Shop OBEX P1 Docs P2 Docs Learn Events
Mystery L298 issue — Parallax Forums

Mystery L298 issue

T ChapT Chap Posts: 4,223
edited 2013-10-06 16:22 in General Discussion
I have a Seeed Studio motor sheild from Radio Shack. A few days ago, I connected it to a car lock motor, powered by a 12VDC 1.5am power supply. A Propeller sends a 3v3 signal to the input 1 to move the lock one direction, another input returns the lock. The first day, I ran over 30,000 cycles overnight. The next day, made some hardware changes and left it to run all night, working perfectly for the first hour. The next day, the lock is not moving, but is making the attempt to move but acts like it has no power. So I connected a brand new L298 driver(identical), it is doing the same issue, acts like it doesn't have enough power. So I connect the inputs to the Seeed Studio motor shield direct to 5V instead of the Prop pin, no change. I connect the motor wires direct to the power supply leads and it works great. This makes no sense. Any suggestions on what to check for? I can't check now but tomorrow I will test the cables for a bad ground between the Prop system ground and the supply to the motor driver.

Comments

  • ercoerco Posts: 20,256
    edited 2013-10-01 18:10
    Did your car lock motor fail?
  • T ChapT Chap Posts: 4,223
    edited 2013-10-01 18:30
    No, as I mentioned if I connect the motor direct to 12V it works fine. I have replaced the L298 driver with the same low power symptom. It seems like the L298 is not powering the motor fully.
  • Mike CookMike Cook Posts: 829
    edited 2013-10-01 18:45
    Doesn't the l298 have something like a 2 volt drop? Will your motor run properly with the reduced voltage?
  • T ChapT Chap Posts: 4,223
    edited 2013-10-01 18:48
    The motor ran 30,000 cycles just fine on it, then for unknown reasons barely moves. Replacing the L298 doesn't fix it.
  • kwinnkwinn Posts: 8,697
    edited 2013-10-01 19:18
    30,000 cycles is would be more than most car locks see in ten years of use. Probably enough wear and increased friction on it that it needs more voltage/current than the L298 can provide.try measuring the current draw with a 0.1 ohm series resistor in series with the battery and then with the L298 driving it.
  • Duane C. JohnsonDuane C. Johnson Posts: 955
    edited 2013-10-01 19:22
    Try replacing the door lock thing.
    While 30,000 cycles sounds OK over the life of a car door doing it in 8 hours seems a bit much.
    I think these things have a max duty cycle spec. If driven to much the windings could get hot and have shorts.
    Shorted windings in motors tend to make them not run.

    Duane J
  • T ChapT Chap Posts: 4,223
    edited 2013-10-01 19:27
    Duane, the motor works perfect though direct to 12V. But I will try that.
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2013-10-02 00:57
    T Chap wrote: »
    No, as I mentioned if I connect the motor direct to 12V it works fine. I have replaced the L298 driver with the same low power symptom. It seems like the L298 is not powering the motor fully.

    Ummm, have you really considered that fact that the L298 h-bridge provides 2 diode voltage drops in the process of doing its job? Provide 14 VDC to the H-bridge and see what it does.

    If you carefully read the pdf for the L298, the two transistors involved seem to require 1 Volt each at saturation... so with a 12 volt input, you really are only driving the motor/coil at 10 volts at best.

    And after that, if you are stalling the motor/coil with a substantial load, you may be going beyond the ability to deliver adequate power to the motor.

