Shop OBEX P1 Docs P2 Docs Learn Events
Prop controlling electric wheelchair motors? — Parallax Forums

Prop controlling electric wheelchair motors?

eagletalontimeagletalontim Posts: 1,399
edited 2012-09-24 13:28 in Propeller 1
I have an old wheelchair that I tore apart and assembled to an old go-cart frame just because I was bored, but recently, I wanted to see about adjusting the turn speeds in the controller and accidentally shorted out the main brain with a meter probe :( Now that the controller is dead, I want to see about building my own but I am not sure what MOSFET's / Transistors I can use to drive the motors. In the controller, there are IRF1010N's which are rated at 60A but the controller gets really hot after about 5 minutes of driving it around. What I am hoping to find is an already assembled H-Bridge that can handle 100A which can be controlled by the Prop directly without other chips needed. If that is not available, I would need to build an H-Bridge, but I don't know what Mosfet or transistors I can use. Any help on this?

EDIT : Voltage is 24V powered by 2 wheel chair batteries.

Comments

  • jmgjmg Posts: 15,173
    edited 2012-09-23 15:59
    eagletalontim wrote: »
    What I am hoping to find is an already assembled H-Bridge that can handle 100A which can be controlled by the Prop directly without other chips needed. If that is not available, I would need to build an H-Bridge, but I don't know what Mosfet or transistors I can use. Any help on this?

    BIG modules tend to be 3 phase, and not cheap. Often they are only the FET Brick, so you still need a Driver IC.

    Probably best to get a Motor Driver Controller from one of the usual suspects (Allegro, Infineon, OnSemi, IRf etc), and add your own FETS.
    Ideally one with a inbuilt charge pump, so you can drive to 100% duty - and one that can set a Current limit is easy to control.

    Then you can choose FETS to suit, and you can get sub-milliohm FETS for sub $1 in volumes, and parallel as needed.
    ( those IRF1010N look to be 11 mOhms )
  • W9GFOW9GFO Posts: 4,010
    edited 2012-09-23 17:02
    I thought this sounded familiar;

    http://forums.parallax.com/showthread.php?139811-R-C-5-channel-radio-and-receiver-to-control-24v-DC-motor
  • eagletalontimeagletalontim Posts: 1,399
    edited 2012-09-23 17:26
    I forgot about that thread :P I did build that, but the H-Bridge I built can only handle 7A which is way under the 100A I need.
  • W9GFOW9GFO Posts: 4,010
    edited 2012-09-23 17:57
    There were several motor controllers mentioned in that thread that are appropriate for wheelchair sized motors. I am skeptical that you actually need 100 amps but perhaps you want to run the motors excessively hard. Also do make sure that the brakes are not dragging.

    I was able to adequately drive my 120 lb wheelchair based robot on 12V using HB-25s (300W). It was easy however to create a situation where the controllers would shut down. Using 40A 24V (960W) controllers will allow over three times the power to the wheels which I expect will be more than enough for whatever I need it to do.
  • frank freedmanfrank freedman Posts: 1,983
    edited 2012-09-23 18:54
    eagletalontim wrote: »
    I have an old wheelchair that I tore apart and assembled to an old go-cart frame just because I was bored, but recently, I wanted to see about adjusting the turn speeds in the controller and accidentally shorted out the main brain with a meter probe :( Now that the controller is dead, I want to see about building my own but I am not sure what MOSFET's / Transistors I can use to drive the motors. In the controller, there are IRF1010N's which are rated at 60A but the controller gets really hot after about 5 minutes of driving it around. What I am hoping to find is an already assembled H-Bridge that can handle 100A which can be controlled by the Prop directly without other chips needed. If that is not available, I would need to build an H-Bridge, but I don't know what Mosfet or transistors I can use. Any help on this?

    EDIT : Voltage is 24V powered by 2 wheel chair batteries.

