Best way to get "Clean" power to a computer ou

[Guest]

I need to have clean power for a computer outlet. What's the cleanest way to run a computer outlet? I can home run to the main panel (meter/main combo) or to a subpanel. What are the tricks for making a clean computer circuit? 20-amps, 110-volt. Do isolated ground outlets help?

 

[dereckbc]

Re: Best way to get "Clean" power to a computer ou

Dedicated with 3.3% or less voltage drop.

 

[hurk27]

Re: Best way to get "Clean" power to a computer ou

If you are running N/M then the circuit is already isolated . if you are running it in conduit then run a ground conductor and use a isolated outlet but unless there is going to be net working or connection to a cable modem there is not realy any reason to isolate the outlet as it will not protect against surges or brown outs only a good active UPS or line conditioner will do this and give total protection.

 

[jrdsg]

Re: Best way to get "Clean" power to a computer ou

agree with hurk above, there's no substitute for an adequately-sized ups system.

take it from someone that regularly consoles clients who have lost valuable data or whose servers do not reboot properly after an unexpected brownout or blackout. it's no fun to to have to say "i told you so..."

 

[hurk27]

Re: Best way to get "Clean" power to a computer ou

Keep in mind that there is two types of UPS systems the first only supply house current untill there is a supply loss then it switch's to battery this type does not isolate from dirty power from the house circuit.
The second is the active inwhich the UPS is allways on battery power and there is only a charger to make up the loss power in the battery's
This is allways isolated from the house current and is the best way to protect your system. and there is no switching.

 

[dereckbc]

Re: Best way to get "Clean" power to a computer ou

Wayne, actually there are more than two types.

There is a standby type which you mentioned. The load operates on commercial AC until there is an outage at which point it switches to the battert inverter.

Then there is a standby interactives with a static switch. This type operates same as above except the inverter is connected to a staic transformer which performs some limited voltage regulation.

Then you get into a commercial class called double conversion UPS. These units have a rectifier-batterry-inverter. The load is always being supplied from the inverter.

From there you get into more versions of the dual conversions UPS. They have different names called dual conversion with staic bypass, redundant, and parallel redundant. These types are the large scalled units used in data centers where no outage can be tolerated.

Then there is another class called mechanical and flywheel UPS. These are the motor-generator types. Where an AC or DC motor turns a generator. The fly-wheel has an AC motor turning a flywheel that turns a generator and provides short term ride through of momemtary outages. Then there are units that have a rectifier charging batteries, then a DC motor turns a generator for long-term outages.

 

[Guest]

Re: Best way to get "Clean" power to a computer ou

Thank you for all that input. If I go NM is there any advantage to going with isolated receptacles? In that case I would keep the ground for the computer circuit isolated all the way to the subpanel or meter/main-combo panel. Is there any advantage to keeping the computer ground isolated all the way back to the point of neutral/ground bonding (in the meter/main)?

Also, what do you people know about "approximation of sine wave"? Are there any readily available UPS's that don't use Approximation of Sine Wave? AoSW can be hard on computers. Some computer manufacturers will void their warranty if they find out the system was on a AoSW UPS. That's all I know about AoSW. Can you fill in my blanks?

THANKS!

 

[dereckbc]

Re: Best way to get "Clean" power to a computer ou

NM cable from the panel that contains the N-G bond is already an IGR. Why pay for it twice?

 

[bennie]

Re: Best way to get "Clean" power to a computer ou

Are you sure the utility power is dirty?

 

[hurk27]

Re: Best way to get "Clean" power to a computer ou

Bennie
Around here it is as the IRC only requires that the voltage be maintained over a period of time not on a given instance. and when I put a chart recorder on line it shows a lot of spikes and dips and this is from the meter socket. the voltage around here can vary from 118 to as high as 132 depending on what time of day it is.

 

[ronaldrc]

Re: Best way to get "Clean" power to a computer ou

If the voltage did fluctuate from 118 to 132 would it effect a good regulated power supply of our solid states devices?

I don't think it would most solid logic circuits work from a well regulated pos. 5 volt source.

