This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

What the hell is going with RST on msp430g2553

Other Parts Discussed in Thread: MSP430G2553, MSP430F5529, MSP430G2001

After succesfull debug of my project with msp430g2553 (not launchpad) I faced with a problem on standalone run.

Due do datasheet, /RST pin pull-uped with 47k resistor to VCC. No success.

Than I pulled-down /RST to GND with 2,2nF capacitor without any effect.

After some googling I also soldered 0,1 uF capacitor between GND and VCC but still circuit don't work out of debug.

During the debug /RST has voltage level 3,1 V, in standalone run - 0. It is obvious that problem is there, but how to pull it up correctly?

UPD I also tried capacitors 100n, 4,7uF and resistors 1k, 10k, 0k (direct wire) and source circuit from battaries but with no luck. In debug mode program works perfectly in any of such RC's.

As a programmer I use Launchpad f5529. 

  • This may sound strange or even insulting. But based on what your posting, you did not provide power to it when it is standing alone.

    msp430g2553 is a "low-power device", not a "no-power device". It needs power.
  • Hehe, good comment :) But standalone power was OK. Firstly I tried with Launchad's power pins, same as used during debug (didn't disconnected circuit at all), than I used two different batteries with 2.9v and 3.57v voltage. I'll post a schematic for more details..
  • So this is the schematics. It is very simple application: using DS18B20 temperature sensor MCU periodically measures temperature and if it higher than needed sends a kind of "TURN ON"command  to my A/C via IR LED. As I mentioned above everything works perfect during the debug mode.

    Thanks for your reply.

  • The schematics look fine. But you did not show the battery.

    When all connected in the stand alone mode.

    1. What is the measured voltage between the battery + and – terminals while in use?
    2. What is the measured voltage from a to b?
    3. What is the measured voltage from c to b?
    4. What is the measured voltage from d to b?

    Your first posting says: During the debug /RST (d) has voltage level 3,1 V, in standalone run - 0.

  • 1. The transistor in the schematic exchanges its `e' and `c'?
    2. Why not use the temperature sensor in 2553 directly just like TI demo?
    3. Can the port drive HL1 without the transistor?
    4. With no RST switch, are R2 and C1 needed?
    5. Can the internal DCO/VLO replace the Cr1?

    After all, in the schematic, is everything except 2553 and HL1 necessary? And of course, the battery as pointed out by old_cow_yellow.
  •  OCY, thanks for your reply. Here is fresh measurements:

    ab - 3,57 v

    cb - 0 v

    bd - 3,39 v. And this looks very strange! During my first measurements it was always 0. May be I was mistaken yesterday somwhere, or it was because the bad soldering... very strange. I checked the circuit several times.

    So now I can state that RST pin works as it should. But the problem remains the same and nothig happens out of debug. 

    Let me say some words about debug. I have an IR-analyzer made with AVR mcu which catch IR pulses and print timings over rs232. This circuit I use few years to decode RCs and make my own controllers for home automation. So out of debug it catches nothing, my problem is not sounds "I CAN'T TURN ON MY AC", just "my MCU circuit does'nt works %-)

  • Hi Sergii!

    What type of battery are you using? Is it capable of providing enough current? Look at R4 - without the internal output resistance of P3.3, this path would already draw about 30mA. And HL1 doesn't have any current limiting resistor at all. Maybe while debugging you got additional power from the FET, but the battery has too high input resistance. So when the transistor turns on, the supply for your circuit drops, causing a reset, over and over again. You should add a resistance for HL1 and you should also increase the resistance of R4. If the current gain isn't enough, use a Darlington NPN or a FET. Do you have a bench supply? I would test if the circuit works with a higher current capable supply.

    And add some decoupling capacitors to the MSP's supply pins - a 10µF + 0.1µF in parallel would be OK.

    VT1 has a wrong orientation, but I think you already noticed that.

    Dennis

  • small_stone_head, thanks for you interest and questions.

    The transistor is used simply as a digital switch, because of high LED current consumption - 50mA in my case, so I risk to burn our the MCU if connect it directly. Connection throug resistor will decrease distance of remote control.
    I prefer use external temperature sensor because of its precision and some added feautures (like using 2 and more sensors or internal alarm signal).
    The R2C1 needed to pull RST pin higher. Before it happens MCU can not start.
    Cr1 is used for accurate and useful calculation of time delay. Most time of working cycle MCU should be in LPM3 mode, and the crystal is used to count seconds with low power consumption. Few added pennies are not the case :)

    Regards, Serge.
  • Hi Dennis! Thanks for your comment.

