We haven't been able to determine this. Where we're seeing it is on a
pair of Exfo test sets. We're trying to qualify a new 50 Mbps
dedicated circuit using RFC 2544 testing using the Exfos. Tests are
failing, reporting lost frames. The confusing thing is that our
counters on both radios and switches are fairly clean. The radios
report zero "defective blocks". The only errors I am seeing that I
wonder about show up on both the radio and ethernet sides. On one
radio, I see 38 "TX length error frame count" errors since Saturday on
the ethernet port. I also have 38 "RX length error frame count"
errors on the radio. On the other side, I have 39 "RX length error
frame count" errors on the ethernet. None on the radio side.
However, I question whether these errors are related to what I'm
seeing in any way. I tested another leg of this circuit the other
day. Identical radios. It tested clean, but was also showing some of
these same errors.
Craig
Quoting David Milholen <dmilho...@wletc.com>:
Are these frame losses on the Radio side or the ethernet side?
On 4/16/2016 10:32 AM, Craig Baird wrote:
So a few months ago we purchased an 11 GHz PTP820S 2+0 link. We
recently installed it, and it appears to be working fine, aside
from a little bit of frame loss that we are investigating. While
looking into this frame loss issue, I stumbled across something
that concerns me. On Cambium's support forum there is a post that
states that when dealing with 2+0 links both radios must be in the
same sub-band. There is no explanation of why this is the case.
In our situation, the radios are in separate sub-bands. When we
did the frequency coordination, the only two 80 MHz channels
available were in different sub-bands. I passed those channels
along to our vendor who worked with Cambium to get a BOM. At no
point did anyone say that this was a problem. So now, fast forward
a few months, and I stumble across this post, and now I'm wondering
what the implications will be. Both links are up and running.
Signal on both is right where it should be (-39 on one, -40 on the
other). Both are running at maximum modulation. There are no
defective blocks shown on the radio interfaces. There is no
indication that this sub-band mismatch is causing any issues, aside
from possibly this frame loss thing. However, if I mute the radios
on one link, the frame loss persists, so I don't think it's related.
In case it matters, the two links are oppositely polarized. On one
side we've got a 2 foot dish with an OMT combining the radios. On
the other side, we've got an 8 foot dual-pol dish.
So I'm wondering if anyone knows why Cambium says that you can't
use radios from different sub-bands. Are we in for trouble at some
point?
Craig
--
