Getting PPS on a modern rack-mounted server? - NTP

This is a discussion on Getting PPS on a modern rack-mounted server? - NTP ; This may sound like an embarrassment of riches, but here is my problem: I have some rack-mounted servers (IBM System x3650 to be precise) that I need to synchronize with a PPS signal. The problem is, of course, that the ...

+ Reply to Thread
Results 1 to 15 of 15

Thread: Getting PPS on a modern rack-mounted server?

  1. Getting PPS on a modern rack-mounted server?

    This may sound like an embarrassment of riches, but here is my problem:
    I have some rack-mounted servers (IBM System x3650 to be precise) that I
    need to synchronize with a PPS signal. The problem is, of course, that
    the machines have only one serial port, which I want to use as a serial
    console, and of course they have no parallel ports.

    Any suggestions? Can I use a PCI serial card? Will that have acceptable
    jitter?

    Thanks,

    /ji

  2. Re: Getting PPS on a modern rack-mounted server?

    John Ioannidis said the following on 05/29/2008 03:10 PM:
    > This may sound like an embarrassment of riches, but here is my problem:
    > I have some rack-mounted servers (IBM System x3650 to be precise) that I
    > need to synchronize with a PPS signal. The problem is, of course, that
    > the machines have only one serial port, which I want to use as a serial
    > console, and of course they have no parallel ports.
    >
    > Any suggestions? Can I use a PCI serial card? Will that have acceptable
    > jitter?


    A PCI serial card should work fine. The main jitter source is the
    interrupt latency which shouldn't be much different on a PCI card than
    on the motherboard port.


    If you need to distribute the same PPS signal to a number of boxes, I
    designed a board which is being sold as a kit by TAPR (we have
    arrangements with someone who can assemble them for a fee, if required)
    that provides up to six TTL level PPS signals, as well as two RS-232
    compatible signals. Check http://www.tapr.org/kits_tadd-3.html.


    John

  3. Re: Getting PPS on a modern rack-mounted server?

    > Date: Thu, 29 May 2008 15:10:35 -0400
    > From: John Ioannidis
    > Sender: questions-bounces+oberman=es.net@lists.ntp.org
    >
    >
    > This may sound like an embarrassment of riches, but here is my problem:
    > I have some rack-mounted servers (IBM System x3650 to be precise) that I
    > need to synchronize with a PPS signal. The problem is, of course, that
    > the machines have only one serial port, which I want to use as a serial
    > console, and of course they have no parallel ports.
    >
    > Any suggestions? Can I use a PCI serial card? Will that have acceptable
    > jitter?


    John,

    You may not need a PCI serial card.

    Almost all rack-mount system with a single serial port also have a
    header for the second serial port. I have many of these, though they are
    SuperMicro, not IBM. The physical connector is a a bulkhead plate that
    fits in the same place as a PCI card would be mounted.
    --
    R. Kevin Oberman, Network Engineer
    Energy Sciences Network (ESnet)
    Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab)
    E-mail: oberman@es.net Phone: +1 510 486-8634
    Key fingerprint:059B 2DDF 031C 9BA3 14A4 EADA 927D EBB3 987B 3751

    _______________________________________________
    questions mailing list
    questions@lists.ntp.org
    https://lists.ntp.org/mailman/listinfo/questions

    --- StripMime Report -- processed MIME parts ---
    multipart/mixed
    multipart/signed
    text/plain (text body -- kept)
    application/pgp-signature
    text/plain (text body -- kept)
    ---

  4. Re: Getting PPS on a modern rack-mounted server?

    In article <483EFFAB.6030303@tla.org>,
    ntp@tla.org (John Ioannidis) writes:
    >This may sound like an embarrassment of riches, but here is my problem:
    >I have some rack-mounted servers (IBM System x3650 to be precise) that I
    >need to synchronize with a PPS signal. The problem is, of course, that
    >the machines have only one serial port, which I want to use as a serial
    >console, and of course they have no parallel ports.
    >
    >Any suggestions? Can I use a PCI serial card? Will that have acceptable
    >jitter?


    I'd expect it to be as good as the built-in serial port.

    If you discover otherwise, please let us know.

    --
    These are my opinions, not necessarily my employer's. I hate spam.


  5. Re: Getting PPS on a modern rack-mounted server?

    jra@febo.com (John Ackermann N8UR) writes:

    >John Ioannidis said the following on 05/29/2008 03:10 PM:
    >> This may sound like an embarrassment of riches, but here is my problem:
    >> I have some rack-mounted servers (IBM System x3650 to be precise) that I
    >> need to synchronize with a PPS signal. The problem is, of course, that
    >> the machines have only one serial port, which I want to use as a serial
    >> console, and of course they have no parallel ports.
    >>
    >> Any suggestions? Can I use a PCI serial card? Will that have acceptable
    >> jitter?


    >A PCI serial card should work fine. The main jitter source is the
    >interrupt latency which shouldn't be much different on a PCI card than
    >on the motherboard port.


