Sunteți pe pagina 1din 14

Portable and Distributed Power for Telecommunications

Jimmy Godby Telecom Power Consultants PO Box 271004 Littleton, CO 80127 jimmygodby@TelecomPowerConsultants.com

Portable and Distributed power for Telecommunications

Contents
1. INTRODUCTION ................................................................................................................... 3 2. INTEGRATED POWER ......................................................................................................... 3 A. RECTIFIERS ................................................................................................................... 4 B. BATTERIES .................................................................................................................... 4 3. VOLTAGE CONSIDERATIONS ........................................................................................ 5 4. BATTERY RESERVE .......................................................................................................... 5 5. DISTRIBUTION.................................................................................................................... 6 SECTION 2: DISTRIBUTED ARCHITECTURE POWER .......................................................... 6 A. BATTERIES ................................................................................................................... 7 B. C. D. E. F. G. H. LOAD ALGORITHMS ................................................................................................. 7 DISTRIBUTED BAY LOAD REQUIREMENTS ...................................................... 8 BAYS TO BE FED ......................................................................................................... 8 DISTRIBUTION ............................................................................................................ 9 RECTIFIERS ............................................................................................................... 10 AC CABLE ROUTING ............................................................................................... 11 GROUNDING FOR DISTRIBUTED POWER ........................................................ 12

I. CONTROLLER ............................................................................................................... 12 6. DISTRIBUTED POWER CONSIDERATIONS .............................................................. 13

1. INTRODUCTION This document discusses distributed power for telecommunications. In todays rapidly changing environment, there is an ever increasing need for a distributed power solution. With todays available technology in both DC power plants and batteries, there is a real economical ability to provide these unique solutions. The implementation of a Distributed Power Solution is a solution that grows distributed architecture power as opposed to BDFBs in a traditional central office. This session will provide insight into the engineering rational behind this solution. These will include space, capacity, and size considerations for the AC required, DC plant with distribution, and associated batteries. This presentation will discuss two aspects of each of the above solutions. The first aspect is the application. When would you use a distributed architecture bay, where should you use it, and how would you implement the solution? The second aspect will tackle design and sizing of that solution. 2. INTEGRATEDPOWER Several suppliers have small (one or two shelf switch mode rectifier plants. For the purpose of this document, the various power systems will be labeled as Micro Power Systems (MPS) for the duration of this document. The MPS is the heart of an integrated power solution for distributed architecture. The MPS provides control and rectification along with intelligent communication via the customers intranet for the distributed architecture plant. In keeping with the tradition of all true power engineers and technicians, the load is just that. It is something hanging off the end of the power plant. In these discussions, the nature of the load is only important as to the answer to these questions: What is the acceptable voltage range of all of the components of the load to be fed from this power plant? What are the acceptable parameters of the DC output (noise, ripple, etc)? What are the acceptable parameters of feedback on the AC CIRCUITS? Of course, to be a complete DC power plant, there must be some form of energy reserve. That reserve will also be discussed.

A.RECTIFIERS The heart of a DC plat is the rectifier. As a user, there are several primary considerations when looking for a rectifier. They include conformation to standard requirements (NEBS GR63 CORE from Telcordia), NEC requirements, UL listing or tested, etc. Exactly which requirements are to be considered and how they are applied are beyond the scope of this presentation and will not be discussed further. The Voltage Requirements are actually on both sides of the rectifier: First, what AC voltage levels are going to be required? (208, or 240 VAC, and phase single or 3 phase AC service.) 480 VAC is not a viable solution for this small application. On the output side, do you need 52.08 or 52.80 VDC to match the existing centralized plant, or can you use a 54 VDC solution as the equipment fed from the distributed architecture bay is isolated from the central office bulk power supply. Of course rectifier density is the real key to viability of these applications. When determining the maximum rectifier size, you must consider whether you have one or several rectifiers per shelf to be powered from the same AC feed. The two parameters here are the input circuit breaker size and the ac cable size. Done incorrectly, the input could call for a non standard AC breaker which will make your AC feed much more expensive. The second consideration is the AC cable size. When large AC feeds are required in very limited space, just running the AC cable can be a real issue. The MPS shelf is designed to eliminate the problems associated with routing and connecting the AC input cables. B. BATTERIES The first technology to be looked at is Valve Regulated Lead Acid (VRLA) batteries. This is a known item. These cells require periodic testing and physical inspection, they have a much higher energy density that traditional flooded batteries, and have a predictable failure mechanism. (They are prone to thermal runaway when subjected to high temperatures. This can result in a release of hydrogen sulfide and be hazardous to people and equipment.) The real problem with VRLA batteries is that they are not maintenance free, but maintenance proof. Placing this technology in a distributed architecture bay would result in an embedded technology that would require quarterly maintenance visits
4

