Shop OBEX P1 Docs P2 Docs Learn Events
MOSFET motor driver — Parallax Forums

MOSFET motor driver

Daniel M.Daniel M. Posts: 14
edited 2008-10-26 03:57 in General Discussion
I am currently trying to build a small 10A motor driver from a few NPN MOSFET transistors. Whenever I bring the gate pin high, the voltage connected to the drain is about 30% lower than the voltage coming out of the source. In this case 12v in and 9v out. Is there likely something wrong with the transistors? I've tried a few of them but all give the same results(IRL520N). If not, do I need to find a different type of transistor to provide a lower power loss? Thanks!

Comments

  • Chris SavageChris Savage Parallax Engineering Posts: 14,406
    edited 2008-10-23 19:35
    This thread is being moved from the SX Forum to the Sandbox.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Chris Savage
    Parallax Engineering
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2008-10-23 19:50
    Can you post your schematic? Something in your description just doesn't add up.

    -Phil
  • Daniel M.Daniel M. Posts: 14
    edited 2008-10-23 20:27
    Here's a schematic of what I have hooked up. I also made sure that the signal ground was linked to the 12v supply ground. It seems almost too simple to me, which is why I suspected that there's just something I'm missing.
    motor.JPG
    532 x 475 - 16K
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2008-10-23 23:11
    Does the motor run when it's connected across your 12V supply? What is the DC resistance between motor leads? When the gate is "high", what is the actual voltage measured between gate and source right at the transistor leads? Try replacing the motor in your circuit with a 1K resistor. What happens to the drain voltage then when the gate switches "high"?

    Here's a hunch: Your connection to ground may not be solid enough, causing the MOSFET's source pin to rise slightly when it's conducting, which lowers the effective gate drive. This, in turn, causes the MOSFET not to saturate, leaving you with too much VDS and, probably, a very hot transistor.

    -Phil
  • Chris SavageChris Savage Parallax Engineering Posts: 14,406
    edited 2008-10-24 14:56
    You are also missing a 10K pull-down resistor on the Gate.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Chris Savage
    Parallax Engineering
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2008-10-25 07:43
    The IRL520 is rated at "absolute maximum" continuous drain current of 10A @25'C and that is with 10V applied to the gate, not 5V. This part is not such a great part and the "logic-level" gate drive is always a bit of a gimmick, check the datasheet and see how much current they quote for low voltages (think microamps). Your post sounds confusing as the source should be at ground potential (0V) and the drain connects to the negative terminal of the motor with the positive terminal tied to +12V. Your schematic does not indicate which is the drain and which is the source either. The only safe way to drive these MOSFETs is usually at higher gate-source voltages using some level translator or MOSFET driver, I usually just use LM358 opamps as drivers tied to the motor supply.

    It's usually a good practice to include a "current-limit" resistor in the gate drive, not so much for the MOSFET as it is voltage driven and doesn't need the resistor but more for limiting fault currents from the motor side from vaporizing the micro. Also, a pull-down resistor is not normally necessary as the I/O pin pulls it high or low but what happens when the pin is floating as is the case during reset or if it hasn't been programmed as an output? In this case any stray voltages will tend to partially turn on the poor MOSFET and end up cooking it before you realize it, so include a 10K pull-down from any I/O pin pronto.

    *Peter*
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2008-10-25 07:56
    Peter Jakacki said...
    The IRL520 is rated at "absolute maximum" continuous drain current of 10A @25'C and that is with 10V applied to the gate, not 5V.
    Peter, you must be thinking of the IRF520. The IRL520 is fully rated for a gate voltage of 5V. In fact, the Vishay datasheet for this part doesn't even mention a VGS of 10V in its electrical characteristics chart or graphs.

    I'm still betting on the bad ground scenario.

    -Phil
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2008-10-25 08:08
    Phil,

    Datasheet for IRL520N from IR says: ID @ TC = 25°C Continuous Drain Current, VGS @ 10V 10A
    www.irf.com/product-info/datasheets/data/irl520n.pdf
    Normally the logic level MOSFETs will operate from logic levels but just don't expect the full ratings from them, it just doesn't happen.

    But what you say about the ground probably adds up, very very bad ground.

    *Peter*
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2008-10-25 17:47
    Peter,

    Different manufacturers, apparently, have different ideas about what constitute "full ratings". In point of fact, looking at Figure 3 in the Int'l. Rectifier doc, one will see that the knee in the transfer characteristics curve occurs well before VGS = 5V. The "on" state resistances are 0.22 ohms for 5V and 0.18 ohms for 10V — not a huge difference, especially when comparing specs with the "non-logic" IRF520. Vishay's approach to this particular part is better, I think: spec it up to VGS = 5V, and let it go at that. I say this because the IRL520 is probably not a transistor one would choose to use with a 10V gate drive anyway.

    The lesson here, for anyone choosing a MOSFET, is to read the electrical characteristics and examine the graphs carefully. It's simply not enough to accept the ratings given in the part description. An even bigger challenge in picking a "logic" MOSFET comes when the logic "high" is only 3.3V. Then you really have to comb the datasheets for likely candidates!

    -Phil
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2008-10-26 02:44
    Phil,

    The datasheets might differ somewhat but this is the original spec from IR. As you said you have to read the datasheet carefully and there are always those little gotchas like "VDS = 50V, 20us pulse width", Tj = 25'C etc. What happens when you want to drive continuously (more than 20us) at low gate drive and high currents is that the substrate heats up and worse still it's uneven ( in fact the conventional MOSFET is made up of 1,000s of paralleled mosfet cells ) so this can degrade the part over time and certainly the datasheet never really covers real-life expectations. Designers beware.

    I can't imagine driving any MOSFET directly from 3.3V logic (5V maybe just) and expecting it to handle 10A motor loads. I used to be more optimistic but I choked on magic smoke too many times. In fact for 10A motors I have used the IRF3205 rated at 110As for commercial product, driving them at over 10V, and even paralled to achieve peace of mind. Automotive power supplies will vaporize a wrench so what do you think will happen to a tiny sliver of silicon?

    All said Phil, I hope I don't seem to be disputing with you but I'm saying my piece from experience for what it's worth.

    *Peter*
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2008-10-26 03:57
    Peter,

    I don't think we're in disagreement — especially not over the necessity to read MOSFET datasheets thoroughly! smile.gif There are some MOSFETs out there rated for 2.7V gate drive, but very few with sub-ohm "on" state resistances at that drive level. The Fairchild NDS331N is one of those rare specimens.

    -Phil
Sign In or Register to comment.