Here is an interesting topic to discuss about the bulk computer workstations usage in business/commercial/Institutional environments, (where 30, 40,....100 desktops are networked, and used everyday) and their power demand calculation. My friends and peers are telling that it is enough if we calculate the demand load per NEC's receptacle load factor calculation method. I am perplexed; comparing the actual, potential and continuous load demand, those computer systems would cause on the building power system.To my understanding /observation , in no single office/commercial environment, the computers are switched OFF until close of the day, once started in the morning. I mean to say, they are mostly continuous load on building power system, and how is it justified to calculate the computers load demand factor on par with the regular receptacle load demand, just because they are plugged into receptacles outlets ? Most of us know that a desktop system consumes somewhere between 350W to 700W including monitor power. Guess 50 or 60 systems @ an average 500W per unit working 6 to 8 hours per day! Don't we need a code amendment?
Computer workstation load and demand load computing
- Thread starter vkvgpal
- Start date
- Status
jwelectric
Senior Member
- Location
- North Carolina
Re: Computer workstation load and demand load computing
Why?
Are these not pluged into a duplex?
Was the duplex calculated in the load?
Why?
Why?
Are these not pluged into a duplex?
Was the duplex calculated in the load?
Why?
petersonra
Senior Member
- Location
- Northern illinois
- Occupation
- engineer
Re: Computer workstation load and demand load computing
most desk top computers do not use anywhere near the amount of power their power supplies are rated for. and many companies are now using lcd screens which are relatively low power.
on top of that most computers now have pwoer saving features included in them.
my guess is that realistically you are looking at more like 200W per workstation.
most desk top computers do not use anywhere near the amount of power their power supplies are rated for. and many companies are now using lcd screens which are relatively low power.
on top of that most computers now have pwoer saving features included in them.
my guess is that realistically you are looking at more like 200W per workstation.
wizzracing
Member
Re: Computer workstation load and demand load computing
The Power supplies on Computers are inverted and can be as high as 180volts at the rails.
This is the reason for a IG being installed with a normal ground. I see to many just tie them in with the normal ground wires which defeats the purpose. It should be run to a ground rod seprate from the service grounding lug.
The Power supplies on Computers are inverted and can be as high as 180volts at the rails.
This is the reason for a IG being installed with a normal ground. I see to many just tie them in with the normal ground wires which defeats the purpose. It should be run to a ground rod seprate from the service grounding lug.
petersonra
Senior Member
- Location
- Northern illinois
- Occupation
- engineer
Re: Computer workstation load and demand load computing
You seem to be avocating a position that creates a potential hazard. And a very deadly one at that.
[ August 04, 2005, 08:20 AM: Message edited by: petersonra ]
Really? And just how would the OCPD manage to function in such a situation as you propose were a ground fault to occur?
You seem to be avocating a position that creates a potential hazard. And a very deadly one at that.
[ August 04, 2005, 08:20 AM: Message edited by: petersonra ]
wizzracing
Member
Re: Computer workstation load and demand load computing
QUOTE]Really? And just how would the OCPD manage to function in such a situation as you propose were a ground fault to occur?
You seem to be avocating a position that creates a potential hazard. And a very deadly one at that. [/QB][/QUOTE]
The extra IG is installed to stop just a situation it's used for ACP in case the power supply fails for what ever reason and returns a full 180volts back threw the neutral.
The AC is being converted to DC but can return 180volts back too the neutral or ground. Computer PS are not all the same hence the highier prices for good ones. Since most PC's are shipped with cheap ones there is a highier chance of them failing.
This is the reason for the extra IG being installed. It needs to be run to seperate grounding lugs, Then to it's own ground rod.
I seen this demonstrated and the voltage spike was 170volts on it. I know the amps are low but 1amp is all it takes to get hurt.
QUOTE]Really? And just how would the OCPD manage to function in such a situation as you propose were a ground fault to occur?
You seem to be avocating a position that creates a potential hazard. And a very deadly one at that. [/QB][/QUOTE]
The extra IG is installed to stop just a situation it's used for ACP in case the power supply fails for what ever reason and returns a full 180volts back threw the neutral.
The AC is being converted to DC but can return 180volts back too the neutral or ground. Computer PS are not all the same hence the highier prices for good ones. Since most PC's are shipped with cheap ones there is a highier chance of them failing.
This is the reason for the extra IG being installed. It needs to be run to seperate grounding lugs, Then to it's own ground rod.
