I'm a hobbyist at best, and am trying to overcome the learning curve as far as pull-up and pull-down resistors, N channel transistors and P channel transistors, etc, etc.

I've got a USB Type-C charge module that absolutely will not work correctly unless it has a hard start. It barely ever activates when I give it a soft start (slowly ramping input voltage up from 0). It seems as though it needs a minimum of a 5V hard start for it to power-on correctly. Which stinks, because when the module pulls a load, it will pull the input voltage from my wind turbine down to 4V. If the wind suddenly dies down, the module will drop below 4V and kick off, and it WILL NOT turn back on again unless it receives a 5V hard start like previously mentioned. Due to this, I need something that can switch it on and off with a wide window of hysteresis; at least 1V.

The big issue here is that it's only powered by a wind turbine. So, 99.9% of the time, it's activated with a soft start and never a hard start.

So right now, I'm currently designing a circuit in EasyEDA to allow hard-starting only. It involves using a TPS3700 to sense for a turn-on voltage, and a turn-off voltage. To be safe, I want a turn-on voltage of 5.2V, and a turn-off voltage of 3.8V

I chose the TPS3700 because it has separate high and a low sense features.

I was originally going to have the TPS3700 drive an SR Latch which would switch on a MOSFET, but I also want current limiting and overheat protection, so I went and used a TPS2595 in place of the MOSFET(The TPS2595 has an internal FET).

The challenge I'm facing right now, is the "Enable" pin on the TPS2595. This pin has an on/off threshold voltage of 1.2V, and a max voltage of 7V. Since my wind turbine puts out anything between 3.5V and 18V, I connected a voltage divider to the pin to keep the voltages within spec.

So my main question is:

Is it better to drive the "Enable" pin with an N channel transistor, which would pull the voltage down ahead of the divider through a resistor? Or, would it be better to use a P channel transistor ahead of the divider?

Also: The TPS2595 datasheet says the "Enable" pin must not be left floating.

For reference, here's a picture of the TPS2595 being turned on/off by an N channel pulling to ground. Am I doing that correctly?

Here's a link to the TPS2595 datasheet: https://www.ti.com/lit/gpn/tps2595

Here's a link to the TPS3700 datasheet: https://www.ti.com/lit/gpn/tps3700

  • 2
    \$\begingroup\$ Please add a link to all relevant data sheets. It's called an NPN transistor and not an N channel transistor. MOSFETs have channels (N or P type) but, BJTs are either NPN or PNP; yours is an NPN. What active voltage range is the DC input from the turbine? Max value needed also. \$\endgroup\$
    – Andy aka
    Mar 16, 2023 at 12:01
  • \$\begingroup\$ Just added links. The voltage range from the turbine is anywhere from 3.5V to 18V. To add, someone else commented about using a Zener diode to prevent from going over the max voltage for that pin. I'll probably just go with that. Thank you for the help! \$\endgroup\$
    – soconnoriv
    Mar 16, 2023 at 18:04

1 Answer 1


The MMBT3904 (bipolar NPN transistor) should work well as you have it, you may just want to add a resistor (maybe 1k-10k) in series with the base (depending on how it is being driven). An actual N-Channel MOSFET could also be used in place of the NPN (with just a pull down resistor on the gate).

Using a PNP bipolar transistor or a P-Channel MOSFET on the high side of the line would add some complication as these would require a high signal to be in the off state.

To protect the EN line from a possible over voltage just place a 5V Zener diode on the line to GND. With that change you would not need a precision resistor divider, but still use R8 and R13 with values selected just to give the minimum voltage for EN. R10 shouldn't even be needed in any case.

  • \$\begingroup\$ Thanks Nedd, I never thought of using a Zener to prevent from going outside the voltage window. Much appreciated! \$\endgroup\$
    – soconnoriv
    Mar 16, 2023 at 18:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.