• SPLD output current protection

    From bitrex@21:1/5 to All on Mon Feb 10 17:57:06 2025
    I have an SPLD with about 8 terminals exposed to the user I'd like to
    protect against over current. They're very low-speed outputs (10s of Hz
    at most), don't really want to spend a bunch for a chip like the L6374
    as an intermediary, or the board space for a bunch of transistors to
    discrete limit...

    The SPLD has a max per-chip-side Vdd input current of 45 mA at 85 C TJ
    and 22 mA at 110 TJ, and can sink 86 mA to ground per side, 41 at 110 TJ.

    The outputs are configured as push-pull or open drain, depending on the firmware/application, either way they can source or sink 40mA in
    isolation. But no more than two outputs per side at a time will be in a
    state such that they can sink/source current either way.

    The uP I'm using has a very nice 10 bit differential ADC (9 bits if used
    in bipolar mode) I'm thinking about just rigging it to two 4051s to
    "scan" a current sense in each output line to implement the overcurrent,
    can set different thresholds based on whether it's configured push-pull
    or OD, but what might conservative limits be? And what rate to scan?

    The SPLD is acting as a simple power supervisor and has a PWR GOOD line
    to bring up the uP. In turn it gets a clock from the uP, I can set up
    the SPLD to set all those outputs hi-z if it ever loses the clock (i.e.
    uP hangs) as an extra layer of protection..

    --
    This email has been checked for viruses by AVG antivirus software.
    www.avg.com

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From piglet@21:1/5 to bitrex on Mon Feb 10 23:07:01 2025
    bitrex <user@example.net> wrote:
    I have an SPLD with about 8 terminals exposed to the user I'd like to
    protect against over current. They're very low-speed outputs (10s of Hz
    at most), don't really want to spend a bunch for a chip like the L6374
    as an intermediary, or the board space for a bunch of transistors to
    discrete limit...

    The SPLD has a max per-chip-side Vdd input current of 45 mA at 85 C TJ
    and 22 mA at 110 TJ, and can sink 86 mA to ground per side, 41 at 110 TJ.

    The outputs are configured as push-pull or open drain, depending on the firmware/application, either way they can source or sink 40mA in
    isolation. But no more than two outputs per side at a time will be in a
    state such that they can sink/source current either way.

    The uP I'm using has a very nice 10 bit differential ADC (9 bits if used
    in bipolar mode) I'm thinking about just rigging it to two 4051s to
    "scan" a current sense in each output line to implement the overcurrent,
    can set different thresholds based on whether it's configured push-pull
    or OD, but what might conservative limits be? And what rate to scan?

    The SPLD is acting as a simple power supervisor and has a PWR GOOD line
    to bring up the uP. In turn it gets a clock from the uP, I can set up
    the SPLD to set all those outputs hi-z if it ever loses the clock (i.e.
    uP hangs) as an extra layer of protection..



    Ptc thermistor or polyfuse?

    --
    piglet

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From bitrex@21:1/5 to piglet on Mon Feb 10 18:24:20 2025
    On 2/10/2025 6:07 PM, piglet wrote:
    bitrex <user@example.net> wrote:
    I have an SPLD with about 8 terminals exposed to the user I'd like to
    protect against over current. They're very low-speed outputs (10s of Hz
    at most), don't really want to spend a bunch for a chip like the L6374
    as an intermediary, or the board space for a bunch of transistors to
    discrete limit...

    The SPLD has a max per-chip-side Vdd input current of 45 mA at 85 C TJ
    and 22 mA at 110 TJ, and can sink 86 mA to ground per side, 41 at 110 TJ.

    The outputs are configured as push-pull or open drain, depending on the
    firmware/application, either way they can source or sink 40mA in
    isolation. But no more than two outputs per side at a time will be in a
    state such that they can sink/source current either way.

    The uP I'm using has a very nice 10 bit differential ADC (9 bits if used
    in bipolar mode) I'm thinking about just rigging it to two 4051s to
    "scan" a current sense in each output line to implement the overcurrent,
    can set different thresholds based on whether it's configured push-pull
    or OD, but what might conservative limits be? And what rate to scan?

    The SPLD is acting as a simple power supervisor and has a PWR GOOD line
    to bring up the uP. In turn it gets a clock from the uP, I can set up
    the SPLD to set all those outputs hi-z if it ever loses the clock (i.e.
    uP hangs) as an extra layer of protection..



    Ptc thermistor or polyfuse?

    I'd rather use something KISS and actively monitor as a backup for sure,
    but I haven't used those devices on low current output lines before, do
    you have a suggestion for a suitable part? Thanks


    --
    This email has been checked for viruses by AVG antivirus software.
    www.avg.com

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From bitrex@21:1/5 to piglet on Mon Feb 10 18:26:19 2025
    On 2/10/2025 6:07 PM, piglet wrote:
    bitrex <user@example.net> wrote:
    I have an SPLD with about 8 terminals exposed to the user I'd like to
    protect against over current. They're very low-speed outputs (10s of Hz
    at most), don't really want to spend a bunch for a chip like the L6374
    as an intermediary, or the board space for a bunch of transistors to
    discrete limit...

    The SPLD has a max per-chip-side Vdd input current of 45 mA at 85 C TJ
    and 22 mA at 110 TJ, and can sink 86 mA to ground per side, 41 at 110 TJ.

    The outputs are configured as push-pull or open drain, depending on the
    firmware/application, either way they can source or sink 40mA in
    isolation. But no more than two outputs per side at a time will be in a
    state such that they can sink/source current either way.

    The uP I'm using has a very nice 10 bit differential ADC (9 bits if used
    in bipolar mode) I'm thinking about just rigging it to two 4051s to
    "scan" a current sense in each output line to implement the overcurrent,
    can set different thresholds based on whether it's configured push-pull
    or OD, but what might conservative limits be? And what rate to scan?

    The SPLD is acting as a simple power supervisor and has a PWR GOOD line
    to bring up the uP. In turn it gets a clock from the uP, I can set up
    the SPLD to set all those outputs hi-z if it ever loses the clock (i.e.
    uP hangs) as an extra layer of protection..



    Ptc thermistor or polyfuse?


    I'd rather use something KISS and actively monitor only a backup for
    sure, but I haven't used those devices on low current output lines
    before, do you have a suggestion for a suitable part? Thanks


    --
    This email has been checked for viruses by AVG antivirus software.
    www.avg.com

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From bitrex@21:1/5 to bitrex on Mon Feb 10 18:32:42 2025
    On 2/10/2025 6:26 PM, bitrex wrote:
    On 2/10/2025 6:07 PM, piglet wrote:
    bitrex <user@example.net> wrote:
    I have an SPLD with about 8 terminals exposed to the user I'd like to
    protect against over current. They're very low-speed outputs (10s of Hz
    at most), don't really want to spend a bunch for a chip like the L6374
    as an intermediary, or the board space for a bunch of transistors to
    discrete limit...

    The SPLD has a max per-chip-side Vdd input current of 45 mA at 85 C TJ
    and 22 mA at 110 TJ, and can sink 86 mA to ground per side, 41 at 110
    TJ.

    The outputs are configured as push-pull or open drain, depending on the
    firmware/application, either way they can source or sink 40mA in
    isolation. But no more than two outputs per side at a time will be in a
    state such that they can sink/source current either way.

    The uP I'm using has a very nice 10 bit differential ADC (9 bits if used >>> in bipolar mode) I'm thinking about just rigging it to two 4051s to
    "scan" a current sense in each output line to implement the overcurrent, >>> can set different thresholds based on whether it's configured push-pull
    or OD, but what might conservative limits be? And what rate to scan?

