For many years, the de facto standard for event time tagging in a substation
has been ± 1 millisecond. This could easily be achieved in commercially
available sequence of events recorders and fault recorders, as all of the events
were hard wired into one box, with one clock. There was no issue of
synchronizing clocks between several IEDs. This was the background when the
Requirement Specifications for Substation Integrated Protection, Control, and
Monitoring were being developed in 1995 and 1996. No one was willing to forgo
that ± 1 ms time tagging standard, and so it became a requirement when RP3599
was published in December 1996. Recognizing the well-known measurement principle
that the measuring device must be one order of magnitude more precise than the
measured result, RP3599 requires that the clocks in all time tagging IEDs be
synchronized to ± 0.1 ms. The same requirement is specified in the draft
versions of IEC 61850 and IEEE P1525. RP3599 calls this “time synching”, but we believe
that the term “clock synching” is more descriptive. Regardless of what it is
called, the requirement has been in RP3599 since 1996, but is being almost
universally today ignored by most IED vendors.
This is so easy to check. Do as your editor did at the Utility Initiative
demonstration in Andover, MA last September. In the demo room, he asked each IED
vendor “What is the least significant bit that may be set (by any means) in
your IED’s clock. One admitted that his IED’s clock could only be set to the
nearest 10 ms. A few said their clocks could be set to 1 ms, but only with an
IRIG-B port, and not all their models include such a port. Only one vendor has
clocks that were settable to 0.1 ms or better. As an interesting sidelight, the
IEEE Synchrophasor Standard requires sampling to be synchronized to the nearest microsecond.
If you are expecting that, in the future, some of your relay IEDs might be
called on to produce synchrophasor data, check on how precise their clocks may
be set. Your editor urges you to do your own survey, and Email us your results.
An additional issue is what clock synching precision can be achieved over a
10 or 100 MB Ethernet LAN. According to the experts who were present at the
demonstration, it now appears that the best that can be realized is clock
synching to about 3 – 5 milliseconds over the LAN. Since this is more than an
order of magnitude worse than RP3599 requirements, one would think that all
vendors would be planning to include an IRIG-B port in every time tagging IED,
but such is not the case. Again, we urge the utilities to do their own survey to
avoid surprises later on. We’d like to see these results, too.
Early on, it was the hope of many utilities that the substation LAN could be
used to transport clock synching signals to all IEDs from a substation master
time source. It now seems clear that even a relaxed requirement of ± 1
millisecond over the LAN cannot be achieved even if the IED had the
capability of such clock setting precision – and some do not. Thus, a
direct IRIG-B timing port (preferably fiber) will be needed to meet even this
relaxed requirement, yet not all IEDs include this as a hardware option.
It’s time the utilities make their needs known loud and clear. Tell us
your requirements, and we will help.
WiseOwl™