    In all honesty, I tend to think of the L298 as a toy motor driver. Bob Blick's H-bridge with TIP120/TIP125 drivers will deliver power to bigger motors 3amps at 24 volt = 720 watts. Think about it, 75 watts is roughly a mere 1/10th of one horsepower.
  • T ChapT Chap Posts: 4,223
    edited 2013-10-02 15:24
    I built an hbridge with IRF540's using a mosfet driver today. It works but it not super strong as with connecting the motor direct to the 12v(actual 13.8) wall mount supply. I put my volt meter in the path to the motor and set up the meter to 10amps, it reads 2.5 under load. The power supply states it is a 1.5amp supply. I am going to look for a new supply tomorrow with at least 3 amps and see what happens.
  • phatallicaphatallica Posts: 64
    edited 2013-10-02 19:13
    T Chap wrote: »
    I put my volt meter in the path to the motor and set up the meter to 10amps, it reads 2.5 under load. The power supply states it is a 1.5amp supply. I am going to look for a new supply tomorrow with at least 3 amps and see what happens.

    1) L298 will sustain a max of 2A continuous or 2.5A repetitive pulsed (much more than your previous thread). The new measurements suggest that the design may be marginal with L298. Realistically, the current was probably a bit lower with the voltage drops.

    2) The previous thread also suggested 1 second duration 10-30 times a day. I think Duane J may be on to something - maybe not the motor coils, but mechanical bits could also suffer under the constant heating. Care has to be taken with highly accelerated testing so as to not create failure modes that would not otherwise exist. A 10% or so duty cycle would help to prevent unrealistic heating anomalies and still simulate about a year of use per day.

    3) You mention that the actuator was "attempting to move". Measuring the voltage output to the motor will let you know if the supply is drooping too much. If the output is, say 2V under the supply voltage, then the actuator maybe needs more voltage. If the output is, say 5V under the supply voltage, then it seems that the supply cannot keep up with the current. If it is ~0V output, then the circuit is NG.

    If you change the supply and all is good, then awesome.
  • ercoerco Posts: 20,256
    edited 2013-10-02 20:54
    IIRC, that L298N shield was on sale recently at the Shack for ~$7, a good deal. China clones are exactly half that price and likely half the quality. :)http://www.ebay.com/itm/Dual-H-Bridge-DC-Stepper-Motor-Drive-Controller-Board-Module-Arduino-L298N-O0-/300962651400?pt=LH_DefaultDomain_0&hash=item4612c59d08

    Per Mike & Loopy, there is a lot of voltage drop through a 298. I have used the cheap China ones with success in a higher voltage (30V), low current app on a robot chassis. Percentage-wise, that 2-volt drop is less noticeable at higher voltages.

    Edit: I just saw that the Shack also carries another L298 shield for $2.48: http://www.radioshack.com/product/index.jsp?productId=12333768
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2013-10-03 08:25
    I think it could be said that the unit has required more voltage after substantial use. If it is a brushed motor situation, the brushes would easily be on their last legs.

    Also consider the reality of automotive 12 volts. It is more than likely to be at 13.8 if the battery or higher if the battery is on full charge or being charged by the alternator.

    We keep running into newbies that don't understand the automotive charge circuit can run as high as 14.2 under normal conditions, and that 12 volts is a rather silly low end figure meant to indicate something... maybe the voltage at sub-zero starting conditions. All batteries suffer from sub-zero weather.

    You do have to admit that opening and closing electric door locks 30,000 cycles is something like 10 time per day for roughly a year if you figure 50 weeks of 6 days per week ( rest on Sundays and two weeks vacation).

    I can seem to see where this is NOT normal wear.

    MosFets certainly make the H-bridge much less wasteful and eliminate the voltage drop issue. But one always has to provide some extra voltage when using an H-bridge. The best solution is to have motors that run at 24 and 36 volts, because the wiring in less bulky and the voltage drops affect power less.
  • Buck RogersBuck Rogers Posts: 2,185
    edited 2013-10-03 08:39
    I think it could be said that the unit has required more voltage after substantial use. If it is a brushed motor situation, the brushes would easily be on their last legs.

    Also consider the reality of automotive 12 volts. It is more than likely to be at 13.8 if the battery or higher if the battery is on full charge or being charged by the alternator.

