• State Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 1 answer
  • Subscribers 78 subscribers
  • Views 1125 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

MSP430G2452: "MSP430: Error connecting to the target: Unknown device"

Errol Kowald
Intellectual 540 points
Part Number: MSP430G2452
Other Parts Discussed in Thread: MSP430F2013, MSP-FET, UNIFLASH

(Almost too embarrassed to state that title, as its featured countless times since the MSP430-FET arrived.)

My problem:

I have used the MSP430F2013 and its tiny cousins for years & have seen the infamous MSP-FET message "Error connecting to the target: Unknown device" a few times as well, and have pretty much worked out how to get around it for that chip. A recent upgrade required a couple more ports though, so the 20 pin MSP430G2452 was pressed into service. I expected to drop that chip into the circuit as an almost direct replacement. The SBW should have been almost exactly the same. To head off any communication issues I did not load the 2n2 reset capacitor (C1 below) for debugging with the MSP-FET, but to no avail.

What have I tried?

  • CCS11 operating in Ubuntu 20.04 was installed to replace CCS10
  • tried Uniflash 6.1 but it could not automatically detect the device
  • SBW connections to the MSP430G2452 were checked again and confirmed
  • 3.44Vcc is applied externally to the custom board & the MSP-FET is connected to suit - in the same manner as with previous MSP430F2013 projects (no change)
  • With the MSP-FET connected, but not communicating, pin 16 (sbwtdio) is at 3.44V & pin 17 (sbwtck) is at 1.26V [I expected higher]
  • I haven't put the oscilloscope on it yet
  • I haven't tried to upgrade the MSP-FET firmware, and CCS11 hasn't attempted to do this for me. (I am not sure if this is necessary)

There is a 10uF filter capacitor at the output of an LT3012 voltage regulator that is not shown

Do you have any suggestions?

  • 0
    Intellectual 540 points

    I connected the MSP-FET to a very tiny MSP430F2013 test board, and ran Uniflash 6.1, which detected that sub family, but not the specific device. On choosing it manually (onchip) I was able to flash the chip successfully.

    I then connected to my 'G2452 boad and Uniflash detected that sub family. The only choice I had though was MSP430G2452 (bootloader). At that point I had no idea what to do to flash the chip, but suffice to say, the MSP-FET is now capable of accessing the G2452 chip. It couldn't do this previously. Somehow, getting it to talk to a 'F2013 enable it to subsequently talk to the 'G2452

    Back to CCS11, the FET now communicates with the 'G2452. It didn't previously - unknown device!

    Something very strange is happening with the MSP-FET that I do not understand.

    Any ideas anyone?

  • 0 in reply to Errol Kowald
    TI__Intellectual 1130 points

    Hi Errol,

    Have you change another G2452 board, to make sure Is it the board's problem or not？ If not, you can try it on windows system to check for system.

    Best

    Joe

  • 0 in reply to Joe Wang(SE)
    Intellectual 540 points

    Hello Joe,

    I did try another board, without success. The SBW interface is so simple though, it'd be hard to muck it up

    Last night I shut CCS11 down and powered the hardware down, including the FET, but not Linux. This morning on starting it all up again the FET would not recognise the G2452 from CCS11. The solution was to shut CCS11 down and run Uniflash 6. It did find the G2452 family. On bringing CCS11 back up, it worked again. I am now debugging the code.

    This is similar behaviour to that I've seen before with earlier F2013 projects. CCS unable to recognise the MSP430 until after Uniflash was run. There is something different in how the 2 programs run the MSP-FET, and whatever Uniflash does tips the FET enough to get CCS working with it.

    It reminds me of a floating input. If the charge is high enough, it is a '1' otherwise it is read as '0'

    I haven't tried WIN yet, but will in the next few days

  • 0 in reply to Errol Kowald
    TI__Intellectual 1130 points

    Hi Errol,

    Glad it works again. You can debug you code now.

    If you have question, you can reply to us.

    Best regards

    Joe

  • 0 in reply to Joe Wang(SE)
    Intellectual 540 points

    I wish I knew WHY it worked again. Currently I can only call it magic.

  • 0 in reply to Errol Kowald
    TI__Intellectual 1130 points

    Hi Errol, 

    If you don't have question now, I will close this e2e.

    Best and Thanks

    Joe 

  • 0 in reply to Joe Wang(SE)
    Intellectual 540 points

    To my knowledge, no-one has offered a plausible explanation for the difference in behaviour between CCS & Uniflash when operating under Linux, despite it being evident for many years. If CCS could adopt the process used by Uniflash in dealing with the MSP-FET, the problem would be resolved.

    On 10 November 2020, MSP Team Member JD Crutchfield acknowledged the difference, suggesting that the issue could be improved by slowing the MSP-FET baud rate. This helped with the reset cap problem, but made no inroads elsewhere

    I will mark this as resolved, although it isn't. 

**Attention** This is a public forum