Diodes Failing

Discussion in 'The Projects Forum' started by Maketronic, Jul 4, 2011.

  1. Maketronic

    Thread Starter Active Member

    Mar 21, 2009
    49
    0
    Hi all

    I have been asked by a colleauge to work on a pwm dc motor speed control for a 55lb electric trolling motor.
    My friend claimed that the motor draws 25amp however online specs say 50 amps.

    When I built the speed controller I built it in a power electronics board (5x parallel mosfets and diodes) and a controller board ( a 12f683 micro with an analog to do the pwm work).

    Originally my friend had a crude pwm generator switching some mosfets which failed when a back emf diode failed short circuit and now I hear that it has 'let the smoke out' again and I suspect that my new board has done the same again.

    I have 5 x IRLF34N Mosfets switching in Parallel (each with there own 10 ohm gate resistor) and to protect against back emf I have 2 x diodes in parallel with the motor. (the diodes arent anything fancy 1N series if I remember correctly)

    It seems to be that the back emf diodes will be happy to work for an hour or so then they fail short circuit.

    How do I choose suitable back emf diodes so I will not keep having these problems?

    I dont know if it makes any difference but the switching frequency is around 15Khz

    Many Thanks

    Bruce
     
  2. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    I think you meant IRLZ34 MOSFETs.

    While the MOSFETs have a fairly low Rds(on) - 50m to 70m depending on Vgs - giving you ~360mV drop with 30A current flow for 5 in parallel, the reverse-EMF diodes will have roughly the same amount of current going through them at ~30% to ~70% duty cycle, but the voltage drop across the diodes will be significantly greater.

    Your reverse-EMF diodes need to have the same current capacity as the MOSFETs, but their heat sinking needs to be even better. Using power Schottky diodes will help somewhat, as Shottky diodes generally have a lower Vf than fast recovery or standard diodes.

    You wouldn't want to use a standard diode; it needs to be at least a fast-recovery diode. It's not the turn-on time that's the problem; it's the turn-off time.

    Why don't you post your schematic diagram the way it is right now?
     
    Maketronic likes this.
  3. #12

    Expert

    Nov 30, 2010
    16,325
    6,818
    "1N series" doesn't really give enough information. 1N5819 is a 1 amp Shottky diode, so it can't be said that there are no Shottky diodes that start with 1N, but the problem is most probably what wookie said. Believe him. He's smart about these things.
     
    Maketronic likes this.
  4. Maketronic

    Thread Starter Active Member

    Mar 21, 2009
    49
    0
    Hi Guys

    Thanks heaps for your help, sorry it is so long since I got back on here, I forgot to subscribe to my own post so I thought nobody had replied.
    I have not got means of hosting either a jpeg of pdf of my schematic at the moment but I will post it once I have organized hosting.
    It has come as quite a suprise to me that the diode has to be capable of carrying the full current.
    I have purchased the following diode of eBay, hoping it will suffice. http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=350471833836

    Once again Many Thanks

    Bruce
     
  5. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    If you click the "Go Advanced" button near the bottom of the page, you'll get a new screen that has a "Manage Attachments" button near the bottom. If you click on that, it'll allow you to upload certain types of files to the website.

    I much prefer .png to most other graphic file types; it's usually compact, doesn't need any software to view other than a browser, and is not "lossy" like .jpg format files are.

    .pdf's are OK, but I have to download them to view them, and load them into a viewer, so that's extra time and steps.
     
    Maketronic likes this.
  6. Maketronic

    Thread Starter Active Member

    Mar 21, 2009
    49
    0
    Thanks Sgt Wookie

    I have uploaded it as a pdf i'm sorry because I couldnt seem to get a image clear enough to read values etc.

    The diodes I used were small recctifier diodes either 400v or 100v at 1A so very insufficient for this project (they had popped like popcorn)

    We learn the hard way though dont we.
     
  7. CDRIVE

    Senior Member

    Jul 1, 2008
    2,223
    99
    These plots may interest you. As you can see, a reactance wants to give back what it took. The Plots are in Amps.
     
    Last edited: Jul 8, 2011
    Maketronic likes this.
  8. Maketronic

    Thread Starter Active Member

    Mar 21, 2009
    49
    0
    Awesome Thanks CD Drive.

    I guess the energy dissipated in heat through the diode is simply its foward voltage drop x current passing through it? I guess this too is why we want to use Shottky diodes?

    Just as an aside, what happens if we use a cap in parallel with the motor as far as back emf and pwm?

    Regards

    Bruce
     
  9. wayneh

    Expert

    Sep 9, 2010
    12,147
    3,057
    By definition, yes. Water falls over a cliff, it loses energy in proportion to how far it falls (voltage) and how much water (current).
     
  10. Jaguarjoe

    Active Member

    Apr 7, 2010
    770
    90
    The discharge time of the inductor is longer with low voltage diodes.
    Some inductive ignitions use very high voltage zeners (350v) which allows the coi primaryl to dump fast, collapse the field fast and generate the spark.
    Fuel injectors need fast closing time so they use higher voltage diodes too, like ~47v.
     
