• Locked Locked
  • Replies 14 replies
  • Subscribers 79 subscribers
  • Views 5309 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.

Part Number Decoder is Out-of-Date!

Andy Neil
Guru 35465 points
Other Parts Discussed in Thread: MSP430FE423, MSP430FW423, MSP430L092, MSP430BT5190, MSP430F5438, CC2560, MSP430F5438A

http://www.ti.com/lsds/ti/microcontroller/16-bit_msp430/getting_started.page#decoder

 

Memory Type: Wot - no FRAM?!

  • Expert 2505 points

    I augmented this a little last year when trying to characterize the different chips for mspgcc support.  Some extra details are in the decoding below, but in practice I think this decoding scheme no longer works: both the FRAM and BlueTooth-optimized MCU identifiers are inconsistent with it since digraphs are now used as single identifiers at the point where memory type and end-equipment optimization used to appear.

    MSP 430 F G 4 6 19 S I QZW R
    -+- -+- + + + + -+ + + -+- +
     |   |  | | | |  | | |  |  |
     |   |  | | | |  | | |  |  + Optional: Distribution format
     |   |  | | | |  | | |  |    R = Tape & Reel
     |   |  | | | |  | | |  |  
     |   |  | | | |  | | |  +--- Package
     |   |  | | | |  | | |    
     |   |  | | | |  | | +------ Optional: Temperature range (C)
     |   |  | | | |  | |         I = -40 to 85
     |   |  | | | |  | |         T = -40 to 105
     |   |  | | | |  | |                          
     |   |  | | | |  | +-------- (UNDOCUMENTED) Packaging variant MSP430x31x
     |   |  | | | |  |           S = 48-pin SSOP (vs standard 56-pin SSOP)
     |   |  | | | |  |
     |   |  | | | |  +---------- Family (memory size and peripheral configuration)
     |   |  | | | |
     |   |  | | | +------------- Series of similar function
     |   |  | | |
     |   |  | | +--------------- Generation
     |   |  | |                  1 = 1xx
     |   |  | |                  2 = 2xx
     |   |  | |                  3 = 3xx
     |   |  | |                  4 = 4xx
     |   |  | |                  5 = 5xx
     |   |  | |                  6 = 6xx
     |   |  | |                  0 = Undocumented (custom chip? MSP430C091)
     |   |  | |
     |   |  | +----------------- Optional: End-equipment optimized
     |   |  |                    G = Medical
     |   |  |                    E = Energy Meter (MSP430FE423)
     |   |  |                    W = Electronic Flow Meters (MSP430FW423)
     |   |  |
     |   |  +------------------- Memory Type
     |   |                       C = ROM
     |   |                       E = EPROM
     |   |                       F = FLASH
     |   |                       G = FLASH (Value Line device)
     |   |                       P = OTP
     |   |                       L = No non-volatile memory (MSP430L092)
     |   |
     |   +---------------------- 430 MCU Platform
     |
     +-------------------------- Processor Family
                                 CC = ChipCon Radio
                                 MSP = Mixed Signal Processor
                                 XMS = Experimental Silicon

    Non-conformant parts:

    msp430bt5190 -- BT apparently denotes a part designed to work with the
    CC2560 Bluetooth radio chip.  This is an msp430f5438 with the BT stack
    already programmed in.

  • in reply to Peter Bigot
    Guru 227245 points

    Peter Bigot said:
    msp430bt5190 -- BT apparently denotes a part designed to work with the
    CC2560 Bluetooth radio chip.  This is an msp430f5438 with the BT stack
    already programmed in.

    Not 'programmed-in', but with a (undisclosed) hardware difference that is checked in the BT stack library and verifies that the license fee for the stack has been paid for this device. Sort of a BT 'dongle' bit somewhere.

  • in reply to Jens-Michael Gross
    Expert 2505 points

    Could well be, but the TI contact I asked last year told me "this device is identical to the F5438 just got number differently as it has the BT stack already programmed in it." so that's what I documented.

  • in reply to Peter Bigot
    Guru 227245 points

    Peter Bigot said:
    Could well be, but the TI contact I asked last year told me "this device is identical to the F5438 just got number differently as it has the BT stack already programmed in it." so that's what I documented.

    Maybe your contact used a sloppy expression.

    The MindTree BT stack comes as library and is included into the binary. It is not 'already programmed'. It's possible that the MSP comes preprogrammed with the demo code, but when you want to do your own applicaiton, teh BT code is erased liek any othe rapplication code and replaced by yours.