The regulator is fed a much higher voltage than this and it is regulated down to 5 volts it continuously monitors the voltage and regulates accordingly.

I know there are multiple voltage sources in a computer or a server but I don't think that would be a problem.

Now as far as spikes and the like that would be a problem.


Ronald

[ August 26, 2003, 07:00 PM: Message edited by: ronaldrc ]

 

[hurk27]

Re: Best way to get "Clean" power to a computer ou

Ron
It all depends on the rise time of the spike as weather or not the spike will be picked up as objectionable noise and weather the device can ignore it. And in the case of high spikes it would depend on the amount of control that can be placed on the voltage from a good line conditioner. 118-132 is tolerable with the right kind of conditioning but most people mainly consumers would not know where to begin when it starts to cause problems with electronics. and many don't even know that there is a problem. they just take the computer back to the store and get a new one. I personally have had where one computer would run without any problems and a different one would freeze all the time on the same circuit. After doing a good diagnostic check of the computer. I had to suspect the power quality. in which I charted the circuit and found that there was a high level of noise on the circuit. and it was traced back to a bad contact in a breaker that was arcing just enough to produce the noise but the other computer handled it fine. and the problem would not even been found except for this computer crashing. What was good was I was able to see if the triplet line conditioner that I use would clean this up so I tried it and it did, the computer ran fine with it inline, so now I carry one in my tool kit for computers. for testing to see if it is a power problem.

 

[ronaldrc]

Re: Best way to get "Clean" power to a computer ou

Hurk I just don't believe a voltage flunctuation between 118 and 132 would give any problem unless there was a power supply problem.

I don't no that much about processors but I do no they work in the voltage range somewhere between 1 volt for which is considered off low and 5 volts high or on and if the processor got a spike where it should have been a low volt it would probably show an error and close down that program.

Ronald

[ August 26, 2003, 08:07 PM: Message edited by: ronaldrc ]

 

[dereckbc]

Re: Best way to get "Clean" power to a computer ou

I get tickled when I see post like this one. There are a lot of myths and misconceptions about power quality. There is no such thing as clean reliable power from a utility company. There is not a one utility that will guarantee their service, and for good reason, it is not possible.

Many believe by using an IGR all their problems will go away only to find out it did nothing or made the problem worse. Others will think to add a TVSS device at the equipment and expect miracles, only to be disappointed. While others will go down and buy a cheap UPS and expect problems to go away, only to learn later the only thing it protected against was a 5-minute outage and still has noise problems.

There is no magic system that is going to provide you with clean uninterruptible power. It takes a combination of systems with proper design to give you clean power up to without interruption 99.9% of the time.

A UPS can go a long way if the proper unit is selected. The problem is the type you would choose. The inexpensive ones are what is called a standby UPS. These units have a small battery charger, battery, inverter, and some sort of switching mechanism. Under normal operating conditions the equipment is connected to commercial line power and waits in “stand-by”. In the event the power fails for a prescribed number of cycles it switches to battery-inverter. They offer no power line conditioning. These types of units are what you find in residents and small businesses connected to a PC. They range in size from 200 VA up to 1200 VA.

The next step up is called “stand-by interactive” UPS. It is similar to the above unit except the inverter is on line and connected to a ferroresonant transformer with a static bypass switch. During normal operation the equipment is connected to commercial AC power like above, but the inverter does provide some voltage regulation in a boost and buck manner. Again if a failure of power occurs the battery-inverter take over. These units are also available in sizes popular for residential from 200 VA up to a few KVA. Some data processing departments will use these type in a rack mounted model to save money and cut corners.

The next step up is the “double conversion” UPS units used by company’s and people who care about PQ. A double conversion UPS input power is first converted to DC. The DC is used to charge batteries and to constantly operate the inverter at full load. The load is completely isolated from the input. There are lots of extras that can be optioned like a ferroresonant output isolation transformer for further power line conditioners, static bypass for maintenance, parallel redundant modules for extra protection in the event the primary rectifier or inverter fails. These units range in size from 2KVA up to 750 KVA The larger unit will use PDU’s for distribution. A PDU is a step-down/isolation transformer with power line conditioning and alarms.

