Power spikes in new building - utility won't admit fault

[gozoliet]

We've just finished construction on our new office building, and have noticed the lights flickering frequently throughout the day. Had the utility company to measure and test everything and they claim it's within our building.

On several occasions I've measured a roughly 8V drop (for example from 116V to 108V) that lasts 5-10 seconds on a120V outlet. I have visually confirmed that one of those drops happened when the lights flickered.

I'm waiting to get my hands on a data logger to measure at the panels. In the meantime I managed to convince the utility to send me their data log from when they measured just inside our meter. I'm seeing a lot of 20V spikes on the 340V (see graph) during business hours, which probably coincide with what I'm measuring at the outlet.

Utility acknowledged the spikes on their graph, but claim that they still fall within their "allowable range" so shouldn't be a problem. I concede that the power can move within a wide range, but the sudden fluctuations seem like a problem to me. Can anyone comment?

 

[davidbeach]

Motor loads starting at the same time as your fluctuations?

 

[itsmoked]

There are a lot of possibilities for this. You should probably get someone on site to flush out your quarry. I doubt it's the utility. Can you go next door and ask them about their power? If it was the utility it would likely be effecting them too.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[magoo2]

You might ask the utility if they are willing to set another meter inside your building and, at the same time, set one at a nearby facility served from a different transformer. If they both show similar variation, then it's on the utility side. If not, then you better look inside your building.

 

[cranky108]

Humm, Office Building, 120V, 20V spikes on the 340V.

First the 120v and 340v don't seem to go togather.

Second, lights flickering, and you have single phase or three phase service? If the office building is very large, the lights probally aren't on 120V.
Common feature maybe air conditioning, electric heat, old copier, or something else.

Third the power is within normal range, so a surge/sag protective device might be helpful.

What type of lighting? Some are more sensitive than others.

 

[magoo2]

We've got 277/480 V and 120/208 V in buildings in the U.S. I think Canada has some 346/600 V, so I'm guessing you've got 600 V at the main service panel and 120/208 V downstream from that. So I think they go together.

 

[gozoliet]

Thanks for the feedback. This is a roughly 12000 square foot one-story building, in Canada. We have 347/600 coming in which then goes to our transformer then to panels. Utility has been on site and measured in underground, and at the transformer and they say OK, but they're probably doing one-time measurement and not seeing the spike. They mention loose neutral as possibility. Electricians have been on-site and re-tightened everything.

Only motors are 4 roof-top units which are just doing heat right now, no A/C. Everything else is lights and computers.

Lights which are dimming are incandescent/ regular light bulbs. Fluorescent don't noticably flicker. Neighbors haven't noticed, but they also don't have incandescent.

Until I get someone in with another data-logger only concrete data I have is from a UPS which seems to miss times when logging, and the utility's data logged which is attached. Otherwise I"m caught in the middle of a finger-pointing game and trying to pull at any thread I can find.

THanks!

 

[jghrist]

You get the same % drop on the 120 and the 347. This indicates the problem is in the 347/600V system. Or maybe on the utility primary system. It could be a loose neutral on the utility side. Looking at all three phases will help. If it's a loose neutral, the dip in one phase will be accompanied by an increase in another other phase or phases.

Try to get the utility to measure the primary of their transformer at the same time as the secondary to see if the problem is on the primary.

Also measure the current at the same time as the voltage to see if the dip is caused by a high momentary current on your system.

 

[gozoliet]

Fortunately the utility measured the current at the same time. Here's some data from the logger at one spike. The logger captures min/avg/max during the 10 second interval, so I'm providing min. I guess the biggest question - could that surge cause dimming lights?

08:55:20 V1=341V, V2=344.9V, V3=343.4V / I1=23.81A, I2=21.06A, I3=21.06A

08:55:30 V1=328.8, V2=333.2, V3=330.4 I1=20.15 I2=20.15 I3=20.15

08:55:40 V1=316.8 V2=321.4 V3=318.9 I1=18.32 I2=20.15 I3=18.32

08:55:50 V1=335.9 V2=340.1 V3=338.3 I1=21.98 I2=21.6 I3=21.06

 

[waross]

It would be a little more helpfull if you added MAX current and normal current.
You need to see if you have current peaks coinciding with voltage dips. Normal current and maximum current, normal voltage and minimum voltage.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[westom]

If claiming a dimming incandescent bulb is a spike, then they are laughing at you - not taking you seriously.

View your numbers. No spike exists. So nobody will take you seriously. Voltage has dropped 9% from what it must be. Voltages dropping more than 5% are harmful to electric motors and must not exist.

So where did they measure those voltages - exactly. Inside the transformer. On a wire outside of the transformer. One the wire where it enters your meter. On the wire from meter to breaker box. On a ... literally you must be that exacting if you want a useful answer here or to have them take you seriously.

Incandescent bulbs are a perfect benchmark for finding defective wiring. Major appliances must power on and off - and no incandescent bulb must change intensity. Those intensity changes are typical of defective building wiring. Major intensity changes imply a serious human safety issue.

But you have significant and unacceptable voltage sag across all three phases when currents are near zero (20 amps).

Because so many consumers have been brainwashed with nonsense about surges, then electrical people often just zone out. Ignore those educated by retail salesmen. Could be what happened.

You have a voltage drop - not a spike or surge. As long as you use words promoted by myth purveyors, then some electrically informed will ignore you. Use correct terms AND include numbers. For example, your voltage sagged 9% below spec. That is unacceptable because low voltage is harmful to electric motors and because properly installed wiring does not create such voltage drops.

 

[dpc]

You have a voltage sag or dip. Not a surge. And yes, this is what is causing the lights to dim. It appears to be fairly balanced across all three phases.

If utility can measure voltage on primary and secondary of the transformer at the same time, as already suggested, the problem can be isolated further.

FWIW, in the large majority of these cases (but not all), the problem is on the customer's side of the meter, not the utility. So keep an open mind.

As David Beach first suggested, I still suspect some internal load is drawing high current for a short time, causing high voltage drop.

David Castor

http://www.cvoes.com

 

[jghrist]

Do you have any phase-to-neutral loads on the 347/600 V system? If all loads are 600V transformers (3Ø delta or 600V 1Ø) and 3Ø motors, then a loose neutral is not your problem; you won't have any neutral current anyway.

The apparently balanced drop on all phases is inconsistent with a neutral problem also. I say "apparently" because you could have unbalanced dips on different phases at different times (not too likely).

 

[cranky108]

Where are the Incandescent bulbs on, voltage wise?

You also said computers, which leads to the question of how many?
The term loose neutral, makes me think that if you have bunches of computers, and reduced neutral conductor, you could have a harmonic problem.

 

[gozoliet]

You guys are really helpful. Being in between the electricians and hydro I am sure they are both trying to swat me away and not have to fix anything. Thank you for the correct terms - I'm a computer engineer, so the closest I've been to this stuff is textbooks.

OK. The correct problem is a voltage drop causing dimming incandescent bulbs.

The meter reads by Hydro were taken between our Meter and the first distribution panel at 347/600 which goes to roof-top units and to our transformer. The lights are connected to 120V panels. There are about 30-40 computers, at the time of these readings I think maybe a total of 4kva of UPS (just the servers).

We did identify a strong sudden increase (30Amp) in current from our building, but these happened outside of work hours (between 10pm-6am) and the voltage at the time did not dip. The voltage only happen Monday-Friday between 6am/10pm. Our office is open 7am-8pm.

I can send the Powerview data file (190MB) if anyone is better at reading those than I am.

For those who want min/average/max of both volts and currents.. take two:

11:45:00 V1=342.9/343.4/344.1 V2=347.3/348.0/348.5 V3=346.5/346.9/347.6 I1=30.22/31.14/36.63 I2=26.56/27.47/32.97 I3=21.06/21.98/22.89
11:45:10 V1=331.0/343.6/345.1 V2=336.3/348.2/349.6 V3=335.0/347.3/348.7 V3=335.0/347.3/348.7 I1=27.47/31.14/37.55 I2=24.73/26.56/36.63 I3=18.32/21.06/21.98
11:45:20 V1=318.7/339.7/343.6 V2=324.2/344.5/348.4 V3=322.9/343.2/347.4 I1=27.47/32.05/37.55 I2=25.64/29.30/36.63 I3=19.23/21.06/24.73
11:45:30 V1=339.6/340.1/340.5 V2=344.1/344.7/345.2 V3=342.9/343.4/344.1 I1=32.05/32.05/37.55 I2=29.3/30.22/34.8 I3=20.15/21.06/22.89

My next step will be to recontact Hydro to see if they can do further testing.

 

[westom]

Posts from magoo2 and ighrist discuss your likely problem. Notice variation occurs on all three voltages simultaneously. That implies no one wire in your building or no one load is causing problems. Implies the common power source (transformer or utility wires) is undersized for the load (yours and all neighbors).

Your current variations are near zero. But voltage varies significantly. That should not happen if utility service is properly sized.

Based upon how I read your data, current increases when voltage increases. Or current increases causing no voltage variation. If problem were in wires or transformer, then current should increase causing voltage to drop. I am wondering if the utility's voltage stabilizer (for the area) is somehow adjusting erratically - especially due to massive power consumption variations by a nearby consumer. Normally current increases coincide with voltage drops. I don't see that trend in your numbers. This further directs blame to the utility side. Their job is to maintain voltage with all current changes. Apparently they are not.

And neighbors would also be suffering same voltage variations.

But to make this problem more interesting - that many computers consuming a greater percentage of total power implies a more reactive load or harmonic problems. Especially if computers are clone type with inferior power supplies. Solution would be easier if power factor numbers for each phase were also provided/measured. Highly reactive loads mean wires and transformers must be sized slightly larger than is standard.

If you have a highly reactive load, also consider power factor compensation equipment. For example, if you power factor was .5 (it would not be that bad), then you are paying for twice the amount of power that you are actually consuming. If power factor was adjusted to 1, then your energy consumption decreases by half. This paragraph is about something 'also learned' so that you might address a second potential problem as well as address the original problem.

Meanwhile, your power if varying 'out of spec'. Surge/sag device is obviously not a solution.

 

[tinfoil]

In your last post, you mentioned a 'first distribution panel'. Does that mean that there are other 347/600V panels?
The lion's share of your building's load (and therefore current) may not be captured by your meter.

Is this part of a 'campus' of buildings supplied by a common utility transformer? Are these buildings owned by a common landload? Is the primary network supplying your building owned by the utility or by the landowner?

I ask these questions to point out that the problems may be 'external' to your service yet still not the responsibility of the local utility.

Reviewing your data (particularly the slow timescale where 30 seconds is elapsing during the 'voltage sag'), it looks like some larger-than-designed load is starting up. Is there non-office industries very near yourt facility?

 

[rockman7892]

What is the acceptable level of voltage drop on any bus within a facility during motor starting? I have often heard that 80% drop is acceptable on all buses, including the bus that the motor is being started from? Is this correct?

 

[dpc]

80% will generally cause problems. NEMA contactors are supposed to hold in down to 85% voltage, IIRC. Below that, all bets are off. At 80% voltage for more than a fraction of a cycle, you can expect contactor chatter, drop-out and other exciting things.

David Castor

http://www.cvoes.com

 

[rockman7892]

dpc

Your comments about contactors dropping out make sense. What about % drop as far as motor fields dropping out or decaying to where the motor shuts off?

Sounds like if you are saying that 85% is too low so would a value more like 90% drop be a minimum value.

What about the % drop for the actual motor that is being started?