Page 1 of 9 123 ... LastLast
Results 1 to 10 of 82

Thread: Pros and cons of different motor types

              
   
   
  1. #1
    Señor Member podolefsky's Avatar
    Join Date
    Oct 2010
    Location
    Boulder, CO
    Posts
    3,473
    Post Thanks / Like

    Pros and cons of different motor types

    AC induction, DC series, PMAC, PMDC...there are a lot of motor types out there. I'm starting this thread to nail down the pros and cons of each. I don't expect to come away with any absolutes like "type X is the best". More like "if you have this setup, these expectations, and this budget, you'd do well with type X, or maybe type Y"


    This site claims:

    "[AC induction are the] Third most popular electric car motors, but if all motors cost the same, these would be the hands-down winner. Why? Because they're better. (There, yes, I said it. I took a stand."

    MetricMind says essentially the same thing.

    And this writeup compares induction and BLDC (aka PMAC), stating:

    "Back when I had hair on my head and carried a slide rule, there were lead acid batteries, DC brush motors, and contactor controllers. Today, none of these remain (including my hair)."

    Then basically says PMAC are good for efficiency (think hybrid), AC induction are good for high-end performance (think Tesla).


    But, we know that PM-DC motors are very popular for electric motorcycles. AC induction looks superior on paper, but it depends on the batteries, gearing, and all sorts of other factors. The Killacycle uses series motors.

    Given all this, how should someone decide what is best for them? That's what this thread is about.
    - Noah Podolefsky -
    The GSX-E

  2. #2
    Señor Member podolefsky's Avatar
    Join Date
    Oct 2010
    Location
    Boulder, CO
    Posts
    3,473
    Post Thanks / Like
    [EDIT - CAUTION] The graphs below are not to be trusted. They aren't comparing equivalent systems, for a variety of reasons explained in this thread. I'm leaving them up, but readers should not take these to mean that AC is worlds better than DC. (There may be other reasons to think so, but not these graphs.)

    ____________________________________________


    Here is my first piece of data to fuel this fire. It compares the D&D ES-15-6 (series) vs HPEV AC-15 (drop in replacement induction). Also shows graphs with 300 and 180A limits (5c and 3c for 60Ah).

    Note that this is using ElMoto spreadsheet 2.18, straight from manufacturers spec, no current limits imposed (apart from what is in manufacturer spec). Otherwise, all bike specs are the same (weight, drag coefs, etc.)

    From manufacturer spec, current at max torque is: AC 550A, DD 480A

    Also note that AC curves drop off after ~90mph (6000 rpm) because I don't have data for that area. Please ignore this part of the graph.

    (One more thing: I have shifted the curves to the left - I believe this is correct so that kW curves cross through zero at 0 mph. The lines should extend to zero, but they are cut out of the spreadsheet chart)





    Last edited by podolefsky; 18 March 2011 at 1943.
    - Noah Podolefsky -
    The GSX-E

  3. #3
    Señor Member podolefsky's Avatar
    Join Date
    Oct 2010
    Location
    Boulder, CO
    Posts
    3,473
    Post Thanks / Like
    And one more where DD and AC torque-at-wheel are matched through different gearing, all else equal and no set current limits.

    - Noah Podolefsky -
    The GSX-E

  4. #4
    Moderator Nuts & Volts's Avatar
    Join Date
    Aug 2010
    Location
    Columbus, OH
    Posts
    1,667
    Post Thanks / Like
    Ok here is my take. I'll try to keep it concise.

    Your 3 main (available and widely used) types of motors are Brushed DC, AC Induction (ACIM), and Permanent Magnet AC (PMAC). To directly relate these motors you must also consider the controller being used.

    First lets assume unlimited awesome controllers for all motors and the motors are rated for the same power. These motors should have about the same constant torque power curve until BEMF limits RPM, then the torque falls off as field weakening is implemented. The falling curve depends on the control method used. Thus over a normalized speed range the power out of any of these three motor types will vary only slightly (maybe more when efficiency is included).
    Construction wise the ACIM is good because it uses no magnets which allows it to operate at higher temperatures, higher RPM (different bemf function i assume), and a potentially a lower cost. ACIM has been around for a while in industry so they are somewhat well "explored". I believe that ACIM can more easily be over-powered past their rated peaks due to the high temp limit. A brushed DC will suffer from efficiency losses in the commutator brushes. This construction makes them simpler design. Magnets limit the maximum temperature and power of the motor. This is also true of a PMAC motor, but these motors gain efficiency by losing the brushes. Power density (both by volume and weight) of all three can be fairly similar depending on how much money you through at it. I have seen that in general the power density goes from high to low in the order of PMAC, ACIM, & brushed. Efficiency generally follows this order too, but some brushed DC can be more efficient than ACIM. Cost (same power motor) from low to high is generally brushed, ACIM, then PMAC. Remember this ^^^ analysis is with an unlimited, perfect controller.

