Showing posts with label DSL. Show all posts
Showing posts with label DSL. Show all posts

Saturday, January 8, 2011

DSL/Phone Wiring

Starting off the New Year right is one of my priorities in life. For me, that often means implementing a fun project, typically involving technology and/or something around the house, car, bicycle or all of the above. This year was no exception and I decided that it was time to put together all the parts I had accumulated over the past several months to update my telephone wiring. To be clear, this is *not* about ripping out all the wiring behind the walls in my house. It's just to improve the "visible" stuff in the unfinished basement utility room area. There were 4 primary goals for this project:

1. Clean up the spider web phone wiring mess just inside the house.
2. Directly connect a DSL jack to Telco Network Interface Device (NID).
3. Implement whole house phone line surge protection.
4. Clean up the alarm system phone line connection.

In addition to these enumerated goals, a side benefit was to gain knowledge and experience about proper wiring supplies, techniques and equipment.

It helps to understand the solutions if the specific problem areas are described in more detail. The following paragraphs will do this ... check them out to see if you may have some of the same issues or concerns about "doing it right" (or at least "doing it better than it was" ).

POTS (Plain Old Telephone System) Wiring
As was common in the mid 90's when this house was built, standard 4 wire (Red, Green, Yellow, Black) telephone cable was installed. Not very resistant to cross talk and capable of generating quite a bit of noise, I was uncomfortable with using it for my DSL connection to the Telco (Qwest, in this case) NID outside the house. The inside wiring was done in the "Star" configuration where each wire pair was connected to the NID and then run directly to a room (or area) of the house. In the destination area, it might be hooked up in series (daisy chained) but the star technique still allowed for isolation of connection problems to a small number of phone jacks. Good idea, but it felt like a better idea to bring a single line into the house (and see the DSL section, below) for a very short distance, subsequently distributing it via multiple cables to the various rooms.

DSL
The aDSL (Asymmetric Digital Subscriber Line) 7Mbps connection had been working OK but there were occasional times (confirmed by Qwest technical support) when somewhat high "dropped bit" rates were encountered. Not knowing the source of this, it felt important to do the best possible job of eliminating local house wiring as an issue. For this reason, and in tandem with the goal to bring a single cable set in from the NID, 4-UTP (Unshielded Twisted Pair) Cat 5E cable was chosen to bring the connection in to the house. While good cable alone doesn't make for a good connection (connectors, jacks, proximity to power and noise carrying lines and installation technique are all contributors), using the correct cable *is* a key requirement for a good, solid connection and it's a straightforward thing to do. The idea was to run as directly as possible from the NID to a jack specifically used only for connecting to the DSL modem.

The "normal" DSL setup for a residence is usually done on the cheap and often consists only of connecting Telco supplied low pass filters to all the active "phone" connections to keep the high frequency DSL signal noise out of voice conversations and prevent the low frequency (voice and DC) signals from attenuating the DSL signal. Not wanting to put a filter on each phone in the house, the original wiring job fed the input lines to a pair of DSL filters before distribution to the rest of the house (with the DSL line running directly out to the NID). These small filters, however, are not designed to operate this way and may be overloaded, depending on the number of phones in the house that simultaneously ring. The alternative chosen was to implement what's called a DSL-POTS splitter. This device is mounted (depending on its specific design) at the NID or just inside the building and it filters and splits the input line's signal into its voice and DSL components. It does cost a bit ($35 or so for this one) but it provides a relatively cheap and elegant alternative to appending a filter to each phone. In this case, the Wilcom PS-15-I3S (mounts inside the residence) xDSL-POTS splitter was chosen and two were installed to allow for possible future easy movement of the DSL connection to Line 2 (from the standard Line 1) as well as the option to have 2 independent DSL lines active simultaneously. Another benefit is the redundancy factor--for single line DSL, should one of the splitters fail, it would be a very quick procedure to switch to the other, avoiding more than minimal down time.

Surge Protection
Similar to the issue (see above) regarding DSL/voice filters, the proliferation of phone line surge protection devices was something that needed addressing. Another good argument for bringing a single set of wires into the house from Telco, the idea was to run them through a surge protection device *before* distribution through the rest of the house phone wiring. The device chosen to accomplish this was the Open House ChannelPlus Telephone Entry Module (model H611). The H611 is designed to mount in a service center enclosure box (Open House model H318 used here) and provide surge protection for up to 4 incoming Tip/Ring pairs. In addition, it provides an RJ-31X jack to allow line seizure capability for an alarm system. Since the DSL line bypasses the voice line surge protection provided by the H611, a separate, surge protection device (TrippLite's Traveler 100Bt) was added at the DSL jack to protect the DSL modem (and downstream devices). Simple, effective (it is hoped) and now the whole house is protected from some degree of line surge without having to incorporate multiple surge protection devices.