I seen this demonstrated and the voltage spike was 170volts on it. I know the amps are low but 1amp is all it takes to get hurt.
petersonra
Senior Member
- Location
- Northern illinois
- Occupation
- engineer
Re: Computer workstation load and demand load computing
I don't see how the IG does anything to protect anything at all in the case you are suggesting.
IGs are not run to an isolated ground rod, but back to the point at which the main bonding jumper is connected at the service point.
I don't see how the IG does anything to protect anything at all in the case you are suggesting.
IGs are not run to an isolated ground rod, but back to the point at which the main bonding jumper is connected at the service point.
- Location
- Lockport, IL
- Occupation
- Semi-Retired Electrical Engineer
Re: Computer workstation load and demand load computing
No. It's already covered in 90.1(B). The NEC gives us the minimum for safe operation. It does not promise that the system will be efficient or that the owner will like the results. If you perform a calculation using the 180 VA per duplex, and don't do any thinking beyond that calculation, you might end up with an unhappy owner. But that is not the NEC's fault.
- Location
- Lockport, IL
- Occupation
- Semi-Retired Electrical Engineer
Re: Computer workstation load and demand load computing
The only purpose of the Equipment Grounding Conductor (EGC) is related to a short circuit fault from an energized conductor to the external metal parts of a component. It will provide a low impedance path for ground fault current to return to the source, so that the breaker will trip and terminate the ground fault event.
The Isolated Ground (IG) wire would also perform that function during the same ground fault event. But there is no need for it to perform that function. Rather, it's only purpose is to provide a reference for the electronic circuits to see a "zero potential" point that is not associated with any other EGC or any other ground bar. It is run from the load back to the main panel, via any intermediate panels, without being connected to any other conductor. But it is, as petersonra has said, connected at the point within the main panel that the Neutral and Ground Buses are bonded to each other.
The only purpose of the Equipment Grounding Conductor (EGC) is related to a short circuit fault from an energized conductor to the external metal parts of a component. It will provide a low impedance path for ground fault current to return to the source, so that the breaker will trip and terminate the ground fault event.
The Isolated Ground (IG) wire would also perform that function during the same ground fault event. But there is no need for it to perform that function. Rather, it's only purpose is to provide a reference for the electronic circuits to see a "zero potential" point that is not associated with any other EGC or any other ground bar. It is run from the load back to the main panel, via any intermediate panels, without being connected to any other conductor. But it is, as petersonra has said, connected at the point within the main panel that the Neutral and Ground Buses are bonded to each other.
Re: Computer workstation load and demand load computing
Of course Charlie. This is about safety issue too. I will explain with a simple example. Suppose a design business office having 100 workstation computers, out of which at least 70 will be serving practically continuous in a typical weekday. Even if we take the minimum power of 250VA per w.stn., as suggested by Petersonra, the total load is 17,500VA. This needs a 65A feeder #6 to safely feed the load at 208volts 3phase power. (17.500VA/208x1.732 x 125%=60A). If the same 17,500VA actual, continuous demand load is calculated per NEC Receptacle demand calculation basis, you will end up with 13,750VA only.(10000VA @100%+balance@50%). That means a #8 feeder would be enough (13,750VA/208x1.732x125%=47.5A),and practically that #8 feeder will be feeding 60A load, whatever time it persists. Is not a safety issue? An average owner may be happy when you show them low installed cost,but he may not be definitely happy, when it later turns out to be either a change order or a premature failures. Will they? And then, I have still my own doubts about the power demand by individual workstations. Normally design offices, especially, graphic design offices use high end systems that consumes more power than 200W.
Regarding IG power my gusess is that IG is used more to reduce electrical noises, that affects the electronic parts and signals. Of course that protects the specified equipment through its reference ground, but not separated at source.
Pal
Of course Charlie. This is about safety issue too. I will explain with a simple example. Suppose a design business office having 100 workstation computers, out of which at least 70 will be serving practically continuous in a typical weekday. Even if we take the minimum power of 250VA per w.stn., as suggested by Petersonra, the total load is 17,500VA. This needs a 65A feeder #6 to safely feed the load at 208volts 3phase power. (17.500VA/208x1.732 x 125%=60A). If the same 17,500VA actual, continuous demand load is calculated per NEC Receptacle demand calculation basis, you will end up with 13,750VA only.(10000VA @100%+balance@50%). That means a #8 feeder would be enough (13,750VA/208x1.732x125%=47.5A),and practically that #8 feeder will be feeding 60A load, whatever time it persists. Is not a safety issue? An average owner may be happy when you show them low installed cost,but he may not be definitely happy, when it later turns out to be either a change order or a premature failures. Will they? And then, I have still my own doubts about the power demand by individual workstations. Normally design offices, especially, graphic design offices use high end systems that consumes more power than 200W.
