UPS with automatic bypass

[don_resqcapt19]

A UPS for a critical industrial control system failed and dropped the power to the control system. There was no power outage, just a failure of the internal parts of the UPS. Is there a way to prevent the loss of power to the loads when this happens?

 

[steve66]

I thought most UPS's had an static switch that acted as an internal bypass. It the inverter or any of the electronics failed, the static switch would switch and pass unconditioned power from the input to the output. Or did the static switch fail?

Of course there are UPS's you can parallel for redundancy. Some have to communicate to sync, and if they loose communications, you can still have issues. But I would assume this would be much more robust than a single UPS.

 

[ron]

I agree with Steve.

You need a UPS that has an internal static transfer switch (and a wrap around maintenance bypass). When the internals of the UPS go up in smoke, then it goes to static bypass. If you need to protect for an entire UPS failure (including the internal bypass), then consider two UPS's serving an external static transfer switch (or multiple). The static switch transfers within the ITIC CBEMA curves so the downstream equipment should never know what happened upstream, even if it was a UPS failure. Of course the static switch in that case is a single point of failure, but you will need to decide where to stop the paranoia (I never stop).

 

[Twoskinsoneman]

Funny just had this happen but the static switch didn't function, one of the SCRs on the convertor failed and was passing AC voltage onto the DC bus, All 6 inverters went died, convertor, convertor control board, to sections of capacitors banks.... all dead. Lots of that fun "electronics are burning smell". Main CB to UPS tripped, DC battery CB tripped, Static switch CB did not trip.

Our system has a redundancy though, both UPS systems feed PDUs which manage the loads.

 

[don_resqcapt19]

This UPS does not have a static bypass and is of the single conversion type. It is probably about 15 years old. Is the static bypass something that was not common 15 years ago?

 

[don_resqcapt19]

... Of course the static switch in that case is a single point of failure, but you will need to decide where to stop the paranoia (I never stop).

You only stop when the client runs out of money

 

[steve66]

This UPS does not have a static bypass and is of the single conversion type. It is probably about 15 years old. Is the static bypass something that was not common 15 years ago?

Don:

I don't have any idea if older UPS's had the static switch or not. But I assumed we were talking about a large UPS - like a floor mounted UPS. I think almost all of those have the static switch, and I think almost all of those are double conversion.


Maybe this just a small UPS that sat in a control panel somewhere??

 

[don_resqcapt19]

Don:

I don't have any idea if older UPS's had the static switch or not. But I assumed we were talking about a large UPS - like a floor mounted UPS. I think almost all of those have the static switch, and I think almost all of those are double conversion.


Maybe this just a small UPS that sat in a control panel somewhere??

This is a 18kVA Ferrups unit by Best Power Technology.

 

[ron]

I think that model was available with or without an internal bypass.

Those are still manufactured by Eaton for people that insist on a direct repalcement, but are very outdated.

http://powerquality.eaton.com/Products-services/legacy/best-ferrups-info.asp?CX=3

 

[jdsmith]

Approximately 75% of industrial UPS's installed today are solid state rather than the Ferroresonant type like the Best power Ferrups units. Much of the remaining 25% that is Ferroresonant is for nuclear plants. I don't know the nuclear world with their different specifications and such, but they are the only ones that are still using predominantly Ferroresonant units.

The two best solid state UPS's for industrial usage are Cyberex and Ametek (formerly Solidstate Controls, Inc., or SCI). Each of those manufacturers uses input and output isolation transformers, they have static switches as a standard feature, and optional wrap around maintenance bypass switches that isolate the UPS with hard contacts (a big drum switch) and therefore qualifies as a LOTO means.

In the last 3 years I've installed two 40 kVA redundant units, two 10 kVA redundant units, and I have two 225 kVA redundant units on order now. One thing to note with the redundancy is that the two units should NOT be paralleled, because during a fault condition (computer/PLC/DCS power supply shorting out, whatever) all of the other loads will see a power dip while the protective devices clear. UPS's are an extremely weak source of fault current, so this dip can be substantial. Model a UPS in Easypower/SKM/ETAP sometime and do a transient analysis to see how big the power dip is - you may be surprised.

My plant's specification for UPS system performance is that the system has to supply continuous power to one source on dual-fed loads (two power supplies on the utilization equipment) while one piece of equipment at a time is down for maintenance and one other unusual event occurs. The other unusual event might be a fault, UPS static switch failure, power dip, etc. This performance requirement forces us to have two UPS's, each with a regulating transformer bypass source.

There are a few good IEEE papers out there on UPS system design. If you'd like copies or any other specific information about manufacturers, a good industrial spec, etc. PM me.

This Saturday I will be attending a meeting for IEEE 1714 - Recommended Practice for Industrial UPS Systems. Great timing with this thread to get my brain thinking about these things!

 

[don_resqcapt19]

I think that model was available with or without an internal bypass.

