High capacitive load

[Rusttree]

I just had a PCB made for a new project I'm working on. Essentially, a uC turns a load on or off through a P-channel FET. The circuit is powered by USB. The uC is powered directly from the USB 5V. There's a 100mΩ shunt resistor in line with the source to current sense. See the attached pseudo-circuit.

When the load is turned on, the initial capacitance of the load causes a transient drop in the 5V source. The shunt resistor doesn't help the situation, either. Unfortunately, the drop is enough to brown-out the uC. No other components in the circuit are sensitive to the voltage drop, just the uC.

I'm trying to think in terms of saving the PCB if I can. Would a large, low-ESR capacitor across the uC power pins be a practical solution here? If so, how do I choose the capacitor value (other than just going with the biggest one I can get)?

Is there another clever technique I should consider?

Thanks!

 

[crutschow]

A few possibilities:

Add a large capacitor across the μP pins as you suggested. The cap should be much larger than the capacitive load you are driving. Also add a 0.1μF ceramic if you don't already have that.

Add the capacitors with a small resistor in series to the μP power (resistor value such as to cause no more than a few tenths of a volt drop at the μP supply current draw). That will allow the use of a smaller capacitor.

Connect the μP power to the other side of the 100mΩ resistor.

 

[Mr RB]

Yep the caps are important!

Also, almost every commercial FET driver I have seen has a small resistor between the logic output pin and the FET gate, it's good for the FET and the micro. Somewhere around 22 ohms to 47 ohms is normal but if you are not needing to switch at a high frequency even 100 ohms is fine.

 

[Jaguarjoe]

That resistor squelches gate ringing.

 

[WTP Pepper]

Most micros will have inputs as Hi-Z when coming out of reset so you need a resistor from the gate of the FET to the supply so it defaults to the FET being off and your high capacitive load is isolated.

From the reset state of the micro, use a PWM output to gently charge the load instead of just banging 0V on the FET causing the dip in voltage.

If your micro cannot do PWM, you can write your own routines to pulse the gate drive momentarily following reset allowing the capacitive load to charge gently to say Vcc/2 and then go fully on. What micro are you using?

Depending on how capacitive the load is, you 100mohm resistor maybe adding to the problem as it cause the power supply to the CPU to droop.

 

[Mr RB]

That resistor squelches gate ringing.

In the context of the user's problem the gate resistor reduces peak currents into the gate caused by the gate capacitance. Those peak currents might be what is causing the micro Vdd to drop, although it is also likely the main load current itself is causing some other PSU issue that is affecting the micro Vdd.

 

[WTP Pepper]

Nonesense, Your last sentence is more likely.

 

[alec_t]

If your micro can work with a Vdd ~ 4.3V then supply its Vdd pin via a diode, and have a large reservoir cap between the Vdd pin and ground.