PQ problem?

Status
Not open for further replies.

JES2727

Senior Member
Location
NJ
A call-center is having some power quality problems that I am unable to diagnose. Occasionally, some of the computers will just re-boot, as if the power has been interrupted. They are not all in the same general location, nor are they all fed from the same panel. It seems to be happening more frequently in recent weeks. It does not affect all of the computers in the area, just two or three at a time. Then it might happen again in another part of the office. I have taken amperage readings of all the circuits in both panels, all three phases, and the neutrals. I have checked the terminations for tightness. I have taken temperature readings of the breakers and the neutral bar. I have found nothing out of the ordinary. The fact that more than one panel is affected and they are supplied by different transformers might indicate that the problem is coming in from the power company. But then it would affect all of the computers at the same time. Instead, it seems to occur randomly, at different times to different workstations.
I'm hoping someone here can offer me some input that may help me resolve this. Thanks.

John
 

charlie b

Moderator
Staff member
Location
Lockport, IL
Occupation
Retired Electrical Engineer
I agree. This sounds like malicious mischief.

Until it is resolved, I suggest an email to all employees recommending that they save their work more frequently than they normally would.
 

ramsy

Roger Ruhle dba NoFixNoPay
Location
LA basin, CA
Occupation
Service Electrician 2020 NEC
JES2727 said:
Occasionally, some of the computers will just re-boot, as if the power has been interrupted. They are not all in the same general location, nor are they all fed from the same panel.
A strategically placed data logging volt meter could verify pwr drops.

If deliberate pwr. interuption is not happening, a voltage sag without UPS protection can also occur by dipping below the ride-through capability of switch-mode power supplies, which reset or lockup during low voltage events.

Ride-through minimums can vary between mfg's. Look behind the PC / Copier / fax / Printer, etc., or pwr. supply for nameplate minimum voltage. Most labeled minimums are close to 108vac, corresponding to the ANSI C84.1 STD, but not all.
iwire said:
My Dell PC says it can run fine with voltages from 90 to 136 volts without changing any settings. :)
One of mine is 110vac, while another has a 90vac minimum also.

If recorded low-voltage events correspond to re-boots of those nameplates with sensitive minimums, (ie) 110vac, and panel, feeder, to service entrance shows similar sags, then look at utility xfmr issues.

If there's no such recorded low-voltage events at panels or service, checking outlet voltages of the failing equipment could find a branch problem. Evidence for branch loading v-drop is also considered when CMV's (outlet N-H) exceed 4.0vac, and may need balancing. CMV's > 2.0vac can also indicate harmonics on balanced branches.

Bulding wiring could also be an issue, if EOL (end-of-line) branch voltages hover around 110vac with balanced loads. If this was my business, I'd ask the owner to either replace the rest of my consumer electronics, with brand new 90vac nameplates, or replace the wiring.

To the extent the nameplate compliance of NEC 110.3(B) includes equipment-voltage limits, branch circuits that fail that minimum-voltage target can not be used with such equipment. Someone should be responsible for either the branch circuits, or a UPS system for my equipment. Ups's exist for multiple or individual PC's.
 

catchtwentytwo

Senior Member
Re: PQ problem?

JES2727 said:
They are not all in the same general location, nor are they all fed from the same panel. It seems to be happening more frequently in recent weeks. It does not affect all of the computers in the area, just two or three at a time. Then it might happen again in another part of the office. .............................................................. The fact that more than one panel is affected and they are supplied by different transformers might indicate that the problem is coming in from the power company. But then it would affect all of the compurs at the same time. Instead, it seems to occur randomly, at different times to different workstations.

Again, this all points to someone "playing around". Before investing a lot of time and money, try the "low-tech" approach. Either lock the panels or figure a way to make it obvious if they are tampered with. Do that without a lot of fanfare and limit the number of people who know about it. This doesn't appear to be a global issue so it will be difficult to install monitoring on every branch circuit. Even if you could, it won't help beyond documenting the time if someone is randomly flipping the breakers .

Try creating a timeline that shows what dates and times the prior events happened. That may narrow it down to a specific shift.
 

tonyi

Senior Member
Not all power supplies are equal. ex. Older early 90's IBM supplies will tolerate an amazing amount of hi/lo voltaging abuse and keep running that would have almost any other machine power cycling.

If this remains a mystery, take a look at the vintage of the machines being affected. During the Pentium II/III era there was a bad batch of electrolytic capacitors that came out of Taiwan. Almost all manufacturers were/are affected in certain models by the bogus caps.