But even a quality UPS will not carry through a long-term outage. For that you need a generator. Oh well I will quit rambling on, sorry to preach.

 

[hurk27]

Re: Best way to get "Clean" power to a computer ou

Ron
Your right most newer computers with auto switching power supplys will take a wide range of input voltage as they are set up to run at 120 (US) or 220 (Europe) without any switches thrown by the user but this is at a continous voltage if these power supplys had a voltage that fluxuated and if the fluxuation was faster that the power supply could keep up this would probly cause a power good signel to fail and shut down the computer. but I agree as I said a 118 to 132 would probly not pose a problem to most electronics if the fluxuation was at a slow enough rate. but there are some electronics that would be sensitive to it and it would be a problem. the arcing in a breaker is the kind of noise that one computer work fine but another froze up and when put in the same inviroment with a line conditioner (triplet) or an (ESP)it ran flawlessly but it still would lock up with the noise reinserted.
we could get back into common mode noise but that has been discussed to death already.

 

[hurk27]

Re: Best way to get "Clean" power to a computer ou

Dereck
That is very true as I see it every day people will buy plug in surge protectors and think they are protected againest even lightning. as most are led to belive this by the manufactures them selves as this is the first wording on the packages they come in. but in the real world we know differant as we see the end result.
the FTC has relaxed many law as to what a manufacture can clame as truth and now we have power amps that clame 1,000 watts of power with only a 20 amp fuse in the power lead at 12 volts we got air compressers that have a 5 HP rating but only draw 3 amps. so we live in a very mis-leading world and our goverment now alows it.

 

[karl riley]

Re: Best way to get "Clean" power to a computer ou

Hey Dereck, I am silently learning a lot from you. Keep it up, please!

And also the rest of you guys.

Karl

 

[Guest]

Re: Best way to get "Clean" power to a computer ou

Thank you for all that input. How about AoSW?

Originally posted by awwt:
<snip> Also, what do you people know about "approximation of sine wave" (AoSW)? Are there any readily available UPS's that don't use Approximation of Sine Wave? AoSW can be hard on computers. Some computer manufacturers will void their warranty if they find out the system was on a AoSW UPS. That's all I know about AoSW. Can you fill in my blanks?
THANKS!

 

[dereckbc]

Re: Best way to get "Clean" power to a computer ou

There are two ways to generate a sinewave in an electronic UPS; inherent sinewave generation, and approximate generation. Approximate generation is just as the name implies.

There are several ways to generate an approximate sine wave such a a triangle wave, square wave, or digital reproduction. The problem lies in the filter method. They can be noisy.

In any event you would need to look into the technical specs of the UPS to determine the output sine wave. Some will tell you "pure sine wave", or "sine wave" to indicate inherent generation. The approximation units use different words like "stepped approximation", or modified sine wave as examples.

Typically the approximate generation method is used on the small inexpensive home models. Personally I would recommend a MGE "Pulsar Evolution" series line interactive UPS as a minimum, or "Pulsar EX" series dual conversion UPS for maximum performance in SOHO applications.

[ August 27, 2003, 09:45 PM: Message edited by: dereckbc ]

 

[Guest]

Re: Best way to get "Clean" power to a computer ou

Originally posted by dereckbc:
NM cable from the panel that contains the N-G bond is already an IGR. Why pay for it twice?

There is a subpanel in the room where the IGR's are going. The main panel with the N-G bond is 90' away. For voltage drop reasons I think it would be better to put the IGR breakers in the subpanel. The IGR insulated ground is allowed to pass through the subpanel, but I won't be able to pull that off with just NM unless I go all the way from the IGR to the main and put the IGR breakers in there. I have my basic answers now for IGR, but I'm moving on to deeper questions about what happens when equipment is plugged into the IGR. Please post any new answers in the new thread I started here:
New thread by awwt "Ohm's Law and Isolated Ground Receptacles".

I've got the basic mechanics down, but now want to move on to deeper physical properties of IGR.