Shop OBEX P1 Docs P2 Docs Learn Events
FIRE AND SMOKE on the MSR1 Prop chip — Parallax Forums

FIRE AND SMOKE on the MSR1 Prop chip

rpdbrpdb Posts: 101
edited 2010-02-17 22:49 in Robotics
Well, I got my new board hooked up only on three pings (0,1,2)·and the hbridge to two motors (powerwheels). The motors are new and run at· about 0.8 to 1.2 Amps with a surge of 3.2 Amps. The specs for the MRS1 board specs the hbridge for 3 Amps and the specs for the the hbridge chip specs 3 amps per channel with a surge of 5 amps. The h-bridge chip also has over temp and over current protection / shutdown built in.

I ran the code from the stingray store page and the pings were working and the motors were turning (on a stand on the bench, no load). I·approched the L and·R pings and the opposite wheel slowed as expected, I walked in front of the center ping and pings stopped pinging (no lights), then both·motors oscillated back and forth and then.....

to my horror....bright orange flame and factory installed smoke began leaking from the Propeller chip right next to the Vss and Vdd pins.

I am only guessing that a very high load on the h-bridge is causing not a shut down of the h-bridge but is causing the on board power supply to lose regulation to the 3.3V and 5V and supply and supply to much voltage to the whole board.

Has anyone else had this experience? I will talk to tech support, but not until·Tuesday as·Moday they are on holiday.

When I went to·Parallax·on Friday (see "WARNING" thread) I did pick up an extra MSR1 board.

I was hoping that someone here might have already·altered the·PingsOnAStingray code to use a PWM.Servo method instead of the PWM.DutyCycle method, so that I can run my motors on my HB25s instead.



Post Edited (rpdb) : 2/15/2010 7:06:33 PM GMT

Comments

  • P!-RoP!-Ro Posts: 1,189
    edited 2010-02-14 01:56
    I know how you feel, I had to make a pit stop at radio shack today because I blew up a necessary transistor. Your best bet on finding a problem is probably looking for a short, or the possibility that the voltage was too high and destroyed the regulator protecting your prop. Good luck.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    PG
  • mugurmugur Posts: 105
    edited 2010-02-15 10:33
    listening to rpdb i wonder if i should ever connect any motor to my MSR1 board [noparse]:([/noparse]
    My motors barelly eat 150mA on load (each) but reading what happend to you, i wonder what is wrong with the MSR1 as i already have a MSR1 board replaced.

    @rpdb: you think you can edit your post title and change MRS1 to MSR1 (this is the real name) so we can help the search function of the site?

    Post Edited (mugur) : 2/15/2010 12:08:40 PM GMT
  • CannibalRoboticsCannibalRobotics Posts: 535
    edited 2010-02-15 14:38
    I'm thinking a power up testing procedure needs to be developed by the manufacturer. If there are on-going issues with controllers, power supplies or circuit elements then we need to have an 'engineered' process to ferret them out.
    Lots of people have paid allot of money for those smoke containment devices. (Including me.)

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Signature space for rent!
    Send $1 to CannibalRobotics.com.
  • mugurmugur Posts: 105
    edited 2010-02-15 14:46
    i subscribe to this. i am sure parallax is testing the MSR1, but maybe other kind of testing is needed.
  • ercoerco Posts: 20,256
    edited 2010-02-15 16:35
    Solidarity, Parallax Brothers!

    Go easy on Parallax, it may be a board flaw, or it may not. It is always possible that a user wiring error caused the problem. We've all released the magic smoke at some point. Even with overcurrent/overheat protection, there are still "perfect storms" of miswiring and crossed code that can bypass any chip's chances of saving itself.

    I guarantee you Parallax is max attentive to your post and are looking for solutions already. Per your previous post, their' customer service is second to none. Remember that they already replaced one board for you. Afterward, you sang their virtues from the treetops. They will help you more than anybody else will. Hope you and they can get to the bottom of the problem quickly.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    ·"If you build it, they will come."
  • mugurmugur Posts: 105
    edited 2010-02-15 16:44
    More than true. But still, can be a bunch of weak regulators?
  • kwinnkwinn Posts: 8,697
    edited 2010-02-15 16:53
    Don't you just hate it when the smoke gets out and the mirrors have nothing to work with. Many years ago I used to service some equipment that was notorious for having the power supply blow and do that to almost all the chips in the chassis (~12 boards with 20-30 chips each).
    The problem was with the power supply design. Really really bad.
    A 5V zener and 3 transistors for +5V
    A 7V zener riding on the 5V supply with 3 transistors for the +12V
    A 12V or 36V zener riding on the 12V supply with 3 transistors for 24V or 48V
    A 146V or 120V zener riding on the 24V or 48V supply with a 2n3440 for the 170V nixie tube supply

    Not too hard to see where this is going. The 2n3440 shorts, the 24/48V pass transistor shorts, the 12V pass transistor shorts, the 5V pass transistor shorts, and 200+ TTL chips let the smoke out.