Sabotage sounds like a more likely bet at the moment though.
 

charlie tuna

Senior Member
Location
Florida
power conditioners are only as good as the battery backup capacity vs the connected load. if the connected load exceeds the capacity of the battery ------- you have no backup!!! and if you don't have a data recorder or more monitoring the troubled areas, you are just rolling the dice -- never can you go to your customer and tell him what the problem is and how to fix it.......
 

JES2727

Senior Member
Location
NJ
peteo said:
It seems to be happening more frequently in recent weeks.

So what changed?

Good question. My only thought on this is that perhaps the Air Conditioning is cycling more often with the change of seasons, and maybe this is causing some disturbance.
I do not think it is the work of a prankster. I will return to the sight this week and check the minimum voltage ratings on the nameplates of the affected equipment. I will also try to get more detailed information, specifically how many users are affected and how often. It looks like I may have to bite the bullet and invest in a Data Logger, as I agree that I am just "rolling the dice" at this point.
Thanks to all for the input.

John
 

hardworkingstiff

Senior Member
Location
Wilmington, NC
Whenever I have "ghosts" type problems, I ask my customer to start a log. Time, date, which machine, weather, etc.

They (customer) need to give you good information to trouble shoot this.

Also, once they start documenting, the problem seems to change. This way you chase the correct problem instead of perceived problem.

Good luck
 

dereckbc

Moderator
Staff member
Location
Plano, TX
OK lets assume it is not a prankster for a minute. There is only one effective way to catch the problem, other wise you are shooting in the dark hoping to hit something. Rent you a data logger or two. Put one at the main, and another on a branch panel having problems.

Another thing I would check is how are the PC's being networked. Are they using ethernet and interconnected by cat-5 (look out for cat-5 with a shield called STP), or some older technology that uses coax, RS-232 with a ground signal return. IF networked by standad CAT-5 or 6 on UTP (no shield), OK. If ground is involved, you might be onto something
 

charlie tuna

Senior Member
Location
Florida
how many customers do not realize that the batteries on these power conditioners need to be replaced after a couple of years -- maybe sooner if they have been functioned alot. the small ones should just be replaced when you consider the cost of the batteries and the shipping costs. when a problem arises concerning power conditioned circuits it is usually a "serious" problem to the customer and the first thing that must be determined is where the problem is. this could be a problem that is two years old -- something that causes a voltage sag -- and all this time the power conditioner has been acting to correct the problem -- and now the batteries can't produce enough energy to fill in the "sag". identify the circuit and install a data logger and allow it to record until the customer see's the problem a few times. download the information and you've found one piece of the puzzle. next, you might have to backtrack to the branch circuit panel -- or it's feeder monitoring all phases and the neutral. if you suspect an outside problem you might go right to the service main. and if your customer has a problem outside his facility, he'll need the data information to discuss the issue with the power company. loose connections can occur inside circuit breakers were it cannot be seen. intermittent arcing occurs driving equipment crazy. neutral connections can also be a major problem especially if harmonics are present. the cost of good data loggers is another thing, but if you are serious about solving the problem professionally ---you need the right tools!
 

peteo

Senior Member
Location
Los Angeles
It would be very difficult to see the actual power input at the workstations with handheld testing equipment. Best bet is never forget that everything has inputs and outputs, and what you can test easily (to eliminate) will get you to the real answer faster than trying to narrow down research into a suspcted or even known issue. How friendly and together are these people, as hardworking said, information will get you there; if they work with yellow stickies can the affected users note the time of day and exactly what they were doing when it happened, for example.

Having said that, an internet search using words such as 'PC' 'reset' 'office' etc., may bring others' experience. PC power supplies have inputs via cubicles (?) and surge supressors, and an IEC connection which is made to fall out if needed. 115VAC and NEMA receptacle issues are seldom 'only' momentary. Without getting into LVPS and PC issues, the question here would be whether all the affected equipment was made by the same manufacturer or not.
 

ramsy

Roger Ruhle dba NoFixNoPay
Location
LA basin, CA
Occupation
Service Electrician 2020 NEC
dereckbc said:
.. (look out for cat-5 with a shield called STP), or some older technology that uses coax, RS-232 with a ground signal return.
How do network-cable grounds reset systems, and how do we test for it? (ie) Where to put the data logger, meter, etc..
 
Status
Not open for further replies.
Top