanks, I'll also look for boost CC drivers with slightly higher output voltage capability.

Here's the basic circuit I've used in the past. I don't remember what component values I used. It was around 18 years ago.

[Zero999]

Now I've found a bit of time. Here's a schematic showing the suggestions above.

The zener diode is 51V.

The MOSFET is rated to 45V, but its gate is biased at 12V and is in series with the internal switch, so it will have less than 45V across it, when off.

The Schottky diode needs to be rated to more than 200mA.

The schematic can be simplified by adding n=11 in the value2 field of the LED, rather than drawing 11 separate LEDs.