Regarding IG power my gusess is that IG is used more to reduce electrical noises, that affects the electronic parts and signals. Of course that protects the specified equipment through its reference ground, but not separated at source.
Pal
- Location
- Wisconsin
- Occupation
- PE (Retired) - Power Systems
Re: Computer workstation load and demand load computing
I don't usually determine circuit sizes but why is there any issue if you know that the circuits are not for general use.
If you know the actual loading then 210.11(A) says to provide enough circuits.
220.14(A) says size each circuit based on the known specific loads served.
The rest of 220.14 addresses methods for "guessing" at loading of general use receptacles.
I don't usually determine circuit sizes but why is there any issue if you know that the circuits are not for general use.
If you know the actual loading then 210.11(A) says to provide enough circuits.
220.14(A) says size each circuit based on the known specific loads served.
The rest of 220.14 addresses methods for "guessing" at loading of general use receptacles.
jimwalker
Senior Member
- Location
- TAMPA FLORIDA
Re: Computer workstation load and demand load computing
I would put no more than 4 work stations on a line.Average load will be less than 300 per station.Yes thats over kill.Now stop and think what happens if that breaker trips,everyone on that circuit just lost all there work unless they have a ups.I am sure they would rather just pay you once than pay every time it trips.Are you after a happy customer or just saving a few bucks ? You could maybe get as many as 10 to run.Can you risk it.The size of the power supply is only the available watts not what it normally pulls.Also nothing says they will not have a few other loads like desk lamp,radio,paper shreder.
I would put no more than 4 work stations on a line.Average load will be less than 300 per station.Yes thats over kill.Now stop and think what happens if that breaker trips,everyone on that circuit just lost all there work unless they have a ups.I am sure they would rather just pay you once than pay every time it trips.Are you after a happy customer or just saving a few bucks ? You could maybe get as many as 10 to run.Can you risk it.The size of the power supply is only the available watts not what it normally pulls.Also nothing says they will not have a few other loads like desk lamp,radio,paper shreder.
Re: Computer workstation load and demand load computing
Jim
The issue is not in designing the known load (computer w.stns=200w,or 300w, or 400w) for the receptacles branch circuits, but preciously on computing the total demand load for the panelboards feeding those receptacles load. Refer 220-13(2002) or 220-44(2005), and guess if that demand calculation and designing of the feeder size will work for a totally computer oriented offices where hundreds of computers (example: system furniture), plugged in to receptacles, works 8AM -5PM, every day if not more.
PAL
Jim
The issue is not in designing the known load (computer w.stns=200w,or 300w, or 400w) for the receptacles branch circuits, but preciously on computing the total demand load for the panelboards feeding those receptacles load. Refer 220-13(2002) or 220-44(2005), and guess if that demand calculation and designing of the feeder size will work for a totally computer oriented offices where hundreds of computers (example: system furniture), plugged in to receptacles, works 8AM -5PM, every day if not more.
PAL
- Location
- Lockport, IL
- Occupation
- Semi-Retired Electrical Engineer
Re: Computer workstation load and demand load computing
I would start the calculation in accordance with 220. Then I would add whatever VA is needed to account for the fact that there are more computer loads than 220 would normally take into account. You can always add load to a 220 calc. You can't give the owner less load than a 220 calc says to give.
I would start the calculation in accordance with 220. Then I would add whatever VA is needed to account for the fact that there are more computer loads than 220 would normally take into account. You can always add load to a 220 calc. You can't give the owner less load than a 220 calc says to give.
jwelectric
Senior Member
- Location
- North Carolina
Re: Computer workstation load and demand load computing
vkvgpal
I see by your profile that you are a consultant. This is good I suppose. A consultant is a person that you go to and ask questions, right? Well that makes all of us consultants due to the fact that you have asked us a question.
The way that I am reading your first post is that you have a problem with a duplex being calculated at 180va each and plugging in a 700 watt computer.
Each and every one of those receptacles art rated at 15 amps each and will not be over loaded.
