Voltage step change, likely cause

Status
Not open for further replies.

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
090404-1624 EST

Today on my voltage plot from about 1400 to 1615 EDT there was a fairly consistent square wave modulation of the 120 V line voltage of very close to 1 V. The total period of one cycle was about 2 minutes, and mostly 1 minute high and 1 low. However, over the two hour period there was variation in the high and low periods, and it was not present the entire time. This does not originate from my load. Remote possibility that my neighbor could be the cause, but I am guessing it originated from the grid.

This ended at 1620 EDT and has held within +/- a couple tenths of 124.0 since then. Current time is 1732 EDT.

If it was a voltage regulator would one see this type of oscillation? I would assume that voltage regulators are designed to avoid tap hunting. If it is not a regulator problem are there any types of industrial loads that might produce this type of variation with this period? It was very much a square wave change meaning within 1 to 2 seconds for the transition from high to low. My voltage measurement is quantized in 1 second averages.

.
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
090404-1709 EST

hurk27:

I am in a residential neighborhood and have one neighbor on my transformer. It would take something much larger than a battery charger on the primary side of my transformer to cause this type of variation.

We have no giant arc furnaces close by, they are 30 miles away, and I think these run a continuous arc. We do have what has been claimed to be the world's most powerful laser. However, that is instantaneous power, and I doubt the average power is all that much to show up as an effect on the grid. Further it might get all its power from the University power plant.

.
 

mivey

Senior Member
I doubt it is a voltage regulator. Does not sound like a regulator test either as it would not take two hours like you described. It sounds like load. If your neighbor had a big shop, you would notice (his name is not Tesla is it? :D).

What is your distribution voltage? What else is on your feeder? I don't remember the saw mill I was at last time taking a minute to cut the log, but I guess it would depend on what was being cut.

Any car crushing plants nearby? Any similar loads? Even a big enough load on the same substation bus would be suspect.

Do you know the highside voltage at the substation? This would give an idea of how stiff your system might be and if any load above your substation transformer might cause problems.

Go to the substation and ride out the feeder if you feel froggy. If you don't feel like riding the feeder out, just ask the POCO and they can tell you the big loads fed from the same substation bus.
 

Doug S.

Senior Member
Location
West Michigan
Is your neighbor trying to "sell" power pack to the POCO on a homemade setup? :grin:

I've noticed I see all sorts of (what I see as ) anomalies when I watch the POCO power. I imagine it could (or is) a full time job tracking some of them down... ?
 

winnie

Senior Member
Location
Springfield, MA, USA
Occupation
Electric motor research
How much load on the secondary of your transformer would be needed to give a 1V drop? How large a transformer is it, and how heavily loaded?

I'm wondering if something like an electric stove element (which cycles on and off on the order of 10's of seconds to minutes; listen closely and you'll hear them cycle) would be enough of a load to cause the 1V drop.

-Jon
 

mivey

Senior Member
He said it was not his load. Maybe the neighbor. What size transformer do you have and we can run a regulation calc.

Usually, you might expect residential load to pick up after 5 pm. This step change started at 2 pm and ran to 4:20 pm, which makes me think it was an industrial load somewhere. It would be nice to see more history and if it was isolated to around 2:00 every day/weekday, I would suspect some process that was started to gear up for the second shift or end of day.
 

mivey

Senior Member
As an approximation of the secondary load you are looking for, given the following transformers and 1 volt drop, this is the load:
25 kVA : 11.2 kW
37.5 kVA : 18.2 kW
50 kVA : 26.3 kW
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
090404-2210 EST

mivey:

charlie has estimated my transformer at 25 KVA. Photos of this are at:
www.beta-a2.com/MISC_TMP_PHOTOS_files/PICT2540A.jpg
www.beta-a2.com/MISC_TMP_PHOTOS_files/PICT2541A.jpg
www.beta-a2.com/MISC_TMP_PHOTOS_files/PICT2544A.jpg
www.beta-a2.com/MISC_TMP_PHOTOS_files/PICT2545A.jpg proof of delta source
www.beta-a2.com/MISC_TMP_PHOTOS_files/PICT2546A.jpg shows transformer primary

I agree it seems like a load.