    >
    >If you need to distribute the same PPS signal to a number of boxes, I
    >designed a board which is being sold as a kit by TAPR (we have
    >arrangements with someone who can assemble them for a fee, if required)
    >that provides up to six TTL level PPS signals, as well as two RS-232
    >compatible signals. Check http://www.tapr.org/kits_tadd-3.html.
    >


    Why BNC connectors?



    >John


  6. Re: Getting PPS on a modern rack-mounted server?

    Unruh said the following on 05/29/2008 05:00 PM:

    >>
    >> If you need to distribute the same PPS signal to a number of boxes, I
    >> designed a board which is being sold as a kit by TAPR (we have
    >> arrangements with someone who can assemble them for a fee, if required)
    >> that provides up to six TTL level PPS signals, as well as two RS-232
    >> compatible signals. Check http://www.tapr.org/kits_tadd-3.html.
    >>

    >
    > Why BNC connectors?


    Because most time-and-frequency related test gear uses coaxial cable for
    interconnects, and BNCs are the most commonly used connectors. For
    really precise measurements, connectors that are threaded rather than
    pressure-fit, like SMA or TNC, are preferred. But the BNC is pretty
    universally used.

    The TADD-3 does have two RS-232 level outputs on headers that mate with
    ribbon cable to DB-9 adapters, but the performance (specifically, the
    pulse rise and fall times) is much worse than the TTL levels because the
    RS-232 spec calls for relatively slow transitions to minimize RFI. We
    use the fastest version of the MAX-232 level converter chips, but it's
    still orders of magnitude slower than the rise-time of the coaxial
    outputs (which are a couple of nanoseconds).

    John

  7. Re: Getting PPS on a modern rack-mounted server?

    jra@febo.com (John Ackermann N8UR) writes:

    >Unruh said the following on 05/29/2008 05:00 PM:


    >>>
    >>> If you need to distribute the same PPS signal to a number of boxes, I
    >>> designed a board which is being sold as a kit by TAPR (we have
    >>> arrangements with someone who can assemble them for a fee, if required)
    >>> that provides up to six TTL level PPS signals, as well as two RS-232
    >>> compatible signals. Check http://www.tapr.org/kits_tadd-3.html.
    >>>

    >>
    >> Why BNC connectors?


    >Because most time-and-frequency related test gear uses coaxial cable for
    >interconnects, and BNCs are the most commonly used connectors. For
    >really precise measurements, connectors that are threaded rather than
    >pressure-fit, like SMA or TNC, are preferred. But the BNC is pretty
    >universally used.


    >The TADD-3 does have two RS-232 level outputs on headers that mate with
    >ribbon cable to DB-9 adapters, but the performance (specifically, the
    >pulse rise and fall times) is much worse than the TTL levels because the
    >RS-232 spec calls for relatively slow transitions to minimize RFI. We
    >use the fastest version of the MAX-232 level converter chips, but it's
    >still orders of magnitude slower than the rise-time of the coaxial
    >outputs (which are a couple of nanoseconds).


    Since my GPS is only good to just under a usec, and since servicing of
    interrupts is also at that level, it does not really seem to matter. The
    program I wrote was to run off the parallel port interrupt, but it seems to
    have about 2usec jitter on readout.
    Ie for computer timing situations, a few nsec is overkill. How much cheaper
    would a board with say 200ns output be? Is there a circuit diagram of that
    kit anywhere?




    >John


  8. Re: Getting PPS on a modern rack-mounted server?

    Unruh said the following on 05/29/2008 06:36 PM:

    > Since my GPS is only good to just under a usec, and since servicing of
    > interrupts is also at that level, it does not really seem to matter. The
    > program I wrote was to run off the parallel port interrupt, but it seems to
    > have about 2usec jitter on readout.
    > Ie for computer timing situations, a few nsec is overkill. How much cheaper
    > would a board with say 200ns output be? Is there a circuit diagram of that
    > kit anywhere?


    The rise time of the MAX-232 chip is about 2.5us; there are some chips
    available that support 1 MBPS operation, and have much faster slew
    rates, but they violate the RS232 spec which requires a maximum slew
    rate of 30v/us (or about 330ns for +/-5v).

    The TADD-3 manual, which includes schematics, is at
    http://www.tapr.org/~n8ur/TADD-3_Manual.pdf

    On the TADD-3, there's essentially no added cost for the performance --
    we just used 74AC chips which have the best performance of TTL logic
    chips. Of course, the input source needs to be quiet enough to make
    that performance usable.

    John

  9. Re: Getting PPS on a modern rack-mounted server?

    Unruh wrote:

    > Since my GPS is only good to just under a usec, and since servicing of
    > interrupts is also at that level, it does not really seem to matter. The
    > program I wrote was to run off the parallel port interrupt, but it seems to
    > have about 2usec jitter on readout.