Those are still manufactured by Eaton for people that insist on a direct repalcement, but are very outdated.

http://powerquality.eaton.com/Products-services/legacy/best-ferrups-info.asp?CX=3

I am not sure what they will pick for a replacement. These units have performed very well for them over the years and this was the only time that they have had an issue. I will have some input on the selection, but not much.

It would seem to me that an manual bypass would be easier with a this type as when on utility power there is no way that input and output can be out of sync like can be the case with dual conversion units. Also I believe that there is some energy savings using single conversion units as compared to dual conversion units.

 

[weressl]

Approximately 75% of industrial UPS's installed today are solid state rather than the Ferroresonant type like the Best power Ferrups units. Much of the remaining 25% that is Ferroresonant is for nuclear plants. I don't know the nuclear world with their different specifications and such, but they are the only ones that are still using predominantly Ferroresonant units.

The two best solid state UPS's for industrial usage are Cyberex and Ametek (formerly Solidstate Controls, Inc., or SCI). Each of those manufacturers uses input and output isolation transformers, they have static switches as a standard feature, and optional wrap around maintenance bypass switches that isolate the UPS with hard contacts (a big drum switch) and therefore qualifies as a LOTO means.

In the last 3 years I've installed two 40 kVA redundant units, two 10 kVA redundant units, and I have two 225 kVA redundant units on order now. One thing to note with the redundancy is that the two units should NOT be paralleled, because during a fault condition (computer/PLC/DCS power supply shorting out, whatever) all of the other loads will see a power dip while the protective devices clear. UPS's are an extremely weak source of fault current, so this dip can be substantial. Model a UPS in Easypower/SKM/ETAP sometime and do a transient analysis to see how big the power dip is - you may be surprised.

My plant's specification for UPS system performance is that the system has to supply continuous power to one source on dual-fed loads (two power supplies on the utilization equipment) while one piece of equipment at a time is down for maintenance and one other unusual event occurs. The other unusual event might be a fault, UPS static switch failure, power dip, etc. This performance requirement forces us to have two UPS's, each with a regulating transformer bypass source.

There are a few good IEEE papers out there on UPS system design. If you'd like copies or any other specific information about manufacturers, a good industrial spec, etc. PM me.

This Saturday I will be attending a meeting for IEEE 1714 - Recommended Practice for Industrial UPS Systems. Great timing with this thread to get my brain thinking about these things!

I'll see you there. (PCIC, right?)

In my experience batteries are the weakest part of the UPS as there are literally hundreds of cells connected in series and any one cell failure can disable the backup supply. Cellwatch with double banks seem to work.

I have NOT followed the recommendation of EATON with their parallel redundant configuration because they feed a SINGLE bus that is a common point of failure, but rather two parallel UPS's with internal static bypass and mechanical wraparound maintenance bypass that is carried to the user point through two separate distribution system. The DCS's do have two internally separate power supplies and each is supplied by a separate UPS sourced distribution system. Where the 'user' does not have a dual input, a small static switch is placed just before the input. The objective is to maintain the separation as close to the user as possible to avoid the common point of failure. High MTBF rack mount static switches, up to 40A, are available. Double up your single input devices.

Each UPS bypass input is from a different source and the input is fed from a normal and a generator supplied alternate source with an ATS. That way the battery is only sized for about 10-15 minutes and the generator starts replenishing the batteries after only a short use.Remember that the battery life is greatly affected by the number of discharges it experiences. The Cellwatch assures that the infant mortality casuality batteries can be replaced and that the dual banks allow the replacement of individual cells without jeopardizing the unavailability of the standby source. put in a 100kVA and 45kVA system like this and many-many others in simpler configuration.

In summary; carefully engineer your system and use the failure analsys over the entire distribution system that can affect the availability of your system. Don't supply unneeded, non-critical components from the UPS as it needlessly increase your failure probability without any benefit.

Manufacturers do have MTBF data, both calculated and when you press them they will give you the actual failure rates of their installed base per type unit. You need to look at the UPS failure rate alone, then how often static bypass operated to maintain the critical power.

 

[brian john]

I have worked in on and around UPS systems for 30 years all large UPS I have ever worked around has static Bypasses.

Maybe the static switch failed.

An external static dual source PDU would resolve your issue or parallel UPS?s.

It is only money.

 

[don_resqcapt19]

Brian,
I don't see anything in the manual that indicates this unit has a static bypass. The only bypass that is mentioned is the optional manual bypass to permit service work on the UPS.

 

[don_resqcapt19]

Thanks for all of the replies. I will try to make sure that the replacement units have a automatic static bypass as well as the maunal service bypass.

 

[Little Bill]

WE had a large UPS unit in my former place of employment. I think it was around 20 years old and it only had the manual bypass. Since nothing critical was running at night or weekends, this worked well enough. Alarms which drove the office workers near it to sound their own alarms. I can't remember the brand but I do remember how fun it was to swap the batteries out.:happyno: