Talk:Uninterruptible power supply/Archive 2

Archive 1

Archive 2

Archive 3

Picture

Can we find a photo of a UPS that's not filthy?

— Preceding unsigned comment added by 74.94.15.114 (talk) 20:41, 28 May 2007 (UTC)

Double online conversion Overview

I'd like to do an overview of various parts of UPS systems but am not sure how technical I should be. I don't want to modify the main page as I'd like feedback. This is a semi response to the question about fault finding of three phase UPS systems.

The main experience I have is Double online conversion UPS systems from 3kVA to 650kVA single module and afew MVA in parallel configuation. All units are static based. I work with various manufactures and products. As desgn methology varies in companies I list lists the basics parts of the UPS and then below listed some ways they are constructed.

BULIDING BLOCKS

A UPS is made from the following parts:

Rectifier

Battery

Inverter

Static switch/transfer switch

Rectifier

The rectifier converts AC to DC and is rated and the full load of the dc inverter current and the battery recharge current. This is the first part of the "double conversion" process.

Regulated

The DC voltage is kept at a constant level using thyristors (6-pulse, 12-pulse) or for example switching of transistor based components (IGBTs) to regulate the DC. This can also be done with the use of diodes/thyristors used as diodes and then the unregulated voltage converted by boosting or chopping to a regulated level. Most good specs quote input THD 5% or 10% rather than 12pulse or 12pulse with 11% & 13th. Details of current THD and total current at 50% and 100% are good to ask. (I have seen a specification for a "12pulse" UPS on a single phase input 10kVA UPS)

The regulation of the DC voltage is done for a specific reason as this is more expensive. The regulated DC may be used to charge the battery or boost the DC voltage to enable modulation of the Inverter waveform in transformerless UPS systems.

Non-regulated

This design is not common but can be seen on smaller UPS still sold as "online". The DC voltage is approximately 1.414 times the input voltage. As variations occur on the input supply this effects the DC voltage. Generally based on Diode technology but requires some form of pre-charge circuit for the DC. The DC voltage into the inverter is unregulated.

The battery is connected to the DC bus by a diode. The diode only conducts when the input voltage drops below the battery voltage. When the supply voltage drops the battery supplies current into the inverter through the diode. In this design the battery charger is separate from the rectifier. Some frequency converters without batteries also use this design. Some UPS are designed with unregulated rectifers to reduce harmonic distortion on the mains. (ie. the spec looks better)

Battery

The battery is usually connected directly to the DC bus. A circuit breaker/fuses is used for protection both at the UPS and battery. The battery current and voltage are checked to limit the maximum recharging current and voltage.

Some UPS do not have the battery permanently connected to the DC bus (some claim that this reduces life of the battery). A thyristor can be used to connect the battery when needed. The UPS therefore needs a reservoir of power for the inverter during this action.

The Battery supplies the power for the Inverter until end of autonomy. Most modern UPS systems use Sealed lead-acid. Sealed mean that during the normal expected operation the battery does not vent or require topping up. It there is a build up in pressure within the battery a valve on the cell opens to release pressure. In smaller UPS they are mounted on their side.

The battery is configured in "strings" of cells. The UPS is design for a specific number of cells in series to create a DC voltage. A 12Volt battery "Block" contains six cells of 2volts. 192cells is not uncommon on transformer based designs (UK 415Vac), or more or transformerless. Parallel strings can be connected to increase runtime and/or add redundancy. If one cell goes open circuit the string can no longer supply power. Parallel battery string does not automatically mean redundancy as it must be able to supply full rated power.

Transformerless UPS have a positive and negative string (Eg. 61+61) which act as one string with a center point as neutral.

Ah

The Ah (Ampere-Hour)is the sizing of the battery and quoted at a discharge rate. The longer the time taken to draw power from a battery the more energy can be supplied. c10 mainly for europe c20 for the UK. Eg: 100Ah at c20 can supply 5Amps for 20hours. If the same battery is supplied at C10 then it may only be regarded as a 90Ah battery. A 100Ah battery will only be able to supply 100amps for a fraction of an hour.

Charger

The battery charger keeps the batteries charging at approximately 2.25V per 2volt cell. This is different for various types of battery and manufacturer. The charging current on a sealed lead-acid should be 10-15% of the Ah.

