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 On Fri, Sep 9, 2016 at 7:10 PM, George Skorup <[email protected]> wrote: > So, what do you think? Based on my test, it's definitely the sync pulse > disrupting the GigE data, at least on the 7/8 pair I'm guessing. > > Is it possible that we have HV marked modules that are in fact not HV? I > assume you have two different assembly lines for regular and HV? Or do you > just do batches of one or the other? Either one would probably rule that > out. > > It would really suck to have to pull out the SS's to get GigE on the 450 > when we finally get 30MHz channel support. No surge suppression at all > scares me. 20MHz channel width really doesn't need gigabit which is why > I've never bothered with it. > > Maybe Forrest would be willing to send you one of his new > PowerInjector+Sync modules and a SyncPipe for debugging? > > And with the 450i taking power on all four pairs, I could see the sync > pulse causing the SS to clamp on the 10/100 data pairs, too. Meaning > 100Mbps won't work either. Just speculation, but it seems logical. > > What else could you do other than let say 90 volts though instead of, what > is it on the HV now, 65 or 70 volts? Or maybe some slightly slower reacting > diodes? But I guess either case opens up the potential (pardon the pun) for > damage. > > > On 9/9/2016 7:16 PM, Chuck McCown wrote: > > You are correct, the POE version adds the transformer. > > *From:* George Skorup <[email protected]> > *Sent:* Friday, September 09, 2016 5:51 PM > *To:* [email protected] > *Subject:* Re: [AFMUG] Gigabit Ethernet on Cambium 450i only running > 100Mbps > > Uh oh. I figured it out. It was as I suspected. The SS is clamping due to > the sync pulse. Maybe the change that Forrest made with the Rev i0 > SyncInjectors/PowerInjector+Sync because of the 450i's finicky handling > of the pulse over long runs? I have no idea. So much for GigE + sync over > power I guess. I don't have any Rev i0 Gigabit SyncInjectors to test with, > but I assume I'd get the same result. > > I set the AP back to negotiate GigE. Turned off the sync on the injector > port feeding that radio. Rebooted the AP. Link stays up at gigabit with no > problems. Then I turned the sync pulse back on and the link dropped almost > instantly. > > 18:33:42 interface,info ether6 link down > 18:33:49 interface,info ether6 link up (speed 1G, full duplex) > 18:33:53 interface,info ether6 link down > 18:33:55 interface,info ether6 link up (speed 1G, full duplex) > > .... sync pulse back on > > 18:35:31 interface,info ether6 link down > 18:35:34 interface,info ether6 link up (speed 1G, full duplex) > 18:35:35 interface,info ether6 link down > 18:35:42 interface,info ether6 link up (speed 1G, full duplex) > 18:35:43 interface,info ether6 link down > 18:35:46 interface,info ether6 link up (speed 1G, full duplex) > 18:35:47 interface,info ether6 link down > 18:35:52 interface,info ether6 link up (speed 100M, full duplex) > > The GigE-APC's I have at this site are definitely HV's. That's all I have > ever ordered. They are all marked HV in sharpie on the handle. I'm feeding > that cluster with regulated 24VDC from a RSD150C-24. > > On 9/9/2016 6:34 PM, George Skorup wrote: > > Uh.. hmm. I have a 5.7 450 sector running at a tower without any customers > on it yet. It's powered and timed from a new GigE PowerInjector+Sync. > GigE-APC-HV at the bottom. No SS on top. ~225 feet of Shireen 1042 cable. > > I just logged into it and set it to Auto 1000F/100F/etc. and rebooted. > Does the same shit. > > 18:14:07 interface,info ether6 link down > 18:14:20 interface,info ether6 link up (speed 1G, full duplex) > 18:14:21 interface,info ether6 link down > 18:14:26 interface,info ether6 link up (speed 1G, full duplex) > 18:14:27 interface,info ether6 link down > 18:14:30 interface,info ether6 link up (speed 1G, full duplex) > 18:14:31 interface,info ether6 link down > 18:14:40 interface,info ether6 link up (speed 100M, full duplex) > > I also just set it back to Auto 100F (disabling gigabit negotiation) and > it does this... > > 18:21:34 interface,info ether6 link down > 18:21:47 interface,info ether6 link up (speed 1G, full duplex) > 18:21:48 interface,info ether6 link down > 18:21:52 interface,info ether6 link up (speed 100M, full duplex) > > So, looks like the boot loader brings it up at 1Gbps, then the config > makes it fall back to 100Mbps. > > This is also the first time that I tried messing with GigE on the 450 > since updating from 13.2.1 to 14.1.2. > > This is a regular old RB1100. I have an Exalt ExtendAir G2-11 and an AF24 > each running on a GigE-POE-APC and gigabit works fine. Chuck would have to > say for sure, but I don't think there's much electrical difference between > the GigE-APC-HV and the GigE-POE-APC. Besides maybe the gas-tube backup > suppression and GigE transformer. > > Me thinks this has more to do with Cambium. > > On 9/9/2016 6:10 PM, George Skorup wrote: > > I upgraded two CRS125-24G-1S-RM's to 6.36.3 (current branch). Both brand > new out of the boxes. Upgraded boot firmware to 3.33 on both. Reset both to > defaults. Ether24 to ether24 linked up at 1Gbps FDX. Ran MT bandwidth tests > between them. No issues. > > Then I put a GigE-APC-HV in between with a 50' patch cable on one side and > a 10' on the other side. Tried with and without the ground attached. Ran > more bandwidth tests. No link drops. No errors. > > So... I have no idea why you're having problems. The ones I have are > definitely HV Rev A1's. > > On 9/9/2016 2:24 PM, Matt wrote: > > I have a couple CRS125's on my bench that I can mess with this afternoon if > I have time. > > Do you have a GigEAPC-HV? My setup was CRS125 -<~50 foot > UTP>-<GigEAPC-HV>-<20 foot UTP>- CRS125 > > > > > -- *Forrest Christian* *CEO**, PacketFlux Technologies, Inc.* Tel: 406-449-3345 | Address: 3577 Countryside Road, Helena, MT 59602 [email protected] | http://www.packetflux.com <http://www.linkedin.com/in/fwchristian> <http://facebook.com/packetflux> <http://twitter.com/@packetflux>