    We keep running into newbies that don't understand the automotive charge circuit can run as high as 14.2 under normal conditions, and that 12 volts is a rather silly low end figure meant to indicate something... maybe the voltage at sub-zero starting conditions. All batteries suffer from sub-zero weather.

    You do have to admit that opening and closing electric door locks 30,000 cycles is something like 10 time per day for roughly a year if you figure 50 weeks of 6 days per week ( rest on Sundays and two weeks vacation).

    I can seem to see where this is NOT normal wear.

    Hello!
    Right!
    Those electric door-lock motor-actuators are not like the electric motors that run some other items in a vehicle. Also the voltage across the storage battery in the car is usually between 10.8 to 13.8 to as high as 14 even. It requires voltage regulators to get everything set correctly. And even so there is drift. The big problem with those things is that they expect to be activated only to lock and unlock car doors. Typically at the start and stop of each trip.

    So don't expect much from an albatross like that.
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2013-10-03 08:45
    My dad bought a Buick Skylark convertable in 1965 with electric windows and door locks.

    We soon discovered when you leave the top down and it rains that you have no electric windows or door locks. Whoever had the bright idea to put all the switches on in the armrests on the top surface just made it so easy for rain to destroy all the switches.
  • T ChapT Chap Posts: 4,223
    edited 2013-10-03 09:49
    The scrutiny for high reliability is because it will cost 1k to 2k to access the motor and latch if it went bad. The good thing about having this experience is that out of it came a new design that relocates the motor and optical limit sensors to a remote location that is accessible. I found a push pull cable at Mcmaster that is 6' long, very heavy duty and low friction. Now the electronics will be easy to access and the stress is gone for the failure possibility.

    The design of the custom Hbridge is still in the works, but will use IRF540 and a mosfet driver IC.


    As for the door lock motor, the nice thing about it is that in a power outage, the latch can still be moved manually since the door lock motors do not lock the mechanism up.

    Thanks for the advice.
    320 x 240 - 36K
    320 x 240 - 27K
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2013-10-04 06:00
    Back in the 1960s, I think everyone (meaning everyone male) wanted to remove the manual doorlocks (they replaced them with one that were impossible to grasp). They you took your date to 'watch submaries in the moonlight' and locked all the doors from your master panel... so she couldn't runaway.

    Very politically incorrect for 2013... but that was indeed what electric doors were about back then.

    These days, it is about locking all the doors with your remote. Nobody messes with the pulls any more.
  • T ChapT Chap Posts: 4,223
    edited 2013-10-06 13:37
    The motor I was using had taken a hit with the overnight non stop testing. I replaced it and the L298 board driver it but the there is not quite enough power. Even using a custom Hbridge with IRF540's it is lacking. I have tried with a 3amp 13.8 supply, and a 2amp 19V supply. I like the compactness of the car lock motor, plus the fact that it is a rack and pinion style setup so that with no power the motor with spin easily. I am looking for any suggestions on something that I can replace this motor with. The throw is 1.1" on this motor, the motor has flags mounted on that that pass into optical limit switches on each end of travel, so the limit switches kill power at each end. I prefer not to use a stepper for several reasons. I suppose I could build my own rack and pinion or a course pitch lead screw that would still allow the motor to spin if the power is off.

    I should mention that speed is important, ideally .5 second up to 1sec for the motion. Also, experimenting with solenoids was not good, in that there is a knock at each end due to inertia slamming the device into the fixed stops. Typically when the latch hits the stops, it bounces back and landing on a fixed position after the bounce is not easy.
    640 x 478 - 117K
  • Too_Many_ToolsToo_Many_Tools Posts: 765
    edited 2013-10-06 16:22
    My thoughts also.
    kwinn wrote: »
    30,000 cycles is would be more than most car locks see in ten years of use. Probably enough wear and increased friction on it that it needs more voltage/current than the L298 can provide.try measuring the current draw with a 0.1 ohm series resistor in series with the battery and then with the L298 driving it.
Sign In or Register to comment.