I am working on a project which requires three different voltage levels - Motor that can run at 7-12 V. Hall sensor that can run at 5-25 V. nRF24L01+ that can run at 2-3.6 V and AVR (arduino) which can run at 2-5 V.

Since battery life is a major concern, I think voltage regulators (specially the linear ones) will be a bad idea. For motor, I am planning to use a LiPo battery - 11.1 V rated. It'd be great if remaining circuit can be run on 3V. This way I can use AA cells for them. The only blocker is Hall sensor. I saw one which can run at 2.5-25 V but most of them run at 4.8-25V so I am not sure whether I'd be able to get low power version in local market.

As such, I am left with two choices:

  • Use AA cells to get 3V which will take care of Arduino and nRF. Use boost converter to power Hall sensors.
  • Use 2xAA cells to power up Arduino and nRF and another 2xAA cells added in series with the previous ones to get 6 V and use it for Hall sensor.

Which one would be a better idea?

Also, it'd be great if you can suggest me some non-latching type Hall sensors in 3 pin sil package, if possible with working voltage range starting from 1.5 or 2 volts.

  • \$\begingroup\$ What hall sensor are you looking at? \$\endgroup\$ Commented May 23, 2015 at 22:15
  • \$\begingroup\$ Why not use a boost regulator from the motor supply for the hall sensor AND a buck regulator for the low voltage stuff? \$\endgroup\$
    – Andy aka
    Commented May 23, 2015 at 22:52
  • \$\begingroup\$ @Ignacio - Currently I am planning to go for W134 hall sensor in TO92 package if I can get it. Otherwise I am not sure. Probably I will pick whichever is available in local market. The only criteria is - it should be non latching type. Means it should turn ON when magnet is brought closer and turn OFF when magnet is taken away. \$\endgroup\$ Commented May 23, 2015 at 23:04
  • \$\begingroup\$ @Andyaka - The only reason I don't want to do it that way is power dissipation in those regulators. I have to save as much battery juice as possible. However I am not an expert. Do you think using regulators will be a good idea. I need a battery life of at least six months out of 5000 mAh battery. \$\endgroup\$ Commented May 23, 2015 at 23:08
  • \$\begingroup\$ Do you really think powering part of the circuit from a LiPo and the rest from AA batteries is an optimal solution? \$\endgroup\$ Commented May 24, 2015 at 14:36

2 Answers 2


In the end, the most simple solution is usually the best. Let's try to sort your needs out.

1) You need a beefy power supply for a motor, 12V-ish. Fine, you have this sorted out already.

2) You need to power the Hall sensor. Why not power it directly from 12V, if it's 5-25V rated? There can be 2 minor difficulties there :

  • Connecting Hall sensor output to mcu input. If the output is open collector, that's obviously fine as-is. Otherwise, depending on the operation frequency and the mcu input's impedance requirements, the 12v logic out -> 3.3V mcu in connection could be as simple as a large enough resistor, input is then clamped by mcu input diodes.

  • Current draw from hall sensor at 12V. That's up to you to determine if it is acceptable. If you need to power it off when dot doing measurements, that's a bit more involved but can be done easily.

3) MCU power can be done simply with a 3.3V buck regulator. Google "low quiescent current buck regulator" returns LT3991 as the second hit, Iq=2.1µA for 12V->3.3V. Then it's just a matter of putting the mcu and peripherials to sleep most of the time, which you'll need to do anyway.

In the end, I think there is NO WAY a second battery with holder can be more efficient for a given battery life, be it weight-wise, volume-wise, and even price-wise, than a simple switching regulator and slightly larger main battery. The only exception I can possibly see is if you need galvanic isolation and low RFI - totally off topic here.


If your hall-effect sensor is (are) digital (as opposed to linear), you can power it directly from the motor supply. Current consumption is very low and most digital hall-effect sensors have an open-drain output that goes LOW when the sensor is active.

I'm assuming that your motor current is significantly higher than anything else in your system. If it's not important that everything keeps running when the motor battery goes flat, I'd use a SMPS buck-converter to drop the motor battery down to 3.6 Vdc and run both the RF module and the AVR controller from that.

You can add a small backup battery pack that switches over automatically if you do need to keep the RF module and AVR running after the motor battery has died. Use 4- "AA" cells in series and use diodes to sum both battery packs into the input of the buck converter.

  • \$\begingroup\$ 3.6V is a very odd voltage to run logic at. That's probably the absolute max from the NRF datasheet, and not intended to be an operating voltage. \$\endgroup\$ Commented May 24, 2015 at 14:37
  • \$\begingroup\$ Choose whatever voltage that you think is appropriate that will power both devices. I like to be at the high end of the range so as to maximize the dynamic range of any A/D inputs. \$\endgroup\$ Commented May 24, 2015 at 15:06
  • \$\begingroup\$ If you run at absolute maximum voltage, any deviation or tolerance in the regulator will put you outside the absolute maximum - not a good idea. \$\endgroup\$ Commented May 24, 2015 at 15:18
  • \$\begingroup\$ Even if it's important to keep mcu/communication running when the battery is "flat", you can just stop powering the motor when the battery is "nearly flat", keeping just enough juice for the mcu/communication. \$\endgroup\$
    – Nicolas D
    Commented Aug 5, 2015 at 7:04

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.