I am working on an Arduino controlled servo circuit that steps the servo (standard R/C servo) in either direction in response to button push. Everything is working great except for ONE thing. The servo "jumps" to the default position at power up, and this will put too much strain on my mechanism. What I want to do is soft start the servo - but because the servo needs to share it's ground with the Arduino in order for the control pulse to work, I need the soft start to supply the current to the load, not sink it. I tried a (N-channel p75nf75) power mosfet with the gate connected to an R/C network consisting of a 47K resistor and a 100ufd cap between Vss and ground, with the source connected to ground and the gate to the connection between the resistor and cap. The load (servo) is connected between vss and the drain. Works perfectly - except I can't control the servo because the control signal doesn't get grounded. I tried putting the load between source and ground, with the drain connected directly to vss, but the voltage to the load changes all over the place depending on the load current/resistance. Is there a simple circuit that will allow me to "soft start" the servo with the power controlled on the + side, and the servo connected directly to ground?? I'm a former auto mechanic and semi-retired computer tech with all kinds of projects I'd like to get done and my electronics theory is about 50 years out of date!!!

 

With 5v going from the gate to the source of the MOSFET it can definitely conduct current. However, the p75nf75 is spec'd out mostly with a Vgs = 10v. To get the specified Rds(on) values in the datasheet, the part must be fully turned on with a greater voltage.

Another option is to look into a "logic level" MOSFET, which will be fully turned on at lower voltages.

 

A schematic is always a much easier way to describe a circuit so many can understand it clearly.

 

With 5v going from the gate to the source of the MOSFET it can definitely conduct current. However, the p75nf75 is spec'd out mostly with a Vgs = 10v. To get the specified Rds(on) values in the datasheet, the part must be fully turned on with a greater voltage.

Another option is to look into a "logic level" MOSFET, which will be fully turned on at lower voltages.

That doesn't solve my problem with needing the ground though - I need the "load" between the FET and ground, and I need thecurrent/voltage to ramp up smoothly - which it does with the circuit I am using - but I can't use it because of the ground configuration. I'll try to draw up the circuit, scan it in, and post it.

 

Fully turning the MOSFET "on" would/could reference the servo to the same ground as the arduino. Assuming they are connected.

 

OK, I scratched together some diagrams The first is what I have done, the second is what I need. I need the power to the servo to slowly ramp up, over several seconds, from 0 to V+

 

Fully turning the MOSFET "on" would/could reference the servo to the same ground as the arduino. Assuming they are connected.

The circuitry in the servo needs to be at the ground potential of the Arduino AT ALL TIMES. Otherwise the servo goes bonkers. The power to the servo needs to be limited at initialization to prevent the servo from pounding the linkage as it jumps to the predetermined initialization position. When powered down, the linkage can move the servo to any position, so it is not possible to predetermine the position (without adding more linkage and a feedback pot, and more code on the Arduino)

 

The circuitry in the servo needs to be at the ground potential of the Arduino AT ALL TIMES. Otherwise the servo goes bonkers. The power to the servo needs to be limited at initialization to prevent the servo from pounding the linkage as it jumps to the predetermined initialization position. When powered down, the linkage can move the servo to any position, so it is not possible to predetermine the position (without adding more linkage and a feedback pot, and more code on the Arduino)

I guess TECHNICALLY what I need is a "high side" soft start, as mu device needs to be on the "low side"

 

How about a p-channel mosfet?

 

Basically puts the switch on the HIGH side.

 

How about a p-channel mosfet?

I have a FQP27P06 available, but how do I implement the "fade in" on a P-Channel? I was thinking that might be the ticket - but it looks like I'll need a 2 stage circuit of some sort to act as an "inverter" because a P channel is ON with no signal, and needs to be pulled negative at the gate to turn it off?

 

With the gate signal HIGH the device is "off." Pulling the gate LOW turns it "on."

Swap the resistor and capacitor to implement circuit similar to what you already have. Capacitor goes across gate-source still.

 

P channel source is connected to 5v.

 

I have a FQP27P06 available, but how do I implement the "fade in" on a P-Channel? I was thinking that might be the ticket - but it looks like I'll need a 2 stage circuit of some sort to act as an "inverter" because a P channel is ON with no signal, and needs to be pulled negative at the gate to turn it off?

If it is the old servo that runs off of pulse width you may have a problem as the electronics inside the servo needs almost full voltage to work.
You might think about some stored energy for the adrino on power off to allow it to return the servo to the center.

 

With the gate signal HIGH the device is "off." Pulling the gate LOW turns it "on."

Swap the resistor and capacitor to implement circuit similar to what you already have. Capacitor goes across gate-source still.

So you say just replace the N-Channel with a P-Channel, and flip the cap and resistor, as well as the mosfet and connect the load between the drain and ground?? Sounds simple enough - I'll grab the p-channel and see how it works.

 

because a P channel is ON with no signal, and needs to be pulled negative at the gate to turn it off?

No.
A standard enhancement-mode P-MOSFET is OFF with Vgs=0 just like an N-MOSFET.
But Vgs for a P-MOSFET must be negative to turn it ON.
This can be done for a high side P-MOSFET by connecting the source to V+ and the gate to ground.
In other words just turn your N-MOSFET circuit upside-down with a P-MOSFET.

 

Here's the LTspice simulation of a P-MOSFET soft-start circuit that adds a Miller integration capacitor (C2) to give a slow linear(ish) ramp startup (about 2s for the R and C values shown).
Without C2 the ramp is only about 0.6s.

If you want to change the capacitor values, maintain the ratio of C1 being about 10 times the value of C2.

The P-MOSFET must be a logic-level type device with the Rds(on) specified at a Vgs of 5V or less.

 

Here's the LTspice simulation of a P-MOSFET soft-start circuit that adds a Miller capacitor (C2) to give a slow linear ramp startup (about 2s for the R and C values shown).

If you want to change the capacitor values, maintain the ratio of C1 being about 10 times the value of C2.

The P-MOSFET must be a logic-level type device with the Rds(on) specified at a Vgs of 5V or less.

View attachment 126886

So if I want to increase the ramp-up time, I could put C1at 47 and C2 at 4.7, keeping the resistor at 470K, or would it be better to increase the resistance of R2?

 

So if I want to increase the ramp-up time, I could put C1at 47 and C2 at 4.7, keeping the resistor at 470K, or would it be better to increase the resistance of R2?

Either would work although if the resistor gets too high, leakage currents can start to be a problem.

Below is the sim for 47μF and 4.7μF.

 

Either would work although if the resistor gets too high, leakage currents can start to be a problem.

Below is the sim for 47μF and 4.7μF.

View attachment 126893

Thanks a lot - I'm picking up the P-Channel FET this morning - likely be this evening before I get to breadboarding the circuit.