Hi,
I am trying to do a current LED driver using TPS40192. I need use PWM for dimming the output. I tested some ideas, but I didn’t get a good result.
Features
• 10-V to 14-V input range
• 13.5-A fixed output, with 200Hz PWM dimming (0-100%)
Could I use the EN pin to do this? I think that the startup sequence time (6 ms) does not accept 200Hz.
There is another way to do this?
Danilo Americano de Carvalho
Danilo,
As you note, the soft-start sequence for the TPS40192 is not going to support PWM dimming via the Enable pin well, however there are a couple of other possibilities:
1) Provide a pull-up on the FB pin such that the programmed LED current is 0A when the PWM circuit is LOW. This can be done by:
Here R1 is sized with Rfb such that when the P-channel FET is ON (PWM DIM = Low) Vfb = 0.7V with the ouptut of the Isns amplifier = 0V or R1 = Rfb * (Vreference - Vfb) /Vfb.
In this way, the voltage across Risns is programmed for 700mV / Isns Gain when PWM DIM is high and 0mV when PWM DIM is low.
This will almost definately trigger PGOOD to go low during a transition unless the edge at R1 is slowed to within the bandwidth of the error amplifer. This will also rely on the error amplifier to respond to the feedback change and PWM Dim needs to be low enough on a LOW to turn ON the P-channel FET but high enough on a HIGH to turn the FET OFF. Typically this will require a 5V PWM signal and a 2.5 - 5V refeence voltage. A TL431 shunt regulator can help produce a 2.5V reference from the IC's on board LDO available at BP.
2) For lower voltage PWM DImming signals, it may be necessary to use an N-channel FET to drive current into the FB pin. Here I would recommend a diode be added, though this introduces error into the programming of the 0A setting which can slow the recovery by over-driving the error amplifier.
Risns and the Isns Gain non-inverting amplifier are set-up to provide 700mV at FB when PWM DIM is high and the diode prevents the added circuit from sinking current. R1 is sized as before, but the effective reference voltage is reduced by the forward drop of the diode @ Vfb / Rfb (this current can be pretty low, so it helps to look up the diode drop at that current so you're not over-driving the error amplifier too much.
You could add a second resistor between the FET and the R1 & diode junction with a capacitor to ground to slow the rising and falling edge of the PWM signal into feedback, but make sure the diode will not be forward biased when the MOSFET is driven ON by the PWM.