    It really pays to check:
    The regulator output voltage at no load and full load with a dummy load.
    The current draw of motors, solenoids, and other peripherals with a current limited power supply.
    The driver circuit and load with a current limited supply.

    Once this is done wire it up and recheck it before connecting the prop board or whatever is controlling the system.
  • ercoerco Posts: 20,256
    edited 2010-02-15 16:59
    If today was April 1, I might assume that MSR1 board was an acronym for Maximum Smoke Release 1.

    Bad joke, but sometimes a little humor helps at dark times.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    ·"If you build it, they will come."
  • rpdbrpdb Posts: 101
    edited 2010-02-15 21:20
    Today I am working on graphing out a full set of measurements on the motors load charachtaristics, as well as testing the current draw of the pings I am using.

    I'll be headed down to Parallax Tuesday to meet with tech support and if possible the folks at engineering, and hopefully come up with a failure analysis conclusion.

    Human error as far as miswiring is the most frequent cause and I am the first to admit I have made mistakes before. I have triple checked my wiring more that thrice and am testing my test equipment. I am checking ground loops, isolation, back EMF etc.

    I hope I will be able to duplicate and find the problem so that I can share with all in these forums as how to avoid such smoke leaks.

    I will test that the LT3501 can maintain full-load regulation and that the L6205 H-bridge exhibts its' shutdown mode for temp and current overload conditions.

    Please Kids, don't try this at home. I am a trained professional, but more importantly only an hour away from Parallax for hands on tech support and RMAs.

    The folks at Parallax are great supporters of their products, and the axiom holds true there that "if you love what you do for a living, you'll never have to 'work' a day in your life".
  • DufferDuffer Posts: 374
    edited 2010-02-15 23:25
    I don't know if this will help, being anecdotal rather than measured and documented, but I've completely stalled both motors in both forward and reverse for up to 15 seconds at a time without ill effects (I had some low clearance issues in my test area). This has happened using PWM (up to 100%) and when setting the motor direction pin high as in the motor test code.

    Most of the actual "running" being tested also uses a standard servo·continuously panning a PING sensor using the SERVO command of PWM_32_V2.spin (used for both the motor PMW and the PING servo). I've not experienced any significant heating of the Propeller chip or other components and certainly no·smoke or flames·as yet. My battery supply is a 7.2V NiMH that puts out over 8V when fully charged.

    Duffer

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Any technology, sufficiently developed, is indistinguishable from magic.· A.C. Clark(RIP)
  • rpdbrpdb Posts: 101
    edited 2010-02-16 00:51
    Yo Duffer,

    I would like to know if someone has tested and has the load specs for start/stop/stall current plotted over voltage and load on the stingray motors.Just for a benchmark.



    Post Edited (rpdb) : 2/16/2010 1:05:14 AM GMT
  • DufferDuffer Posts: 374
    edited 2010-02-16 19:45
    rpdb,
    This is the best I could do. I don't have the tools and equipment anymore to do this kind of testing properly. This graph is using an RS-232 connected digital multimeter sampling at 1Hz and running the motor test code below.
    repeat
      dira[noparse][[/noparse] 27..24 ] := %1111 ' Set P24 Through P27 To Output
      outa[noparse][[/noparse] 26 ] := 1 ' Right Motor Forward
      waitcnt(clkfreq * 2 + cnt) ' 2 Second Pause
      outa[noparse][[/noparse] 26 ] := 0 ' Right Motor Stop
      waitcnt(clkfreq * 2 + cnt) ' 2 Second Pause
      outa[noparse][[/noparse] 27 ] := 1 ' Right Motor Reverse
      waitcnt(clkfreq * 2 + cnt) ' 2 Second Pause
      outa[noparse][[/noparse] 27 ] := 0 ' Right Motor Stop
      waitcnt(clkfreq * 2 + cnt) ' 2 Second Pause
    

    The peaks at ~200mA is the motor running with no load. Peaks at ~500mA are at ~half stall (wheel is loaded to half of no-load RPM). Peaks at ~1A is full stall (motor loaded to full stop). All tests were run using a switching power supply at 7.5V.

    During about 5 minutes of testing, including a number of full stall periods of up to 15 seconds each, none of the board's components were even warm to the touch.

    Duffer

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Any technology, sufficiently developed, is indistinguishable from magic.· A.C. Clark(RIP)

    Post Edited (Duffer) : 2/16/2010 11:45:02 PM GMT
    562 x 412 - 53K
  • rpdbrpdb Posts: 101
    edited 2010-02-17 22:49
    Thanks Duffer,

    That will help alot!

    I'll have to postpone my trip to Parallax, got the flu.
Sign In or Register to comment.