and frequent battery replacement. Furthermore, due to the required maximum bay width and depth in equipment line-ups, VRLA technology would require the use of smaller capacity batteries and two bays of these batteries to provide the required capacity where one bay of lithium batteries would suffice. NICAD technology is mentioned here just to observe that the technology is not being ignored. At the same time, the technology is not being considered. Cost and disposal considerations are the primary reasons that the NICAD technology is not being discussed. 3. VOLTAGE CONSIDERATIONS The real issue here is that flooded lead acid batteries that have been traditionally used in telecommunications float service operate between 52.08 and 52.8 volts per string. Where VRLA batteries are used in telecommunications, the power plant is designed to float at around 54.4 volts. The concern is not at the minimum volts per string, but the normal float voltage of the plant. This concern is due solely to the fact that some older equipment will not work at the 54 volt level... Specifically some of the older switch technologies are upper voltage level limited at 53 volts. If all of your equipment is less than 10 years old, the higher voltage is probably not an issue. However, it is necessary to ensure that all older equipment will function at these higher voltages before any transition can be considered. 4. BATTERY RESERVE Battery reserve is mentioned here because it is a critical engineering element. Short battery reserves can be acceptable with an auto start/auto transfer stand by engine/alternator for the essential loads at the site. With an engine/alternator and 3 or 4 hours of battery reserve, we will see that it is possible to provide all of the power (rectification) needed, the distribution, and the battery reserve required in a relatively small space. It the case of distributed power, that would be all of the equipment in a single bay as opposed to a bay for the plant and an additional bay for batteries. Reality here is that distributed power is generally not going to be a viable option in a Community Dial Office (CDO) environment. Due to the small size of CDOs,
5

generally if there is an existing power plant, there are no economics in the distributed architecture technology because the cost effectiveness of this solution only comes in to play with long DC cable runs to BDFBs. The solution in a small site would be to replace the traditional power plant with a one or two bay distributed architecture plant and recover the space currently occupied by the traditional plant and its associated flooded batteries. 5. DISTRIBUTION The decision to use fuses or breakers is a customer preference. Breakers can be reset, but over time may weaken and need to be replaced. When a fuse fails it always needs to be replaced so the deterioration clock starts over every time a fuse is replaced. Fuse/breaker positions for distributed architecture will be in the 0 to 70 amp range. This will be discussed more thoroughly later. SECTION2:DISTRIBUTEDARCHITECTUREPOWER Distributed Architecture is essentially a power plant that exists in the equipment space that a traditional BDFB would. It will have all of the components of a complete power plant batteries, distribution, rectifier, and controller. The following will discuss each of these and develop a typical size for the distributed architecture power bay. This type of bay was originally proposed by Ed Silverman in the early 1980s. However, the VRLA batteries available at the time were not reliable enough nor did they have the power density required for this solution. The key developments that make this approach practical are the highly dense MPS rectifier plant, lithium batteries. The critical physical requirement for this bay is that it be no deeper than 15 inches so that the bay will fit in a standard relay rack aisle. In addition (for the American RBOC market), the bay should be 23 wide. As we develop the distributed architecture model, we need to thank Curtis Ashton from Qwest Communications who contributed the fact that the average transport equipment bay draws approximately 7.5 amps so that each A or B load to any bay
6

will be approximately 4 amps. (NOTE: This low drain is largely due to the fact that there are a lot of vacant slots in each bay. If a site has an appreciably higher drain per bay, a proportionally smaller number of bays will be fed than described below.) A. BATTERIES Since the batteries are a critical element, and Lithium batteries are required in order for this solution to be maintainable, because of rapid technology advances, it will be necessary for the reader to research lithium battery options. Because of the circuit density of the rectifier plant used, the distributed architecture bay will have approximately 2/3s of the frame space available for batteries. The capacity required will need to fit in this 2/3 of a bay in order for distributed architecture to be a viable solution. B. LOAD ALGORITHMS Earlier I stated that Curtis Ashton of Qwest provided informaiton that the average load on a typical toll bay in a RBOC central office was 7 amps. This average was derived across an entire central office where a large number of bays with various loads are averaged. This accounts for areas where a bay or several bays may pull 1 amp or less added with bays that draw 40 amps. With a centralized plant, if the load increases in one are of the building while decreasing in another, the net result is a wash to the power plant. With distributed power plant architecture, it is essential to look at a smaller universe. It is also critical to look at growth patterns in the equipment to be served. If the load in an office moves from one area to another, then one distributed plant can be essentially deloaded while another is overloaded. By the same token, if the distributed architecture plant is designed with a 200 amp maximum capacity and say 20 bays, if 5 bays draw 40 amps each, it will be necessary to place an additional distributed architecture power plant(s). (More about this later.) The engineering questions that come out of this situation are: 1) Does bay space need to be reserved for additional distributed architecture power bays? 2) If so, how much space should be reserved for additional power?

C. DISTRIBUTED BAY LOAD REQUIREMENTS A key element to the decision of Distributed Power or not is the load that each bay will draw from the plant. This load can be measured as the Average and Worst Case drains as shown in the slide. The question that you have to ask yourself is Is the drain that I see on this bay now all that I can expect on the bay? Some specialized bays are required for certain equipment types. The office might need only one or two shelves of this equipment, so although the bay has a capacity of 20 or 30 amps the actual load may never exceed 5 amps. Other bays may start with only one or two amps of load and then eventually blossom to 40 amps. (If a distribution bay is gong to draw 60 or more amps, it would probably be more efficient to feed that bay directly from the centralized power plant provided capacity and distribution fuse/breaker space exists on the centralized plant. As the decision is made to use distributed power, careful consideration must be made to determine not only the initial load on the power plant from the bay, but also the projected load over the life of the plant. To that end, for this section we will assume that each bay draws a total of 10 amps 5 amps each for the A and B feeds. This should be a safe average in the smaller universe to be served by the distributed power plant. D. BAYS TO BE FED The goal of this technology is to provide the most economical power. To that end, it was first necessary to determine the amount of battery capacity available in 2/3 of a distributed architecture power bay. As the battery technologies being discussed are still in development, the battery capacity of approximately 200 amps is available with the lithium batteries that have been researched. The second step was to determine the amount of current drawn by each bay. Based on estimates above, for the purpose of this discussion we will use 5 amps each for the A and B loads on the equipment bays for a total of 10 amps per equipment bay. W we will use a battery capacity of 200 amps for 4 hours. From the above, we can calculate that the 200 amps will provide power to 20 bays or to the A or B side of 40 bays.

E. DISTRIBUTION Since this is a single bay solution, there is limited distribution space. In a 7 foot relay rack, when you remove the top and bottom space for iron work, there are 6 feet of space available for equipment. Reserving 2/3 of that space for batteries and 3 inches for the DC plant leaves approximately 1 feet of space for distribution The decision to use fuse or circuit breaker distribution again is a decision specific to the customer. However, the decision can be made easier now because there are panels available that will accept either. Distribution fuses and breakers are sized to protect the cable at maximum load to the bay. These fuse/breakers are generally 0 to 60 amps. There are several manufacturers that can provide the required distribution. An acceptable solution will provide a total of 44 zero to seventy amp fuse positions. NOTE: Distributed architecture provides a significant savings in DC distribution cable over a centralized power solution. There are two areas of savings. First, the need for large cables from the centralized power plant to the BDFB does not exist. Second, and providing an even greater savings, is the cable from the distributed power plant to the load. As was discussed earlier, there could be as little as a .5 V drop from the distributed plant to the load. However, this requires relatively large cable, and is not necessary. Because of the discharge characteristics of the Lithium battery and the fact that the load equipment will work down to 42 volts, it is possible to use much smaller cable (and smaller fuses) from the distributed architecture power plant to the load than would allowable from the BDFB solution. This results in a significant savings in cable with distributed architecture. (Fuse prices do not generally vary by size so there is no real dollar savings there. However, care must be taken to correctly size the fuses.) All traditional BDFBs are rear access. With this configuration, the fuse or breaker is in the front while cabling is done to the rear. There are three things that happen with traditional BDFBs. First, the DC feeder cable and all distribution cables are all wired in the back. This can result in cable congestion and routing problems with a large number of loads in a BDFB. Second, with the cable termination in the back and the fuse or breaker position in the front there is a possibility of misidentifying a position and placing the wrong size fuse or breaker resulting in a power failure to equipment when the only problem is the fuse is too small. Third, this configuration requires rear access to the bay.
9

