High power mosfet reliability (BL - ESC)

Discussion in 'General Electronics Chat' started by Shagas, Sep 20, 2014.

  1. Shagas

    Thread Starter Active Member

    May 13, 2013
    802
    74
    Hello

    I was wondering how reliable a power mosfet would be if used continuously at 35% of its max rating with peaks at around 65%.

    The reason I'm asking is because I'm currently doing various things with my by-parts assembled quadcopter and I'm curious as to how reliable the motor controllers are to know if I can fly over water etc.

    My motors are rated at a 21 amp max. I don't know if this is a peak maximum or stall current , I've yet to check with an ammeter (I don't have one that buff yet but I'm going to get one soon).
    My ESC's are rated at 30amps. I'm using these ones

    I've attached a schematic of a typical BL - ESC.

    Questions:

    - Do components like mosfets vary in quality? Are they crap cheap chinese ones and good quality ones?
    - As long as I don't come close to their max ratings can I trust that they won't give out on me for no good reason?

    Notes: After a full flight the ESC are barely warm to the touch. The motors are probably around
    40-45 Degrees C.
     
  2. Shagas

    Thread Starter Active Member

    May 13, 2013
    802
    74
    Aaaand it's not letting me upload a file :/ Some issue with Adblock
     
  3. faley

    Member

    Aug 30, 2014
    94
    13
    Hello Shagas. I'm guessing that this is for some sort of RC aircraft. I'm not an expert there. However, I was a military pilot for some years and have worked in industrial controls/electronics for far more years. So, I'll throw out some "points to ponder" and hopefully others can fill in the blank in a more decisive manner.

    As far as controller and semiconductor reliability- you will have few problems if the quality is top notch. China? It's like the early days of Korean and Japanese production- there's good stuff and there is junk. Perhaps others will provide recommendations.

    If this is an aircraft, the question is, "Is it fixed wing or rotary?" In essence, you will find that your motor will be subjected to more stresses if it is a rotary wing application.

    Regarding MOSFET's, in my experience, output devices are usually IGBT's, Darlingtons, GTO's, SCR's, and the like. MOSFET's, as a final output, may not be the best choice for rapid heavy loading/unloading, continually varying loads, and unpredictable transients, particularly in the case of rotary wing. However, they will likely work OK as a driver.

    Hopefully, others will chime in.
     
    Last edited: Sep 20, 2014
  4. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,669
    804
    i was pondering a quadcopter for a while, and I chose the desgin from these guys: http://ng.uavp.ch/FrontPage
    The bldc driver seems nicely done, and they provide the sourcecode for the atmega chips.
     
  5. MaxHeadRoom

    Expert

    Jul 18, 2013
    10,498
    2,364
    If it is running cool then it most likely is running well below the rating of the ESC and motor, when run flat out they usually heat up fairly fast.
    If there is a danger of water immersion you need to keep it covered, but the down side there is with cooling, water is never good with electronics.
    When using these with various motor combo's, it is good to get hold of a current/voltage/wattage monitor for testing and set up, these can be had on ebay for very cheap.
    There is also a couple of Utube videos showing how they are used.
    Max.
     
    Last edited: Sep 20, 2014
  6. Shagas

    Thread Starter Active Member

    May 13, 2013
    802
    74
    faley : It is an RC aircraft , as I mentioned. A multirotor to be precise.
    kubeek : Looks like a decent project. I decided to make my own thing because I got tired of spending time figuring out what is going on in other peoples platforms. That being said , I'm going to keep using commercial off the shelf BL controllers like the ones that I linked. I was just wondering how reliable the electronics inside them were.
    MaxH.R : Yes I'm going to order a decent high power RC wattmeter and test the current that they are running at.
     
  7. faley

    Member

    Aug 30, 2014
    94
    13
    My apologies Shagas. It appears it's time for a new pair of glasses on this end. To be honest, my original post was simply intended to draw attention to your query. For what it's worth, you may as well disregard my comments. The problems you face are very much different from those in the realm of my experience.

    Nonetheless, hopefully you've gained some insight and possibly inspiration.

    All the best.
     
    Shagas likes this.
  8. #12

    Expert

    Nov 30, 2010
    16,247
    6,744
    In the crude form, reliability is all about heat inside the transistor. If you have top quality transistors (Low RDSon) being driven with an excellent PWM chopper where the rise and fall times are...let's say 1% of the "on" time, your transistors are going to run cool, regardless of the power the motor is experiencing. This idea (almost) divorces the relationship of transistor stress compared to power delivered. The heating will still be proportional, but the multiplier can be made very low with proper design of the driver circuits.
     