Alarm System
While the alarm system (Rokonet Wisdom) was a DIY home project, it works very well and is designed to seize the phone line to call out to user programmed phone numbers (ostensibly to an alarm monitoring service but, for this system, just my cell phone). In order to do this, the alarm system must be connected as the first "phone" in system and the RJ-31X jack is designed to make that easy to do. The challenge here was to take out the old 4 wire cable connecting the alarm system and connect it to the H611's RJ-31X jack with standard (3-UTP) 24 gauge twisted pair wiring.

In summary, the problems above were to be addressed and the implementation accomplished is described below. Pictures of the detailed connections in the H318 Service Center Enclosure tell the story.


This picture shows the white H318 Service Center Enclosure box mounted in its final location between two wall studs in the basement utility room. The two black components mounted near the top of the box are the Wilcom PS-15-I3S DSL-POTS splitters for Line 1 and Line 2, respectively. The horizontal rectangular unit below the splitters is the H611 Telephone Entry Module. The blue Cat 5E cable entering the Enclosure at the top left side carries the Line 1 and Line 2 wires from the Telco Network Interface Device outside the house. The other blue Cat 5E cable exits the Enclosure at the top right side and goes directly to the DSL modem wall jack.

The leftmost white cable (on the H611) is a Cat 3 UTP from the "Phone" outputs of the two splitters and it connects via the H611's 110 punchdown connector for (input from the CO) Tip and Ring on Lines 1 and 2.

The middle white cable is another Cat 3 UTP connected to the H611's RJ-31X connector and it runs to the alarm system in the upstairs entryway to the house.

The rightmost white cable leaving the H611 consists of the surge protected Line 1 and Line 2 outputs from the H611 and it runs up and to the left of the Enclosure where it connects to the rest of the house's phone wiring.

Finally, the green wire running horizontally across the top of the Enclosure is simply a 10 gauge ground wire.


This is a closeup of the Open House H611 Telephone Entry Module. Here you can see (from left to right) the 110 connector for lines 1-4 (note that lines 1 and 2 are connected), the RJ-45 telephone disconnect plug, the RJ-31X alarm connector (at the bottom) and the house telephone output on the right side with Line 1 and Line 2 connected via the RJ-45 plug (which is connected in parallel to the alternative output 110 connector). For maximum connection flexibility, RJ-45/RJ-11 plugs were used wherever the option was available. Note the "Surge Warning" lights associated with each of up to 4 telephone lines that the H611 might be protecting. Each line is protected to 50 Joules and 4000 Amps. If the surge level is high enough, surge protection will be lost and the surge event will be shown as an illuminated LED associated with the appropriate line(s). Repair or replacement of the H611 module would then be required. Regarding connecting wires to the 110 connector--be sure to invest in a 110 punch down tool for this task. The 110 is a fairly sophisticated Insulation Displacement Connection (IDC) system and trying to get by with, for example, a screwdriver and/or needle nose pliers is unlikely to seat the wire properly, resulting in damaged contacts and possible future failure of the connected phone line(s). Finally, the module's RJ-45 jacks are wired to the TIA T568A standard.

The H611 module connects to the Service Enclosure with metal tabs in the back designed to fit the vertical columns of device mounting holes in the Enclosure.


This picture shows the Wilcom xDSL-POTS splitters mounted and connected in the Service Enclosure. The two units are facing each other to allow use of the RJ connections while minimizing the wire length needed as Line 1 and Line 2 are broken out from their respective cables to be sent to the individual signal splitter devices. The splitter on the left handles Line 1, the one on the right handles Line 2. Since these devices are not designed to mount directly into the Enclosure's pre-drilled holes, Velcro was used to attach them and to give some additional flexibility regarding precise location.

Testing
Integrity of the wiring was verified by inspection and by alternating the Line 1 supplied DSL signal between both the Line 1 and Line 2 splitters. Given this setup utilizing the RJ plug connections, line swapping to accomplish this test was trivial. Additionally, although this residence does not currently have regular POTS service from Qwest, a Phone Labs Dock-N-Talk was attached at the Network Interface Device to provide Tip and Ring in parallel with the DSL signal from the Telco. The phones and alarm system in the house were successful at dialing out. But full confidence in the alarm's line seizure capability will not be achieved until such time as regular Telco service can be used to verify it. Most DIYer's working on this kind of project probably have regular Telco phone service so that test is something that will normally be easily accomplished and it should not be neglected.


Finally, the new Service Enclosure all buttoned up and operating as planned. Much neater looking than the random wires that were floating around in the basement. For future reference, all device documentation as well as a rough schematic and URL reference to this blog are included in an envelope inside the Service Enclosure.

It is hoped that the information presented here is useful to others that may be mulling over the issues of telephone line surge protection, DSL/POTS performance and possible problems with handling an alarm system's dialing requirements. Again, this is just what one individual did and there may be better solutions ... but perhaps it's a good starting point.

All comments are welcome and please don't be put off because comments are reviewed before appearing in the article.

Thursday, December 23, 2010

Qwest DSL--Throttling?

