• AVR Freaks

Hot!AT24CM02 - 1 MHz operation at VCC = 1.8 V?

Author
DanCooper
New Member
  • Total Posts : 3
  • Reward points : 0
  • Joined: 2019/12/04 16:47:13
  • Location: 0
  • Status: offline
2019/12/05 18:02:41 (permalink)
0

AT24CM02 - 1 MHz operation at VCC = 1.8 V?

Hello,
I have the AT24CM02 on my board, running at 1.8 V. I noticed the datasheet only guarantees operation up to 400 kHz at 1.8 V, with 1 MHz reserved for >= 2.5 V.
My system programs the AT24CM02 at 400 kHz, but later reads back the contents at 1 MHz. This has actually been working fine, which leads to my questions:
Is 1 MHz read-back OK at 1.8 V, and the limitation is just that you can't write at 1 MHz?
Or is there some condition, such as extreme temperature for example, that would cause a 1 MHz read to fail, and I shouldn't rely on 1 MHz read-back always working?
Thanks in advance for any insight,
DC
#1

8 Replies Related Threads

    ric
    Super Member
    • Total Posts : 28019
    • Reward points : 0
    • Joined: 2003/11/07 12:41:26
    • Location: Australia, Melbourne
    • Status: online
    Re: AT24CM02 - 1 MHz operation at VCC = 1.8 V? 2019/12/07 00:09:33 (permalink)
    0
    If the datasheet doesn't specify it, then the manufacturer doesn't guarantee it will work over the entire temperature range at the full range of voltages.
    If you choose to ignore those limits, then you either have to do your own testing over all possible conditions, or you are using your customers as guinea pigs.
     
     

    I also post at: PicForum
    Links to useful PIC information: http://picforum.ric323.co...opic.php?f=59&t=15
    NEW USERS: Posting images, links and code - workaround for restrictions.
    To get a useful answer, always state which PIC you are using!
    #2
    hudejun
    Electrical engineer
    • Total Posts : 15
    • Reward points : 0
    • Joined: 2019/12/05 08:48:06
    • Location: 0
    • Status: offline
    Re: AT24CM02 - 1 MHz operation at VCC = 1.8 V? 2019/12/07 03:05:46 (permalink)
    0
    If manufacturer guarantee 400KHz at 1.8V, you must use it.
    Of course, it can work with 1Mhz.
    But, when read bulk data, it will be not stable.
    For exam, there are 40MHz and 50MHz version for the same PIC32 MCUs.
    If you init system 50MHz for the 40MHz version, it can work sometimes.
    But, when use special functions such as interrupt and another complex resources, it does not work.
    It is my experience.
    #3
    DanCooper
    New Member
    • Total Posts : 3
    • Reward points : 0
    • Joined: 2019/12/04 16:47:13
    • Location: 0
    • Status: offline
    Re: AT24CM02 - 1 MHz operation at VCC = 1.8 V? 2019/12/07 15:34:41 (permalink)
    0
    Hi Ric,
     
    Agreed; I know all bets are off if you operate outside the parametric table specifications, and I wouldn't suggest a final product be designed this way (this is just a proto board). Sometimes the rules can be bent a bit if we have insight on the why behind the spec limitation, and can intelligently try to avoid the operation conditions that cause it to fail.
     
    I'm not sure if the team that supports the AT24CM02 looks at this forum, but even if not, I was hoping that perhaps some folks might have experience with pushing serial EEPROMs a bit past their limits and know what the failure modes look like.
     
    I'm not sure if I should expect an invalid read out of a memory location, or a physical violation of setup up and hold times, or if the EEPROM stops responding entirely. If it's a temperature-related limitation, does it respond properly again once the junction temperature cools back off, or it requires a power cycle to restore operation?
     
    Since I haven't seen a failure, I don't know what to look for.
     
    hudejun, appreciate you sharing your experience with MCU speed ratings. When you mention bulk data reads won't be stable, are you thinking that perhaps a buffer gets full / isn't able to keep up with the 1 MHz read-out? When this occurs, would the EEPROM NACK a read to catch-up, and then presumably a re-read from the host would pick right up where it left off?
     
    DC
    #4
    ric
    Super Member
    • Total Posts : 28019
    • Reward points : 0
    • Joined: 2003/11/07 12:41:26
    • Location: Australia, Melbourne
    • Status: online
    Re: AT24CM02 - 1 MHz operation at VCC = 1.8 V? 2019/12/07 15:56:18 (permalink)
    0
    Of course you can push the specs a bit, but then when it fails, you can NOT predict exactly HOW it will fail.
    There might be a single bit (high/low) error.
    The value might be shifted one bit.
    The value might be read from an adjacent address.
    Once it has failed, it will PROBABLY recover when it cools down, but that can't be guaranteed.
     
     
     

    I also post at: PicForum
    Links to useful PIC information: http://picforum.ric323.co...opic.php?f=59&t=15
    NEW USERS: Posting images, links and code - workaround for restrictions.
    To get a useful answer, always state which PIC you are using!
    #5
    Mysil
    Super Member
    • Total Posts : 3678
    • Reward points : 0
    • Joined: 2012/07/01 04:19:50
    • Location: Norway
    • Status: offline
    Re: AT24CM02 - 1 MHz operation at VCC = 1.8 V? 2019/12/07 16:25:04 (permalink)
    5 (1)
    Hi,
    I am sure that hardware engineers have tested the chips outside specifications,
    to see what the margin for correct operation will be,
    and thay do not only test that is 'working fine', they would test all instructions, all peripherals,
    and the whole memory over entire temperature range.
    But even if they were reading this forum, I would Not expect them to give out any details,
    outside of what is in the datasheet.
     
        Mysil
    #6
    hudejun
    Electrical engineer
    • Total Posts : 15
    • Reward points : 0
    • Joined: 2019/12/05 08:48:06
    • Location: 0
    • Status: offline
    Re: AT24CM02 - 1 MHz operation at VCC = 1.8 V? 2019/12/08 03:04:16 (permalink)
    0
    Yes, I think output data will be not correct sometimes.
     
    For the Temperature and any other conditions, Support team does not guarantee 1MHz with 1.8V.
     
    Also, if you have method to detect bad value from EEPROM data array, you can re-read data from that position.
     
    But, I think this process will takes more delay than the 400KHz.
     
    I think you'd like to use 1MHz due to speed.
     
    But, Error detection&correction process will be slow.
    #7
    Mysil
    Super Member
    • Total Posts : 3678
    • Reward points : 0
    • Joined: 2012/07/01 04:19:50
    • Location: Norway
    • Status: offline
    Re: AT24CM02 - 1 MHz operation at VCC = 1.8 V? 2019/12/08 09:07:44 (permalink)
    5 (1)
    Hi,
    In my previous message,
    I did overlook the hint that this was about a EEPROM memory chip with I2C interface.
     
    There may be 2 different limitations about operation at 1.8 V.
    one is internal reading of the memory array,
    and the other is about I2C communication at low voltage and high frequency.
     
    To specify the memory device as suitable for 1 MHz operation at 1.8 V supply,
    hardware engineers must test that it is able to work reliably with I2C maximum bus capacitance,
    at that combination of signaling frequency and supply voltage.
     
    Look up: 'I2C-bus specification and user manual' UM10204 on the NXP website.
    400 pF may be up towards 10 meter of suitable wiring,
    or shorter wiring and a lot of devices connected together.
     
    If your bus have short wiring and few devices connected,
    then it may work satisfactory, in your conditions.
    Use a Oscilloscope to study signal quality,
    with attention to what signal detection levels VIL and VIH are for a bus with 1.8 V supply voltage.
     
        Mysil
     
     
    #8
    DanCooper
    New Member
    • Total Posts : 3
    • Reward points : 0
    • Joined: 2019/12/04 16:47:13
    • Location: 0
    • Status: offline
    Re: AT24CM02 - 1 MHz operation at VCC = 1.8 V? 2019/12/13 18:07:31 (permalink)
    0
    Thanks for the suggestions everyone. Mysil, as you mention it's possible that part of the VCC / speed spec is due to rise/fall time limitations at 1.8V. In my case, it's a short point-to-point I2C between the EEPROM and host, so the bus capacitance is likely < 20 pF.
    If I happen to create a repeatable case where it fails at 1 MHz, I'll report back what I observe. But otherwise, we can probably consider this thread 'closed', assuming no engineers from the manufacturer wish to weigh in.
     
    DC
    #9
    Jump to:
    © 2020 APG vNext Commercial Version 4.5