    Now to apply this to more real world this current controller technologies the pictures widens. The cost ranking stays the same for the most part as you include controllers, but high power ACIM and PMAC are fairly close in cost. Brushed controllers are cheap and fairly simple in operation. ACIM and PMAC controllers are mostly similar to each other and are more costly and complicated. So when we look at performance in the real world the brushed DC is commonly limited by controller power and quality of build (ie cooling). A Netgain 11 brushed DC motor has hit over 500HP, so DC is not weak by any means. The main advantages of AC is more fined tuned control. This makes for smoother control and more efficient operation, but this still can be limited by controller choice as well.

    So for ultimate power I believe ACIM is the best because it can physically handle more heat generation, this means higher current and more torque created. A PMAC is close, but must be better designed to put out the same numbers.

    Ok so I believe PMAC is the best choose for power per dollar per pound per availability. DC has its place in raw power, but efficiency is down and drivability (control is limited). ACIM is harder to find motorcycle size, but is very popular for electric cars.

    Basically if you made cost equal you could create a very similar AC and DC system. But high cost, low wait brushed DC motors dont really exist to my knowledge, so AC it is

    As always let me know where I went wrong, let the discussion begin
    Last edited by Nuts & Volts; 17 March 2011 at 2002.
    Whats under my tank may shock you!!! R6 Build, Motor Thread , Sorry excuse for a blog/
    OSU IOM Build Thread, OSU Current webpage

    "Noah ducks and runs away with Ted. N&V joins in and we have a 3 way tranny fest.
    Hmmm...that didn't come out right..." -podolefsky

  5. #5
    Senior Member
    Join Date
    Oct 2010
    Location
    North America
    Posts
    749
    Post Thanks / Like
    Quote Originally Posted by podolefsky View Post
    Here is my first piece of data to fuel this fire. It compares the D&D ES-15-6 (series) vs HPEV AC-15 (drop in replacement induction). Also shows graphs with 300 and 180A limits (5c and 3c for 60Ah).
    Well you're doing something wrong or getting bad data to start with. Assuming both motors (and associated controllers) have similar efficiencies, which I think is the case to about 5%, then with the same voltage and current limit for the two, peak power will be approximately the same magnitude. Which is clearly not the case on your plots. So something is amiss.

  6. #6
    Señor Member podolefsky's Avatar
    Join Date
    Oct 2010
    Location
    Boulder, CO
    Posts
    3,473
    Post Thanks / Like
    I think N&V has it mostly right. I'll add some more to the discussion, based on what I've learned here and there. I'll also try to be concise.

    We need to split out DC brushed into series and PM. PM is more efficient because you don't need to excite the armature field. In the motors that are out there, it seems to be ~80% for series vs ~90% for brushed PM. The commutator adds maintenance, but brushed PM is actually more efficient even than ACIM, since ACIM also has to excite the armature field. (Efficiency will vary with construction - I'm going based on the specs for common EV motors out there).

    Brushed PM will do regen fairly easily, whereas series will not (without essentially turning it into sepex). Sepex will do regen, but not as well as PM or ACIM. Sepex gets marginally higher RPM than a similar series motor because of field weakening, but the real reason to do sepex is regen or simple reversing (both of which are of questionable value on a motorcycle).

    DC series and PM suffer from BEMF at relatively low RPM, whereas ACIM is more limited by mechanical failure (armature flying apart). (This may also be true for PMAC, but I'm not sure.) This gives ACIM higher RPM capabilities (for similar size motor). Both ACIM and PMAC use complex controllers that excite the field windings in sequence, but that is where the similarity ends. AC induction uses sine waves, which is why they are used widely in industry - they will run off of 3-phase AC power without an inverter. To run from batteries, the inverter turns DC into 3-phase AC. For speed control, the controller only needs to know the motor RPM, but not position. Torque is generated when the applied field frequency differs from the induced field frequency, or put another way, the field frequency is higher than the rotor RPM.

    PMAC requires the controller to know the position of the armature magnets. They will only work with a controller made specifically for the motor, and need Hall effect sensors to figure out the rotor position. The controller sends pulse waves to the field windings in a specific sequence. They are generally more efficient than AC, again because the armature field does not use any current. More limited in RPM because the magnet material carries more mass and is harder to keep in one piece.

    DC series and PM can produce maximum torque from zero RPM, whereas ACIM needs to induce a field first. ACIM also tend to require higher current for the same torque because what the motor sees is not the peak DC current from the battery, but instead the RMS current to the motor which goes as DC_current / sqrt(2).

    Whew...as usual, happy to be corrected and / or further informed.
    - Noah Podolefsky -
    The GSX-E

  7. #7
    Señor Member podolefsky's Avatar
    Join Date
    Oct 2010
    Location
    Boulder, CO
    Posts
    3,473
    Post Thanks / Like
    Quote Originally Posted by lugnut View Post
    Well you're doing something wrong or getting bad data to start with. Assuming both motors (and associated controllers) have similar efficiencies, which I think is the case to about 5%, then with the same voltage and current limit for the two, peak power will be approximately the same magnitude. Which is clearly not the case on your plots. So something is amiss.
    I agree that it seems weird.