    The SPLD is acting as a simple power supervisor and has a PWR GOOD line
    to bring up the uP. In turn it gets a clock from the uP, I can set up
    the SPLD to set all those outputs hi-z if it ever loses the clock (i.e.
    uP hangs) as an extra layer of protection..



    Ptc thermistor or polyfuse?


    I'd rather use something KISS and actively monitor only a backup for
    sure, but I haven't used those devices on low current output lines
    before, do you have a suggestion for a suitable part? Thanks



    My head starts to hurt trying to interpret some of these polyfuse
    datasheets, I feel like hard limiting using software is almost simpler
    than trying to figure out what thing to use..

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From john larkin@21:1/5 to bitrex on Mon Feb 10 19:41:38 2025
    On Mon, 10 Feb 2025 17:57:06 -0500, bitrex <user@example.net> wrote:

    I have an SPLD with about 8 terminals exposed to the user I'd like to
    protect against over current. They're very low-speed outputs (10s of Hz
    at most), don't really want to spend a bunch for a chip like the L6374
    as an intermediary, or the board space for a bunch of transistors to
    discrete limit...

    The SPLD has a max per-chip-side Vdd input current of 45 mA at 85 C TJ
    and 22 mA at 110 TJ, and can sink 86 mA to ground per side, 41 at 110 TJ.

    The outputs are configured as push-pull or open drain, depending on the >firmware/application, either way they can source or sink 40mA in
    isolation. But no more than two outputs per side at a time will be in a
    state such that they can sink/source current either way.

    The uP I'm using has a very nice 10 bit differential ADC (9 bits if used
    in bipolar mode) I'm thinking about just rigging it to two 4051s to
    "scan" a current sense in each output line to implement the overcurrent,
    can set different thresholds based on whether it's configured push-pull
    or OD, but what might conservative limits be? And what rate to scan?

    The SPLD is acting as a simple power supervisor and has a PWR GOOD line
    to bring up the uP. In turn it gets a clock from the uP, I can set up
    the SPLD to set all those outputs hi-z if it ever loses the clock (i.e.
    uP hangs) as an extra layer of protection..

    What's an SPLD?

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From bitrex@21:1/5 to john larkin on Mon Feb 10 22:59:23 2025
    On 2/10/2025 10:41 PM, john larkin wrote:
    On Mon, 10 Feb 2025 17:57:06 -0500, bitrex <user@example.net> wrote:

    I have an SPLD with about 8 terminals exposed to the user I'd like to
    protect against over current. They're very low-speed outputs (10s of Hz
    at most), don't really want to spend a bunch for a chip like the L6374
    as an intermediary, or the board space for a bunch of transistors to
    discrete limit...

    The SPLD has a max per-chip-side Vdd input current of 45 mA at 85 C TJ
    and 22 mA at 110 TJ, and can sink 86 mA to ground per side, 41 at 110 TJ.

    The outputs are configured as push-pull or open drain, depending on the
    firmware/application, either way they can source or sink 40mA in
    isolation. But no more than two outputs per side at a time will be in a
    state such that they can sink/source current either way.

    The uP I'm using has a very nice 10 bit differential ADC (9 bits if used
    in bipolar mode) I'm thinking about just rigging it to two 4051s to
    "scan" a current sense in each output line to implement the overcurrent,
    can set different thresholds based on whether it's configured push-pull
    or OD, but what might conservative limits be? And what rate to scan?

    The SPLD is acting as a simple power supervisor and has a PWR GOOD line
    to bring up the uP. In turn it gets a clock from the uP, I can set up
    the SPLD to set all those outputs hi-z if it ever loses the clock (i.e.
    uP hangs) as an extra layer of protection..

    What's an SPLD?