    That's why people complain that for building their own projects they need to buy a full IAR license: Since the BT stack only comes as precompiled library, and is larger than the free IAR code size limit, they need to pay for a full unlimited IAR version in order to use the BT.

    From information I got here in this forum (from a TI employee), it seeems that the BT library checks for an intentional hardware anomaly that separates a BT from a real 5438.
    I don't know what difference this is, maybe just the part ID in the TLV structure.

  • in reply to Jens-Michael Gross
    TI__Prodigy 90 points

    Hi All,

    I just wanted to point you to a new blog post that includes the most recently updated part number decoder:

    http://e2e.ti.com/support/microcontrollers/msp43016-bit_ultra-low_power_mcus/b/msp430_news/archive/2011/12/09/new-msp430-part-number-decoder.aspx

    Hope this answers your questions!

  • in reply to Stephanie Hsu
    Guru 227245 points

    Stephanie Hsu said:
    I just wanted to point you to a new blog post that includes the most recently updated part number decoder:

    Thanks for the info.

    One thing I wondered on a second look:

    the 'L' marking is listed as 'non-volatile memory'. Didn't it stand for 'low voltage device'? The L092 does have volatile memory only and requires extrernal non-volatile memory for booting (FLASH won't run on such a low voltage)

  • in reply to Jens-Michael Gross
    Expert 2505 points

    Jens-Michael Gross said:

    One thing I wondered on a second look:

    the 'L' marking is listed as 'non-volatile memory'. Didn't it stand for 'low voltage device'? The L092 does have volatile memory only and requires extrernal non-volatile memory for booting (FLASH won't run on such a low voltage)

    I believe that says "no non-volatile memory".

    I, in turn, noticed that the P memory type for OTP got dropped.  There are a few P chips (e.g., msp430p315) that are still listed in the devices list I get from TI for msp430mcu.

    On the whole, this is a great improvement and help, and I thank TI for updating it as now I don't have to maintain my hacked-up text illustration any more....

    Peter

  • in reply to Peter Bigot
    Guru 227245 points

    Peter Bigot said:
    I believe that says "no non-volatile memory".

    You're right, my eyes. Maybe I thought I was just squinting when reading this double-negation.
    Since the key is headline memory type/device type, I think the 'low voltage' is the dominant attribute of the L devices rather than the fact that they don't have non-volatile memory (and in fact, some L devices have - just not for program execution and only for high-voltage operation, at least during boot)

    Peter Bigot said:
    I don't have to maintain my hacked-up text illustration any more..

    Hey, it was a nice piece of work.

    Peter Bigot said:
    There are a few P chips (e.g., msp430p315) that are still listed in the devices list

    Officially, the 3x family is discontinued. But maybe there are some special customers who still order significant amounts or have a supply contract with TI that isn't expired yet.

  • in reply to Jens-Michael Gross
    Expert 2505 points

    Jens-Michael Gross said:

    I don't have to maintain my hacked-up text illustration any more..

    Hey, it was a nice piece of work.[/quote]

    Thank you.

    Jens-Michael Gross said:

    There are a few P chips (e.g., msp430p315) that are still listed in the devices list

    Officially, the 3x family is discontinued. But maybe there are some special customers who still order significant amounts or have a supply contract with TI that isn't expired yet.[/quote]

    I don't think the details in the PN decoder should be limited to currently available chips; ideally it will cover the entire product history, so it can be used to identify the characteristics of some legacy hardware found gathering dust in a desk drawer.  I would look to the decoder to answer "What does this P mean" if I encountered one of those chips in the wild.  (Though it would also help if "OTP" was defined somewhere in the diagram.)

    Peter

  • in reply to Jens-Michael Gross
    TI__Prodigy 90 points

    Jens-Michael Gross said:

    Since the key is headline memory type/device type, I think the 'low voltage' is the dominant attribute of the L devices rather than the fact that they don't have non-volatile memory (and in fact, some L devices have - just not for program execution and only for high-voltage operation, at least during boot)

    Thank you for bringing this up. We originally defined "0" as the low voltage series and wanted to keep the "Memory Type" category specifically memory. However, if this causes confusion, we'd be more than willing to modify "L" to low-voltage.