Then in a later post you come in with 100 computers and do the feeder load for all at one time. Are you plugging a computer in every duplex you see in the office space?
A twenty amp circuit will calculate at 2400va and will handle up to 3 of your 700 watt computers with power left over.
If you are planning on putting a 100 computers in one spot and don?t tell the electrician until after the job is over then shame on you.
Tell him before the job starts then he will address the problem at that time and there will be no problem when these 100 computers are turned on. You can?t just rent an office space and expect it to be ready for 100 computers.
As to your IG going to a rod by itself. A good electrical consultant should know Ohm?s law, E=IxR. 120 volts and 25 ohms of resistance would be 4.8 amps and not enough to open a 20 amp circuit. I think this ground rod should be bonded to the service in order to have a low resistance so the breaker will open.
Now where do I send the bill?
vkvgpal
I see by your profile that you are a consultant. This is good I suppose. A consultant is a person that you go to and ask questions, right? Well that makes all of us consultants due to the fact that you have asked us a question.
The way that I am reading your first post is that you have a problem with a duplex being calculated at 180va each and plugging in a 700 watt computer.
Each and every one of those receptacles art rated at 15 amps each and will not be over loaded.
Then in a later post you come in with 100 computers and do the feeder load for all at one time. Are you plugging a computer in every duplex you see in the office space?
A twenty amp circuit will calculate at 2400va and will handle up to 3 of your 700 watt computers with power left over.
If you are planning on putting a 100 computers in one spot and don?t tell the electrician until after the job is over then shame on you.
Tell him before the job starts then he will address the problem at that time and there will be no problem when these 100 computers are turned on. You can?t just rent an office space and expect it to be ready for 100 computers.
As to your IG going to a rod by itself. A good electrical consultant should know Ohm?s law, E=IxR. 120 volts and 25 ohms of resistance would be 4.8 amps and not enough to open a 20 amp circuit. I think this ground rod should be bonded to the service in order to have a low resistance so the breaker will open.
Now where do I send the bill?
- Location
- Massachusetts
Re: Computer workstation load and demand load computing
wizzracing welcome to the forum we have a good time here learning new things.
Never run the IG to an isolated ground rod.
It is dangerous, a code violation and pointless.
As petersonra pointed out the IG conductor should be connected to the electrical system at the point of the main bonding jumper.
This could be at the buildings service disconnect or a transformer or disconnect supplying a separately derived system.
This first section pretty much spells out how IGs are to be connected.
2002 NEC
2002 NEC
Bob
wizzracing welcome to the forum we have a good time here learning new things.
NO! absolutely not!
Never run the IG to an isolated ground rod.
It is dangerous, a code violation and pointless.
As petersonra pointed out the IG conductor should be connected to the electrical system at the point of the main bonding jumper.
This could be at the buildings service disconnect or a transformer or disconnect supplying a separately derived system.
This first section pretty much spells out how IGs are to be connected.
2002 NEC
Then there is this general requirement.
2002 NEC
wizzracing I can not urge you enough to rethink your thoughts on IG practices.
Bob
Re: Computer workstation load and demand load computing
Jwelec
I am sorry to note that, probably, you did't grasp my first posting as well as the subsequent. All my concern were how you guys (I mean the professionals with total awareness of NEC)arrive at a reasonable load demand factor against 220-13/220-44, in a kind of situation, to decide on the feeder size. [Also, I suppose you know that you can't load a twenty amps outlet to its headswinging 20A rating(=2400VA)]. Thanks. PAL
Jwelec
I am sorry to note that, probably, you did't grasp my first posting as well as the subsequent. All my concern were how you guys (I mean the professionals with total awareness of NEC)arrive at a reasonable load demand factor against 220-13/220-44, in a kind of situation, to decide on the feeder size. [Also, I suppose you know that you can't load a twenty amps outlet to its headswinging 20A rating(=2400VA)]. Thanks. PAL
- Location
- Massachusetts
Re: Computer workstation load and demand load computing
Actually you can.
You can't?
Actually you can.
jimwalker
Senior Member
- Location
- TAMPA FLORIDA
Re: Computer workstation load and demand load computing
For how long ?
For how long ?
- Location
- Massachusetts
Re: Computer workstation load and demand load computing
179 minutes, then off, then back on for 179 minutes etc.
179 minutes, then off, then back on for 179 minutes etc.
- Status