Bizarre voltage regulator problem (unstable) on MSP430 LaunchPad board

Discussion in 'General Electronics Chat' started by tom66, Sep 1, 2010.

  1. tom66

    Thread Starter Senior Member

    May 9, 2009
    2,613
    213
    I connected an LED up to my LaunchPad board to test out the features of it.

    But I'm getting the weirdest of symptoms from what seems to be an unstable voltage regulator.

    The voltage regulator (3.3V) will continue normally for minutes, but suddenly, its output will drop by about 200mV, and it will start swinging wildly about the place. This is measured from the anode of the LED I have connected - I do not see the effect on the cathode (which deepens the mystery further.) The effect will often happen if the reset button is pushed, or if the regulator is touched.

    I assume since the LaunchPad was professionally designed the designer considered the effects of the voltage regulator. It's a tiny TSSOP-8 package, but only has to drop 5V USB to 3.3V.
     
  2. beenthere

    Retired Moderator

    Apr 20, 2004
    15,815
    282
    Does the regulator get hot? How is the LED hooked up? If the cathode goes to ground, it would be a great surprise to find a voltage varying on it. Are there any decoupling capacitors with iffy solder?

    This is true, but only if the company did not tell him to somehow make the circuit cheaper by a certain amount.
     
  3. tom66

    Thread Starter Senior Member

    May 9, 2009
    2,613
    213
    The regulator is cool to the touch.

    The boards are only $4.30 each, so perhaps they did cheap out on the capacitor?

    I'll have a video up in a few minutes showing what I mean.
     
  4. Potato Pudding

    Well-Known Member

    Jun 11, 2010
    684
    92
    In one of the Texas Instruments PDF Abstracts, MSP430 Software Coding Techniques, they describe some things that you need to be aware of for coding that can prevent instability.

    It works best with consideration of the design featuring low power usage.

    You need to prevent code from taking too long to initialize, so that the Watchdog timer goes into a reset loop. You can either play nice with the watchdog or just disable it if it is not needed.

    You need to allow time for the oscillators to stabilize. Too fast a code boot will cause glitching, so for shorter programs start them with a delay of about 250ms especially for the Low Frequency Oscillator which you might be addressing for an LED flasher.

    Certain functions can only be handled with low level commands. So if your programs are compiled from C you need to know what needs to be called with direct instructions.

    Frequency is voltage sensitive but that should be sorted in the Launchpad board.

    Programming is also voltage sensitive which I am certain is sorted in the launchpad board.

    If you have read the launchpad datasheet and you are certain you didn't overload the output then before you give up try this version of doing a blinking LED from the launchpad learning community.
     
  5. tom66

    Thread Starter Senior Member

    May 9, 2009
    2,613
    213
    Here's a video of what I'm talking about. Sorry it took so long, but the wireless signal in my room is really weak. (I'm going to get a booster/range extender soon.)

    http://www.youtube.com/watch?v=ImIbUEMUJ3Q

    The code essentially consists of turning the LED on, then running an infinite loop. I did this as a 'hello world' experiment.

    The LED had a 330 ohm resistor, it was quite dim.

    Anyway I fixed this problem, I don't know how, but it isn't happening at all any more. But, I am curious, what would cause such bizarre behaviour?
     
  6. Potato Pudding

    Well-Known Member

    Jun 11, 2010
    684
    92
    That looks like it was as likely to be a wiring problem with the breadboard.

    If you don't see it again I would guess that the launchpads regulation is not the problem.

    It would be easy to suspect they had 4 year olds in Mongolia putting these Dev boards together when they are only $4.30. Most commenters say they are good enough to make you want more which is likely the idea of pricing them so low.

    It's like Printers for $29 and ink refills for $85.

    They have lots of toys to go with the Launchpads and other Dev boards for form factors other than PDIP, that are higher priced.

    I think they also just plan to sell a few Million of the MPS430 chips with getting Devs and techheads everywhere to try them as the biggest step.
     
  7. Potato Pudding

    Well-Known Member

    Jun 11, 2010
    684
    92
    They don't suggest that you drive LEDs direct from the outouts at all.

    They say to use the MPS430 ouput to drive a transistor and use the transistor to switch the LED.

    These are power scrounger devices rated at only about 1Ma per output. If you must drive an LED direct I think you would have to use multiple outputs in parallel to get say 10mA.

    With 330 Ω and only 3.3V it would be very dim.
     
Loading...