    Last edited: Sep 20, 2014
  9. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,669
    804
    Another factor of reliability will be the voltage stress on the transistor from induced back emf. If the spikes are close to the rated voltage the reliability may suffer.
     
    Shagas and #12 like this.
  10. MaxHeadRoom

    Expert

    Jul 18, 2013
    10,498
    2,364
    My Parrot Drone quad copter uses a 1000ma/hr 10C LiPo battery which means it can run at 10amps continuous current.
    Only lasts for around 12minuits on a charge, however.
    Max.
     
  11. Shagas

    Thread Starter Active Member

    May 13, 2013
    802
    74
    Yes , the back emf might be something to look for. I wonder how much I can find out when I scope the motors under various loads/impulses.
    In my (very limited) experience even tough mosfets can be very brittle to even brief overvoltage/overcurrent.
     
  12. #12

    Expert

    Nov 30, 2010
    16,247
    6,744
    It depends on how fast your scope is. However, how well you can measure this becomes almost irrelevant if you simply include voltage limiters on the circuit. These can be snubbers, zener diodes, or Transient Voltage Suppressors (to name a few).
     
  13. ErnieM

    AAC Fanatic!

    Apr 24, 2011
    7,386
    1,605
    You want to check two things: the voltage stress the part sees, and the temperature it sees.

    Voltage has been discussed well here. Temperature has been ignored. The lower the temperature the longer the device will work. Let it run cool and it will run for years. Cook it and it will die in minutes.

    Even a small heatsink can do wonders, especially when you have moving air. Too bad you don't have room for fans on a flying machine.

    If you could say mount a fan near each motor you could cool the transistor very nicely. ;)
     
  14. jpanhalt

    AAC Fanatic!

    Jan 18, 2008
    5,671
    899
    I do not fly quadcopters, but I do fly some electric. Our club is fairly large for the area and several members use HK ESC's without problems. I believe your ESC is rated for 30A continuous. Operating at 35% of that should be no problem.

    The one piece of advice I can give has to do with ensuring you do not exceed either the motor's, ECS's or battery's ratings. Those ratings are the "allowable" rating. It is not a limit of what they may do, particularly for the motor and battery ratings. It is possible to exceed them easily and cause failure. As just one example, you could take your ESC, your motor, and a LiPo battery of appropriate voltage. Now, if the propeller is too big and/or has too great a pitch, you will easily exceed one or more of those limits and get a failure.

    Years ago when I did my first electric conversion of a sailplane, I made that very mistake. Sure, I went to the Neu Motors' site and did the calculations. On paper, the system was OK (barely). The glider performed beautifully and climbed almost straight up -- that was the problem. I was over stressing the battery during a steep climb-out and puffing it about every 3rd flight. A smaller prop solved the problem. Bottom line advice, as you assemble your quadcopter consider starting with somewhat smaller (lower power) props to be sure everything works as predicted. The correct propeller is a very important part of the system.

    John
     
  15. Shagas

    Thread Starter Active Member

    May 13, 2013
    802
    74
    @jpanhalt.
    I am not without experience in the realm of RC equipment.
    The props are slightly undersized for the motors , which have a max current rating of 70% of the ESC max rating and all that is powered by
    a 45-90C LiPo which who's continuous rated amperage is not exceeded even at peak (4 x) motor performance.
    So you can see that the ratings stand Lipo > ESC > motors as they should be and I took care to insure that when I was choosing the parts.
    Thanks for the advice though.

    I just did a quick scope test and I'm getting around 23-25 Vpp readings. Those are readings with the Gnd attached to one of the phases and probed at one of the other phases.
    I didn't see any excessive inductive kickbacks other than the usual stuff. Tested using a 60Mhz 500MS/s scope .

    I don't know what the ratings of the Mosfets are though , I'd have to dissect the ESC for that.

    @ErnieM I'm planning to test the temperatures once I place the final props on the quad . As I mentioned earlier , right now I'm using slightly undersized ones. I was thinking of physically loading the quad with an extra 800 grams or so
    and giving it a full flight , monitoring the onboard temps in real time by and RF link.
     
  16. Sensacell

    Well-Known Member

    Jun 19, 2012
    1,124
    266
    I would say that your MOSFETS are the last thing to worry about.
    I would focus on things like wiring and connectors, these are MUCH more likely to fail and down your flying machine.
    Vibration is the enemy.
     
  17. alfacliff

    Well-Known Member

    Dec 13, 2013
    2,449
    428
    here in the aerospace industrial world, more mosfets fail after a thunderstorm than any other time. if they are runing cool, they will last a long time. volotage spikes seem to be the main cause of failure.
     
Loading...