Shop OBEX P1 Docs P2 Docs Learn Events
How do the big boys and girls design “one battery” robot systems? — Parallax Forums

How do the big boys and girls design “one battery” robot systems?

vanmunchvanmunch Posts: 568
edited 2014-08-21 06:43 in General Discussion
Hey everyone, I’ve been working on a couple of new robots and I’ve been having some problems with the motors pulling too much current and causing the logic voltage to drop. My “go to” solution is to just add a second battery that only the logic uses, but I’ve been a bit more stubborn than normal and I’ve been trying a number of things during the past couple of weeks without success. I’ve added progressively larger caps across 3.3 to GND (1000uf), Vin to GND (3300! uf), and I even tried a up/down 3.3 voltage regulator from Pololu that feeds directly into the 3.3V plane on the Propeller Proto board. They’ll work for a 5-10 seconds, but then it brown’s out. Using a second battery to power the motors and logic has been the only reliable solution that I’ve found….

So now I’m on the ropes and about to give over to the temptation of the extra battery so before I give in I thought that I would once more throw myself onto the collective wisdom (?) of the forums and ask for suggestions or ideas on how you use one battery to power everything in a robot and avoid the dreaded brown out!

Thanks for your time!

Dave

Up/down regulator: http://www.pololu.com/product/2122

Comments

  • ctwardellctwardell Posts: 1,716
    edited 2014-08-20 11:22
    Dave,

    There a lot of variables here.

    My sumo bots from last years Ohio expo and my fire fighter from about 10 years ago used a single battery for both motors and logic with no problems.

    Probably the biggest thing is to make sure the battery has sufficiently low internal resistance compared to what the motor will draw under load. That will help reduce sudden drops in voltage beyond what your DC/DC convertor can compensate for.

    Ideally the battery voltage will be high enough to insert a diode between the battery and the input capacitor to your regulator, this lets the capacitor handle regulator load instead of trying to help power the motors.

    It might also help to use a DC/DC convertor with a 5V output and then use LDO 3.3V linear regulators for the logic, that is what my Sumo bots have.

    C.W.
  • GadgetmanGadgetman Posts: 2,436
    edited 2014-08-20 12:34
    My thought...

    Up the Voltage!

    you're working with a 6V pack or something, right?
    Go for a much higher voltage pack, and add a switching DC-DC converter to take it down to the voltages you need.

    Motors drawing 1A at 6V will draw 0.35A from an 18V battery pack
    (Assume some loss in the converter)
  • SRLMSRLM Posts: 5,045
    edited 2014-08-20 12:56
    I assume that you're using lead acid batteries? Working on a quad project with LiPos we never had any problems with brownouts, even with the big 30A motors.
  • Beau SchwabeBeau Schwabe Posts: 6,566
    edited 2014-08-20 13:56
    The Discharge characteristics of your Power supply can play an important roll also... when I was in Prosthetics, LiPos were not available to the public. We were in the migration stage from NiCd to NiMh.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2014-08-20 14:15
    It also matters a lot how the grounds are connected. Each of the three circuit sections should have it's own dedicated ground return line directly to the battery. You do not want motor ground-return currents running through your logic or analog sections, for example.

    -Phil
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2014-08-20 14:45
    Beau Schwabe (Parallax) wrote: »
    I spent 10 years in the robotic research and development industry for prosthetics and what you describe is a common problem. One of the most common solutions is to divide your power into 3 main groups off of your main supply. One dedicated to your micro controller and NOTHING else, another dedicated to your motors, and lastly one dedicated to anything your micro controller needs to talk to ... i.e. sensors, motor controller (<- but not the motor itself), etc.

    To do this a common "diode capacitor filter" can be used for each stage with the exception of the motors. Motors usually get the full source power.

    Note: Notice the use of two diodes on the voltage regulator so that the voltage drop of the diode is factored out when supplying power to individual components.

    This is basically the same thing I do expect I diode + cap before the regulator and then let the regulator do it's job best. What I also do is soft-start big motors and do what I can to reduce power on logic during that time so that the logic will not be starved too long of power. This also means detecting motor stalls or brown-outs and taking appropriate action.

    Also as Phil said, seperate grounds grounds grounds :)
  • John BoardJohn Board Posts: 371
    edited 2014-08-20 14:59
    The Diode-Cap system worked for me when I had the same issue on my RoboCup robot (Thanks Peter Jakacki!)
  • vanmunchvanmunch Posts: 568
    edited 2014-08-21 05:42
    This is great! Thank you everyone for your help. This has always been something that I would like to reliably fix.

    After reading everyone's suggestions I think my problem is that I'm drawing the power for my motors from the regulated 5V line on a propeller protoboard. I'm guessing that it's drawing too much current and starving the 3.3V regulator that's "downstream" from it. {5AA alkaline batteries (~7.5V) to 5V regulator to 3.3V regulator} I'll try separating the power supplies using a diode capacitor filter from the higher voltage of ~7.5V and make sure each one has it's own GVD line to the battery.

