audio & video latency, packet loss, jitter, surge protec

About six months ago we approached one of the regional school district’s in our area about providing connectivity to the schools for their video classroom courses. previously they were paying around $900 a month per connection for a single T1. The regional school district is made up of 8 smaller municipal school districts in 8 towns and 2 college branches in our area that we had POP’s located in. This allows them to provide classes that they normally would not be able to provide due to low enrolment. They are using video conference equipment at each school. All video feeds link to a special switch at one of the college’s that allows them to control witch schools participate in certain classes. My thinking on it was that if a T1 could handle it so could Canopy and for the most part it has. This is how you find out how good your network really is. Most of the issues we have had really wouldn’t be noticed with internet traffic, but with full duplex audio and video you will. Latency, packet loss, network jitter all have to be low to nonexistent. A standard command prompt ping -t test might not show any issues and if any of the traffic is UDP with no retransmit or accountability capabilities you really notice especially with a congested link when the UDP has to fight for space with TCP. Using a tool like ping plotter or what I use is Solarwinds enhanced ping. It charts the latency, packet loss, and network jitter on a graph making problems much more visible. Rf is our main business and I am pretty much a perfectionist when it comes to installation and equipment so there is no shoddy work but I am just one person, and yes I do screw up occasionally. I guess my point is all bases were covered I thought. All installations were done per Motorola specs and considering we are a Motorola MSS service center and Mobile VAR all work that we do for Motorola must meet R56 standards and it has become standard practice. Anyway this is already getting longer then it should so ill get to the point. After searching and searching latency really never was an issue, packet loss and jitter were. The cause turned out to be the CAT 5 surge protection. I have tried the Motorola 300ss, 300ss verB, Transtector, Hyperlink, and others with no success. We talked to a Motorola Engineer on the design team of the 300ss who acknowledged there is a problem especially if you use a suppressor and shielded cable and connectors. The shielded cable and connectors has been known to cause the suppressor to trip. His recommendation was to use shielded cable and connectors and not to use the surge suppressor. After removing all the suppressors at all the POP’s everything is running fine. We still use the 300ss at customer locations so we are not liable for their equipment getting damaged and we run the risk of losing our own equipment but have no other choice. Another piece of good info. On the CMMmicro the 24 volt power that feeds the Canopy radios is all tied together so a short can cause a problem for all or any radio in the cluster. The same goes for the surge suppressors. One bad one on a CMM can cause problems for any or all too. If anybody has any other solution I would love to hear about it. Sorry for the long post.

So you were having issues with the Transtectors on the AP’s at your pops?


Can you go into detail? All of our pops use the spec’d transtector supressors and full R56 grounding (MSS here as well so you know the drill) and I have not had issues.

nice post

Is it better to use transtector like protection or use shielded cable only and ground the shield? the more I read, the more I get confused

I have always used STP and the suggested Transtectors and haven’t had an issue…


YMMV though…

I didn’t think I had an issue either, But as the system gets loaded more and more you start to see things that were not their before especially with the use of video and Voip. The first indication of a problem was on a project for a customer where we were installing a SmartLink wide area trunking network. A few of the sites were linked together through our Canopy network. For our area the customer is a large corporate electric and natural gas utility company and have engineers on staff who designed the system. The sites that were effected were set up to use a MultiTech MultiVOIP gateway to send analog voice channels between sites and a modem plugged into the rs232 port of the smartlink and the other end of the modem plugged into one of the MultiVOIP channels for sending data. For the record I do not like this setup at all. The customer started to complain about packet loss and surely it had to be on our network. My boss wasn’t seeing any problems with a standard ping -t test so he gave me a call. Sure enough looking at the counters in the MultiVOIP’s major packet loss. About the same time we were starting to bring online the school project that had to go through the same links. Starting to troubleshoot I found that the data links using the modems were wanting to run all the time. when packet loss would happen it would drop the modem link and would not come back up until a reset of the modems. At one site I saw entries in the event log of the BH’s that looked like an occasional loss of sync. I checked the cables, tried replacing the shielded connectors no luck. Replaced cables still not it. changed out BH still not it. CMM showed 9 visible and 9 tracked sats, 3d fix, no errors. PS voltage good. When touching one of the cables that plugged into the CMMmicro I saw all the GPS lights go off, then come back on. Turned out to be a cold solder joint on the main board of the CMM. Not taking any chances I installed a new CMM. What do you know the packet loss was still there. Using enhanced ping to monitor the link allowed me to ping all devices between sites in the order of the physical location and I could see the loss on the Ethernet port of one of the BH’s further down the line. I checked out everything at that site just like the previous, but it wasn’t until I bypassed the surge protector that the loss went away. During turning up the school video project I found a few more sites with the same issues that wouldn’t normally show up.

Nepalken the only advice I can give you is to follow Motorola installation procedures. All I can do is inform you of my situation and what it took to resolve it. I am putting myself at risk of losing equipment and will not suggest anyone else do the same. You will have to determine that yourself.

We do install surge protectors at the customer location to protect their equipment and not make us liable in the event of an issue.

Nepalken

From Motorola for subscriber installations

"Shielded & Unshielded Cabling
The preceding design guidelines are based on using unshielded twisted pair (UTP) cabling. One may ask, “Could a shielded twisted pair cable be used and thereby eliminate the need for a 300SS discharge unit?” A search for braid shielded 10BaseT cabling was unfruitful. There exists foil
shielded, Cat 5, 100BaseT cabling that includes a 24 AWG drain wire, e.g., Foiled Shielded Twisted pair (FTP). This cabling is considerably more expensive than UTP, e.g., $0.50 per foot versus $0.30 per foot. Also, would grounding the 24 AWG drain wire with a 10 AWG copper
wire at the entrance of the dwelling satisfy the NEC 810-20 requirements of using a metal shielded cable that is effectively grounded? This is not an easy question to answer and there are probably differing answers depending on the expert one might ask. NEC 810-21 states in part:
“A lightning arrester is not required if the lead-in conductors are enclosed in a continuous metal shield, such as rigid or intermediate metal conduit, electrical metallic tubing, or any metal raceway or metal-shielded cable that is effectively grounded. A lightning discharge will take the path of lower impedance and jump from the lead-in conductors to the metal raceway or shield rather than take the path through the antenna coil of the receiver.”
In the case of FTP, the drain wire will need to withstand the lightning current without melting, i.e., fusing open. If the drain wire fuses open, then the current may follow the twisted pairs into the building. A 24 AWG copper wire will fuse at 8,500 amps from a surge with a 1 microsecond
by 70 microsecond waveform. More than 80 percent of all direct lightning strikes have a current that exceeds 8,500 amps as seen previously in Figure 2. Therefore, given a direct strike to the Canopy SM, we would expect the 24 AWG drain wire to fuse open 80 percent of the time thus
creating the possibility that the lightning current will travel through the twisted pairs into the dwelling. Hence, the reliance on a 24 AWG drain wire to meet the intent of NEC 810-21 is questionable. Based on these results, it is mandatory to have the SMMB1 (Canopy mounting bracket) grounded with a 10 AWG copper wire connected by the most direct path to the ground bonding point (where the ground of the AC power service utility entry is, or eight foot ground rod). This would better assure that lightning would take the 10 AWG wire route to earth ground and most likely not fuse open and meet the NEC 810-15 section. It is a good practice to have the SMMB1 mounting pipe located even or higher than the top of the Canopy SM. Use at least a 10 AWG copper wire from the Canopy Surge Suppressor 300SS ground lug to the same ground bonding
point as above or have a separate eight foot ground rod so as to meet NEC 810-21.

And for AP’s and BH’s

Motorola recommends the use of shielded outdoor cables that adhere to Category 5 and 5e (TIA/EIA 568-B) standards for installation of the Canopy system AP and BH modules. Motorola has designated Best-Tronics Manufacturing as an authorized dealer of cables that meet our
rigorous specifications. Cables of this nature are also available from various cable manufacturers and are suitable for use with the Canopy system. The Best-Tronics Category 5e (Wireless Internet Service Cable CA-0287S) Shielded Outdoor Plenum Cable is an enhanced screened twisted pair cable for use in outdoor horizontal cabling
systems. The cable exceeds Category 5 and 5e electrical characteristics and consists of 24 AWG solid bare copper insulated conductors assembled into four tightly twister pairs with an overall foil and drain enclosed in an overall jacket.
The Downside of Using Unshielded Cable
If shielded cable is not used in AP or BH module installations, the integrity of the Canopy system may potentially be compromised with problems occurring such as:
• Inability to keep Global Positioning System (GPS) synchronization.
• Failure to properly negotiate 10 or 100 BaseT connections.
In addition, without properly shielded cables the signal may become distorted and customers may experience interruptions to their service. In the end, the only resolution may be to install the shielded cable hence requiring additional costs for replacement cable and installation.