    I tried to run this circuit with Launchpad's supply pins, CR 2032 battery, Nokia battery BL-5CA (3,58v), 2 AAA elements without any luck.

    To check your suggestion about current draw I totally removed transistor, 100 Ohm resistor from its' base, IR LED and even DS18B20 sensor and changed the program to send some number through UART periodically. The result is still the same - the device works only during the debug session. And yes, I added decoupling capacitors to the MSP's supply pins - a 10µF + 0.1µF in parallel. Ufff, this is kinda strange))

  • Did you try if a simple blink LED program works?
  • I just tried it after your comment. So the result is expectable: blinks only in debug mode.
    As a final solution I think disassemble all, sing some mantras, pray to some gods and start again. Your avatar is perfect to this situation :)
  • Are you using the MSP-EXP430G2 LaunchPad? If yes, when powering the board from the battery, do you remove the jumpers between the emulation side and the processor?
  • Nope. I have another Launchpad with msp430f5529 and use just it's FET to load firmware into msp430g2553. This last controller is in 28 TSSOP package and soldered to circuit. In Launchpad jumpers between FET and main board disconnected (3,3v, GND, TST, RST). Every flashing is succesfull, I don't see any problem here. I also did it with MSP Flasher 1.3.9 and it was succesfull too.
  • Could you add a photo of your complete hardware setup?
  • Sure. Now it looks dirty after adding that all capacitors, resistors and LED, but I still trying to find the solution.

  • Oh, OK...I was hoping one could see the traces and their connections...
  • Yes, I understand that it is hard to see anything and requires some paintings. But if assume that everything was soldered up to schematics, do U have any idea what could cause the problem? I am going to develop a new PCB and solder it again 'coz see no solution with this one.

  • When even the LED blink program only works when the debugger is connected, then, when having it disconnected, you have no power at your circuit or the reset pin is somehow not connected properly (floating).
  • Sergii Matsiievsky said:
    Nope. I have another Launchpad with msp430f5529 and use just it's FET to load firmware into msp430g2553. This last controller is in 28 TSSOP package and soldered to circuit. In Launchpad jumpers between FET and main board disconnected (3,3v, GND, TST, RST). Every flashing is succesfull, I don't see any problem here. I also did it with MSP Flasher 1.3.9 and it was succesfull too.

    I am confused now. You are now saying: "... I don't see any problem here ...", and "... and it was successful too."

    Does "successful" here include the G2553 runs standing alone after that?

    P.S. some debugger or MSP Flasher may be able to flash the code correctly, but may lock-up the G2553 in a non-cooperative way. afterwards. You need to completely power-down and then power-up the G2553 in order to run the code standing alone.

  • old_cow_yellow said:
    I am confused now. You are now saying: "... I don't see any problem here ...", and "... and it was successful too."

    Does "successful" here include the G2553 runs standing alone after that?

    P.S. some debugger or MSP Flasher may be able to flash the code correctly, but may lock-up the G2553 in a non-cooperative way. afterwards. You need to completely power-down and then power-up the G2553 in order to run the code standing alone.

    I was misundersood, succesfull were only flashings, not standalone runs. 
    I am quite confused with so simple application. Also adding/removing components with much soldering-unsoldering made the PCB looks like spaghetti (see pics). I decided design new PCB with additional res. and caps and solder another 'g2553. This will do everything clear where is the problem. Also your notation about possible lock by debugger also will'be checked. Thanks a lot for you replies.
    Regards.
    Serge.
  • Quote from p.161 in davies08_MSP430 Microcontroller Basics:

    ``This is a classic example where any problem will probably not arise during early testing. The connections to the debugger usually include the RST/NMI pin and the output impedance of the interface pod is often low enough to suppress any noise picked up on the board. The sensitivity to noise appears only when the debugger is disconnected.''
  • The connection at RST is the one I would also check first if the MSP does not start to run the application. Maybe the capacitor is connected to GND, but the resistor does not have a proper connection to DVcc, so RST is floating with an unconnected debugger...?

  • I have the same problem with an MSP430g2001 custom board.   I can download a program using CCS via TI FET, single step, and all is well.   When the spy by wire emulator is disconnected, the code does not run and based on the power consumption it looks like the processor is held in reset.

    47K pullup with a 2.2nF per the TI ref des and app notes.

    Any suggestions?

**Attention** This is a public forum