    That is about the time an 8Bit ISA buscycle takes +- .
    Is the jitterdistribution a gauss bell or a box?

    uwe

  10. Re: Getting PPS on a modern rack-mounted server?

    Kevin Oberman wrote:
    >
    > Almost all rack-mount system with a single serial port also have a
    > header for the second serial port. I have many of these, though they are
    > SuperMicro, not IBM. The physical connector is a a bulkhead plate that
    > fits in the same place as a PCI card would be mounted.


    Some MB only have a header for an IRda module.
    Then TX and RX are inverted TTL levels, no handshakes.

    uwe




  11. Re: Getting PPS on a modern rack-mounted server?

    John,

    The PPSAPI was designed so that any serial device, including the
    console, could be used for the DCD signal. Our problem with Sun Netras
    is that the DCD signal is not supported. Something like a SIIG serial
    card should work just fine.

    Dave

    John Ioannidis wrote:

    > This may sound like an embarrassment of riches, but here is my problem:
    > I have some rack-mounted servers (IBM System x3650 to be precise) that I
    > need to synchronize with a PPS signal. The problem is, of course, that
    > the machines have only one serial port, which I want to use as a serial
    > console, and of course they have no parallel ports.
    >
    > Any suggestions? Can I use a PCI serial card? Will that have acceptable
    > jitter?
    >
    > Thanks,
    >
    > /ji


  12. Re: Getting PPS on a modern rack-mounted server?

    Uwe Klein writes:

    >Unruh wrote:


    >> Since my GPS is only good to just under a usec, and since servicing of
    >> interrupts is also at that level, it does not really seem to matter. The
    >> program I wrote was to run off the parallel port interrupt, but it seems to
    >> have about 2usec jitter on readout.

    >That is about the time an 8Bit ISA buscycle takes +- .
    >Is the jitterdistribution a gauss bell or a box?


    >uwe


    It seems to be a "bell". A bit hard to tell since the clock only has a usec
    resolution, so one is working near the quantization limit. But it is
    definitely larger than 1usec, and nearer 2.5-3usec.

  13. Re: Getting PPS on a modern rack-mounted server?

    David L. Mills wrote:
    > John,
    >
    > The PPSAPI was designed so that any serial device, including the
    > console, could be used for the DCD signal. Our problem with Sun Netras
    > is that the DCD signal is not supported. Something like a SIIG serial
    > card should work just fine.


    Could you elaborate. Which model of Netra? Are you saying that the
    model of Netra you are talking about has a serial port, but that
    serial port does not have DCD connected?

    Brian Utterback

  14. Answer: PPS on a rack server

    A few days ago I had asked this list for assistance with the following
    problem: my rack-mounted IBM x3650 servers only have one serial port,
    and the data I was feeding them from my GPS clock were interfering with
    the serial console feature, the BIOS in particular. Thanks to several
    people who replied to my original message, I now have a working setup.
    Here it is:

    1. A Spectracom 9283 netclock feeds *only* the PPS signal to the
    servers. Pins 1 (DCD) and 5 (GND) of a DB-9F connector are connected to
    the core and shielding of the coax (through T connectors, and the cable
    is terminated with a 50 Ohm terminator from the old thinlan days!). The
    signal seems to be strong enough to drive the DCD pin, but for longer
    runs I would use this: http://www.tapr.org/kits_tadd-3.html.

    2. NTP is used to get the time reference from the Spectracom.

    3. The harder part was figuring out what to put in /etc/ntp.conf. Here
    is what I have (the address of the Spectracom is not the real one, of
    course)

    server 192.2.200.2 minpoll 5 maxpoll 5 prefer # 9283

    server 127.127.22.0 minpoll 5 maxpoll 5 # PPS
    fudge 127.127.22.0 flag3 1

    Of course, the kernel is configured with option PPS, and /dev/cuad0 is
    symlinked to /dev/pps0.

    That's it! The machines appear to be within a few microseconds of each
    other, so I'm happy.

    4. I can still use the serial console, after a fashion -- the machines
    support IPMI, so I can use serial-over-lan with ipmitool to get console
    i/o (for the BIOS, which is all I care about) without actually needing
    to plug anything to the serial ports. Of course, what I really should
    have done is paid the extra $1K or so and gotten the RSA-II cards for
    the boxes, but that's another story.

    Thanks again to all who replied

    /ji

  15. Re: Getting PPS on a modern rack-mounted server?

    Brian,

    I haven't seen the machine; it allegedly is a 1U and was installed as a
    replacement for a Sun 10. My report was based on a staff message. I do
    know the connector is a modular jack with four pins.

    Dave

    Brian Utterback wrote:

    > David L. Mills wrote:
    >
    >> John,
    >>
    >> The PPSAPI was designed so that any serial device, including the
    >> console, could be used for the DCD signal. Our problem with Sun Netras
    >> is that the DCD signal is not supported. Something like a SIIG serial
    >> card should work just fine.

    >
    >
    > Could you elaborate. Which model of Netra? Are you saying that the model
    > of Netra you are talking about has a serial port, but that serial port
    > does not have DCD connected?
    >
    > Brian Utterback


+ Reply to Thread