LM3410 and Multiple LED Strings

[modern_messiah]

Hi all,

I have used the LM3410 LED boost driver multiple times in the past, but I am busy doing something different and would like to know if the circuit below would function as I desire.

Basically I want to use the LM3410 to power 3 LED strings at different times. The LED strings will be driven at the same current, but because of the differing LEDs between each string, the forward voltage will differ. I need a way to properly switch between which string is being driven. I am doing this so that I can use 1 driver circuit instead of 3 separate circuits.

My solution is relatively complicated, so I'm open to improvements.

The biggest problem I face is ensuring that the output from the LM3410 boost circuit has no connection if an LED string is not connected. This is because the LM3410 circuit may be receiving power without a load connected. This should never happen, but I am using a mechanical switch to move between each string so the chance is there. The only way I can see to solve this issue is by isolating the output capacitor from everything until an LED string is activated. So in the circuit above you have 3 on/off inputs. Only one of these can be activated at a time:

- VIN is providing power to the boost circuit. No FETs are on. Nothing is connected to the boost circuit's output. The LM3410 is not doing any boosting then.

- FET1 is activated which also activates FET4. This connects LED STRING 1 and C_OUT to the output of the boost circuit.

- FET2 is activated which also activates FET5. This connects LED STRING 2 and C_OUT to the output of the boost circuit.

- FET3 is activated which also activates FET6. This connects LED STRING 2 and C_OUT to the output of the boost circuit.

The reason for the 3 FETs connecting to C_OUT is because if the LM3410 has power and C_OUT is connected but no LED string is connected I don't think C_OUT will last very long before the circuit destroys itself! C_OUT needs to be isolated from the 2 strings that are not connected otherwise it will induce a reverse voltage across the 2 off LED strings which will destroy them.

Apart from needing 6 FETs this looks as if it will work to me. I've identified this FET as possibly being suitable for the job. It's cheap and small so it may be possible to do what I need to do with this.

Good? Bad? Any recommendations would be highly appreciated.

Thanks,

- Matt

 

[goldfist]

The biggest problem I face is ensuring that the output from the LM3410 boost circuit has no connection if an LED string is not connected. This is because the LM3410 circuit may be receiving power without a load connected. This should never happen, but I am using a mechanical switch to move between each string so the chance is there.

In a situation where you need some small load vs no load, but you dont want to have a load that exceeds that of the devices down stream to be powered, I would think that an electrolytic capacitor across the circuit parallel to the load would avoid a backlash when the load is not present if I am reading this correctly. The electrolytic cap will demand a very small load to maintain its charge and so it doesnt go from full load to no load and then back to full load and lash it. When it goes back to full load the capacitor will have a reserve for the load to pull from so that the IC itself isnt getting slammed with a load and no load over and over again but instead maintains the charge of the cap to full voltage, the cap then cushions the immediate demand, and then the IC softly takes on the load of the LEDs ... if there is a condition to where no load is present in the switching process.

If there is concern that a mechanical slide switch can potentially connect more than 1 of the 3 to the power at a time, I could only see this happening if the switch was of poor construction to bridge more than the intended connections which are usually just 2 contacts out of 6 to control 3 on states with no off state. *The way I interpret this is that there is a short period of time in which no load is connected when the contact slides into position to connect the next 2 contacts, where this will cause a loss of load until the next contacts come into contact with the slide contact in which a capacitor can cushion the IC prior to the switch of this non-load condition which then turns into a sudden load condition from the next strand of LEDs.

 

[modern_messiah]

Thanks for the reply goldfist.

I took some advice from a friend and have ended up with this:

Now I know I need to ad some pull down resistors to the FETs. Some of the reasoning behind the above circuit is as follows:

1. There is now an over-voltage protection circuit (R5 and the second zener diode pointing up) to prevent a case where the output capacitor is connected but no LED strings are.
2. According to someone who knows better than me, having two LED strings connected at one time won't be bad. Because the LM3410 is a constant current LED driver, the string with the lowest Vf will be lit. The other string just won't do anything.
3. With the above point in mind I will simply design my switch so that as it moves from one LED string to the next, it will over-lap slightly. This might actually be pointless given I have incoporated the over-voltage protection though.

- Matt