I do not know the location of the substation. There should be nothing but residential and commercial on it. I believe most of the University is fed from their power plant which will include the giant U Hospital complex. I would expect that the University's power plant is a float on the grid.


Doug S:

There are only two of us on one transformer, and he is a professor of business administration. But I am beginning to wonder if their oven may be the source.


winnie:

About 3 KW on a 120 V side produces about 1 V on that side. Based on about 0.5 V with a 1.5 KW load. Virtually all this is transformer impedance.

During most of the time when the square wave voltage variations were occurring we were gone and my load was relatively constant, at least known variations were in the 0.4 KW range and uncorrelated with the voltage variations.

I will check with the neighbors and see if their oven may have been used during the time period. More on this below.


mivey:

I think my comments above answered your questions.

I had some thought of an industrial load, but I am leaning to my neighbor's stove.

I believe my transformer has a higher impedance than you are predicting or it is smaller. 12 A at 120 V is producing about 0.5 V. 24 A and 240 would be about 5.76 KVA.


Additional comments:

I would like to provide plots, but that is difficult at the present time because the software I am using can not provide output to anything except the screen and that even has some problems.

There are software and possibly hardware problems in the "TED" system I am currently using. However, I can manually read thru most of these problems and I do not believe they are the source of what I have described. But it is possible some sort of software bug might be the origin.

I ran some experiments with my oven. It does look like when a moderate size, 3 KW, step function change in load occurs in my house in the positive direction that there is a loss of data transfer for about 1 minute from the main panel electronics to the indicator unit and thus to the collection computer.

Looking at one time period where data was being received every 1 second I had a voltage change from about 123.4 at 1400:29 to 123.2 at 1400:46 to 123.1 At 1400:55, and then 122.6 at 1400:56. It continued around 122.7 to .6 until 1402:51, then changed to 122.8 until 1402:58, then at 1402:59 it jumped to 123.3 and was close to this until 1405:43, then it dropped to 122.0. All this time my load was in the range of 0.43 to 0.68 KW. So here we were in the range of 2.5 to 3 minutes per half cycle of the square wave.

The periods that I mentioned previously had the higher frequency.

It is hard to see the periodicity in the raw data compared to the plots.

My oven from cold to full temperature on bake at 350 deg about 8 minutes and about 3 KW. Off for less time, about 5 min, on for short time and dropped back to a lower level for 5 min. then I changed to broil. I need to run a more controlled experiment now that I see a partial drop back. But maybe something else came on at that time. A refrigerator, freezer, or furnace.

More at a later time.

.
 

mivey

Senior Member
Given the data of about 0.5 volt drop with 1.5 kW load, consider the following configuration:

A 15 kVA transformer & 150 ft of #2 and 1.5 kW load yields a drop of 0.2325v for the transformer and 0.2723v for the service drop for a total of 0.505v. This matches the results at your house.

This configuration will take a secondary load of 2.97 kW to yield a drop of 1. But this is not the transformer alone. It will take a load of 6.45 kW at your neighbor's house (i.e. the transformer alone) for you to see a 1 volt drop.

A similar scenario could be run for a 25 kVA transformer and a 200 ft #2 service drop (yields 0.5v drop with 1.5 kW at your house). In this case, it would take a 3 kW load at your house for you to see a 1 volt drop, but a 11.24 kW load at your neighbor's house for you to see a 1 volt drop.

I have assumed some service configurations but you could fill in these details for a better calculation.
 

wptski

Senior Member
Location
Warren, MI
090404-2210 EST

Additional comments:

I would like to provide plots, but that is difficult at the present time because the software I am using can not provide output to anything except the screen and that even has some problems.
.
gar:

Won't the keyboard "PrtSc" to your clipboard buffer then a paste to a photo editor for trimming work?
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
090405-1048 EST

mivey:

I have 0000 copper and I estimate about a total length of about 80 to 90 ft from main panel to transformer. This is 30 ft or less down the pole, 4 ft into the ground, 40 ft across the yard, about 8 ft up to meter, 6ft back down and into the main panel. At 100 ft of 0000 this is about 0.005 ohms.

See my post #14 at
http://forums.mikeholt.com/showthread.php?t=111313&highlight=0000+voltage+drop+copper+1500

Basically my comment of virtually all the impedance is at the transformer is about correct. There is drop in the meter, the fuse, the breaker connection to the bus bar, and the breaker. Much less in the neutral path.

The "TED" device that I am currently using for monitoring has the current transformers on the two hot lines just above the input lugs in the main panel. The "TED" voltage measurement is only from neutral to its own breaker. Thus, it reads the neutral to phase X's bus.

My pole transformer is about 43 years old. Whether higher impedance transformers were being used then or not I have way to know.


Larry:

I like the tankless idea, but it would show at other times of the day and I would assume a much different time constant.


chris:

What I mean by square wave is the shape of the modulation on the power line voltage. Consider the power line as a carrier frequency of 60 Hz. On top of this carrier is a modulation that makes a small change in the amplitude of the carrier. The square wave modulating frequency is about 1/100 Hz. 1 minute high followed by 1 minute low, a 2 minute period for one cycle. Square wave means it is an abrupt transition from one level to another of two possible levels.

A short sample of data
1720:01 122.5 V ---- my microwave was on for a short time to this point
. ------------------- step change in voltage here
1720:02 123.0 V
1720:59 123.1 V ---- 57 seconds on of the thing I have been talking about
. ------------------- step change in voltage here
1721:00 124.0 V
1721:43 123.9 V ---- 43 seconds off
. ------------------- step change in voltage here
1721:45 122.9 V
1722:58 123.1 V ---- 73 seconds on
. ------------------- step change in voltage here
1722:59 124.0 V
1723:45 123.8 V ---- 46 seconds off
.etc

I have taken some photos of the screen plots, but it will be some time before I add notes.

For the "TED" system to be a reliable system they are going to have to make some improvements. I am not in the sales business for this kind of product, but if I was I would not sell it at the current stage of development. However, for someone that wants to play it is a low cost useful tool.

.
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
090405-1145 EST

I have talked to my neighbor and their oven is gas and therefore they are not the source of the variations.

jghrist:

If it was a timer controlled capacitor bank or some other timer controlled system, then I would have expected a more precise period from one cycle to the next.

I am now fairly certain it originated from the primary lines. The only other likely possibility is a problem in the "TED" device. But the nature of the signal implies it is external to TED.


wptski:

That can be done, but it does not produce the results I want, nor do the photos. I need to process the raw data and create my own plot. That is time consuming.

.
 

cadpoint

Senior Member
Location
Durham, NC
Couldn't it be as simple as someone bringing some big service back on line?
It seems that we always bring up big stuff early Saturday morning, or in the middle of the day Saturday!

I know, I'm on the "C" phase with the Sub-station about 1 mile directly down the street, I can experience flickers or even brown'n on the weekends, and can hear a trannie let loose in a thunder storm. The same can happen with a car smacks a pole and jolts a power pole...
 

mivey

Senior Member
OK. On to the primary, but we need input.

The answers to my questions in #4 would be easiest to get from the POCO. It might interesting for you to see if you could trace your feeder back to the sub. The wire size should increase as you get closer.
 

mivey

Senior Member
gar,

I am going to make a pretty broad assumption about your location, distance to substation, substation transformer, voltage, etc.

If you are trying to identify a suspect load on your feeder, look for something that could have a step load variance of around the 1/2 MW magnitude. If looking for a load on a different substation feeder, look for something that could have a step load variance of around a 1 MW magnitude.
 
Status
Not open for further replies.
Top