    If it ain't a FPGA or a CPLD it's an SPLD. I think.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From piglet@21:1/5 to john larkin on Tue Feb 11 08:58:45 2025
    john larkin <JL@gct.com> wrote:
    On Mon, 10 Feb 2025 17:57:06 -0500, bitrex <user@example.net> wrote:

    I have an SPLD with about 8 terminals exposed to the user I'd like to
    protect against over current. They're very low-speed outputs (10s of Hz
    at most), don't really want to spend a bunch for a chip like the L6374
    as an intermediary, or the board space for a bunch of transistors to
    discrete limit...

    The SPLD has a max per-chip-side Vdd input current of 45 mA at 85 C TJ
    and 22 mA at 110 TJ, and can sink 86 mA to ground per side, 41 at 110 TJ.

    The outputs are configured as push-pull or open drain, depending on the
    firmware/application, either way they can source or sink 40mA in
    isolation. But no more than two outputs per side at a time will be in a
    state such that they can sink/source current either way.

    The uP I'm using has a very nice 10 bit differential ADC (9 bits if used
    in bipolar mode) I'm thinking about just rigging it to two 4051s to
    "scan" a current sense in each output line to implement the overcurrent,
    can set different thresholds based on whether it's configured push-pull
    or OD, but what might conservative limits be? And what rate to scan?

    The SPLD is acting as a simple power supervisor and has a PWR GOOD line
    to bring up the uP. In turn it gets a clock from the uP, I can set up
    the SPLD to set all those outputs hi-z if it ever loses the clock (i.e.
    uP hangs) as an extra layer of protection..

    What's an SPLD?




    Simple programmable logic device?

    --
    piglet

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From piglet@21:1/5 to bitrex on Tue Feb 11 08:58:43 2025
    bitrex <user@example.net> wrote:
    On 2/10/2025 6:07 PM, piglet wrote:
    bitrex <user@example.net> wrote:
    I have an SPLD with about 8 terminals exposed to the user I'd like to
    protect against over current. They're very low-speed outputs (10s of Hz
    at most), don't really want to spend a bunch for a chip like the L6374
    as an intermediary, or the board space for a bunch of transistors to
    discrete limit...

    The SPLD has a max per-chip-side Vdd input current of 45 mA at 85 C TJ
    and 22 mA at 110 TJ, and can sink 86 mA to ground per side, 41 at 110 TJ. >>>
    The outputs are configured as push-pull or open drain, depending on the
    firmware/application, either way they can source or sink 40mA in
    isolation. But no more than two outputs per side at a time will be in a
    state such that they can sink/source current either way.

    The uP I'm using has a very nice 10 bit differential ADC (9 bits if used >>> in bipolar mode) I'm thinking about just rigging it to two 4051s to
    "scan" a current sense in each output line to implement the overcurrent, >>> can set different thresholds based on whether it's configured push-pull
    or OD, but what might conservative limits be? And what rate to scan?

    The SPLD is acting as a simple power supervisor and has a PWR GOOD line
    to bring up the uP. In turn it gets a clock from the uP, I can set up
    the SPLD to set all those outputs hi-z if it ever loses the clock (i.e.
    uP hangs) as an extra layer of protection..



    Ptc thermistor or polyfuse?

    I'd rather use something KISS and actively monitor as a backup for sure,
    but I haven't used those devices on low current output lines before, do
    you have a suggestion for a suitable part? Thanks



    I was thinking of something like the Murata PRG18 series?

    But instead of one PTC per output maybe the mcu can monitor current in the
    spld supply and ground pins. May need some amplification to keep sense
    shunt resistance values low.


    --
    piglet

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From john larkin@21:1/5 to bitrex on Tue Feb 11 07:50:58 2025
    On Mon, 10 Feb 2025 22:59:23 -0500, bitrex <user@example.net> wrote:

    On 2/10/2025 10:41 PM, john larkin wrote:
    On Mon, 10 Feb 2025 17:57:06 -0500, bitrex <user@example.net> wrote:

    I have an SPLD with about 8 terminals exposed to the user I'd like to
    protect against over current. They're very low-speed outputs (10s of Hz
    at most), don't really want to spend a bunch for a chip like the L6374
    as an intermediary, or the board space for a bunch of transistors to
    discrete limit...