    If you need a good reference design for an H bridge driver using discrete components, go to GE site and grab the service documents for an AMX4+ portable rad unit. The system uses a pair of H gridge drivers using the IRF-250 (30A max) to move the thing at a pretty good pace. Its normal rated weight is 1080lb. The system is using internal gel cell batteries stacked to 115V (9 12V 28Ah). But then again, that may be part of your problem. It may be better to use smaller batteries arranged to give you a higher voltage to drive with and enable a lower rated current to the motors. These motors are 120V PM motors and weigh in about 7lb. As I mentioned, the AMX4+ is a half ton of machine. this is a rather nice device from the service point of view in that all the detailed schematics are there so you can also figure out how the drives are controlled by the pre-amp board. Lot of other parts provide good working examples if a bit dated though.

    FF
  • jmgjmg Posts: 15,173
    edited 2012-09-23 19:47
    W9GFO wrote: »
    There were several motor controllers mentioned in that thread that are appropriate for wheelchair sized motors. I am skeptical that you actually need 100 amps but perhaps you want to run the motors excessively hard.

    You may have missed this detail, it is no longer still a wheelchair ! ;)

    and assembled to an old go-cart frame ...
  • W9GFOW9GFO Posts: 4,010
    edited 2012-09-23 19:57
    jmg wrote: »
    You may have missed this detail, it is no longer still a wheelchair ! ;)

    and assembled to an old go-cart frame ...

    What did I say that requires it to still be in wheelchair form?
  • eagletalontimeagletalontim Posts: 1,399
    edited 2012-09-23 21:21
    The reason for the 100A capability is to help with heat and allow for longer "play" time. Me and my kids ride this contraption up and down my backyard which is sloped pretty good. Going up the hill once is about all the other controller could handle before it went into "thermal cool down" mode. Steering was a nightmare as well when on any kind of angle. Building my own controller seems a better way to control this device.

    QUOTE : "Then you can choose FETS to suit, and you can get sub-milliohm FETS for sub $1 in volumes, and parallel as needed."

    Does this mean I can "stack" 2 of these IRF1010N's and make the controller handle more power? If so, I can order a bundle of 10 for about $1 a piece.
    From the spec sheet, it appears that the gate only needs 2 to 4 volts to fully activate the MOSFET. Does this mean they can be run directly from the output of the Prop?

    While tinkering with the old controller, I attempted to activate one of the h-bridge circuits manually by tapping onto the pads where one of the chips was located which I suppose activated the h-bridges based off of input from the main processor. I removed both chips with my hot air station to ensure I would not short anything out. Once I powered it up, I checked to make sure power was where it needed to be then accidentally touched one of the wires I soldered on which caused one of the mosfets to explode :( Apparently these things are very touchy. I did ensure the pads I soldered to were traced to the mosfet's gate and everything was 100% correct. I suppose the natural body resistance caused current to flow through me directly to the gate which was more than it could handle. Next time I will be more careful...time to bust out my wrist strap!
  • jmgjmg Posts: 15,173
    edited 2012-09-23 21:34
    eagletalontim wrote: »
    QUOTE : "Then you can choose FETS to suit, and you can get sub-milliohm FETS for sub $1 in volumes, and parallel as needed."

    Does this mean I can "stack" 2 of these IRF1010N's and make the controller handle more power? If so, I can order a bundle of 10 for about $1 a piece.
    From the spec sheet, it appears that the gate only needs 2 to 4 volts to fully activate the MOSFET. Does this mean they can be run directly from the output of the Prop?

    You can buy better FETs than IRF1010N, but you can parallel like-FETs; if they are on the same PCB, then PCB traces provide some sharing resistance.
    If you are trying to lower power loss, 3.3V gate drive is not going to be enough.
  • vanmunchvanmunch Posts: 568
    edited 2012-09-23 21:43
    I've used the Sabertooth 2X60with great success controlling my ~500 lb watering robot that uses power chair motors. I highly recomend them. They can handle peaks of 120 amps and 60 amps continually.

    http://www.dimensionengineering.com/products/sabertooth2x60
  • eagletalontimeagletalontim Posts: 1,399
    edited 2012-09-23 21:45
    Maybe I am not quite understanding the specs on Mosfets. I have found another Mosfet that is a little more expensive, but has a much higher current handling ability.