    I think I can partly explain it by saying that even though they say 72V, in fact the voltage is not always 72V under load. In the D&D data, under 450A load, voltage sags to about 58V. Efficiency is also much lower than the peak at this low RPM (~68% at 1700 RPM). Those two together result in lower power. So peak power isn't necessarily peak current * peak voltage.

    If you take the data from HPEV for the AC-15, it says 72V, 550A, which should be ~53HP. But their peak HP is ~32 at 2800 RPM. So it's not as simple as P = I_max*V_max. In an ACIM, 72V at the battery is 72 / sqrt(2) V RMS at the motor. Factor in 85% efficiency, and you get about 32HP.

    I need to stare at the source data some more to understand this, though.
    Last edited by podolefsky; 17 March 2011 at 2054.
    - Noah Podolefsky -
    The GSX-E

  8. #8
    Moderator Nuts & Volts's Avatar
    Join Date
    Aug 2010
    Location
    Columbus, OH
    Posts
    1,667
    Post Thanks / Like
    Nice additions Noah! and corrections

    I would like to add also the PMAC controllers dont necessarily need to be made to be motor specific, but PMAC controllers are not as universal as ACIM are. A PMAC like you said needs to know rotor position at all times for best operation. This can be accomplished by a couples of means, hall effect sensors, encoders, or resolvers. Hall effect gives the controller a position and speed once a rotation and the (some) controllers guesses the position in between the sensors being activated. Encoders use a magnet to send a wave function to the controller to give position of the rotor over a full cycle. Anyways a controller can be set up to run any of these sensor types, but hardware and software must be put in place to support all three. In theory I thought an ACIM controller can be used to control a PMAC motor if software allows it, ie the hardware is the same and a lot of software can be similar.

    So AC vs DC - efficiency becomes a factor - PMAC wins, controller cost is a factor - Brushed DC wins, motor cost is not always a big factor - Similiar cost of same power and weight motors, torque control/smoothness is a factor - AC wins, maintenance - AC wins

    Any other factors to add?
    Whats under my tank may shock you!!! R6 Build, Motor Thread , Sorry excuse for a blog/
    OSU IOM Build Thread, OSU Current webpage

    "Noah ducks and runs away with Ted. N&V joins in and we have a 3 way tranny fest.
    Hmmm...that didn't come out right..." -podolefsky

  9. #9
    Señor Member podolefsky's Avatar
    Join Date
    Oct 2010
    Location
    Boulder, CO
    Posts
    3,473
    Post Thanks / Like
    Quote Originally Posted by Nuts & Volts View Post
    Nice additions Noah! and corrections

    I would like to add also the PMAC controllers dont necessarily need to be made to be motor specific, but PMAC controllers are not as universal as ACIM are. A PMAC like you said needs to know rotor position at all times for best operation. This can be accomplished by a couples of means, hall effect sensors, encoders, or resolvers. Hall effect gives the controller a position and speed once a rotation and the (some) controllers guesses the position in between the sensors being activated. Encoders use a magnet to send a wave function to the controller to give position of the rotor over a full cycle. Anyways a controller can be set up to run any of these sensor types, but hardware and software must be put in place to support all three. In theory I thought an ACIM controller can be used to control a PMAC motor if software allows it, ie the hardware is the same and a lot of software can be similar.

    So AC vs DC - efficiency becomes a factor - PMAC wins, controller cost is a factor - Brushed DC wins, motor cost is not always a big factor - Similiar cost of same power and weight motors, torque control/smoothness is a factor - AC wins, maintenance - AC wins

    Any other factors to add?
    Good points about PMAC sensors. I didn't know the ins and outs.

    One small thing, AC motors are actually simpler and less expensive to produce than any commutated motor. It's the controllers that are expensive.
    - Noah Podolefsky -
    The GSX-E

  10. #10
    Senior Member
    Join Date
    Oct 2010
    Location
    North America
    Posts
    749
    Post Thanks / Like
    Quote Originally Posted by podolefsky View Post
    So peak power isn't necessarily peak current * peak voltage.
    Yes it is. Realize that peak voltage isn't going to be open circuit voltage of the battery but the voltage it will have at that peak current. And peak current is current limit.
    (This assumes a reasonable current limit which will not pull the battery voltage to below 50%.)

    In the D&D data, under 450A load
    But your plots used a 180 & 300A current limit.

    If you have the same current limit and battery voltage for each case, then peak motor power will be very close, unless one system is a real stinker. And why would you use that?
    Last edited by lugnut; 18 March 2011 at 0336. Reason: added ( )

Page 1 of 9 123 ... LastLast

Tags for this Thread

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •