I have been tasked with installing one of these systems for a sixteen-channel bi-directional STL. This system was first mentioned here: The 16 channel bi-directional STL system. As some of you pointed out, the unlicensed 5.8 GHz IP WLAN extension was the weak link in this system. It was not an interference issue, however, which was creating the problems. The problem was with layer two transparency in the TCP/IP stack. Something about those Cambium PTP-250s that the Wheatstone Blade hardware did not like and that created all sorts of noise issues in the audio. We installed the Wheatstone Edge Routers, which took care of the noise issue at the cost of latency. It was decided to go ahead and install a licensed link instead of the license-free stuff as a permanent solution.
Thus, a Cambium PTP-820S point-to-point microwave system was purchased and licensed. The coordination and licensing took about three months to complete. We also had to make several changes to our network architecture to accommodate the new system. The PTP-820 series has a mast-mounted radio head, which is the same as the PTP-250 gear. However, for the new system, we used three different ports on the radio to interface with our other equipment instead of the single port PTP-250 system. The first is the power port, which takes 48 VDC via a separate power cable instead of POE. Then there is the traffic port, which uses Multi-Mode fiber. Finally, there is the management port, which is 1GB Ethernet and the only way to get into the web interface. The traffic port creates a completely transparent Ethernet bridge, thus eliminating all of the layer two problems previously encountered. We needed to install fiber transceivers in the Cisco 2900 series switches and get those turned up by the IT wizards in the corporate IT department.
The radios mount directly to the back of the 24-inch 11 GHz Andrew antenna (VHLP2-11) with a UBR100 interface. The waveguide from the radios is a little bit deceptive looking, but I tried not to overthink this too much. I was careful to use the O ring grease and conductive paste exactly where and when specified. In the end, it all seemed to be right.
Not wanting to waste time and money, I decided to do a back-to-back test in the conference room to make sure everything worked right and I had adequately familiarized myself with the ins and outs of the web interface on the Cambium PTP-820 radios. Once that was done, it was time to call the tower company.
One side of these is mounted on the studio building roof, which is a leased space. I posted RF warning signs around the antennas because the system ERP is 57.7 dBm, which translates to 590 watts at 11 GHz. I don’t want to fry anybody’s insides, that would be bad. The rooftop installation involved pulling the MM fiber and power cable through a 1 1/4-inch EMT conduit to the roof. Some running back and forth, but not terrible work. I used the existing Ethernet cable for the management port. This will be left disconnected from the switch most of the time.
The other side is mounted at about 85 feet AGL on a hot AM tower. I like the use of fiber here, even though the tower is skirted, the AM station runs 5,000 watts during the daytime. We made sure the power cables and Ethernet cables had lighting protectors at the top of the run near the dish and at the bottom of the tower as well as in the transmitter room rack. I know this tower gets struck by lightning often as it is the highest point around for miles.
Aligning the two dishes was a degree of difficulty greater than the 5.8 GHz units. The path tolerances are very tight, so the dishes on each end needed to be adjusted in small increments until the best signal level was achieved. The tower crew was experienced with this and they started by panning the dish to the side until the first side lobe was found. This ensured that the dish was on the main lobe and we were not chasing our tails. In the end, we achieved a -38 dBm RSL, the path predicted RSL was -36 dBm so close enough. This means the system has a 25 dB fade margin, which should be more than adequate. While were aligning the transmitter site dish, a brief snow squall blew through causing a whiteout and the signal to drop by about 2 dB. It was kind of cool seeing this happen in real-time, however, strangely enough, the tower crew was not impressed by this at all. Odd fellows, those are.
Currently brushing up on FCC part 101 rules, part C and H. It is always good to know the regulatory requirements of any system I am responsible for. As AOIP equipment becomes more mainstream, I see many of these types of installations happening for various clients.
5 thoughts on “Part 101, Private Fixed Microwave Service”
Another interesting read. Always look forward to see what your doing next that may save me some time down the road…
Just being curious, how big were the latencys with the previous license free hardware ?
Is this a big problem for radiobroadcasting ? The listeners wont notice if the broadcasts are being delayed a few 100msec ?
Just my thinkning ,
Anyways great post and interesting reading your blog, thx
The 5.8 GHz unlicensed link had only a few milliseconds of latency. The problem came when we installed the Edge Routers, which buffer the incoming audio. That added much more, and took away some of the functionality of the Blade hardware.
You could have been writing about our situation!
We’re using an analog STL to am AM site with the exact same 5 kw and skirted tower. Only our STL dish is 300′ just to clear the path.
There’s the real possibility we’ll have to use this AM site as a midpoint to another AM site so we’re looking at possible solutions.
Very few unlicensed IP radios are based on TDM… as we know realtime audio such as wheatnet relay on the jitter being totally stable.
We use Huewai licenced microwave and have wheatnet running over it and it works fantastic… (standard IP88AD blades not the edge blade)