I write this letter in response to “Fusion splicing for critical low-loss applications” (October 2017). I like splice-on connectors (SOCs). I agree with many of the statements in this article. However, I see another side of the story.
SOCs require a fusion splicer. While the price of the splicers has dropped considerably in the past 10 years, for many installers, the price is significant. Is the use of a fusion splicer and SOCs the best choice?
The answer is my favorite: It depends on the total installed cost (TIC). To be a logical business decision, the TIC of using SOCs should be lower than the cost of alternative connector installation methods. If the TIC of the SOCs method is lower than that of alternative methods, the savings pay for the splicer.
This analysis is more complicated than it may appear, as it requires several steps. The first step is consideration of: connector cost, process yield, labor cost, and labor utilization. Labor utilization is the ratio of time spent in the activity of connector installation to the total time required for the connector installation. If the rate of installation by any method is 15/man-hour, but time spent in travel to site, setup and cleanup consumes 25 percent of total time, the actual rate per man-hour is 15x0.75, or 11.25/man-hour.
The second step in the decision to use SOCs is the determination of the number of connectors that the installer needs to install to recover the cost of the splicer. If that number is very high, 20,000 connectors, and the installer expects to install 2,000 connectors per year, SOCs may not be the best choice.
The third step in the decision is determination of how many fusion splicers will the installer need. A single installer requires a splicer. If the completion time for jobs is short, multiple splicers may be required. In such a situation, the total number of connectors that must be installed to recover the cost of multiple fusion splicers increases.
At this time, there are splicer and SOC combinations that have TICs that are both higher and lower than alternative methods.
Of course, there are situations in which other factors predominate the decision to use SOCs. As a first example, the activity is replacement of connectors in an enclosure that has no space for splice trays to be used in pigtail splicing. In this situation, SOCs are the best choice.
As a second example, an activity may require connector replacement in an enclosure with hundreds of pigtailed connectors. Replacement of the pigtails would require removing and replacing old pigtails with new pigtails. Such replacement may take more time than replacing connectors with SOCs.
As a third example, the installation organization intends to perform midspan splicing. Thus, it will have a splicer. In this situation, several activities justify purchase of a splicer capable of installing SOCs.
As a fourth example, low link loss is essential to link operation. This reason, included in this article, is critical in data centers, in which the links are established, but not necessarily permanent. Low loss provides the maximum flexibility for future rearrangement of links without resulting in excessive link loss.
However, for initial installations, there is an alternative that can result in TIC lower than that of SOCs: pigtail splicing. There are several factors that favor pigtail splicing. The first is cost; pigtails range in cost from approximately $5 to $10 per end, while SOCs range in cost from approximately $7 to $18. Even adding the cost of a splice tray, at <$1/end, the TIC for pigtail splicing can be lower than that of SOCs.
The second factor is the yield of pigtail splicing, which is 100 percent. If the first splice is unacceptable, the installer can re-splice the pigtail. Conversations I have had indicate that SOCs have acceptable yield, but not 100 percent. The splicer influences SOC yield; sometimes the fusion splicer “burps,” resulting in a gas bubble or high loss.
As I said in the beginning of this letter, I like SOCs and recommend them to clients in certain situations. But cost considerations do not always favor their choice.
Eric R. Pearson, CFOS/T/C/S/I
President, Pearson Technologies Inc.
PoE to the desktop? Really?
The article “PoE-enabled computing: The next step in the digital building” (September 2017) made me wonder if it was April Fool’s Day.
Can PoE provide the power? The lowest-power normal PCs use 30-100W power and a 21-inch LCD display uses about 90W for 120-190W total. A 21-inch iMac with the low-res LCD uses about 75W, but the retina display version uses 161W. Can these people convince PC manufacturers to build such computers? Will companies buy them instead of laptops?
Practically all workers today use laptops, not desktop computers. Typical users of desktops are power users doing engineering or graphics on workstations that require really high power. This idea could work for laptops that use lower power, I suppose because most are within the 90W limit. But most modern laptops do not have a jack for Cat 5—none of ours do!
Practically all of those laptops people use are connecting to the network using WiFi, or maybe cellular wireless.
Because of building codes, wherever a worker sits will have AC outlets.
Perhaps it’s just another sign that the “Cat 5” community is really grasping for something to do with all that copper wire in their warehouses. Just like the plan to run LED lighting off all those small conductors. All these proposals have also brought PoE to the attention of the people writing electrical codes, because the high power heats up the cables sufficiently to be a potential hazard.
President, The Fiber Optic Association
Editor’s note: Jim Hayes’s letter is derived from an item in The Fiber Optic Association’s October 2017 newsletter.