    Here is the link : http://www.digikey.com/product-detail/en/IRFB3207ZPBF/IRFB3207ZPBF-ND/1300656

    In the specs on that page, it shows : Vgs(th) (Max) @ Id 4V @ 150µA

    Does this mean that a 4V 150uA input at the gate will fully activate the mosfet?
  • eagletalontimeagletalontim Posts: 1,399
    edited 2012-09-23 21:49
    Why are pre-made controllers sooo expensive? The components don't seem to add up to over $30 to $50 so am I missing something here? I can spare $30, $50, even $100 to build my own..... What is the difference between a custom built vs a pre-made?
  • jmgjmg Posts: 15,173
    edited 2012-09-23 22:00
    eagletalontim wrote: »
    In the specs on that page, it shows : Vgs(th) (Max) @ Id 4V @ 150µA

    Does this mean that a 4V 150uA input at the gate will fully activate the mosfet?

    It says :

    VGS(th) Gate Threshold Voltage 2.0 V(Min) ––– 4.0 V (Max) VDS = VGS, ID = 150μA

    Which means the worst case (max) FET could draw only a teensy ID = 150μA, at 4V VGS,
    ID is the DRAIN current, at that Gate Voltage, and the spec is the threshold

    Find the Rds value, which says 3.3mOhms, for 10 V on the Gate. ie you WILL need 10V to drive this FET.
  • frank freedmanfrank freedman Posts: 1,983
    edited 2012-09-23 22:52
    eagletalontim wrote: »
    Why are pre-made controllers sooo expensive? The components don't seem to add up to over $30 to $50 so am I missing something here? I can spare $30, $50, even $100 to build my own..... What is the difference between a custom built vs a pre-made?

    You are paying the cost of development across however many units are expected to be sold plus some profit on the sale plus the cost of stocking parts and support infrastructure the company expects for the device plus any regulatory costs....... etc......... all part of the cost of doing business. Because the developer expects to make a profit, guess who gets to pay a percentage of these costs. Pay da man...... Or spend the money and develop it yourself, putting in the design time, the parts you have let the smoke out of and so on. Which way will cost you less and which way presents the reliability level you can tolerate.

    FF
  • Mark_TMark_T Posts: 1,981
    edited 2012-09-24 13:28
    jmg wrote: »
    It says :

    VGS(th) Gate Threshold Voltage 2.0 V(Min) ––– 4.0 V (Max) VDS = VGS, ID = 150μA

    Which means the worst case (max) FET could draw only a teensy ID = 150μA, at 4V VGS,
    ID is the DRAIN current, at that Gate Voltage, and the spec is the threshold

    Find the Rds value, which says 3.3mOhms, for 10 V on the Gate. ie you WILL need 10V to drive this FET.

    My standard advice for MOSFETs is:

    1) Check the voltage spec is twice the supply voltage (more for inductive loads).
    2) Check the Rds(on) value will lead to a low-enough power dissipation for the currents you'll be dealing with and the heatsinking you'll be using.
    3) Check the Vgs needed to achieve that Rds(on) - if its 10V you can't drive from 5V, if its 4.5V you can drive from 5V (but it'll be dodgy from 3V3).
    4) If using PWM calculate the switching losses and choose a frequency and a gate-drive current that'll ensure the losses aren't dominant (typically
    the device is spending less than about 2% of the time actually switching - for high power loads this is more important - during switching the device dissipates far more power than at any other time.

    You can pretty-much ignore the current rating, if you follow item 2) you already have ensured this will be fine (slight caveat unless using low-duty cycle high-current pulses).
    Don't go by the threshold voltage, follow 3)
    Note that typically the fully-on Vgs will be around 3 to 4 times the Vthr for most devices.
    The total gate charge divided by the gate drive current will tell you approx how long switching takes. 10nC / 10mA = 1us, 200nC / 1A = 200ns etc etc
Sign In or Register to comment.