Distributed architecture frames can be produced that are 100% front access. This allows the user to place the bay against the wall if so desired. Also, with front access, the cable is terminated directly above and on the same side as the distribution fuse eliminating the possibility of placing the wrong size fuse on a power feed. On the same lines, since there is no DC input to the distributed architecture bay, and only a maximum of 40 outputs, DC cable congestion in the bay is eliminated. F. RECTIFIERS Our original goal was to develop a distributed architecture bay. Since the discussion revolves around a single bay, the rectifiers must be chosen to provide the power and redundancy required in the space available. There are generally three rectifier sizes that can be placed in a 15 inch deep relay rack. (All of the solutions will be 208/240 VAC 120 VAC and 480 VAC solutions are not practical for this architecture.) Rectifiers for plants that will meet these requirements are in the 20 to 50 amp range. We will look at 20, 30, and 50 AMP rectifiers. One of the critical considerations is whether to go with N+1 reliability in the rectifier plant or since the two bays cross feed to go with 2N redundancy. If the configuration is going to power both the A and B side of up to 20 bays, the typical solution and one that provides the same reliability as most existing bulk DC plants, N+1 philosophy can be used. If the configuration is a two bay pairing where each distributed architecture power bay provides power to either the A or B load of 40 frames, the 2N rectifier redundancy design looks practical. This would provide added reliability of the system by feeding each of the mated bays from a different commercial AC service entrance. Using separate AC service entrances (usually available only in the largest offices) decreases the probability of a simplex AC failure causing the power system to go on batteries. The 2N solution provides an AC electrical advantage. By using a 2N configuration in a MPS power plant with two rectifier shelves, it is possible to feed each pair of rectifiers with a single AC breaker. This solution results in 4 rectifiers per shelf each shelf balance fed with two AC breakers and a bulk output to the distribution panels.

10

G. AC CABLE ROUTING The first question to ask and this is another first and last question is Is AC available? This is not a capacity question but it is a distribution question. It is the first and last thing to consider when contemplating this technology. It is an economic question. When AC fuse/breaker panels are assigned, there is an assumption that there will be no point in time when all of the circuits draw full load. Therefore typically (you can see this in your home) a 150 amp AC panel will have 300 to 500 amps of fuse or breaker positions. As there is never a time when everything in the home is on and attempting to pull full power, this is not a problem. However, when rectifiers are installed, THERE WILL BE TIMES WHEN ALL RECTIFIERS PULL 100% OF THEIR MAXIMUM DRAIN SIMULTANEOUSLY. Having said that, it is essential that the breaker panel that provides power to the rectifiers be dedicated to those rectifiers and have enough capacity to handle all of the rectifiers pulling 100% - this is not 100% of the breaker, but 100% of the AC rectifier current required at full load. So a 50 amp rectifier running at 208 VAC and 95% efficiency will pull approximately (54*50/.05)/208 = 14.4 amps. NOTE: This is with the rectifier running at 100% not 110% (which some rectifiers can be set to). A breaker panel must be available (this could be installed for the application). If two pairs of distributed architecture power bays are being installed, the panel will need 16 positions. Each position will have approximately 30 amps of drain at full load so the panel must be sized for and capable of 480 amps. Adding 20% for a safety margin means that the source feeding this panel needs to be capable of 576 Amps. (NOTE: This discussion is not based on available fuse sizes, but on projected loads.) At this point it is necessary to go back to the house service entrance. There are two questions to ask here. First, is the site being fed with enough AC to handle the additional load? The second question is Is there a breaker position on the panel with enough available capacity to handle the load? (If as an alternative to powering 4 bays from one panel the decision is made to power 1 bay from each distribution panel, then it will be necessary to have 4 feeds with a minimum of 144 amps each. If it is necessary to replace the house service entrance to install distributed power, then it may not be the best solution.

