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.

Information memory A - Can I erase the mistake?

Other Parts Discussed in Thread: MSP430F2330, MSP-TS430QFN23X0

Hi, All,

I use MSP430F2330.

And I made a catastrophe by erasing the information memory A, that amounts to 700 pcs.

1. Is there any tool to recover?

2. Is there any service regarding the recovery from TI? How much cost?

3. What kinds of calibration data were stored? DCO? and ADC? or others?

As I know, since there are only 2 bytes for each of the (1, 8, 12, 16 MHz), yet 64-8=56 bytes are none to understanding.

Does any body know about it? May you show each one of the fields, its start-address, end-address and attribute, numeral ranges, typical values?

4. Can you suggest anything that how can I do to remedy the unreversible harm?

Can you suggest some methods so that I can stuff some reasonable calibration data back to IM-A?

 

Thanks a lot for everybody's comments.

 

 

  • Hi Ken,

    DCO calibration data (and nothing else) were stored in INFOA memory. Have a look at the data sheet and/or users manual; you will find the details on the addresses used there (http://focus.ti.com/lit/ds/symlink/msp430f2330.pdf).

    Well, somebody pls correct me if I'm wrong, I think that TI is programming the calibration data to INFOA segment during testing of the devices. So, they should have some specialized equipment to do that.

    Since this is not an option for you I would recommend to have a look at the TI MSP430 code examples page (http://focus.ti.com/mcu/docs/mcuprodcodeexamples.tsp?sectionId=96&tabId=1468). There is Code Library section which also has a 'Using the DCO Library' entry (http://focus.ti.com/general/docs/litabsmultiplefilelist.tsp?literatureNumber=slaa336a). This library function can be used to calibrate the DCO during run-time of the users application (or once the app is started, or ...).

    I think it should be possible for you to write your own application based on the library function. Your software should 'calibrate' the DCO to 1/8/12 and 16MHz, store the results in an array or something (in RAM) and - when finished - write the array to INFOA memory. Done!

    Regarding the tools you need: you need an IDE for code writing (CCS or IAR; free version were available from both), a target board with a ZIF socket on and a JTAG-tool (i.e. this set http://focus.ti.com/docs/toolsw/folders/print/msp-fet430u23x0.html). And the last thing you need is some spare time to do a boring job (open ZIF socket --> insert device --> close socket --> download application --> run --> wait until finished -> open socket --> remove device and restart the procedure with another part). So, HAPPY NEW YEAR *g*!

    Rgds
    aBUGSworstnightmare 

  • *********************************************************

    Hi, aB.,

    It seems an option to create an automatic re-calibration tool(i.e., a 32K crystal, the DCO lib, in-system flash programming and a sticky ZIF socket) to restore DCO parameters.

    Thank you for the sugestion.

    *********************************************************

     

     

    And has anybody better or different methods for the recovery?

    And my other questions, has anybody ideas?

    One more again, I have another new question:

    Since after applying the DCO calibration data which is "precisely" calibrated by using the expensive testing device, the outcome still has at most +- 6% offsets according to the data sheet. Why would it be?

     

    Thanks for all the suggestions in advance.

     

     

  • Hi Ken,

    I further though about it and got another idea! Do you have some other parts - in best case from the same production lot - where the DCO values were not erased?

    You could take some parts, read out the DCO values and do a statistic compare on them (production yield should be very low --> variation in calibration data should be nearly 0). Then, you take one of them (which fits best) as a master, take it - togehter with the 700 pieces with INFOA erased - to a programming service and let them do the reprogramming.

    Rgds
    aBUGSworstnightmare

  • You did not lose much in mistakienly erasing the Info-A.

    The factory calibration are useful only if your product satisfies ALL three conditions below:

    (a) cannot use a crystal

    (b) need to run the DCO at 1, 6, 8, or 16MHz

    (c) +-6% accuracy is good enough but +-20% is not

    If you still want to calibrate the chips before you solder them, you can use the TI MSP-TS430QFN23x0 to do it. It cost about $75.

    The calibration prcess only gives you the DCO settings at the Vcc, temperature, and age of the chip to generate the desired frequency. When you use this setting of the DCO at a different Vcc, temperature, and age of the chip, the frequency may be different by 6%.

    Let me put it this way.

    Use the best equipment to "calibrate" how far you have to step on the accelarator peddle of you car to get 25, 35, and 65mph. What kind of speed do you get when you use those "calibration" for how hard to step on the accelarator when you drive the car later?

  • Hi, pals,

    Yes, it has much higher probabilities that the same lot ICs has almost the same calibration parameters. But I thought that perhaps the method we concluded above would be more realistic.

    And it's reasonable that temperature, Vcc, and chip age cause working Freq to vary and excellent tolerance of under 6%.

    In my application, more precise frequency is for it better. Therefore I decide to adopt one of two:

    1. freeze it in store for later other app inevitably utilizing external crystal.

    2. the method by using crystal/DCO-lib.

     

    Thank you guys for good suggestions.

     

     

     

**Attention** This is a public forum