We have a Force 400c link which is exhibiting strange behavior. We are using UBNT AF-5G30-S45 dishes with included 6" coax jumpers and isomax radomes at both ends. Our alignment seems to be right on and the Master end shows an RSSI of -66/0 and SNR of 18/0 (down/up).
What’s with the unbalanced numbers? Could someone help me understand what might be happening here? Damaged antenna/coax? We’ve tried realigning both ends to no avail. I did notice the Slave end is on firmware 5.0.1 while the Master is on 5.1.3 - could this be the entire issue?
As Eric has said, check your rf cables. Could also be a bad antenna and there is still a possibility of a bad radio. If you can hook up another force400c to the dish and the chains make no change then after testing both ends then its a good chance of a bad antenna.
As to which one, make 2 cables and change one pair then the other. If no changes then start swapping another radio in.
These kinds of problems are tedious and time consuming to figure out.
I don’t know if it is the firmware but v4+ firmware has a reputation for being very problematic when you mix and match versions.
What does the slave end show ? Our first PTP550 did this (one chain was 0) however I’m pretty sure in our case it turned out to be a seriously misaimed master.
IP 0.0.0.0, Device Name Unavailable ? -60/0 23/0 yet DS7/DS7 100% / 60% god I hate the ePMP interface…
Wonder why not a single bit above MCS 7. Did you do like me and load a config to these radios from PTP550 (or maybe a AC/N will cause problems also) ? I did that and even though the lying ePMP interface showed everything configured correctly in actuality it was not and I had to change then save settings and then change/save them back to force it actually have the settings the Interface was showing. In my case the pair was locked to MCS 7 (or maybe it was 9) even though the settings showed DS MCS11
The slave radio started rebooting randomly I ended up defaulting both ends and setting them up from scratch. No problems since so looks like loading configs from PTP550 (and possibly other models) can really bork things.
I believe the Ubiquiti AF-5G30-S45 offers a 45 degree slant to the signal that is being transmitted across the two nodes.
I have some experience from the ALGcom antennas that when you make such a slant on both antennas, you will need to reverse the polarity of the RF connections at one end. I will advice you to do the same and see if you will get all the information displayed normally.
RF RP-SMA Connections:
AF-5G30-S45-#1 and F400C-#1: Ch0-to-Ch0 and Ch1-to-Ch1
AF-5G30-S45-#2 and F400C-#2: Ch0-to-Ch1 and Ch1-to-Ch0
We are going to buy these antennas this week/anytime soon and might have to do the same thing. Hoping to learn from you. Would be good to see the F400C perform well with the AF-5G30-S45.
I’m sorry for not updating this topic sooner - the entirety of my problem stemmed from my firmware mismatch oversight. As soon as I updated the firmware to the most current on both ends and performed another alignment, I saw - 54dBm at 8 miles (13km). No polarity reversal needed.
For reference, I’ve attached photos of how I mounted the 400 radios to the Ubiquiti dishes, enabling us to utilize the included 4" jumpers. One of my colleagues mounted his differently on another link, using instead some 3’ coax. His final result was measurably worse than this link.
Thanks for the response given and the pictures shared.
Yes, the firmware should be the same on both sides for the radio communication to make sense.
From the pictures, it looks like you crossed the 4" jumpers already. Not sure if it is what you did on both sides. The UBNT Quick Start Guide actually says: +45° to Chain 0 and -45° to Chain 1. I can see that the CH1 and CH0 on the F400C are on opposite sides with the AF-5XHD, perhaps forcing the cables to be crossed.
So you are confirming that you did the same thing on both side:
RF RP-SMA Connections:
AF-5G30-S45-#1 and F400C-#1: +45° to Ch0 and -45° to Ch1
AF-5G30-S45-#2 and F400C-#2: +45° to Ch0 and -45° to Ch1
But yes, the 4" RF jumpers are fixed to the feedhorn, forcing you to place the F400C as close as possible to the back of the antenna. I like the way you mounted the F400C. Were you able to cover it with the protective shroud that comes with the antenna?
Having a second look at the pictures, I can see that the F400C radio is not mounted on the antenna, meaning the protective shroud would not fit. This means that the F400C radio is flipped around behind the antenna, sitting on the brace metal between the pole brackets. Interesting!
From what I can see, it looks like you did:
RF RP-SMA Connections:
AF-5G30-S45-#1 and F400C-#1: +45° to Ch1 and -45° to Ch0
AF-5G30-S45-#2 and F400C-#2: (No pics)
Also, this picture looks like one side of the link. Did you manage to have picture of the other side?
Thank you very much for the enlightenment and the clarity that is coming from these posts.
Just FYI according to Cambium the ePMP radios work out the polarity regardless of how you have them connected to the rocketdish (or other antenna). We use to just make sure that both ends matched (e.g. both ends crossed or not crossed) but Cambium has said even that doesn’t matter to ePMP and I have intentionally mismatched the ends of links and then put them back and noticed not even the smallest difference in the link (thats been several years ago though as we don’t have any ePMP PTP any more other than a new AX link we just put up).
They don’t even appear to care if one end is H / V while the other is 45° since the Cambium ePMP 2.4Ghz sectors are 45° and the 2.4Ghz client radios are H/V. They cliamed it was to reduce interferrence on the AP end since most interferrence is H/V. I have no idea why they didn’t just go ahead and make the 2.4 clients 45° aslo.
I did not mismatch chain orientations on this link but instead matched them to each other. As Douglas said, there’s no need for this, proven by the great performance of our link.
@Thomolar The cover did not fit over this setup, instead we filled the UBNT pigtail recess with silicone and applied mastic tape to the Cambium radio’s bulkheads the best we could. As I mentioned, taking care to use the shortest possible coax has yielded substantial improvement over an identical link utilizing longer jumpers. The signal loss on this type of cable is too great over distance.
pigtail length matters for long cable lengths, As long as your cables are a good quality cable, the length is not cut to a multiple of the frequency period length and the connectors are properly attached paying attention to ensure the lowest RF injection attenuation then the cables will be the least of your loss worries.
We commonly run cables 4ft long and are seeing VSWR of 1.1:1