Results 1 to 7 of 7

Thread: UPS ELI and PSC

  1. #1
    Full Member
    Join Date
    Oct 2012
    Location
    Durban
    Posts
    66
    Thanks
    0
    Thanked 5 Times in 4 Posts

    UPS ELI and PSC

    I am getting a lot of installations where UPS's are the supply points. All of the UPS's so far do not have any fault levels recorded on their rating plate and I would imagine that this is because the fault level would be determined by the amount of batteries installed in any particular UPS which would vary.

    My questions are:
    1. How is the fault level determined. (assuming the manufacturer cannot answer that question).
    2. The internal configurations are not known so can an ELI test be done on the load side of the UPS and from there work out the fault level through I=V/R.

    One further question not related to UPS's.

    If a fault level for a particular circuit is close or equal to the circuit breaker's KA rating how does installing fuses upstream instead of circuit breakers help reduce the fault level (as explained to me once by an engineer). Perhaps someone with a lot of theory knowledge could explain this as the fault level can be calculated using transformer and cable impedance etc which has no bearing on reducing fault levels with fuses.

  2. #2
    Diamond Member AndyD's Avatar
    Join Date
    Jan 2010
    Location
    Cape Town
    Posts
    4,923
    Thanks
    576
    Thanked 934 Times in 755 Posts
    Depends on the type of UPS and the mode it's in at the time of the fault. If it's in battery mode the fault level would be determined by the components in the output stage such as the IGBT's, not by the batteries themselves. Manufacturers must be able to supply this data, I don't know how you could determine the fault level without either obtaining manufacturers figures or destructive testing. If the UPS has a circuit breaker built into it on the output you could assume that the kA rating of that breaker would be above the fault current the UPS can deliver....but assuming can be dangerous.

    Correctly coordinated fuses are far preferrable to circuit breakers in high fault current applications because the speed that the fuse disconnects under fault conditions is far faster than any mechanical system such as an MCB, MCCB or ACB can achieve. By disconnecting fast enough to prevent the full potential fault current from flowing you effictively limit the fault current to a lower value, this is known as current limiting. By having fuses upstream you can size the kA rating of your downstream circuit breakers according to the fault current the upstream fuse will allow rather than the PFC of the network itself.
    _______________________________________________

    _______________________________________________

  3. #3
    Platinum Member
    Join Date
    Jun 2017
    Location
    Port Elizabeth
    Posts
    1,236
    Thanks
    34
    Thanked 304 Times in 258 Posts
    Hi

    The rule of thumb with fuses has always been anything under 250Amp reduces the fault level to 5Ka

    There are different ratings on fuses and you need to be careful when selecting the fuse - A "aM fuse"will give a different result to a "gG fuse" - Best is to check the rating charts of the fuse you are wanting to select similar to checking the ratings on C/Breakers

    My understanding of Fault currents -
    The max fault current will always happen at the highest point of the waveform/cycle - The closer you are to the source the greater the DC component of the fault that is added to the waveform of the AC cycle

    The further away you move the more capacitance in the cable nullifies the DC component and also reduces the energy of the AC fault through the impedance of the cable .
    The DC component is not normally taken into account in the calculations as it is negligible.

    To reduce the fault current you need to "interrupt" the fault while it is still rising or dropping to the zero point in the 50Hz cycle - A fuses will generally rupture after the zero point of the cycle on the way up before reaching maximum and were generally first choice to protect systems and reduce fault levels downstream so that you could install "cheaper"C/Breakers. Maybe cheaper is incorrect but rather low KA rated C/Breakers

    With the advent of electronics being introduced into c/breakers your CB's become more accurate instead of waiting for Bi metal strips etc to overheat and trip your C/breaker. Current limiting became a lot easier and the accuracy of the electronics could switch the CB off as it crosses the zero point of the wave cycle which meant almost no energy having to be dissipated within the CB - This reduced the size of the circuit breakers as well as improving the accuracy and suddenly C/Breakers have become part of Fault level calculations / determinations.

