Shop OBEX P1 Docs P2 Docs Learn Events
L298 Dead Time — Parallax Forums

L298 Dead Time

AGCBAGCB Posts: 327
edited 2014-01-29 06:51 in General Discussion
I'm working on a balancing project and using a L298 driver module. My question is, do I need to provide protection from shoot through current?

I have researched the L298 data sheet and see nowhere that mentions either that it has this protection built in or that it must be provided in software or something.

To be safe I have so far halted it before a direction change for 8 uS but if it's not needed I could speed up the program a little.
I don't see any delay or anything in others programs.

Thanks for consideration.

Aaron

Comments

  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2014-01-28 09:00
    According to the datasheet, the turn-on delays for the direction inputs are slightly longer than the turn-off delays. This would seem to imply that external protection for shoot-through is unnecessary. If it were my project, though, I would not take that chance, and I would do what you are doing with the enable input.

    -Phil
  • AGCBAGCB Posts: 327
    edited 2014-01-28 10:24
    Thanks Phil
  • Duane DegnDuane Degn Posts: 10,588
    edited 2014-01-28 12:06
    I have several projects using L298N drivers so I was wondering if I should be worrying about this.

    I don't think there's a danger of shoot through with the L298.

    attachment.php?attachmentid=106550&d=1390939420

    From looking at the block diagram, I don't see how both switches on one side of the "H" can be open?

    Am I missing something (again)?

    I thought if both In1 and In2 were set high (or low), it would cause the motor to brake?
    L298nBlock140128a.PNG
    819 x 640 - 77K
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2014-01-28 12:09
    Duane,

    The schematic is not enough to determine this. It depends upon the turn-on and turn-off delays. If the source-current turn-on delay were shorter than the sink-current turn-off delay, for example, there would be shoot-through when In1 or In2 changed state. 'Same for the sink-current turn-on vs. the source-current turn-off. But according to the datasheet, that appears not to be the case. Nonetheless, the timings are so close that I would still disable the switches when changing state on the inputs, just to be sure.

    -Phil
  • Duane DegnDuane Degn Posts: 10,588
    edited 2014-01-28 12:29
    Phil Pilgrim (PhiPi) wrote: »
    Duane,

    The schematic is not enough to determine this. It depends upon the turn-on and turn-off delays. If the source-current turn-on delay were shorter than the sink-current turn-off delay, for example, there would be shoot-through when In1 or In2 changed state. 'Same for the sink-current turn-on vs. the source-current turn-off. But according to the datasheet, that appears not to be the case. Nonetheless, the timings are so close that I would still disable the switches when changing state on the inputs, just to be sure.

    -Phil

    Okay, I think I now understand the original question.

    Thanks for your patients (again).
  • ercoerco Posts: 20,256
    edited 2014-01-28 21:25
    Duane Degn wrote: »
    Thanks for your patients (again).

    Dr. PhiPi's patients have great patience.
  • Duane DegnDuane Degn Posts: 10,588
    edited 2014-01-28 21:32
    erco wrote: »
    Dr. PhiPi's patients have great patience.

    You know that crossed my mind as I typed it but I didn't stop to check.

    Having had a heart transplant and having spent more time waiting for doctors than most, it occurred to me many times that the name "patients" was likely derived from their need to have "patience".

    But I'm not going to call PhiPi "Dr. Phil", he deserve better.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2014-01-28 21:44
    Duane Degn wrote:
    But I'm not going to call PhiPi "Dr. Phil", he deserve better.

    Don't [quack] be so [quack] sure! [quack]
    duck%20quack-thumb-360x252-847.jpg

    -Phil

    (A friend of mine has an iPhone that he's programmed to emit different rings for different callers. When I call him, it sounds like a duck quacking.)
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2014-01-28 23:04
    You might take another peek at the L298 pdf. Source and Sink current turn on/off delays all seem to be sub 3 microseconds. See, Figure 3 for how the time delays accumulate.

    I'd just say the propagation delays in the logic and the transition times of the power devices are represented there as a total. So you might take your 8 microseconds down to 5 microseconds.

    The only other thing is the H-bridge voltage drop can be quite high... about 2 volts. So you input 12 volts, and the motors see 10 volts. Generally, the higher the voltage that you can drive your H-bridge at, the less power proportionally wasted in the H-bridge. Of course, a MOSfet design would be far better.

    And a higher voltage to the motor would also add to the response from a standstill.
  • AGCBAGCB Posts: 327
    edited 2014-01-29 06:51
    Loopy Byteloose wrote: »
    Of course, a MOSfet design would be far better.

    I hadn't thought of that. I do have a LMD 18200 on hand but have not looked at the data sheet yet. I also have a homemade MOSFET driver that I made quite some time ago but don't know if I trust my memory as to how it worked or even IF it worked.

    Oh well, more study and work to keep me busy.

    Thanks for all the replies.

    Aaron
Sign In or Register to comment.