[Linuxptp-devel] EtherCAT and max frequency shift

[Bassam Alsanie]

I have couple of questions.

Can linuxptp and etherCAT coexist?

Regards the maximum frequency shift the clock can support, I have a case
that my custom ptp clock has maximum of 1000 ppm possible tune from the
center frequency. My question is, how the ptp algorithm deal with the cases
where it need to tune the clock beyond the maximum range? Does it cause to
adjust the time again after the servo locked?

Thank you
Bassam,

[Richard Cochran]

Post by Bassam Alsanie
I have couple of questions.
Can linuxptp and etherCAT coexist?

Define "coexist".

Post by Bassam Alsanie
Regards the maximum frequency shift the clock can support, I have a case
that my custom ptp clock has maximum of 1000 ppm possible tune from the
center frequency. My question is, how the ptp algorithm deal with the cases
where it need to tune the clock beyond the maximum range?

The maximum adjustment is a fundamental limitation of the clock, and
there is nothing you can do about it. However, 1000 ppm is enough to
tune most any clock, I would think.

Post by Bassam Alsanie
Does it cause to
adjust the time again after the servo locked?

-ENOPARSE;

Thanks,
Richard

[Bassam Alsanie]

Post by Richard Cochran

Post by Bassam Alsanie
I have couple of questions.
Can linuxptp and etherCAT coexist?

Define "coexist".

etherCAT has its own software stack over the physical layer (mostly at the
data link layer), and I was wondering if I enabled the etherCAT on my
device, can the linuxptp run over etherCAT software stack? as the TCP/IP is
not necessary available when the etherCAT is enabled. Another question
would be, does the linuxptp run (and require) on specific OSI software
layer (such as TCP, UDP or IP)?

Post by Richard Cochran

Post by Bassam Alsanie
Regards the maximum frequency shift the clock can support, I have a case
that my custom ptp clock has maximum of 1000 ppm possible tune from the
center frequency. My question is, how the ptp algorithm deal with the

cases

Post by Bassam Alsanie
where it need to tune the clock beyond the maximum range?

The maximum adjustment is a fundamental limitation of the clock, and
there is nothing you can do about it. However, 1000 ppm is enough to
tune most any clock, I would think.

Post by Bassam Alsanie
Does it cause to
adjust the time again after the servo locked?

-ENOPARSE;
Thanks,
Richard

[Richard Cochran]

Post by Bassam Alsanie
etherCAT has its own software stack over the physical layer (mostly at the
data link layer), and I was wondering if I enabled the etherCAT on my
device, can the linuxptp run over etherCAT software stack?

PTP is either Layer-2, UDP/IPv4, or UDP/IPv6, and so it cannot run
over EtherCAT.

EtherCAT has its own synchronization protocol called "Distributed
Clocks" (DC), and so adding PTP makes no sense at all.

Post by Bassam Alsanie
as the TCP/IP is
not necessary available when the etherCAT is enabled.

PTP does not use TCP in any way.

Post by Bassam Alsanie
Another question
would be, does the linuxptp run (and require) on specific OSI software
layer (such as TCP, UDP or IP)?

See above.

HTH,
Richard

[Keller, Jacob E]

-----Original Message-----
Sent: Monday, November 09, 2015 6:26 AM
To: Bassam Alsanie
Subject: Re: [Linuxptp-devel] EtherCAT and max frequency shift

Post by Bassam Alsanie
Regards the maximum frequency shift the clock can support, I have a case
that my custom ptp clock has maximum of 1000 ppm possible tune from

the

Post by Bassam Alsanie
center frequency. My question is, how the ptp algorithm deal with the

cases

Post by Bassam Alsanie
where it need to tune the clock beyond the maximum range?

The maximum adjustment is a fundamental limitation of the clock, and
there is nothing you can do about it. However, 1000 ppm is enough to
tune most any clock, I would think.

For perspective, many of the Intel parts have less than 700ppm adjustment range and work just fine.

Regards,
Jake