LM3409: Shunt MOSFET PWM with output capacitor

Hello David,

One thing this will depend on is Shunt FET dimming frequency.  There can be an output capacitor when Shunt FET dimming, it becomes switching loss for the Shunt FET.  It will also cause a turn-on delay when recharging the capacitor.  Larger inductor value will make lower current ripple requiring less capacitance.  The other option is to use standard PWM with a lower inductor value.  The turn-on time will be V = L * (di/dt), where V = Vin - Vled.  The turn-off will be the same equation where V = Vled.  If this is your low output voltage application the turn-off will be longer then you turn-on.

What is your PWM frequency?  Output voltage?  Output current range?  Input voltage?

Best Regards,

Hi Irwin,

The PWM frequency has not been specified by my client, but earlier versions of similar products can be switched at frequencies > 10kHz with a turn-on time < 10us.  I expect that a switching frequency of a few kHz would be okay.

The design specifications are as follows:

Programmable current range: 250mA to 4A

LED voltage range: 1.5V to 6.5V (at 90% efficiency, a duty cycle range of 6.6% to 30%)

Input voltage: 24V ± 5%

Maximum turn-on/off times need to be < 20 us or so.

My PWM frequency would be limited to < 300 kHz or so due to the 210 nS minimum on time of the LM3409.

Thank you,

David

Hello David,

I don't see the LM3409 reaching 90% efficiency with low output voltages.  This part become efficient when Vout approaches Vin.  A synchronous part makes more sense for this however the part we have, TPS92641, has the current sense on the low side.  You mentioned synchronous with the LM3409, since it is a P-channel drive referenced to Vin the gate drive circuit becomes a little more complicated.  I don't see an easy solution yet.

Best Regards,

Hi Irwin,

I'll put some design effort into coming up with an appropriate driver for the diode replacement MOSFET and see what I can come up with.

Thanks,

David