You may be aware that there's quite a bit of discussion reflecting concern and actual problems with true available bandwidth for the end users of various broadband delivery systems, including cable, satellite and DSL. I have Qwest DSL (ADSL), which I recently upgraded from 1.5 Mbps (download) to 7.0 Mbps. The key reason I upgraded was because I was seeing quite a bit of lag time with streaming video ("buffering") from various news websites as well as YouTube. The idea I had was that the basic bandwidth being received might have been so close to the edge that any kind of internet load could cause such delays. In addition, as past posts will reveal, I added MagicJack internet calling this spring, further consuming bandwidth on my DSL line. So it seemed like a good idea and the cost was actually a wash if I switched from my local ISP to Qwest's preferred ISP, MSN.

The switchover was very smooth and I went so far as to even buy Qwest's ActionTec PK5000 modem (rather than a 3rd party modem). So I was lined up with Qwest's preferred equipment and ISP, the change was very efficient and there I was one fine day, running at 7Mbps and enjoying it. By the way, no faster DSL speed is available at my location or I may have tried for something like 20Mbps.

But the honeymoon was fairly short lived. Although the interruptions were fewer, I started noticing that there was still quite a bit of streaming video buffering going on. The typical time of day was during the work week between 6pm and 10pm CT with virtually no problems during the daytime unless it was a holiday or, occasionally, on weekends. I started capturing bandwidth data from 3 main online (and free) speed measuring sites: bandwithplace.com, speedtest.net and pcpitstop.com. Sure enough, I was seeing download speeds of well under Qwest's requirement (80% of 7Mbps, or 5.6Mbps). Often speeds were in the under 2Mbps range with the lowest reading to date being .77Mbps (on December 9, 2010 at 7:44pm CT) ... and, yes, these are download readings, *not* upload (which is spec'd at about .75Mbps and is consistently around .65 to .72Mbps).

So I called Qwest and immediately learned a very interesting fact. Their performance criteria of 80% of 7Mbps applies *only* to the data link (in the midwest) from Qwest's servers in Kansas City or Minneapolis (take your pick) to your residence (for example, I'm here in Iowa). So if you want to verify that Qwest is contractually compliant with their requirement on their DSL delivery bandwidth, the only way to check it out is to run their test at http://kansas-city.speedtest.qwest.net/ ... what that means is that your overall performance is not their concern at all (yes, I was told those very words). The only thing they are obligated to provide is whatever bandwidth you're paying for from their server(s) to your residence.

This is new stuff to me and I've not seen it described this way in the forums and blogs I've been reading on the subject. The fact that my performance is good except when it seems that a much larger part of the population is probably logging on to the internet could mean that the internet is just busy and everyone is slowing down ... or not. I have a theory that says it's possible for Qwest to throttle the "input* side of their servers based on total data transfer requests (which will obviously go up during high usage times like the evening) but still provide the full bandwidth paid for on their servers' output side (which is the input to residences). In my case, I typically do see about 6Mbps and, since it should never be lower than 5.6Mbps (.8 * 7), it's certainly acceptable. But why is it that there are many times when I'm seeing 6Mbps from Qwest but at the same time only, say, 1.58Mbps from a measuring site like bandwidthplace.com or speedtest.net?

When Qwest actually came out to my home, they verified that the modem was properly installed, the wiring was all good and everything in my local network was working fine. But I have also actually seen bandwidth data from Qwest showing download speeds as low as 2.1 to 2.7Mbps. It's quite rare, but did give me the reason for a no-cost visit from them while they checked it out. They had no answers for me but when the problem occurs (and I'm keeping the measurements in a spreadsheet) they also want me to verify the internal modem status on my LAN at 192.168.0.1. It always has shown 7168Kbps for download and 896Kbps for upload but I'm also now collecting that data.

So the bottom line here is that it looks suspicious to me that I can get excellent performance from Qwest but dismal overall bandwidth when loading actual internet data. And, after all, it's overall performance that counts. If I'm getting throttled 1Mbps data over a "bigger pipe" (i.e. 7Mbps DSL), it's still only 1Mbps to my system. And what I'm wondering is if there are other folks out there that see similar performance, particularly Qwest DSL customers. Any 20Mbps customers out there that have similar issues? It's discouraging to me to think that I could probably go back to 1.5Mbps service and have everything work just as well ... but that's about where I am today, based on what I see at my computer and the data I've collected regarding this issue.

To be "fair and balanced", I do see where it makes sense for Qwest to have contractual terms that limit their performance liability to something they can control (i.e., Kansas City or Minneapolis to you). Otherwise, they'd probably be getting hammered by folks about every underpowered server system on the internet where the output bandwidth may be 1/2 Meg or so. But the performance hits I regularly see (pretty much every day) are enough to at least make me think that DSL might be a poor choice, at least from Qwest, for consistent download bandwidth operation. Wish there were some alternatives, like fiber optic ... but cable is my only other choice and it's more expensive, plus there seem to be similar bandwidth issues there, according to some comments I've heard and read.

Please comment if you have more information on this. Thanks everyone!