Not throwing any blame around here. Just trying to help. Like I said,
I'm not real worried about GigE on the regular 450 APs... yet. Would be
nice to find a solution to this problem so we can eventually run gigabit
when the 5GHz finally gets approved for 30MHz channel bandwidth. And of
course Cambium get a stable official build done with 30MHz support.
Probably not too far off from that at this point though, maybe by the
end of the year.
Then there's the 450i. And the 450m down the road. The 450i can take
2-pair or 4-pair power. If you're using the aux port for POE output to
another radio, then you really need to power it on 4 pairs. The 450m's
60 (or was it 70?) watts will no doubt demand all four pairs. And both
need 56VDC. 24/29.5 is fading away.. finally. And I'm guessing Cambium
will make the regular 450 AP fade away too. We haven't deployed any
450i. No need yet. Probably just end up jumping to the 450m where we
need the capacity.
Anyway, my point is, 4-pair power plus the sync pulse + GigE-APC/SS-HV I
foresee being bad for 10/100 as well as gigabit. Cambium set it up real
nice on the 450i though. It'll take either polarity on each pair. So I
suppose you could set it up to put positive on 4/5 & 7/8, negative on
1/2 & 3/6, then maybe 10/100 would still work at least.
If you and Forrest want to collaborate on it or something, that would be
cool. If you come up with a new GigE-APC prototype to resolve this, I'd
be more than willing to give you real world results.
Then there will be 1588v2 timing. Which is all fine and good. But I
won't be ripping out sync-over-power when this becomes available. But
everything is becoming "I wants my netflicks, now!" So I really want
every AP having at least two full-time timing sources going forward if
possible.
On 9/10/2016 4:37 PM, Chuck McCown wrote:
I am wondering if I need to spin a Cambium only surge suppressor.
-----Original Message----- From: George Skorup Sent: Saturday,
September 10, 2016 3:23 PM To: [email protected] Subject: Re: [AFMUG]
Gigabit Ethernet on Cambium 450i only running 100Mbps
My site is a GigE PowerInjector+Sync. No changes to the jumpers. Took
it out of the box, plugged it in. The injector shows 23.9 volts (from
a RSD 48 to 24 converter) on PowerA. PowerB has nothing connected. Two
5GHz 450 APs on ports 1 & 2, two 3.6GHz 450 APs on ports 3 & 4. ~225
feet of shielded cable to each radio.
I would assume the same. Probably would be worse on 48/56v. I wouldn't
expect this on 24V systems though. I believe Matt said his setup is a
GigE SyncInjector, GigE-APC-HV's and PTP-450i's, which would
definitely require 48-56VDC. 450i's need at least 40 volts at the
radio, IIRC.
I just confirmed this at another site where we added a GigE
PowerInjector+Sync with two 5.7 450 APs and one 3.6 450 AP. Slightly
less cable, about 185 feet. The 3.6 AP doesn't have an in-line SS yet.
Radio direct to injector port. Ethernet directly to a RB1100AHx2 port.
I just powered that AP back up and switched it to negotiate 1000F.
Main status page shows 1000Base-T FDX and receiving sync on the power
port. MT port shows 1Gbps FDX. No errors. This one is fed similarly,
just regulated 24VDC on PowerA. The two 5.7 APs are on older L-com
10/100 suppressors, so obviously I can't test gigabit on those.
So yeah, I can pretty much confirm what Matt is seeing. The sync pulse
+ GigE-APC-HV in line appears to be the culprit. Am I remembering that
Chuck said the GigE HV's clamp at 69 volts on all pairs? Seems like
that should be enough headroom for 24VDC + the sync pulse. Perhaps the
cable length is pushing it over the edge.
I believe negotiation takes place on the 10/100 pairs only? Then the
radio's PHY has a feature called Ethernet@Wirespeed. If it doesn't see
data (or in this case loses I guess) on the 4/5 and/or 7/8 pairs, then
it gives up on GigE and falls back to 10/100 negotiation. That's why
the link is bouncing at 1Gbps a couple times at boot and then settles
on 100 FDX. SS clamps due to the sync pulse. No gigabit for u!
On 9/10/2016 4:30 AM, Forrest Christian (List Account) wrote:
Just curious... what voltage are you feeding into the injector?
I wonder if maybe if you're starting at 56V, then adding a bit of
inductive surge effects due to the cable and sync pulse thing. I
could see this surging into 60 volts for sure... not sure where
Chuck's suppressors clamp at.
-forrest