    Some more details that I should have organally posted are:
    The motors that I'm using are two of these http://www.pololu.com/product/2208 and one of these http://www.pololu.com/product/993
    I'm controling two of the motors with one of these http://www.pololu.com/product/2135 and I'm controlling the thrid motor using PWM with a T0-251AA Mosfet http://www.mouser.com/ProductDetail/Fairchild-Semiconductor/RFD14N05L/?qs=sGAEpiMZZMshyDBzk1%2fWi1qxDdW%2fXBi25Mt%2fEZi9Ir0%3d (I have a flyback diode connected between the motor leads for reverse voltage spikes.) It just needs to spin in one direction.

    On a related note, I had always thought there was less power was wasted by arranging your voltage regulators in order so that Vin goes into 5V regulator which then goes into your 3.3V regulator as opposed to going Vin to 5V regulator and Vin to 3.3V regulator. I thought that directly going from Vin to 3.3V means that you waste more energy as heat?

    Thanks again for everyone's help!

    Dave

    PS @ SRLM Don't worry, I'm using those lead acid batteries you gave me at the Boston Propeller expo in a different robot. I didn't even realize it at the time, but they are exactly what I needed and are the correct size and voltage! Thanks again! (those are connected using HB-25 motor controllers)
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2014-08-21 06:06
    vanmunch wrote: »

    1. After reading everyone's suggestions I think my problem is that I'm drawing the power for my motors from the regulated 5V line on a propeller protoboard. I'm guessing that it's drawing too much current and starving the 3.3V regulator that's "downstream" from it. {5AA alkaline batteries (~7.5V) to 5V regulator to 3.3V regulator} I'll try separating the power supplies using a diode capacitor filter from the higher voltage of ~7.5V and make sure each one has it's own GVD line to the battery.

    Some more details that I should have organally posted are:
    The motors that I'm using are two of these http://www.pololu.com/product/2208 and one of these http://www.pololu.com/product/993
    I'm controling two of the motors with one of these http://www.pololu.com/product/2135 and I'm controlling the thrid motor using PWM with a T0-251AA Mosfet http://www.mouser.com/ProductDetail/Fairchild-Semiconductor/RFD14N05L/?qs=sGAEpiMZZMshyDBzk1%2fWi1qxDdW%2fXBi25Mt%2fEZi9Ir0%3d (I have a flyback diode connected between the motor leads for reverse voltage spikes.) It just needs to spin in one direction.

    2.On a related note, I had always thought there was less power was wasted by arranging your voltage regulators in order so that Vin goes into 5V regulator which then goes into your 3.3V regulator as opposed to going Vin to 5V regulator and Vin to 3.3V regulator. I thought that directly going from Vin to 3.3V means that you waste more energy as heat?

    1. You are joking me surely, I thought you at least had the motors connected to the input power, not a regulator, and the logic regulator at that, staged or not.
    Okay, so you see I'm not impressed. Motors should never be powered from regulators, that's what PWM is for. The stall current is only 0.36A so even a baby MOSFET will handle that (I probably would have just used a good NPN though)

    Thanks for supplying the details of the parts you are using.

    2. The same amount of power will always be wasted except it would be distributed over both devices which means that the 3.3V regulator dissipates less power plus many 3.3V regs these days are rated for only 6V max input or so. Once again, never hookup inductive loads to logic regulators, give them their own regulator if they really need one or at least filter and clamp the line etc.
  • PJAllenPJAllen Banned Posts: 5,065
    edited 2014-08-21 06:39
    Why not two batteries?
  • vanmunchvanmunch Posts: 568
    edited 2014-08-21 06:40
    Peter Jakacki wrote: »
    1. You are joking me surely, I thought you at least had the motors connected to the input power, not a regulator, and the logic regulator at that, staged or not.
    Okay, so you see I'm not impressed. Motors should never be powered from regulators, that's what PWM is for. The stall current is only 0.36A so even a baby MOSFET will handle that (I probably would have just used a good NPN though)
    .

    No joke, and I'll even let you call me Shirley. :)

    [video=youtube_share;0A5t5_O8hdA]

    It didn't even dawn on me, since you can power 2-4 servos off of it and there is even a nice little spot for 4 servos to plug into.

    The robot is running off of Vin just fine now.

    I really appreciate your and everyone's help. :)

    Dave
  • vanmunchvanmunch Posts: 568
    edited 2014-08-21 06:43
    PJ Allen wrote: »
    Why not two batteries?

    Part of it is because I just want to be able to make it work and part of it is to same space (The robot can fit in you hand.).
Sign In or Register to comment.