design question about "circulating currents"

Status
Not open for further replies.
J

jcd

Member
i have a customer (non engineer) who really wants reliability. he is wanting me to route two feeders to his ITT room, one from one service entrance UPS and another independant feeder from a second service entrance UPS. (each UPS has a transformer.) The two feeders will feed two panelboards (one main panel, and one redundant) in the ITT server room, and the panelboards will feed racks (i.e., each circuit from one panel will be duplicated by the second redundant panel, and routed to the same rack).
i have explained the possibility of circulating currents being introduced in the racks/equipment. in my explaination i stated that feeding the same equipment from two separately derived systems would introduce circulating currents to the system.
is there somethign in the NEC that specifically supports not routing equpment from two separately derived systems? since the NEC is not a design manual, i do not think there is anything that would state this explicitely. but, the customer is not accepting "bad engineering judgement" as a good enough reason to not do this.
help.
thanks,
jcd
 
It depends on his end equipment in and application like that I think I would use an switching power supply like these http://www.pulizzi.com/ this would prevent any problems from the separately derived feeds. By the way those are just my favorites APC and a few others also make them also. Because they are listed and only allow one source at a time it prevents circulating currents. they also switch fast enough on loss of one source to prevent loss of all computer hardware I use them in highly critical environments. if you need any more info ask away.
 
We have done the same thing for one of our biggest customers
and have never heard of any problems with doing that.
I should say that both UPS's and all the racks are in the same room
and both UPS's are fed from the same Load center. That may be why.
 
Last edited:
Edit: I see your concern now. However, I think its a common design for data centers.

And two feeders are more reliable than one feeer with a transfer switch.

Steve
 
Last edited:
MJJBEE said:
It depends on his end equipment in and application like that I think I would use an switching power supply like these http://www.pulizzi.com/ this would prevent any problems from the separately derived feeds. By the way those are just my favorites APC and a few others also make them also. Because they are listed and only allow one source at a time it prevents circulating currents. they also switch fast enough on loss of one source to prevent loss of all computer hardware I use them in highly critical environments. if you need any more info ask away.
Click to expand...
thanks! so, only allowing one source supply at a time prevents the curculating currents. i thought they would be present (even with one source on, and the other off). thanks for the website. jcd.
 
There for a minute, I thought there would be a parallel path with the ground conductors and the neutral wires.

But I think I've came to reason since then. Each neutral only connects to the ground system in one place. So I don't see any circulating currents.

Each piece of equipment will draw power from only one source. And that current will flow back on the corresponding neutral.

I don't see how this is any different than having two transformers that serve separate loads in any building.

I think you might be visualizing a problem that won't be there.

Steve
 
thanks for the info.
if you plug in the same equipemnt into two sources, there will be circulating currents. you said that its the same as ... having two transformers that serve separate loads in any building. the key words are 'separate loads' -- sure no problem with them being fed from two transformers. but, in racks, with dual feeds to the same equipment, you have to be careful that both sources are not utilized at the same time -- then there will be circulating currents. with an ATS at the rack, this possibility (of plugging in or using two sources) is pretty much eliminated.
thanks again.
jcd.
 
jcd said:
thanks for the info.
if you plug in the same equipemnt into two sources, there will be circulating currents. you said that its the same as ... having two transformers that serve separate loads in any building. the key words are 'separate loads' -- sure no problem with them being fed from two transformers. but, in racks, with dual feeds to the same equipment, you have to be careful that both sources are not utilized at the same time -- then there will be circulating currents. with an ATS at the rack, this possibility (of plugging in or using two sources) is pretty much eliminated.
thanks again.
jcd.
Click to expand...

I think your equipment will take care of this for you. No need to add another ATS. The equipment in the racks will have dual cords, and dual power supplies. The current that each power supply draws from the hot wire of one source must return to that source through the neutral wire of that source.

For the most part, each piece of equipment will pick one supply, and utilize that supply. But that really doesn't matter. Even if the equipment used both sources at the same time, the currents will balance themselves out.

Steve
 
jcd said:
hmmm. i'll have to think about it some more. i appreciate your input. jcd
Click to expand...

We are talking dual corded (dual power supply) equipment?
There should be no circulating currents, because the two sources (power supplies) are not connected together on their AC side - rather on their DC side... The DC will not "backfeed' and produce AC...
 
yes, dual power cords. thanks for the clarification, i think i understand now. one, i was thinking that the circulating currents would be present (even with one source on or being utilized, and the other not) -- i guess i was wrong about that. but, also, in this particular environment, the people pluging in the racks historically have not been paying attention to which outlet or wireway they are using. for example, if there are two wireways, A and B, equipped with receptacles, one on each side of the rack. and each wireway is fed from its own UPS, A or B (these UPS's are not in the rack, they are at the service entrance of the building, each with a transformer). keeping in mind that B wireway is to be utilized ONLY if A becomes deenergized for some reason. i thought that if a person plugged a piece of equipment in the rack to an outlet in wireway A, and in the same rack, inadvertantly plugged another piece of equipment in the same rack into an outlet in wireway B, that circulating currents would be introduced into the system and equipemnt in the racks -- this is my concern. if you still say no problem, then cool, i will let it go. again, thank you for your expertise! jcd
 
jcd said:
i have a customer (non engineer) who really wants reliability. he is wanting me to route two feeders to his ITT room, one from one service entrance UPS and another independant feeder from a second service entrance UPS. (each UPS has a transformer.) The two feeders will feed two panelboards (one main panel, and one redundant) in the ITT server room, and the panelboards will feed racks (i.e., each circuit from one panel will be duplicated by the second redundant panel, and routed to the same rack).
i have explained the possibility of circulating currents being introduced in the racks/equipment. in my explaination i stated that feeding the same equipment from two separately derived systems would introduce circulating currents to the system.
is there somethign in the NEC that specifically supports not routing equpment from two separately derived systems? since the NEC is not a design manual, i do not think there is anything that would state this explicitely. but, the customer is not accepting "bad engineering judgement" as a good enough reason to not do this.
help.
thanks,
jcd
Click to expand...

There were a lot of replies and most of them contained information reflecting on no close familiarity of reliability issues and UPS design. I strongly urge you to turn this over to an engineer who has proven track record in designing such systems.

The two UPS's must be synchronized and designed for parallel operation.

It is called STATIC transfer switch not switching power supply and I would stay away from any company, regardless how nice their website looks, who advocates circuit breaker use over fuses in the case of UPS power distribution.

One of the cardinal rule of redundancy is avoiding common failure point from source to load. Ergo the most reliable system is that has separate source AND load including all components along the way. So if you have a single load that is non-redundant, placing a HIGHLY reliable static directly front of it will provide you with and high degree of reliability. Use generator for Utility reliability increase.

Cyberex is the "mother of all static switches". https://cyberex.com/product.xhtml?Product_ID=ent_DSTS
 
Status
Not open for further replies.
Top