    Before current limiting CB's it was only cable that could reduce your fault levels , and fuses to a certain extent.

    With UPS , everything is electronically controlled so I would expect your output fault level to be reasonable low , but no higher than your incoming fault level.
    The battery fault level would only come into play within your UPS and rule of thumb is 12 times your Amp hour rating of the battery

    As I said at the start , the above is my understanding of fault levels and I don't proclaim to be an expert

  4. #4
    Full Member
    Join Date
    Oct 2012
    Location
    Durban
    Posts
    66
    Thanks
    0
    Thanked 5 Times in 4 Posts
    Thank you for the replies.
    The issue with UPS suppliers is that they do not include the KA rating on their rating plates. I am in the process of contacting one such manufacturer to see whether they can supply the details. I am not holding my breath.

    I understand the fuse drops the KA rating, however if a test is done and lets say once the results are obtained the value still needs to multiplied by 1.73 (SANS three phase) and this reading now becomes higher than the local circuit breaker rating how would one take into account a fuse upstream that would likely drop the recorded value. On the COC the measured value would still have to be recorded. Any thoughts on this.

  5. #5
    Diamond Member Justloadit's Avatar
    Join Date
    Nov 2010
    Location
    Johannesburg
    Posts
    3,479
    Thanks
    134
    Thanked 695 Times in 593 Posts
    Blog Entries
    1
    With the advent of power electronics, and switching speeds of FETs are in the region of 50nSec switching on and about 90nSec(Yes nano seconds) switching off. The fuse has no action if there is a fault on the load side. Long before the fuse has had a chance to react, the current limit circuitry with in the UPS will have switched the FET off. This is the reason that the UPS can run an over current, usually twice the typical standard rating for a predetermined time before shutting down as an overload. The electronics can differentiate between and overload and a short circuit.

    The reason that fuses are used, is for protection in the case that the FETs fail as a short circuit, which simply shorts out the battery, as they are usually connected in a bridge arrangement, and is there to protect the cables and to prevent a fire.
    It is utmost importance that the correct fuses are used in DC circuits. Incorrect fuses can not handle the fault condition and can plate the inside of the glass, and maintaining energy flowing because there is an electrical connection inside the glass tube due to the arcing, a very dangerous situation - think about a welding machine, the melting of the rod is when you pull the tip of the rod away from the metal, and the arc is then created.

    I can assure you that all low cost UPS's will fail, simply because the cost to improve the internal switching circuits driving the FETs would be very expensive, and the current solutions stress out the FETs every time they are switched on. Now consider that some of these circuits use 20KHz PWM(that is 20,000 times per second) to create the sine wave, and you get the idea the amount of stress that these devices take on a regular daily basis.

    This is reason that if you get to a UPS, and the fuses are blown, that the chances are that the UPS is faulty, and not because there was an overload.
    Victor - Knowledge is a blessing or a curse, your current circumstances make you decide!
    Solar pumping, Solar Geyser & Solar Security lighting solutions - www.microsolve.co.za

  6. #6
    Platinum Member
    Join Date
    Jun 2017
    Location
    Port Elizabeth
    Posts
    1,236
    Thanks
    34
    Thanked 304 Times in 258 Posts
    Hi

    The question of PSCC values, calcualtions , testing and cascading has always bothered me / confused me and made me wonder why we are involved in an industry that pays so little for the responsibilities we take upon ourselves.
    This thread has made me do a little bit more thinking then I probable should.

    A single phase system under 100Amp is pretty str forward - Stick the meter on , get a reading and log the results on the COC. ( SANS 10142 8.4.2)

    Over that it starts becoming complicated to an extent and meters should form no part of the calculation, in my view. Generally in SA you will not have a single phase supply over 100Amps which means swopping to a 3 phase PSCC tester.
    According to 8.4.2 note 1 - In a balance 3 phase system you can use a single phase PSCC tester and multiply by 1.73 - It is going to be extremely rare that a 3 phase system is going to be balanced.So i would say no single phase PSCC test on a 3 phase system unless you want to check your calculation.

    If there is Coordination/Cascading of CB's on a 3 phase system ( which is the normally due to budget constraints on all contracts) and you use a 3 phase PSCC - Is it going to be able to pick up the cascading that has been taken into account when the design was done ? I very much doubt it.
    By using current limiting CB's you can go from 150KA fault level to 6 KA fault limit with 3 CB's and not even bother about calculating the cable impendance.
    Based on the assumption that a 3 phase PSCC meter will not give me the accurate results , it is pointless using one and far better to carry out the calculations and reference to the CB manufactures charts for cascading.

    SANS 10142 8.4.3 allows us to use graphs, tables and computer programs from suppliers of the equipment to calculate/ verify the PSCC . AS long as we have the calculation and can produce the results at a later stage we are safe in the event of a problem.The CB manufacturer's are generally happy to sit down and go through the design with you to verify compliance

    Now comes the problem of a UPS being fitted into the mix - SANS 10142 8.4.6.2 - Obtain the PSCC from the supplier of rectifiers etc. - We also need to remember that we have 5.3.5 / 5.4.1 in our arsenal and at the end of the day we have to sign according to 8.5.3(3) all protective devices are capable of withstanding the PSCC
    If the supplier cannot give us the information then he cannot have his UPS connected - I know it sounds simple , but if all contractors/designers adhere to the regulation we would not have a problem as the UPS suppliers would soon learn that they need to supply the info well in advance or explain to the client why his item of equipment is not working.Same goes for generators.

    It may look like we are being obstructive but at the end of the day we have to work by the rules - Same way if you jump a red light because you late for a meeting you will still suffer the consequences , if caught , and pay the fine.

    When using the manufacturesers charts or given a design by the electrical consultant never just sign off at the end of the contract without first checking.We had a case where the design was done on a 300Amp municipal supply , with a 300Amp CB in the kiosk outstide feeding into a main DB with a 300Amp isolator .All checked and verified by the CB manufacturer for cascading.
    The supply was changed to 250Amp and so the kiosk CB was changed to a 250Amp CB now feeding a 300Amp isolator in the main DB - Would assume that everything is still in specification and OK to sign off. Started checking the charts and discovered that the design could no longer be signed off.
    An isolator cannot have a KA rating as it cannot trip under fault conditions.So do not believe the ratings that you see on isolator's and if you look at any manufactures charts it generally has a note next to , or under, the isolator's KA rating saying "to be backup by a suitable short circuit protection device"
    The cascading charts are all written up after suitable laboratory testing.
    According to the charts a 250Amp CB protecting a 300Amp isolator had not been tested and could therefore not be verified to be capable of withstanding the PSCC.We had to change the 300Amp isolator to a 250Amp isolator as we would never have been able to provide proof of compliance.

    Let's think domestic - The municipality provides the single pole CB in the kiosk outside from manufacturer A and the contractor installs an Isolator at the main DB from Manufacturer B - I say that is not allowed as it is not a tested solution by Manufacturer B and if something goes wrong you will not be able to prove compliance.

    Some food for thought.

  7. #7
    Full Member
    Join Date
    Oct 2012
    Location
    Durban
    Posts
    66
    Thanks
    0
    Thanked 5 Times in 4 Posts
    The issue of PSCC leaves a lot of open questions. Without UPS generator etc it still leaves the question of measured values as opposed to cascading etc and with UPS there is the issue of PSCC calculations.

    I notice that according to 8.4.6 the UPS PSC can be calculated using the formulas there but even getting the battery details and quantities can be challenging. Normally the UPS's that I come across are for running factory wide instrumentation complete with emergency shutdown procedures etc (double 225KVA). Now I know trying to get battery details from the clients would be a waste of time. I look at DB's fed from the UPS and know that the existing boards would have been engineered and originally signed off but there seems to be a big hole where previous COC's disappear so referencing that is not an option.

    Any comments from anyone with a similar dilemma.

Did you like this article? Share it with your favourite social network.

Did you like this article? Share it with your favourite social network.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •