Author Topic: Low power servo idle  (Read 4030 times)

snorp

  • Jr. Member
  • **
  • Posts: 62
  • Country: us
Low power servo idle
« on: February 16, 2016, 04:06:17 PM »
I'm using the Moteino to control a servo, and I want it to use very little power while it's not actually controlling it. I have a N-channel MOSFET to control power to the servo, but the signal line seems to draw about 3ma (!) when connected to the Moteino. I got it down to about 500uA by switching the pin to INPUT during sleep, and a 10k resistor in series got me down to 350uA or so. At that point the servo started to experience some control issues, however. How can I get the power usage lower? Is it possible to use another MOSFET for the signal? I'm a software guy by trade, so I'm pretty new to this stuff. Thanks!

Felix

  • Administrator
  • Hero Member
  • *****
  • Posts: 6866
  • Country: us
    • LowPowerLab
Re: Low power servo idle
« Reply #1 on: February 16, 2016, 04:24:59 PM »
Any possibility we could see a schematic of how that N channel is wired to the moteino and servo?

snorp

  • Jr. Member
  • **
  • Posts: 62
  • Country: us
Re: Low power servo idle
« Reply #2 on: February 16, 2016, 05:08:59 PM »
Here's the schematic. It does not show the series resistor on D7, but otherwise that's what I have.

TomWS

  • Hero Member
  • *****
  • Posts: 1930
Re: Low power servo idle
« Reply #3 on: February 16, 2016, 05:27:25 PM »
I'm using the Moteino to control a servo, and I want it to use very little power while it's not actually controlling it. I have a N-channel MOSFET to control power to the servo, but the signal line seems to draw about 3ma (!) when connected to the Moteino. I got it down to about 500uA by switching the pin to INPUT during sleep, and a 10k resistor in series got me down to 350uA or so. At that point the servo started to experience some control issues, however. How can I get the power usage lower? Is it possible to use another MOSFET for the signal? I'm a software guy by trade, so I'm pretty new to this stuff. Thanks!
Looking at your schematic, are you sure you're not drawing 3mA through D7?   If that is output low when you turn off the transistor, current will probably flow through the servo into D7.  Also, it seems as if you may be exceeding the voltage on D7.  The servo is pulled up to 6V but the processor is running at 5V so D7 is probably drawing current even if it is high.

Also, what is the part number of the N channel FET?  What type of package?

Tom
« Last Edit: February 16, 2016, 05:30:33 PM by TomWS »

snorp

  • Jr. Member
  • **
  • Posts: 62
  • Country: us
Re: Low power servo idle
« Reply #4 on: February 16, 2016, 06:06:29 PM »
Looking at your schematic, are you sure you're not drawing 3mA through D7?   If that is output low when you turn off the transistor, current will probably flow through the servo into D7.  Also, it seems as if you may be exceeding the voltage on D7.  The servo is pulled up to 6V but the processor is running at 5V so D7 is probably drawing current even if it is high.

Also, what is the part number of the N channel FET?  What type of package?

Oh, that's entirely possible. Hmm, I didn't think the logic level would really be a problem for the servo. I guess I need to step up the level of D7?

The FET came from Sparkfun, the datasheet is linked from here: https://www.sparkfun.com/products/10213

snorp

  • Jr. Member
  • **
  • Posts: 62
  • Country: us
Re: Low power servo idle
« Reply #5 on: February 16, 2016, 06:38:50 PM »
Actually, is a Schottky diode the right thing to use here?

TomWS

  • Hero Member
  • *****
  • Posts: 1930
Re: Low power servo idle
« Reply #6 on: February 16, 2016, 06:58:13 PM »
Actually, is a Schottky diode the right thing to use here?
Well, a diode and resistor might be the easiest solution.  The diode doesn't have to be a Schottky diode, any silicon diode should work.  Place the diode between D7 and the Servo input pin 2 with the diode anode at D7, cathode at the servo pin.  Place a resistor, like the 10K you've used between pin 2 of the servo and pin 3 of the servo, this will provide a current return path for the PWM pulses when you're driving the servo and the blocking diode will prevent the voltage on the Servo pin from flowing back into the processor.

The voltage drop of the diode won't be an issue because the servo will operate with 3V pulses and you're sourcing from a 5V processor.

I think  your transistor SHOULD be able to handle the servo current  (not exactly 'delicate' that one!) ;)

Tom

snorp

  • Jr. Member
  • **
  • Posts: 62
  • Country: us
Re: Low power servo idle
« Reply #7 on: February 16, 2016, 07:03:29 PM »
Makes sense to me. So I can use any diode that has a forward voltage of < 5v and a breakdown voltage > 6v? That doesn't seem like it would be hard to find. Thanks a lot! I guess I'll head to radioshack for some overpriced (but free shipping) diodes :)

TomWS

  • Hero Member
  • *****
  • Posts: 1930
Re: Low power servo idle
« Reply #8 on: February 16, 2016, 08:17:20 PM »
Makes sense to me. So I can use any diode that has a forward voltage of < 5v and a breakdown voltage > 6v? That doesn't seem like it would be hard to find. Thanks a lot! I guess I'll head to radioshack for some overpriced (but free shipping) diodes :)
Yeah, let me know if you find any that meet that requirement, although, technically, the forward voltage needs to be Vf < (5V - 3.3V), but I suspect you'll still find a few...

Tom


perky

  • Hero Member
  • *****
  • Posts: 873
  • Country: gb
Re: Low power servo idle
« Reply #9 on: March 27, 2016, 10:55:00 AM »
I use high side switching for things like this. In my setup I wanted to drive a standard servo from 4 lithium cells so needed a voltage limiter, the inclusion of a zener and PNP transitor also turns off the high side FET when it gets avove 5.6V. Works very well.
Mark.