A separate battery charger is common in smaller UPS systems. This limits the size of the battery that can be fitted and the runtime of the UPS. It is more efficient but less flexible.

An intergrated charger/rectifier has the advantage of increasing battery recharge current by reducing the load on the inverter.

Inverter

All modern static inverters are IGBT based with pulse width modulation. The IGBT is more fragile than it's predecessor the thyristor but more effient. The inverter switches at high frequency producing a switched DC waveform. Inverter capacitors and the transformer (or a choke on a transformerless) create a low pass filter removing switching noise created by modulation of the IGBTs.

Transformer

In transformer based designs the inductor part of the filtering is done by the transformer. The transformer isolates the DC and AC output. This enable use of one string to generate positive and negative halves of the AC waveform. Inverter capacitors and the transformer create a low pass filter

In transformer based designs the transformer isolates the DC from the AC output. A H-Bridge (single phase inverter)can be used to reverse the Pulse-width_modulation voltage on the input winding of the transformer. This enables use of one string to generate positive and negative halves of the AC waveform with a centre tap. The use of a transformer enables the DC input voltage to the inverter to be designed according to the battery size. (small units may only have a battery of 16, larger may have upto 50). Galvanic isolation operates through the inverter. A rectictier transformer is required for fully galvanic.

Transformerless

The transformerless design requires more batteries to create the AC waveform. The Battery must have a positive string and negative string for each half of the waveform. This is a more effient design. The footprint and weight are reduced. The low frequency transformer noise is replaced by high frequency switching noise from the IGBTs. Transformerless UPS.

Static Switch

The static switch transfers the load between the Inverter and the bypass mains. The static switch is two switches with a commom output. The first input is the inverter and the second is the mains. The Inverter maybe be thyristors or contactors and the bypass is usually thyristors. The bypass is required to be used almost instantly in case of inverter fault so solid state thyristor are required. The use of a contactor on the inverter is to increase effiency compared to a thyristor. Contactors are slightly more prone to failure however a thyristor based version may rely on cooling fans.

Manual Bypass

The manual bypass is used to bypass the UPS for servicing and repair. Most UPS have a manual bypass intergrated. Many customer with larger UPS systems also have an external bypass so that effectively the UPS could be removed without a shutdown. This may be intergrated with a key system to avoid Switching in the wrong sequence(Castell keys).

Bypass Isolation transformer

The main use is to link the neutral into the UPS to the Earth. Neutral is normally linked to earth at the substation. If the UPS neutral can be broken, for example by a generator changeover panel, then it can float from the earth reference. The UPS should not be affected by this however loads on the UPS or surge supressors can be damaged by high earth-neutral voltages. The floating of the earth neutral is mainly dependant on load on the UPS. Djlammi 23:19, 26 August 2007 (UTC)

While this used to be the dominant type of UPS, they are no longer used for common applications. Power factor correcting equipment found in newer computer systems interacts with static ferro-resonant transformers, causing potentially damaging oscillations, and the transformer itself can create distortions which yield power less acceptable than poor quality line AC.

This sounds like more copy from an APC brochure. It's also untrue.

70.228.163.254 05:27, 28 August 2007 (UTC)

I believe this regards the transformer based inverter statements, please correct me if I'm wrong.

I am aware on Issues, particlurly with 1st gen bladeservers, that the inverter must be downrated because of increased heating. A haven't had any experience with voltage distortion. Air conditioning seems to have a greater affect distorting the waveform on starting the compressors. I have seen similiar issues eg. 6&12pulse rectifier resonanting with PFC on the supply and the distortion is dramatic. What affect does the replacement of the transformer have? I have only see the magnetic component changed to an inductor in transformerless.

My company sells transformerless to 100kVA and most companies sell them now. Transformerless generally also has a reduced footprint compared to the transformer based design.

I tried not to say if one technology was better than the other or push my company products. Selection of a UPS is never one size fits all.

I'm based in the UK and the only APCs i've seen here are under desk types.

Ive just Re-read your statment: ferro-resonant transformers - I was refering to use of a transformer only to isolate a H-Bridge. I did put at the top ONLINE, sorry I will change the wording to avoid confusion. Thanks for the input, do any other parts need work? more/less detail? Djlammi 22:04, 29 August 2007 (UTC)

Surge

Should there be a paragraph about power surges (eg lightning)? None of the 9 levels/types of protection refers to surges. I came here to know more about this, and the topic no even being mentioned in the page is a bit strange. I'll look somewhere else (I have read "Storms and lightnings" here, which is useful), but I just thought a word about power surges would be an improvement. — Preceding unsigned comment added by 203.144.32.165 (talk) 01:15, 3 October 2007 (UTC)

Buzz, buzz, buzz

I've been an electrical engineer for over 25 years, and this article makes my eyes glaze over -- too much jargon, buzzwords and stilted phrasing. I pity the reader who doesn't have an engineering background. Some sections need serious rewriting. —Quicksilver^{T @} 22:25, 12 October 2007 (UTC)

INDEED! 24.6.66.193 (talk) 12:42, 27 February 2008 (UTC)

Plug UPS back into itself

Perhaps add a trivia section: what happens when you plug a Uninterruptible power supply/Archive 2 back into itself? Jidanni 00:39, 13 November 2007 (UTC)

The solution is here. In a nutshell, it's not recommended since it would most likely damage the internal circuitry, as the webpage says. As for adding a trivia section, they are discouraged; maybe WP:TRIVIA can give an explanation. --Bruin_rrss23 (talk) 14:18, 13 November 2007 (UTC)

SBS

The article says that Standby Power Supply is shortened to "SBS". What does the B in SBS stand for? ZorroIII (talk) 09:52, 23 November 2007 (UTC)

UPS Designs subsections written with personal opinions

Many of the subsections under UPS Designs discuss the pros and cons, then give the writer's opinion comparing which to buy. For example, "This type of UPS typically costs more than the off-line type, but is worth the additional cost." That line seems more like an editorial than an encyclopedia article. Micaelus (talk) 05:05, 3 January 2008 (UTC)

Internal UPS not safe?

In the article, the following paragraph states that this type of internal UPS is not safe. Why not? Is there a reference, or more detailed information explaining why?

"The first type often requires extra connection wires between the internal UPS and computer's power supply. Some internal UPS of this group output high voltage (110 V - 220 V) direct current (DC) and some output nine-step table wave AC. Neither design is safe or energy efficient. As of 2006, there are only couple companies still selling this type of internal UPS in Australia, Asia and some part of Europe"

— Preceding unsigned comment added by Bseib (talk • contribs) 16:49, 28 March 2008 (UTC)

Sizing

Who can explain the dual sizing that manufacturers sell UPSs by? They offer size in Watts and in VA. Retail customers are expected to select a UPS based primarily on this VA-Watts information. Impossible. A UPS might be sized as 1000 VA and also 670 Watts. This is probably the most important fact that users come to wikipedia to find out - and its not here. —Preceding unsigned comment added by Srcap10 (talk • contribs) 13:53, 10 April 2008 (UTC)

Take a look at the volt amps article and you'll see that this is not a simple question to answer. Most regular people's eyes glaze over when the topic of power factor and apparant power are discussed. I don't see how to approach it in an easy to explain manner, summed up in a paragraph. DMahalko (talk) 09:35, 13 April 2008 (UTC)

Major rework in progress

I really don't like this article. Way over-technical, too much jargon, and refers to complex topics with no discussion or even linkage. But rather than just complain about it, I'm being bold here and trying to fix it. This is a topic I've researched heavily for my own job, so I think I know what I'm talking about. ;-)

I see a need to split off the UPS types from the application. An Outdoor UPS could be Standby/Interactive or Online, so it doesn't merit being called a type of UPS. But is is an application of one.

The same goes for the discussion of dual-redudant UPS power systems and N+1 UPSen. Probably should be a new section just for these applications. 216.56.12.98 (talk) 19:13, 16 April 2008 (UTC)

Doh, stupid Wikipedia logged me out for taking so long to do my complex edits. I am 216.56.12.98. :-(

DMahalko (talk) 19:24, 16 April 2008 (UTC)

Excellent work! —EncMstr 19:50, 16 April 2008 (UTC)

This article needs simplification for the average person who may not want to read long tedious paragraphs.

I am not an expert at runtimes for some of the more obscure UPS versions, but I'm adding info boxes with the hope that more knowledgable people will correct what I've added.

DMahalko (talk) 05:30, 18 April 2008 (UTC)

Outdoor is a location, not an application. —Preceding unsigned comment added by 216.203.6.12 (talk) 19:20, 18 November 2008 (UTC)

Modular UPS Systems

I am trying figure out some information on modular UPS systems and having trouble, and I see little about it in this article. If someone knows a bit about them, that would be great! A few things that it should probably cover are:

• Applications
• Battery differences, if any
• Cost comparison (obviously relative, as every manufacturer will be different)
• Energy efficiency
• Inherent risks and reliability issues
• Internal redundancy

If I think of anything else I will post it. --Dpaulat (talk) 20:02, 27 June 2008 (UTC)

The applications are the same for any typical load you would use for a UPS. They can be 208 or 240 input and output. The battery packs are usually smaller batteries in sets of 10 in series to give 120VDC for the pack, a typical battery is a Yuasa NP7-12. You could have as few as one or as many as you have space for. I've seen units with 24 battery packs.

The internal redundancy is dependant on how many power packs are put in, typically 4kva packs. Generally you can't put more than 4 in one UPS. So internal redundancy will depend on your load. They also have a logic chassis and a redunant one, so you can suffer a logic failure as well. —Preceding unsigned comment added by 216.203.6.12 (talk) 19:25, 18 November 2008 (UTC)

Limitations

This section was recently removed by Thumperward (talk) reasons being the material was unsourced and "speculative." Indeed, it is unsourced, hence there is the tag asking to cite references on the page. However, I disagree that the entire section is speculative, and I do not believe it should be completely removed. Please provide further comment here. --Dpaulat (talk) 12:43, 25 July 2008 (UTC)

I didn't say it was "speculative", I said it was "prescriptive". Wikipedia is not meant to advise or teach. The material which was removed does not merely need "referenced", as if the source were merely omitted by the author; it is evidently personal commentary by the author, and is inappropriate. Wikipedia is not a publisher of original material, and the material removed falls under that category. I'm not opposed to parts of it making its way back in as references are found (it's all there in the page history), but for now the material is unsuitable and does not benefit the article by its inclusion. Chris Cunningham (not at work) - talk 14:17, 25 July 2008 (UTC)

Thanks for the clarification. I redid the edit. I am still learning everything about Wikipedia, but being bold with my edits per the instructions :) --Dpaulat (talk) 15:08, 25 July 2008 (UTC)

Sure thing. WP:BRD is a perfectly appropriate method for collaborative editing here. Chris Cunningham (not at work) - talk 15:20, 25 July 2008 (UTC)

As the author of nearly all of that section, the material can be sourced but it is a major pain to do so since it is highly technical and tends to be buried since UPS companies try to gloss over these details for the general consumer.

Generally I find that sourcing of every single sentence or paragraph in an article is really only needed when other editors think their knowledge is different or better, and since I had butted heads with dear Thumperward at almost exactly the same time as his excising of that section I don't think his removal of that large section was entirely neutral. But if you want cites for all that, I will work on providing them.

I do not see where the removed text was attempting to provide instruction. Lead acid batteries typically fail every 3-5 years and must be replaced and recycled. Large UPS systems require the use of a PDU to handle large currents. Surge strips can damage UPS protection systems. UPS battery capacity estimation must be occasionally recalibrated or it becomes inaccurate. These are all statements of fact, not instruction.

DMahalko (talk) 05:01, 28 October 2008 (UTC)

This entire article is virtually citation-free, and that's not a good thing. I haven't looked at the removed Limitations section, but removing one uncited section when the entire thing is uncited makes little sense. There's a lot in this article that I feel is dubious or salesy. About the only sources are the ones marked "Part of this article was based on a public-domain entry:..." DMahalko: Please work on adding cites to the article as you offered.

Possible Additions

The statement that smaller UPS's can rarely be combined to effectively make one larger one due to synchronization issues is no longer true. Many modern UPS offer this facility.

A small section explaining the correlation between the number of batteries and the back-up time available would be nice and maybe the maintenance issues of the UPS battery bank.