  • in reply to Peter Bigot
    TI__Prodigy 90 points

    Peter Bigot said:

    There are a few P chips (e.g., msp430p315) that are still listed in the devices list

    Officially, the 3x family is discontinued. But maybe there are some special customers who still order significant amounts or have a supply contract with TI that isn't expired yet.[/quote]

    I don't think the details in the PN decoder should be limited to currently available chips; ideally it will cover the entire product history, so it can be used to identify the characteristics of some legacy hardware found gathering dust in a desk drawer.  I would look to the decoder to answer "What does this P mean" if I encountered one of those chips in the wild.  (Though it would also help if "OTP" was defined somewhere in the diagram.)

    [/quote]

    For the P chips, we decided to exclude them from the part number decoder because there are no new P devices being made. We tried to find a balance between the old and new devices without making the decoder all inclusive and too extensive. Hopefully this decoder can be a useful tool that may be applied to a wide variety of MSP430 parts. Any questions with other part number definitions may be posted to the forum for clarification.

    Thank you for reviewing this!

  • in reply to Peter Bigot
    Guru 227245 points

    Peter Bigot said:
    I don't think the details in the PN decoder should be limited to currently available chips; ideally it will cover the entire product history,

    I wholeheartedly agree. I too prefer if a data source tells the whole story, not just the latest version. That's why many people use versioning systems.
    However, the TI philosophy seems to be to extinct outdated information from history.
    At least the errata sheets now contain a section with changes from older versions. However, before this was introduced, you didn't have a chance to look up errata which were present on older silicon but removed with a new revision. You still cannot for information that has been removed before this history was introduced. (e.g. the 1232 errata sheet starts with silicon revision L).

    Even it has been a bug/typo that has been removed, I want to knwo it. because maybe I relied on this typo and not, since it is corrected, I need to know that there was a change that may have impact on my code/design.

    Or when coding a compiler, you'll need to know which workarounds to implement even for the older silicon. Not just for the latest one.

    Stephanie Hsu said:
    We originally defined "0" as the low voltage series

    So it was jsut coincidence, but I was assuming the '09' was referring to the 0,9V minimum operating voltage. Or the 'L09' together.
    You're right, the naming fits the C0 devices with low voltage and ROM. It's just that the 'L' was misleading (I had chosen 'N' or 'V' for 'no non-volatile' or 'volatile only'). But the name is given.
    The key differences are that there is low voltage operation and no way to program the device yourself internally. Boot rom or mask rom, for 99% of the users this makes no difference, since if you need external flash memory, it will need higher voltage anyway, and MASK ROM is only for people with a high production volume.

    When reading about the L092 first, i thought " why isn't there flash? Nobody will start to develop on this baby if it needs external circuitry and can be developed/debugged on the target board"
    My idea was to implement flash - with an internal charge pump. So on boot, flash is operational and the code can be copied to ram, then the charge pump is shut off and flash is deactivated. Reprogrammable, still low power, convenient, and no need fo a high-volume count for the rom versions. And the FLASH version can be 1:1 replaced with a ROM version for mass production.

    Stephanie Hsu said:
    For the P chips, we decided to exclude them from the part number decoder because there are no new P devices being made.

    But there are still available somewhere,. So if someone stubles across it and tries to look it up on the selector, he'll not find it there, maybe asking himself whether this might be a fake chip, a printing error, or whatever.

    Stephanie Hsu said:
    We tried to find a balance between the old and new devices without making the decoder all inclusive and too extensive.

    Understandable, but IMHO a fact sheet should cover everything. Imagine Windows would silently remove links and files from your desktop or harddisk, just because you didn't click on them for some time. (Yes, it notifies for unused items on the desktop, but even if I agree to remove them, they are collected in a folder on the desktop and not deleted)

  • in reply to William Cooper
    Guru 227245 points

    Jens-Michael Gross said:
    Imagine Windows would silently remove links and files from your desktop or harddisk, just because you didn't click on them for some time.

    I just learned that the latest Windoze version really does this. No notification or assistant that you can cancel. Links are simply vanishing (except, of course, those that Microsoft things you should have on your desktop - even if you don't use them and never intend to use them). Looks like my 1.5 years old doomsday vision has become true.

    William Cooper said:
    Here is a new part number decoder for your reference:

    Thanks for the update.

    One remark: "BT" isn't pre-programmed with Bluetooth. It comes with a BT stack license included (and you need an IAR full license to make use of it), but neither has the requied BT hardware included nor does it have (at least AFAIK) anything pre-programmed (which wouldn't make sense, as you have to reprogram the chip anyway to add any useful application). It's just a re-labeled MSP430F5438A with an undisclosed marker that allows the BT stack library to verify the license.

**Attention** This is a public forum