    The SPLD has a max per-chip-side Vdd input current of 45 mA at 85 C TJ
    and 22 mA at 110 TJ, and can sink 86 mA to ground per side, 41 at 110 TJ. >>>
    The outputs are configured as push-pull or open drain, depending on the
    firmware/application, either way they can source or sink 40mA in
    isolation. But no more than two outputs per side at a time will be in a
    state such that they can sink/source current either way.

    The uP I'm using has a very nice 10 bit differential ADC (9 bits if used >>> in bipolar mode) I'm thinking about just rigging it to two 4051s to
    "scan" a current sense in each output line to implement the overcurrent, >>> can set different thresholds based on whether it's configured push-pull
    or OD, but what might conservative limits be? And what rate to scan?

    The SPLD is acting as a simple power supervisor and has a PWR GOOD line
    to bring up the uP. In turn it gets a clock from the uP, I can set up
    the SPLD to set all those outputs hi-z if it ever loses the clock (i.e.
    uP hangs) as an extra layer of protection..

    What's an SPLD?



    If it ain't a FPGA or a CPLD it's an SPLD. I think.

    What part is it? It seems to have power drive beyond what an FPGA
    would have.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From bitrex@21:1/5 to john larkin on Tue Feb 11 11:55:53 2025
    On 2/11/2025 10:50 AM, john larkin wrote:
    On Mon, 10 Feb 2025 22:59:23 -0500, bitrex <user@example.net> wrote:

    On 2/10/2025 10:41 PM, john larkin wrote:
    On Mon, 10 Feb 2025 17:57:06 -0500, bitrex <user@example.net> wrote:

    I have an SPLD with about 8 terminals exposed to the user I'd like to
    protect against over current. They're very low-speed outputs (10s of Hz >>>> at most), don't really want to spend a bunch for a chip like the L6374 >>>> as an intermediary, or the board space for a bunch of transistors to
    discrete limit...

    The SPLD has a max per-chip-side Vdd input current of 45 mA at 85 C TJ >>>> and 22 mA at 110 TJ, and can sink 86 mA to ground per side, 41 at 110 TJ. >>>>
    The outputs are configured as push-pull or open drain, depending on the >>>> firmware/application, either way they can source or sink 40mA in
    isolation. But no more than two outputs per side at a time will be in a >>>> state such that they can sink/source current either way.

    The uP I'm using has a very nice 10 bit differential ADC (9 bits if used >>>> in bipolar mode) I'm thinking about just rigging it to two 4051s to
    "scan" a current sense in each output line to implement the overcurrent, >>>> can set different thresholds based on whether it's configured push-pull >>>> or OD, but what might conservative limits be? And what rate to scan?

    The SPLD is acting as a simple power supervisor and has a PWR GOOD line >>>> to bring up the uP. In turn it gets a clock from the uP, I can set up
    the SPLD to set all those outputs hi-z if it ever loses the clock (i.e. >>>> uP hangs) as an extra layer of protection..

    What's an SPLD?



    If it ain't a FPGA or a CPLD it's an SPLD. I think.

    What part is it? It seems to have power drive beyond what an FPGA
    would have.


    I use these a lot:

    <https://www.renesas.com/en/products/programmable-mixed-signal-asic-ip-products/greenpak-programmable-mixed-signal-products/greenpak-asynchronous-state-machine/slg46537-greenpak-programmable-mixed-signal-matrix-asynchronous-state-machine>

    They're cheap enough in quantity to use as general output expanders and
    the pins have more configurable options than the average uP tends to; push-pull, NMOS open drain, PMOS open drain, 1x,2x,4x drive depending on
    the pin, with 10k, 100k, 1Meg or floating selectable for pull-up/pull down.

    The async state machine is a nice add-on and one can do some e.g. power sequencing with that.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)