A section explaining the modern advanced inverter/charger that can be utilised as a UPS. 'Victron' inverters for example can use a combination of multiple power inputs such as wind turbines, solar panels etc, not just grid power. They can be connected together in series to give a higher VA output or in parallel to give a 3-phase output from a single phase grid input and also give seamless back-up when grid power fails which is the definition of a UPS. If this sounds like an advert, it isn't meant to be, I think Xantrex and a couple of other manufacturers supply a similar intelligent and scalable inverter. I think the relevant thing here is how the line between the 'UPS' and the 'inverter' is becoming increasingly blurred with the more modern offerings available and that alternative power sources can be utilised as an integral part of seamless back-up power systems.

Finally a section on how to correctly size a UPS for a computer system would be helpful for many.

By the way in answer to a previous post above from 'User Srcap10', Kilowatts and VA are the same for a single phase supply (110V USA, 220V Europe). If the supply is three phase then KVA equals KW divided by 0.85. KVA is a unit that is preferred by engineers because it gives consideration to the power factor of the load being driven. Veryunstable (talk) 22:03, 26 August 2008 (UTC)

Shall I edit this article as suggested here or is someone already busy with a major re-write. I know they say 'be bold' but I don't want to cause another ruckus please advise... Veryunstable (talk) 22:15, 26 August 2008 (UTC)

How-to guide

The following is inappropriate for Wikipedia, at least in its current form.

When pairing an UPS with a generator, make sure it is an industrial model, i.e. a UPS that does NOT have a non-power-factor corrected (PFC) "front-end". PFC is important for home, office and building settings but a standby generator's power is too dirty, and UPS Systems from all popular suppliers all have PFC. In this case an industrial UPS is one that has a straight-rectifier front-end, so the quality of the power, in terms of shifting frequency, brush-inducted noise, etc, really doesn't matter. For generator applications above ~20kW, any on-line UPS should work but for small construction-grade gen-sets, UPS suppliers will not recommend using their product in this fashion. In the case Integrated systems that have UPS and standby generator components are often referred to as emergency power systems.

There are three distinct UPS types:

• Off-line: aka, stand-by power system (SPS) or "back-ups", remains idle until a power failure occurs, and then switches from utility power to its own power source, almost instantaneously. Better than no protection at all, fine for PC's, also acts as a surge protector, depending on the model. The popular suppliers, like APC and Tripplite, offer good products. Be wary of very cheap devices.
• Line-interactive. This is an off-line UPS, or SPS, with a power conditioning built in via an automatic-voltage-regulation (AVR) transformer. This is a popular design and does offer some power conditioning but after years of use, most users are not happy with frequent battery changes, which is the downside: the batteries are over-cycled due to excessive battery drain every time the power is boosted or trimmed. Also problematic, these devices can over-charge dead batteries, leading to swelling of batteries, or worse (gasses), if they are not replaced on time, rendering the entire UPS an real hassle. Regarding the switch over time with both this design and the off-line - it is typically between 5-25 milliseconds, or not even .5 seconds. This is fine for servers, PCs and other devices that have a power supply. Cable modems, routers and other gear may be affected by this switchover but this doesn't happen in a typical home/network / datacenter setting. Overall better than a straight off-line UPS? This is how they are marketed but it isn't clear if the AVR feature robs the battery of its life, which adds up to hundreds of dollars over the lifetime of an UPS and lead-acid batteries are not "green".
• On-line: continuously regenerates a new, perfect sinewave output to the protected load via it's internal inverter. With this design, the incoming AC power is rectified, or changed, to DC power. It a properly designed on-line UPS, there is a DC-DC conversion stage which is always powering the inverter (DC to AC). The poorly designed on-line UPS will have a straight AC->DC->AC, which costs less to make. The typical misconception with the on-line topology is that the batteries are always powering the load, so they wear out faster. Just the opposite is true, if the UPS is designed properly. Finally, an on-line UPS with galvanic isolation is an additional feature which protects and "isolates" the load by eliminating problems due to a poor grounding. It also provides protection against all common power problems, and for this reason it is also known as a power line regulator. Power line conditioners are usually the AVR type, so a clear definition is regulation.

Jpatokal (talk) 07:48, 16 November 2008 (UTC)

Performed a quick clean up.