    Maketronic likes this.
  11. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    Here is a simulation of a PWM'ed inductor, which more closely mimics what you are trying to do. Inductor value and current will be different, but the principle is the same.
     
    Last edited: Jul 9, 2011
    Maketronic likes this.
  12. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    It's basically the average current times the average Vf (forward voltage).

    You don't want to use a very large cap at all, less than 1nF/1000pF, usually somewhere between 220pF and 470pF. The small cap will help to suppress peak voltages prior to the diode turning on. If you use too large of a value, you will put a lot of stress on the MOSFET, as surge current during turn-on will be very high.

    You should look at using a TVS/Transorb:
    http://en.wikipedia.org/wiki/Transient_voltage_suppression_diode

    Use one rated for somewhat more voltage than your motor will see across it, and less than the Vdss rating of your MOSFETs. Make sure it has adequate heat sinking.

    By the way, you're driving five MOSFET gates from a single PIC output pin - that's a heck of a load for it. You should really be using MOSFET gate drivers, as the PIC outputs are rated for +/-25mA peak.
     
    Maketronic likes this.
  13. Maketronic

    Thread Starter Active Member

    Mar 21, 2009
    49
    0
    Thanks SgtWookie and RonH

    I had seen tvs transorb diodes since having these problems but I have not yet tried one.

    Thanks for the simulations too. I was going to ask what you did them in and then when I clicked the asc file it opened up in lt spice.
    I will have to spend some time on it and try to figure out how to use it.

    Regards

    Bruce
     
  14. Adjuster

    Well-Known Member

    Dec 26, 2010
    2,147
    300
    Another option for getting faster decay is to put some resistance in series with the anti back emf diodes. Doing so will reduce the average power dissipated in the diodes, as the current decays faster.

    This requires care because you cannot let the voltage rise too high, and of course the resistor will need to be able to handle the power dissipated in it.
     
    Maketronic likes this.
  15. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    When you are PWM'ing a motor, I don't think you want to waste power in resistors. The idea is to keep the current flowing through the diode and the motor during the OFF time of the PWM switch.
     
    Maketronic likes this.
  16. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Thanks for jogging my memory ... this also makes TVS diodes a poor choice, as a lot of power will be dissipated in them.

    Large Schottky diodes are hard to find nowadays, because people are using power MOSFETs as "ideal diodes" instead, which would really be the way to go in this case. The low Rds(on) MOSFETs that are available now sure beat the relatively high Vf of diodes. But, the circuitry gets more complex.
     
    Maketronic likes this.
  17. ifixit

    Distinguished Member

    Nov 20, 2008
    639
    108
    Hi All,

    The 1A diodes are likely currently blowing because the inductive kick back (not motor back EMF) has edges that are way too fast for the diodes to react to, and buy the time they do react there forward voltage has risen too high causing excessive power dissapation and therefore overheating and death eventually. Remember, there are fast edges generated by the brushes making and breaking at an RF rate once the rotor is turning.

    Small, more robust, 3A Schottky diodes are all that is needed for this application to suppress the non-repetive inductive kicks during start-up or rotor-lock situations.

    Use a transorb MOV across the motor at the driver end to suppress all the fast moving edges so diodes and MOSFETs don't have to experience them. A 0.1uF 200V extended-foil film capactor would be a good idea as well, or maybe instead of the transorb.

    Motor back EMF is never suppressed by the protection diode since the "back" EMF is in the same direction (forward) as the applied voltage. When the applied voltage is remove the motor continues to generate a EMF (voltage) as the motor comes to a stop.
     
    Maketronic likes this.
  18. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    Ifixit, this is a PWM application. The motor current flows through the diode (100-DC) percent of the time, where DC is motor duty cycle (percent).
     
    Maketronic likes this.
  19. ifixit

    Distinguished Member

    Nov 20, 2008
    639
    108
    Hi Ron H,

    I am aware it is PWM and my post was based on the circuit posted in post #6.

    I believe brushed, DC motors do not reverse there output when disconnected from the power source as in the case of a toggle switch, or a PWM switch. However, switching the motor will cause reverse EMF from what little inductance is apparent in a rotating armature. These pulses are usually very short (~1uS) with a peak current equal to the armature current at the time. A fast recovery diode can absorb this energy when it conducts in the forward direction once every 67uS (PWM 15KHz). The RMS current will be less than .4 amps.

    The brush noise is in the nS range and will be missed by the diode, hence the need for a capacitor to slow and suppress these fast pulses.

    It has been awhile since I worked with these things so if I have time tomorrow I'll do a few tests with some motor I have in my junk bin. Sometimes my memory suffers from the CRAFT effect:)

    Regards, I'm off to bed,
    Ifixit
     
    Maketronic likes this.
  20. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    I may be all wet, then. I always assumed that a PWM'ed DC motor reacts (no pun intended) like an inductor.
     
    Maketronic likes this.
Loading...