11

H. GROUNDING FOR DISTRIBUTED POWER Bonding and grounding this system is critical to safe and proper operation. Integrated Versus isolated grounding is generally beyond the scope of this paper. However, it is critical to bond and route the AC conduit correctly. The first general assumption is that distributed power will be used for toll room transport equipment as opposed to switch. To that end, distributed architecture will always power integrated equipment. If the distributed power bay is within 6 feet of any isolated equipment bays, it will be necessary to run the AC conduit past and bond it to the ground window. The same situation exists with the DC ground and return. Equipment positioning will be required to determine whether the return on the distributed architecture should be bonded to the ground window, or merely tied to the vertical riser and then back to the OPGP. I. CONTROLLER

It is the nature of these small switch mode rectifier plants that the controller and rectifiers are in integrated unit. The MPS controller must be capable of providing everything from dry contact alarms to a full featured web based interface for managing the power plant and alarms from a centralized location. The advanced features deal with the operation and maintainability of the plant. There are also several questions that must be asked. The first question Does this equipment require remote monitoring? is always yes in todays environment. For functionality it is necessary that monitoring of the distributed power component fit in with the architecture of the users monitoring system. As there are several of these available, this document will hit the highlights. The first monitoring would be a customer that required dry contact major and minor alarms. These would be routed to the site alarm panel or the remote alarm center to provide major and minor alarms for rectifier fail and fuse fail. The next configuration would be an intelligent controller that provided full power plant information to a down-stream intelligent alarm center. Because of the distributed nature of the power, it is critical that any alarm or information (load, battery, discharge state, etc.) include enough information to uniquely identify the power bay. For example, the RBOCS use a CLLI code to define a site ALBQNMMACGO where ALBQ means Albuquerque NM is the state New
12

Mexico, MA is the site location (Main) and CGO is traditionally the switch associated with the power plant. To use this scheme for distributed architecture, the CGO could be replaced with the bay location (ALBQNMM214). Of course, this would only work in a site with less than 10 floors and fewer than 100 bay locations on each floor. The next function is to get the informaiton out. Because ports in an office may be limited, one possibility is to daisy chain all of the power plants together so that there is only one physical appearance for all of the distributed plants at the site. A second solution is to provide each power plant with a unique address and have intranet access to each bay from a private network. Since the concept is for a distributed architecture plant to be truly maintenance free, the down stream alarm should uniquely identify the plug and play part that has failed or the status of the plant. This would include distribution, rectifiers, and especially batteries. 6. DISTRIBUTED POWER CONSIDERATIONS So far this presentation has pointed out the technical issues and sizing considerations to be addressed when making a choice between centralized power and distributed power. One of the first considerations is to determine if converting to distributed power would leave a lot of stranded capacity in the existing DC power plant. This would mean that the rectifiers, batteries, controller, and adequate AC is available for the proposed additional load. The decision in this situation can be simple or thorough. The simple decision is merely one of comparing the cost of feeding from the power board or placing a BDFB VS the costs associated with a distributed architecture bay. The thorough solution would additionally look at the condition of the existing DC plant. For example, if the existing plant had one or two 400 Amp Manufacturer Discontinued rectifiers in service, the customer could remove those rectifiers and replace them with the distributed power option. This would 1) get rid of obsolete equipment; and 2) provide AC breaker capacity for the distributed architecture bays. When the new bays can be fed from the main power board (for example in a small CDO where the outside room dimensions are approximately 40 by 80 feet), there is no need for a BDFB. In this situation, there is only the volt drop to the served equipment, there are no 750 MCM DC distribution cables, and the savings associated with distributed power are questionable. (NOTE: the exception to this would be where one or two bays of distributed power could be used to replace the
13

entire existing DC plant with one or two distributed architecture power bays. This would allow removal of the entire existing DC plant (including batteries) and provide prized growth space in a small CDO. An entire dimension is added when growth of the main DC plant is required. The economic comparison here is much more favorable to distributed architecture. When this is coupled with the need for a BDFB to be placed a considerable distance from the existing power plant, then the economic consideration will favor distributed architecture even more. In the end, the decision to grow distributed or centralized power could be a corporate decision based on reduced maintenance requirements or the increased overall site reliability because the catastrophic failure of one DC power plant would not take the entire site down.

14

S-ar putea să vă placă și