Several sections including "Common Power Problems" and "Technologies" contained either poor English (which is something we should always be striving for in the English version, just as I'm sure non-English speakers are striving to keep their wiki clean), and some references to sourceless issues in Pakistan and otherwise. I'd say the whole article needs a cleanup as this reads more like an Engineering wiki (and a poor one at that) than an encyclopedia article. Scryer_360 (talk) 01:09, 10 May 2009 (UTC)

Rotary section

I made some minor changes to the "rotary" section, as it seems a bit biased against it. Flywheel-based UPS's seem to be the up and coming tech for large datacenters. I'd like to rewrite it, but I dont really have the time. Pentadyne makes the "coolest" ones, carbon fibre spindles using maglev for low friction, significantly higher up-front cost than batteries, but much better efficiency and great density for short-duration (waiting for genset to spool up) applications. Active power is their main competition, and both have made lots of big-name sales to the likes of nasa and DoD, and of course lots of large datacenters. Methodical (talk) 05:43, 27 May 2009 (UTC)

it should be clarified, that a flywheel doesn't make a UPS rotary. rotary UPS inherently use synchronous generators for voltage generation. —Preceding unsigned comment added by 80.128.116.133 (talk) 13:32, 7 December 2009 (UTC)

Powerchute

The intent of that internal link was not advertising. It was germane to the "Communication" section: a simplified version of the program is the basis for Windows XP's UPS service. The connection wasn't visible, but that section needs work anyway. ElPeste (talk) 07:32, 24 December 2009 (UTC)

Multiple redundancy

Is about more than just a power supply. For example, a chandelier hanging by one connection to the ceiling does NOT have multiple redundancy. Often the term is used in architecture and construction, to denote efforts to prevent building collapses in the case of fire.Ryoung122 20:09, 22 April 2010 (UTC)

Line-Interactive UPS Diagram(2).PNG

I replaced the old diagram that showed the autotransformer incorrectly. Cuddlyable3 (talk) 13:21, 5 May 2011 (UTC)

Battery charge voltage?

The description of the line-interactive UPS says:

The main 50/60 Hz transformer used to convert between line voltage and battery voltage needs to provide two slightly different turns ratios: one to convert the battery output voltage (typically a multiple of 12 V) to line voltage, and a second one to convert the line voltage to a slightly higher battery charging voltage (such as a multiple of 14 V).

• No reason for the voltage change is given. I think this is to change from full charging to trickle charging.
• The transformer provides AC that must be rectified to DC for the battery
• The voltages stated are approximate. In both cases the current is deliberately limited.

These notes about battery charging are not exclusive to the line-interactive type of UPS. I propose that we edit the article accordingly. Comments? Cuddlyable3 (talk) 15:22, 8 May 2011 (UTC)

I see an inconsistency here (my underlining):

• To gain the buck/boost feature, all that is required is two separate switches so that the AC input can be connected to one of the two primary taps...
• ...while "boosting" an undervoltage, the transformer output is too low to charge the batteries.

If boost is done at the transformer input taps the output voltage for battery charging is also boosted. (The possibility of switching between output taps is not mentioned.) Cuddlyable3 (talk) 12:43, 3 June 2011 (UTC)

Air-DRUPS Compressed (Air Diesel Rotary UPS)

The first commercially available compressed air batteries were developed and launched 2009 as alternatives to traditional batteries, for industrial applications including standby and Uninterruptible Power Supply (UPS). In 2008/9 these batteries were adopted by companies including National Grid (UK), Telecom Italia (Italy), Eskom (South Africa), ATK (USA) and Harris (USA) Applications include an installation at ATK, which uses surplus compressed air to generate electricity and feed it into the site’s power grid.

There are also available standardised plug-and-play compressed air batteries, in capacities of 3, 5, 10, 20, 100 and 200 kW, and containerised 1MVA Compressed Air Diesel Rotary UPS Systems Air-DRUPS, As these systems have no moving parts in standby mode, it offers greater energy efficiency which increases reliability and reduces standby losses leading to an overall higher efficiency system and lower carbon emissions . — Preceding unsigned comment added by Goodwinaj (talk • contribs) 10:01, 3 June 2011 (UTC)

Comment Andrew Goodwin has a WP:COI with respect to Air DRUPS. See article stating he is a product manager for Pnu Power. Glrx (talk) 15:24, 3 June 2011 (UTC)