07-17-2017 06:35 AM
Whilst commissioning my first AF24 link recently I encountered some problems with the ETH-SP surge protectors.
I had connected an ETH-SP to each AF24 and used UBNT shielded outdoor cable and RJ45 plugs. I had also cable tested each cable & ETH-SP for continuity.
The first problem was I couldn't connect to the slave AF24 from the data or management ports if I had an ETH-SP in the loop. Bypass the ETH-SP and I was able to connect to the AF24. So this was removed and I was able to ping the master AF24, the router and 126.96.36.199.
However, I noticed that occassionally the ping would drop to the router and 188.8.131.52. Very strange. I got a ping going from a PC at the master site to the slave AF24 and two devices at the master end. The ping to these devices all dropped at the same time. After my experience with the ETH-SP at the slave end I bypassed it at the master end and the ping stayed up.
I have AirMax links that have had no problems using ETH-SP.
07-17-2017 12:34 PM
07-17-2017 07:40 PM - edited 07-17-2017 07:43 PM
I use these on all my towers. I've seen all sorts of problems with them too.. here's a summary:-
- Loose connectors. A large amount of undercrimp required to get a reliable connection if, for example, I use a platinum tools ratchet crimper, I cannot use it's default setting - need to back off the ratchet at only about half depth - which pretty much equates to the pins being flat level with the dividers. If I crimp to standard depth even the slightest movement of the cable will cause an outage.
- Failed qualification - a Fluke Cable-IQ fails to qualify at 1gbps even for short runs if these are installed as suggested (one at building entry point and one at top of tower). The fluke metre reports excessive attenuation as the cause. If only one is used they usually qualify ok.
I'm still using them as I bought them in bulk - but I've found I have to deeply undercrimp, zip tie the cables in and use silicon grease to get any semblance of reliability out of them. They're not an easy product to use at all.
07-18-2017 02:46 AM
For anyone that's interested this video shows part of the problem... the depth of crimp... it's possible that I just got a bad batch. What this video doesn't show is the other problem (more than 1 eth-sp cascaded = attenuation fault).