Design Challenge 4: Reliable, Portable Medical Design With Motors

The following is one difficult design problem some former colleagues of mine came across for which they haven’t been able to come up with a good solution. Of course, some semiconductor company may have solved the problem already, but they haven’t seen anything along the lines of what they need. Perhaps you, as savvy electronics gurus on Planet Analog, can help:

    Both sensored and sensorless 3-phase brushless DC motors are in wide use for various medical devices and their uses in portable medical devices are increasing. These types of motors often use between 10 W to 100 W of power and are driven by a DC voltage in the 12 V to 30 V range. The common configuration for the motors is a Y arrangement of three power leads connected to either a positive voltage or ground via six FETs, three on the high side and three on the low side.

    As is common for medical design, it is desirable to have a backup motor controller as well as backup FETs to drive the motor. This requires detecting that the primary motor controller is functional and that none of the six FETs have failed in a shorted or open state. In addition, it would be ideal to check that the backup FETs are also functional as well as the backup motor controller while the motor is operating properly.

    This is a challenging design problem because the failure of a single component, either primary or backup, must not cause the motor to stop for any significant length of time. The amount of time allowed varies depending on the application, but many applications require it to be less than one or two seconds. In addition, the backup circuitry must be able to operate the motor even if a primary FET is shorted, so there must be some way to isolate that component. The good news is that the design has to accommodate only one failed component.

    Finally, portable products require small and lightweight motor drivers. So, the design should not increase the overall size too much. Ideally, the design would not be much larger than duplicating the basic motor drive design twice.

Please contribute your ideas and comments. I believe that this is a real design challenge to be met head-on by our Planet Analog readers. I look forward to your solutions.

10 comments on “Design Challenge 4: Reliable, Portable Medical Design With Motors

  1. Davidled
    June 6, 2014

    Motor controller might require the big trace path of both power and ground in the Board with the huge heatsink surrounding by MOSFET switch (approximately 100V@50 Amp). Then if motor would be stuck during a certain period of time, it could be more challenge to design the circuit for protecting both board and motor. Board Package might not be plastic, but aluminum. It could be more cost. Design engineer would face more challenge task. I am wondering what size of DC motor would be used in most medical application.  

  2. etnapowers
    June 13, 2014

    In case of failure of one of the six FETs  a path to the electric current has to be provided: a paralleled , normally in OFF state, additional FET might work.

  3. D Feucht
    June 21, 2014

    The good news is that the design has to accommodate only one failed component.

    In my experience with motor-drive design for permanent-magnet synchronous (PMS) or “brushless” motors, when a power swtch in the driver bridge shorts, it takes out more than one transistor for Y-configured windings because two switches conduct in series. One way to avoid this is to have six ground-referenced FETs driving six windings, two per phase, of opposite flux polarity. Then the current is unipolar yet the flux remains bipolar. This circuit scheme makes it easier to isolate shorted switches because each has its own current loop, and current sensing for each of the 6 switches can be used to determine a shorted switch. Then an alternative FET (or set of FETs) can be driven instead.

    Another design feature that greatly aids in the determination of power-driver failures is to use winding-sensed (“sensorless”) control. By sensing the motor induced voltage, it is possible to deduce (in software algorithms) what is going on in the driver and motor circuits. Induced-voltage sensing can also be used to anticipate and predict an impending failure. From it, the phase control of the drive is readily determined. By observing both the motor terminal voltages and currents, almost every quantity that affects the motion control system can be observed and deduced.

  4. etnapowers
    June 23, 2014

    “One way to avoid this is to have six ground-referenced FETs driving six windings, two per phase, of opposite flux polarity. Then the current is unipolar yet the flux remains bipolar”


    @D Feucht: nice post thank you. I wonder if the solution that you described might manage the case of simultaneous failure of two or more power switches, due to the increased current on the remaining branches of the motor, caused by the failure of a power switch and finally of a branch.

  5. D Feucht
    June 23, 2014

    With 6 ground-referenced power switches, the observability of the drive is both high and can be implemented with few parts. By sensing both motor terminal currents and voltages on all six windings – not at all an impractical design scheme – essentially everything the motor or its driver are doing can be determined and multiple-part failures can be deduced in software. Recovery from multipart failure with alternative parts might need some additional ideas to implement well, however.

  6. chirshadblog
    June 24, 2014

    d-feucht: I think we do need to give it some more time. Only time can identify the issues and provide new suggestions to overcome them in the future. So time will be the key 

  7. chirshadblog
    June 24, 2014

    @etnapowers: Do you mean to say it works as a mirrored technology ? 

  8. RedDerek
    June 26, 2014

    One thought that came to mind is that there can be a current sense on the input power side and on the return power side of the switches. Top-side current sense would watch for any reverse current that could mean a swith was still open as current could be fed back in to the supply. Watch the bottom side and top side for excessive current draw as well.

    (sorry for the late response as I was out on vacation for 3 weeks and not much electronic communication)

  9. etnapowers
    July 1, 2014

    @Dennis: I agree with you on the feasibility of the control and I think that the optional stock parts, to be utilized in case of failure should be provided during the designing of the motor. Some simulations of failures and some verifications of real cases of multiple failure might be a nice additional practise, to guarantee the reliability of the overall system.

  10. etnapowers
    July 1, 2014

    @chirshadblog, I mean a self-recovering design solution that is robust to the simultaneous failure of two branches of the bridge.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.